Rethinking Homicide: Exploring the Structure and Process Underlying Deadly Situations (Cambridge Studies in Criminology)

This page intentionally left blank

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

Rethinking Homicide Using multiple data sources and methods, this book involves a microhistorical analysis of the nature of change and stability in homicide situations over time. It focuses on the homicide situation as the unit of analysis, and explores similarities and differences in the context of homicide for different social groups. For example, using Qualitative Comparative Analysis, we investigate whether various social groups (e.g., men vs. women, teenagers vs. adults, strangers vs. intimates, Blacks vs. Whites) kill under qualitatively different circumstances and, if so, what are the characteristics of these unique profiles. The analysis of over 400,000 U.S. homicides is supplemented with a qualitative analysis of narrative accounts of homicide events to more fully investigate the structure and process underlying these lethal situations. Our findings of unique and common homicide situations across different time periods and social groups are then discussed in terms of their implications for criminological theory and public policy.

Terance D. Miethe is currently Professor of Criminal Justice at the University of Nevada, Las Vegas. He is the author and coauthor of several books, including Crime and its Social Context, Whistleblowing at Work: Tough Choices in Exposing Fraud, Waste, and Abuse on the Job, The Mismeasure of Crime, and Crime Profiles. Wendy C. Regoeczi is currently Assistant Professor of Sociology at Cleveland State University. She is the author of several articles on homicide and has served as guest editor for a special issue of the journal Homicide Studies on missing data.

i

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

Cambridge Studies in Criminology Edited by Alfred Blumstein, H. John Heinz School of Public Policy and Management, Carnegie Mellon University and David Farrington, Institute of Criminology, University of Cambridge Other books in the series: Life in the Gang: Family, Friends, and Violence, by Scott H. Decker and Barrik Van Winkle Delinquency and Crime: Current Theories, edited by J. David Hawkins Recriminalizing Delinquency: Violent Juvenile Crime and Juvenile Justice Reform, by Simon I. Singer Mean Streets: Youth Crime and Homelessness, by John Hagan and Bill McCarthy The Framework of Judicial Sentencing: A Study in Legal Decision Making, by Austin Lovegrove The Criminal Recidivism Process, by Edward Zamble and Vernon L. Quinsey Violence and Childhood in the Inner City, by Joan McCord Judicial Policy Making and the Modern State: How the Courts Reformed America’s Prisons, by Malcolm M. Feeley and Edward L. Rubin Schools and Delinquency, by Denise C. Gottfredson The Crime Drop in America, edited by Alfred Blumstein and Joel Wallman Delinquent-Prone Communities, by Don Weatherburn and Bronwyn Lind White-Collar Crime and Criminal Careers, by David Weisburd and Elin Waring, with Ellen F. Chayet Sex Differences in Antisocial Behavior: Conduct Disorder, Delinquency, and Violence in the Dunedin Longitudinal Study, by Terrie Moffitt, Avshalom Caspi, Michael Rutter, and Phil A. Silva Delinquent Networks: Youth Co-Offending in Stockholm, by Jerzy Sarnecki Criminality and Violence among the Mentally Disordered, by Sheilagh Hodgins and Carl-Gunnar Janson Corporate Crime, Law, and Social Control, by Sally S. Simpson Companions in Crime: The Social Aspects of Criminal Conduct, by Mark Warr The Criminal Career: The Danish Longitudinal Study, by Britta Kyvsgaard Violent Crime: Assessing Race and Ethnic Differences, edited by Darnell Hawkins Gangs and Delinquency in Developmental Perspective, by Terence P. Thornberry, Marvin D. Krohn, Alan J. Lizotte, Carolyn A. Smith, and Kimberly Tobin Situational Prison Control, by Richard Wortley Early Prevention of Adult Anti-Social Behavior, edited by David Farrington and Jeremy Coide iii

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

Rethinking Homicide Exploring the Structure and Process Underlying Deadly Situations

Terance D. Miethe University of Nevada, Las Vegas

Wendy C. Regoeczi Cleveland State University

with assistance from Kriss A. Drass

v

21:38

cambridge university press Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo Cambridge University Press The Edinburgh Building, Cambridge cb2 2ru, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521832991 © Terance D. Miethe and Wendy C. Regoeczi 2004 This publication is in copyright. Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published in print format 2004 isbn-13 isbn-10

978-0-511-16522-1 eBook (EBL) 0-511-16522-6 eBook (EBL)

isbn-13 isbn-10

978-0-521-83299-1 hardback 0-521-83299-3 hardback

isbn-13 isbn-10

978-0-521-54058-2 paperback 0-521-54058-5 paperback

Cambridge University Press has no responsibility for the persistence or accuracy of urls for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

This book is dedicated to Kriss A. Drass, a friend and collaborator, whose untimely death has left a major void in our personal and professional lives.

vii

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

Contents

List of Figures List of Tables Preface Acknowledgments

page xiii xv xvii xxi

1. Introduction: Studying Homicide Situations Histories of Violent People and Dangerous Times Macro-Quantitative Approaches Subgroup Differences in Homicide Risks and Rates Homicide Situations as the Unit of Analysis Qualitative Comparative Analysis and Homicide Situations Structural Elements Underlying Homicide Situations Data Sources for Studying Homicide Outline of the Current Study

1 4 6 7 8 10 11 12 13

2. Theoretical Development Definitions of the Situation Macro- and Micro-Theories of Change in Homicide Situations Summary

15 16

3. Data and Methodology for Studying Homicide Situations Data Sources for Studying Homicide Data Sources in the Present Study Analytic Approaches for Studying Homicide Situations Qualitative Comparative Analysis Summary

36 36 41 44 50 60

30 35

ix

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH

x

November 11, 2003

CONTENTS

4. The Empirical Distribution of Homicide Historical Trends in Homicide Offender, Victim, and Offense Elements in Homicide over Time Homicide Situations in SHR Data, 1976–98 Conclusions

61 61 72 83 99

5. Instrumental and Expressive Motives in Homicide Situations Instrumental and Expressive Crimes The Structure of Motive-Based Homicide Situations Narrative Accounts of Expressive and Instrumental Homicides Summary of Findings Conclusions and Implications

101 101 104

6. Gender Differences in the Structure of Homicide Situations Patterns of Homicide by Male and Female Offenders Theoretical Perspectives Research Questions Methods and Analysis Results Narrative Descriptions of Homicide Situations Conclusions and Implications

130 131 132 136 137 142 153 160

7. Change and Stability in the Structure of Youth Homicide Trends and Correlates of Youth Homicide Research Questions Methods and Analysis Results Homicide Narratives for Juvenile Offenders Summary of Findings Conclusions and Implications

165 166 168 168 169 179 187 188

8. Racial Differences in Homicide Situations Trends in Race-Specific Homicide Rates over Time Theory and Method in Homicide Research Racial Differences in Homicide Offending Changes in the Nature of Homicide over Time Homicides among Hispanics The Analysis of Homicide Situations Research Questions Data Description and Methods Results

191 192 193 194 196 198 199 199 200 201

116 125 126

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH

CONTENTS

Homicide Narratives Summary of Findings Conclusions and Implications 9. Victim–Offender Relationships Types of Victim–Offender Relationships The Disaggregation of Homicide by Victim–Offender Relationship Trends in the Victim–Offender Relationship over Time Data and Measurement Results of Qualitative Comparative Analysis (QCA) Narrative Accounts of Family and Stranger Homicides Summary of Findings Conclusions and Implications 10. Conclusions and Implications Major Findings Dispelling Myths and Misconceptions Theoretical Advances Public Policy Issues and Violence Intervention Limitations and Future Directions Conclusion References Author Index Subject Index

November 11, 2003

xi 215 223 224 227 228 228 230 231 232 246 253 254 256 256 259 265 272 276 279 281 307 315

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

Figures

4.1. 4.2. 4.3. 4.4. 4.5. 4.6. 4.7. 4.8. 4.9. 4.10. 4.11. 4.12. 4.13. 5.1. 7.1. 7.2. 8.1. 8.2. 8.3. 8.4. 8.5.

U.S. Homicide Victimization Rates over Time Offender’s Sex and Age Offender’s Race Victim’s Sex and Age Victim’s Race Expressive and Instrumental Motives Types of Instrumental Motives Victim–Offender Relationship Weapon Used Multiple Offenders and Multiple Victims Number of Homicide Situations over Time Number of Homicide Cases in Situations over Time Profile of Unique and Common Homicide Situations Homicide Situations and Cases with Unique and Common Motives Proportion of Common and Unique Situations over Time by Age Main Effects across Context [Age Differences] Proportion of Homicides in Common and Unique Situations [Racial Differences] Changes over Time in Attributes Distinguishing Unique Profiles [Racial Differences] Cases Involving Unique and Common Situations for Hispanic and White Offenders Attributes Distinguishing Uniquely Hispanic vs. White Offenders Cases Involving Unique and Common Situations for Hispanic and Black Offenders

page 64 74 75 76 77 78 79 80 81 82 85 86 87 107 174 177 203 208 209 212 213 xiii

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH

xiv Attributes Distinguishing Uniquely Hispanic vs. Black Offenders 9.1. Percent of Cases in Unique and Common Profiles by Relationship Type 9.2. Cases within the Dominant Unique Profiles for Acquaintance Homicide 9.3. Stability and Change in Intimate Partner Homicide Situations 9.4. Unique and Common Profiles for Acquaintance vs. Stranger Homicides

November 11, 2003

FIGURES

8.6.

214 233 236 241 243

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

Tables

3.1. Qualitative Comparative Analysis of Gender Differences in Homicide Situations (Chicago 1965–69) page 56 4.1. Profiles of the 25 Most Dominant Homicide Situations Common across All Three Decades (SHR Data, 1976–98) 89 4.2. Profiles of the 25 Most Dominant Homicide Situations Unique to the 1990s (SHR Data, 1990–98) 95 5.1. Particular Offender, Victim, and Offense Characteristics Underlying Uniquely Expressive, Uniquely Instrumental, and Common Homicide Situations 108 5.2. Profiles of the 10 Most Dominant Situations Unique and Common to Expressive and Instrumental Homicide (SHR 1976–98) 110 5.3. Comparability of Structural Profiles within Subtypes of Expressive and Instrumental Homicides 115 6.1. Coding, Value Labels, and Frequencies for All Variables (Chicago Homicides, 1965–94) 139 6.2. Qualitative Comparative Analysis of Homicide Situations Unique to Male and Female Offenders, 1965–94 144 6.3. Qualitative Comparative Analysis of Stability and Change in Male Homicide Situations over Time 148 6.4. Qualitative Comparative Analysis of Stability and Change in Female Homicide Situations over Time 150 7.1. Age-Specific Distribution of Homicide Cases across Variables over Time (Source: FBI SHR Data, 1976–98) 170 7.2. Qualitative Comparative Analysis of Homicide Situations Common to Both Age Groups, Unique to Juvenile Offenders, and Unique to Adult Offenders (Source: SHR Data, 1976–98) 172 xv

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

xvi 8.1. Coding, Value Labels, and Frequencies for All Variables (Source: FBI SHR Data, 1976–98) [Racial Differences] 8.2. Qualitative Comparative Analysis of Homicide Situations Unique to Black and White Offenders, FBI SHR Data, 1976–98 8.3. Qualitative Comparative Analysis of Homicide Situations Unique to Hispanic, White, and Black Offenders, California 1987–99 9.1. Qualitative Comparative Analysis of Homicide Situations Unique to Intimate Partner/Other Family, Friend/ Acquaintance, and Stranger Homicides, 1976–98 9.2. Qualitative Comparative Analysis of Homicide Situations Unique to Intimate Partner and Other Family Homicides, 1976–98 9.3. Qualitative Comparative Analysis of Stability and Change in Intimate Partner Homicides 9.4. Minimized Solution of the Most Prevalent Profiles of Uniquely Acquaintance and Uniquely Stranger Homicides, 1976–98

TABLES

202

205

210

234

238 242

244

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

Preface

We are exposed to a phenomenal amount of information about homicide in everyday life. It is the focus of local and national news, a primary theme in TV dramas and motion pictures, and the prime material in “true crime” magazines and the widely popular murder mystery. Homicide sells in most Western countries because of its novelty and seriousness. Although statistically the least frequently committed major crime, the public is often totally enthralled and mesmerized by homicide. This fascination with homicide is understandable given that it offends the basic values we place on human life and vulgarizes the presumed civility and moral supremacy of modern culture. Criminologists in the social sciences are not immune to these forces. We are equally captivated and appalled by lethal violence, devoting far more attention to homicide than any other criminal act. The scientific literature on homicide is absolutely enormous. Lethal violence is an omnipresent topic in major criminological journals. Most critical tests of existing theories of crime causation and its distribution have focused on lethal violence because it is widely held that official counts of homicide are more reliable and valid than for any other major crime category. Furthermore, public policies on crime prevention and research on risk factors are disproportionately directed toward violence. Using this vast scientific literature on lethal violence as the background, our goal in this book is to encourage “rethinking” about homicide as it relates to how we describe, explain, and study it. Our rethinking of homicide has lead us to depart from existing traditions in the following three ways. First, we focus on the homicide situation as the unit of analysis, exploring the structure and process underlying these deadly encounters. Second, homicide situations are defined by the nexus of offender, victim, and offense elements in time and space. It is the combination of these elements, not xvii

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

xviii

QC: FCH/FFX

T1: FCH November 11, 2003

PREFACE

their operation in isolation, that provides the context for lethal violence. Third, we apply a relatively new comparative method for the study of combinatorial or conjunctive events like homicides. This comparative method (called Qualitative Comparative Analysis) provides a systematic means of identifying the most prevalent situational contexts for homicide and the distinct combinations of attributes that underlie them. It also allows us to address the question of whether the nature of homicide situations is qualitatively unique for different social groups. The specific focus of this study involves a microhistorical analysis of U.S. homicides over the last three decades. Both national homicide data and narrative reports for select cities are analyzed to examine the following substantive questions: r What are the major situational contexts for U.S. homicides in the last three decades? Have homicide situations become more or less diverse over time? What dominant offender, victim, and offense characteristics are found among the most prevalent homicide situations in each decade? r Has the structure of homicide situations (i.e., combinations of offender, victim, and offense attributes) changed over the last three decades? What are the particular combinations of characteristics that underlie emergent forms of homicide in the 1990s, historically stable contexts, and those that have become relatively extinct over time? r What is the nature of subgroup variation in homicide situations? Are the situational contexts for homicide unique or similar for each of the following subgroups: (a) male vs. female offenders, (b) teenage vs. adult offenders, (c) White vs. Black vs. Hispanic offenders, (d) homicides within instrumental vs. expressive motives, and (e) killings involving strangers vs. acquaintances vs. family/intimate partners? Have the structural profiles of homicide situations for each group changed or remained stable over time? r Are major themes underlying homicide (e.g., character contests, status threats, masculine competitiveness, sexual possessiveness) diffuse across various situational contexts or are they more restricted to particular social groups and context-specific? Our findings about the nature of homicide situations challenge many of the basic assumptions in previous work. For example, we discover that homicide situations are representative of both simple and complex crime events, subgroups kill in both unique and common situational contexts, and the structures of homicide situations exhibit both change and stability over time. Our primary conclusion is that a more complete understanding of homicide involves conjunctive thinking and attention to the interaction between offenders and victims within particular situational contexts. We also

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

PREFACE

QC: FCH/FFX

T1: FCH November 11, 2003

xix

challenge public policy analysts and commentators to redirect their attention to the dominant situational contexts of homicide rather than focusing on the exceptional case. These dominant situations involve lethal disputes with guns that reflect the interplay of gender, race, class, and urban locations. It is hoped that the current research will provide an empirical and conceptual framework for future study of the process and structure underlying interpersonal violence.

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

Acknowledgments

This book is a collaborative project that has benefitted from the efforts of many people. The previous insights of several researchers have provided the theoretical and conceptual framework for our study of homicide situations. Others have given us the general moral and intellectual support that provides a context for thinking about new ideas and maintaining a sense of curiosity about social issues. Our interest in studying homicide situations as the unit of analysis has been motivated by the previous work of Gary LaFree and Chris Birkbeck on the “neglected situation” in criminological research. Jim Short has long recognized the importance of levels of explanation, and his seminal research on group processes and gang delinquency has provided a blueprint for our microsocial study of crime. Our conceptualization of homicide situations is similar to the “crime event” perspective developed by Bob Meier, Les Kennedy, and Vince Sacco. They call for theoretical integration and the wider study of process and structure, which are central themes in the current study. Ken Polk’s research on homicide has provided both a conceptual and analytic framework for integrating and interpreting narrative accounts of lethal violence. We are indebted to Charles Ragin’s work on comparative methods and his incredible insights on the link between theory and method, the value of conjunctive thinking, and bridging the traditions of qualitative and quantitative methods. This study is our attempt to follow in this tradition. This book could not have been completed without the contributions of several people. We especially appreciate the assistance of Rick Rosenfeld who provided us with homicide narratives that were collected as part of the St. Louis Homicide Project and Marc Riedel for providing us with the California homicide data. We would also like to acknowledge the assistance of the Los Angeles and Las Vegas Police Departments for access to their homicide narratives. At Cambridge University Press, we are indebted to xxi

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

xxii

QC: FCH/FFX

T1: FCH November 11, 2003

ACKNOWLEDGMENTS

the support and insights provided by Al Blumstein and David Farrington, the encouragement and competence of Alia Winters, and the copyediting provided by Lee Young. We thank all of you for your help. This book is dedicated to Kriss Drass. Kriss has been a major collaborator in our previous work and was involved in the initial formulation of this book at the time of his death. He developed the initial computer program for Qualitative Comparative Analysis (QCA), introduced us to the enormous potential of this method for studying crime events, and has been a major contributor to the growing application of QCA to large-N studies. More importantly, Kriss fundamentally altered our views about understanding and describing human behavior. This book’s focus on conjunctive thinking and combinatorial methods is a direct result of many daily conversations with Kriss. Finally, we acknowledge the continued assistance of family, friends and colleagues to this project. They have listened to us for years talking about homicide and have provided the moral and intellectual support that piques our curiosity and feeds our passion for research.

21:38

P1: FCH/FFX P2: FCH/FFX CB617-FM CB617-Miethe-v1

QC: FCH/FFX

T1: FCH November 11, 2003

Rethinking Homicide

xxiii

21:38

P1: FCH/FFX 0521832993c01

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 1, 2003

C H A P T E R

O N E

Introduction: Studying Homicide Situations

homicide has been studied from a variety of theoretical and methodological perspectives. Biological, psychological, and sociological theories have been widely used to explain the etiology and epidemiology of violence. Methodologically, homicide has been investigated through both qualitative and quantitative approaches, cross-sectional and longitudinal designs, and across individual and aggregate units of analysis. Despite all of this attention, one aspect of homicide still has not been studied systematically – homicide situations. By this we mean the quintessential convergence of offender, victim, and offense characteristics that define the situational context of homicide and that forms the basis for distinguishing homicides qualitatively. Several authors (e.g., LaFree and Birkbeck 1991; Kennedy and Forde 1999; Meier, Kennedy, and Sacco 2001; Miethe and Meier 1994) argue that the situational context of crime has largely been neglected as a topic of empirical research. However, recent developments in both theory (e.g., the emergence of a criminal event perspective) and method (e.g., the development of Qualitative Comparative Analysis) allow us to address this gap by developing an integrated approach to the study of homicide situations. We will use this approach in the current study to describe similarities and differences in the structure and process of homicide situations across groups and over time. A focus on the situational context of crime, and the application of Qualitative Comparative Analysis (QCA) as a means to study it, is particularly well-suited for criminal events involving violence. Specifically, our concern lies in the convergence of victim, offender, and offense elements that structure violent events. As Wilkinson and Fagan (2001) note, violent crimes are distinct from property crimes in the sense that they consist of interactions between at least two parties, which are frequently characterized by dynamic 1

13:34

P1: FCH/FFX 0521832993c01

2

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 1, 2003

RETHINKING HOMICIDE

exchanges of actions and words. In this respect, violent acts can be viewed as transactions, explainable from a situational approach. Most empirical research on homicide focuses on differences in level, measured either as a homicide rate or as a risk of victimization or offending. This is a reflection of the dominance of offender-based theories as the basic framework for the majority of studies on homicide. Consequently, we know a great deal today about differences in the social and spatial distribution of homicide, changes in homicide rates over time, and the characteristics of typical offenders and victims (see, for reviews, LaFree 1999; Reiss and Roth 1993; Short 1997; Smith and Zahn 1999). What has been neglected in this approach is an examination of why particular homicide situations are more common than others, whether these have changed over time, and the extent of subgroup variation in the situational context of homicide. As Birkbeck and LaFree (1993) argue, it is easier to link social phenomena such as subcultures and parenting styles to criminal propensity than to the situational context of crime. A neglected area within current research on violence involves the lack of a general and thorough description of the conjunction of offenders, victims, and situational elements that result in homicide. The reliance on a separate treatment of these essential aspects that come together to form a lethally violent act has produced a somewhat fractured body of literature that fails to develop a comprehensive understanding of the complete homicide event. Thus, among the key issues that remain unaddressed in the homicide literature is the level of concentration or diversity in the convergence of offender, victim, and situational elements resulting in lethal outcomes. In discussing the limitations of structural theories of violence, Luckenbill and Doyle (1989:422) note that “these theories focus on why certain people are more disposed to violence than others, but they do not specify the situational conditions that channel such dispositions into concrete lines of action.” Situational approaches and criminal event perspectives broaden this focus to include not only offender motivation but victim and situational characteristics as well. That Blacks are overrepresented as offenders as a result of experiencing higher levels of strain, living in areas characterized by greater social disorganization, or holding values conducive to violence does not inform us about whether and how they come to engage in lethal violence in some situational contexts more than in others or why these assaults are disproportionately committed against certain types of individuals. As another example, a well-noted finding in the homicide literature is the far greater involvement of men in homicide than women. While numerous biological, psychological, and sociological explanations have been offered for the higher propensity toward violence on the part of males, a situational

13:34

P1: FCH/FFX 0521832993c01

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

INTRODUCTION

T1: FCH November 1, 2003

3

approach is needed to explain why the context of offending varies across gender. In other words, the crux of the problem is not simply explaining why women commit homicide less frequently than men, but whether the situations in which women kill are qualitatively different from males. A thorough understanding of lethal violence cannot be achieved through an exploration of motivation in isolation from the other fundamental aspects of a criminal event. Individuals who harbor a propensity for violence do not act on these tendencies in all times or places, nor are they directed against a random selection of targets. In contrast to the standard approaches taken to studying homicide, our focus is on differences in type or kind of homicide. We take the perspective that the situational context of homicide can be examined from two separate but interrelated aspects: structure and process. The need to incorporate both of these elements in studies of crime and violence more generally has been asserted by others (see Meier, Kennedy, and Sacco 2001). The current research examines the issues of structure and process in homicide situations in the following ways. First, a comparative method (Qualitative Comparative Analysis) is used to identify the unique and common structural contexts of U.S. homicides over the last three decades. Second, qualitative analyses of homicide narratives from select U.S. cities (Los Angeles, Miami, St. Louis, Las Vegas) are conducted to explore the underlying processual elements in situations of lethal violence. By using these diverse methods on different types of homicide data, the current research is able to answer various questions about the situational context of homicide. For example, is female homicide qualitatively different from male homicide, as some claim? Have new forms of homicide emerged over time? Have some forms of homicide become less prevalent, or even disappeared, over time? An examination of case accounts allows us to address the question of whether the processes underlying incidents of lethal violence vary across different types of homicide. This combination of analytic techniques and emphasis on both structure and process locates the current study within the larger context of examining conditions that facilitate the occurrence of violent acts (see also Meier, Kennedy, and Sacco 2001). We believe that the answers generated by this research complement the existing literature and contribute to a more complete understanding of the complexity of homicide as a social phenomenon. In the remainder of this chapter, we review existing approaches to studying homicide and their limitations, discussing both case studies and statistical analyses. Through the application of the method of QCA, the current research is designed to unite these disparate methodological traditions to better understand qualitative changes in the nature of homicide trends and subgroup variation in homicide risks.

13:34

P1: FCH/FFX 0521832993c01

4

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 1, 2003

RETHINKING HOMICIDE

Histories of Violent People and Dangerous Times Case studies, both biographical and historical, offer rich descriptions of the personal legacies of violent offenders and their acts. All of the most notorious murderers in American history have been immortalized through these life histories and case studies (see Bugliosi 1974; Frank 1967; Nash 1973; Rule 1980; Schwartz 1981; Sullivan and Maiken 1983). Narrative accounts of homicide cover the historical landscape from Brutus’ killing of Caesar, Lizzie Borden’s axe murder of her parents, the shoot-outs by notorious Western outlaws (e.g., Billy the Kid, Jesse James, Cole Younger), lynchings by vigilante groups like the KKK in the post–Civil War period, the robbery and murder sprees of early gangsters (e.g., John Dillinger, Bonnie Parker and Clyde Barrow, “Pretty Boy” Floyd), the mass genocide during Nazi Germany, and the multiple slayings by Jeffrey Dahmer, David Berkowitz (“Son of Sam”), Ted Bundy, and other serial murderers. Biographical accounts of individual offenders are useful for understanding homicide because they often provide rich details of the potential motive, precipitating events, and offense elements underlying these crimes. In other words, they underscore the need to view situations holistically, rather than as a simple sum of individual elements. Although case studies are a mainstay in homicide research, substantive conclusions that derive from them are limited in several respects. First, the findings from case studies are difficult to generalize because they are limited to small samples. Second, these descriptive accounts are often unrepresentative of homicides in a particular historical period because sensational, bizarre, and idiosyncratic killings elicit the most attention. Random acts of excessive brutality by serial killers and strangers are the focus of most case studies simply because of their novelty and severity. Domestic violence is widely regarded as a major situational context for homicide, but the social and legal tolerance of violence against wives and offspring throughout modern Western history ultimately downplays its media attention (see Goetting 1995; Straus, Gelles, and Steinmetz 1980). Third, case studies and personal narratives usually lack the comparative control data that are necessary for determining the relative importance of particular offender, victim, and situational elements in the general production of homicide. Definitive statements about the most dangerous eras in Western history are also often problematic due to the limitations of archival data and diverse conceptions of violence. Nonetheless, most scholars contend that the medieval era was an extremely violent time and that homicide and other types of interpersonal violence have declined sharply throughout Western history over the last six centuries (see Cockburn 1977; Gurr 1989; Lane 1997; Short 1997). Archival court records indicate that lethal brawls and violent deaths by robbers were a common occurrence in medieval England. Murder rates

13:34

P1: FCH/FFX 0521832993c01

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

INTRODUCTION

T1: FCH November 1, 2003

5

in rural areas in this period ranged from 10 to 25 per 100,000, far exceeding the rate of roughly 2.0 per 100,000 for England and Wales in the late twentieth century (Barclay, Tavares, and Siddique 2001; Given 1977; Gurr 1979). In the context of U.S. history, Brown (1979) asserts that, since its early beginnings, violence has been a major fact of American life. His argument is supported by the observation that many of the most influential events in U.S. history (e.g., the American Revolution, the Civil War, the stabilization of the Western frontier, land and labor reform, the struggles for racial equality) involved serious and prolonged acts of violence (Brown 1979:41). Violent crime in the United States is thought to have decreased substantially during the last quarter of the nineteenth and up to the mid-twentieth century, only to have increased dramatically during the last quarter of the twentieth century (Sickmund, Snyder, and Poe-Yamagata 1997). However, the 1990s witnessed a considerable drop in violent crimes known to police, but the number has increased again in the new century (Blumstein and Wallman 2000; FBI 2002). The level of interpersonal violence in particular historical periods in the United States is subject to alternative views. Narrative descriptions and archival records of the “Wild” Western frontier in the mid to late 1800s, for example, often yield contradictory conclusions about the prevalence of violence (see McKanna 1997). Some authors conclude that “as a place of wild lawlessness the frontier’s reputation is largely without substantiation” (Prassel 1972), while other narrative accounts (e.g., Drago 1970; Rosa 1969) convey a far more severe pattern of violence during this time period. Estimates of homicide rates for particular locations range from 4 per 100,000 in Alameda County, California, in 1893 (Friedman and Percival 1981), 8.9 per 100,000 in Caldwell, Kansas, for the period 1879–85 (McKanna 1997), 116 per 100,000 in the “boom and bust” gold town of Aurora in 1877 to 1882 (McGrath 1984), a rate of 160 per 100,000 in Dodge City in 1878, and an incredible 422 per 100,000 in Ellsworth, Kansas, in 1873 (McKanna 1997). Compared to a national rate of about 9 per 100,000 in the 1990s, these estimates of homicide rates for some towns on the Western frontier are simply staggering. Most studies of violent times and places, however, suffer from several major methodological problems that limit the reliability and validity of their substantive conclusions. First, the accuracy of data sources is a serious problem, both in terms of the completeness of the enumerations and the meaning attached to them. Even if a complete census of violent acts were possible, the very definition of what is considered a homicide or violent act is historically ambiguous and inconsistent. For example, the counting of acts like infanticide, abortion, capital punishment, dueling and mutual combat, and self-defense slayings as homicides has varied greatly over time and across different types of historical records (see Riedel 1999). Second, a growing

13:34

P1: FCH/FFX 0521832993c01

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

6

November 1, 2003

RETHINKING HOMICIDE

literature now questions the accuracy of current crime reports and media accounts, claiming that such data are socially constructed for assorted purposes (see Best 1999; Goode and Ben-Yehuda 1994; Jackson and Rudman 1993; McCorkle and Miethe 2001; Zatz 1987). By either inflating or deflating the apparent magnitude of violence in their jurisdiction, officials and organizations have been able to demonstrate accountability, enhance resource mobilization, and increase their relative position among competing interest groups.

Macro-Quantitative Approaches Quantitative studies of homicide patterns and trends have dramatically increased over the last century. Multivariate analytic techniques have been used to examine homicide rates and their correlates across a wide range of aggregate units in both cross-sectional and longitudinal studies. Using national vital statistics on deaths and police data on known offenses, temporal changes in homicide rates within particular countries have been well documented. Simple graphic representations and more sophisticated time series techniques have been widely used to model changes in national homicide rates over time and their correlates (see Brantingham and Brantingham 1984; Chen 1996; Cohen and Felson 1979; Gurr 1989; Holinger and Klemen 1982; LaFree and Drass 1996; Silverman and Kennedy 1993; Zahn and McCall 1999). An extensive body of research has also accumulated on change and stability in homicide rates for particular cities over time (see Block and Block 1980; Block and Christakos 1995; Boudouris 1970; Chilton 1987, 1996; Dobrin, Wiersema, Loftin, and McDowall 1996; Lane 1979; Monkkonen 2001; Wilbanks 1984; Zahn and Jamieson 1996). These types of statistical studies contribute to our approach by helping us to identify variables that describe important structural characteristics of homicides. However, these studies generally focus on the main effects of individual variables across situations rather than on the interactive effects of combinations of variables within situations. Therefore, they tend to ignore the importance of context and do not adequately capture the complexity of homicide situations. Previous quantitative research on homicide rates has essentially been studies of the level of violence. In other words, these studies are concerned with whether the level or amount of homicide has increased, decreased, or remained stable over time. The absence of comprehensive longitudinal data prior to the twentieth century and the limitations of the available time series (e.g., UCR data, National Mortality Files) have contributed to the ongoing debate about the magnitude of these level differences in homicide over time (see Riedel 1999; Zahn and McCall 1999). An equally problematic

13:34

P1: FCH/FFX 0521832993c01

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

INTRODUCTION

November 1, 2003

7

issue, however, concerns the lack of attention given to changes in the situational context of offending over time. A fundamental question that needs to be addressed is whether lethal violence in more concentrated in particular situations during some time period(s) than others. Furthermore, while rates of male homicide offending may remain stable over a particular period of time, for example, this does not preclude the possibility that the situational context of lethal violence by males has undergone considerable change. Whether the types of situations in which men kill have become more diversified or concentrated cannot be deciphered by simply tracking their homicide rates over time. Moreover, policies suited to addressing one of these patterns may not be appropriate for others. Without an understanding of the full range of patterns of change and stability the most efficient use of funding for prevention initiatives is unlikely to be achieved.

Subgroup Differences in Homicide Risks and Rates Analyses of level differences also characterize homicide research comparing subgroups in terms of their relative risks of offending and victimization. These subgroups are typically defined on the basis of socio-demographic attributes of the offender or victim (e.g., males vs. females, Whites vs. Blacks, teenagers vs. adults) or the elements of the offense (e.g., instrumental vs. expressive motivations, stranger vs. acquaintance vs. family violence, firearm vs. nongun killings, single vs. multiple offenders). Previous research reveals wide variability in homicide rates and individuals’ risks across different subgroups. According to UCR arrest data in the United States for the 1990s, homicide offenders are disproportionately male, young, and African American. About 90% of homicide offenders known to the police are males. African Americans are clearly overrepresented among known homicide offenders, accounting for more than half of homicide arrestees in the 1990s. More than half of homicide arrestees are under 25 years old (Federal Bureau of Investigation [FBI] 1998). They also tend to be poor and live in economically disadvantaged areas (see Harries 1997; Martinez and Lee 1999; Short 1997; Wolfgang 1958). Homicide victims often share many of the same sociodemographic characteristics as their offenders (see Brearley 1932; Sampson and Lauritsen 1990, 1994; Singer 1981; Wolfgang 1958). Level differences in offense elements involve making comparisons across categories for variables measuring particular characteristics and activities surrounding the criminal act. Accordingly, offense elements include the motivation for the crime, interpersonal dynamics (e.g., the victim–offender relationship, intraracial vs. interracial crimes), the type of weapon used, the number of offenders, the presence of alcohol and drugs, and the spatial and temporal location of the offense.

13:34

P1: FCH/FFX 0521832993c01

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

8

November 1, 2003

RETHINKING HOMICIDE

Based on our analysis of the FBI’s Supplementary Homicide Reports (SHR) for the 1990s, homicides involve a higher proportion or level of the following attributes: arguments and other expressive motives (53%), acquaintance slayings (51%), intraracial crimes (88%), firearms (70%), and single offenders (87%). Other data sources suggest that a majority of homicides involve the presence of drugs or alcohol in the victim, offender, or both parties (see Auerhahn and Parker 1999; Parker 1995; Parker and Rebhun 1995; Wolfgang 1958). The disproportionate occurrence of homicides and other violent crimes in evening hours and on weekends has been established in both earlier and contemporary research (see Miethe and McCorkle 2001; National Centers for Health Statistics [NCHS] 1993; Perkins and Klaus 1996; Wolfgang 1958). Higher levels of homicide have also been observed in urban areas than rural areas and in western and southern states (FBI 1998). With the exception of lower rates of gun homicides observed in other countries, the U.S. pattern of offense characteristics is similar to that found in England, Canada, and other Western countries (see Brantingham and Brantingham 1984; Silverman and Kennedy 1993).

Homicide Situations as the Unit of Analysis Of the range of possible units of analysis for homicide research, the situational context has been relatively neglected. This is not to say that previous research has ignored situational elements of homicide like weapon use, number of victims or offenders, victim precipitation, alcohol and drug use, and the time and location of the crime. Instead, what has been missing is the treatment of the situation itself as the unit of analysis. Such an approach requires that the homicide incident be treated holistically, as a complete composite of offender, victim, offense, temporal, and spatial elements. Previous studies have focused on particular idiosyncratic elements rather than the total combination of elements that define the structure of homicide situations. The situational context of crime has been defined in several ways in past research. One approach treats the homicide situation as the “microenvironment” for crime, involving offender, victim, and physical and spatial elements of the crime (see Davidson 1989; Miethe and Meier 1994). Another perspective considers the situational context as a criminal event or transaction, consisting of beginning, middle, and end stages (Luckenbill 1977; Meier, Kennedy, and Sacco 2001; Sacco and Kennedy 2002). Other researchers view the situation in terms of specific offense elements and, in some cases, limited combinations of crime elements (see Block and Christakos 1995; Block and Block 1992; LaFree and Birkbeck 1991; Miethe and McCorkle 2001; Reiss and Roth 1993). The primary unit of analysis in the current study is the structure of the homicide situation. This structure is defined by combinations of offender, victim,

13:34

P1: FCH/FFX 0521832993c01

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

INTRODUCTION

T1: FCH November 1, 2003

9

and offense elements that underlie homicides. Fundamental offender and victim characteristics in homicide include their gender, race, and age. Basic offense elements of homicide include the motive or circumstance surrounding the crime, the victim–offender relationship, the number of co-offenders, the type of weapon used, and the physical context of the crime. These characteristics define homicide’s structure because they pattern the nature of interpersonal dynamics that is likely to take place in deadly situations. The total combination of these offender, victim, and offense attributes considered simultaneously is what defines the diversity in the structure of homicide situations. For example, if the major offender, victim, and offense elements in a study involve 10 dichotomous variables, there are 1,024 unique combinations of attributes that define the diversity of structures in homicide situations. Adding another dichotomous variable would double the number of homicide situations to 2,048. These combinations of attributes are the unit of analysis for the current study. This conceptualization of the homicide situation is similar to the image of a criminal event or microenvironment that locates potential victims and offenders in a particular situational context. As an illustration of the structures of homicide situations, consider the following narrative descriptions of homicides in Miami in 1980 (Wilbanks 1984): 49-year-old, Latin male victim was killed by his 45-year-old, Latin wife during a domestic argument/fight. The victim was bitten, stabbed, and shot. (case #046) Two Latin male victims (age 23 and 24) were killed in a robbery by three unknown Latin males. The two victims were in a bar when the offenders entered and ordered patrons against the wall. When one patron resisted, shots were fired and the two victims were killed. Three other patrons were also shot but not killed. (case #113) 40-year-old, Anglo female victim and her 37-year-old, Anglo husband had been legally separated after longstanding marital difficulties. They met on a Friday night in a bar in New York City, got drunk together, and decided to fly to Miami for a weekend trip. Upon arrival in Miami, the couple got in a loud argument in their hotel room. Witnesses heard the victim pleading for her husband not to hit her anymore. The victim was found strangled to death the next morning by the maid. (case #148) 17-year-old, Black male victim and the 18-year-old, Black male offender were acquaintances who became involved in an argument over a stolen bicycle. The offender shot the victim with a revolver. The victim had an alcohol level of .11. (case #141) The 34-year-old, Black male victim was shot to death at his own home in a dispute over money owed for drugs. The victim was shot several times. The offender was a 30-year-old, Black male. (case #211)

13:34

P1: FCH/FFX 0521832993c01

10

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 1, 2003

RETHINKING HOMICIDE

Although these homicide incidents share some characteristics (e.g., all of them are intraracial killings, some are domestic assaults), they are all qualitatively unique in terms of their particular structures. This is the case because each of these five homicide situations involve distinct combinations of offender, victim, and offense attributes. For example, the four incidents that involve disputes and arguments as the motive differ from each other in terms of the race of the parties, the weapon used, and the location of the crime. As a group, these dispute situations share many of the same features as the incident of robbery-homicide (case #113), but the robbery-homicide situation is qualitatively different in terms of its occurrence in a bar and the involvement of multiple victims and multiple offenders. It is the systematic investigation of these common and unique structures underlying different types of homicide situations that is the basis for our study of homicide. Compared to other units of analysis in homicide research, an examination of the structure of homicide situations is important for several reasons. First, this situational approach allows us to assess the extent and nature of diversity in types of homicide. For example, we can identify the number of qualitatively distinct structures for any particular time period. Comparing the number and particular nature of these homicide situations over time permits clear inferences about both change and stability in the basic structure of homicide situations. Second, comparisons can be made within and between subgroups (e.g., males vs. females, teenage vs. adult offenders, acquaintance vs. stranger killings) to determine whether some types of homicide have a more complex and diverse structure than others. Analyses of these subgroup configurations over time can also be conducted to identify historically extinct, emergent, and stable patterns of homicides across time. Traditional macro-quantitative analyses and case studies of homicide are not designed to deal with these types of research questions.

Qualitative Comparative Analysis and Homicide Situations An examination of the structure of homicide situations requires the use of an analytic procedure that allows for the investigation of complex interrelationships among sets of variables. It also requires an analytic approach that is case-driven (i.e., focusing on the totality of attributes) rather than variabledriven (i.e., focusing on the effect across contexts). Standard multivariate statistical procedures that estimate main effects or a limited number of interactive effects are not able to handle the complexity of such an analysis. Over the past decade, the method of Qualitative Comparative Analysis (QCA) has been used increasingly for comparative studies in the social sciences (Amenta and Halfmann 2000; Amenta, Carruthers, and Zylan 1992; Amenta and Poulsen 1994; Coverdill, Finlay, and Martin 1994; Drass and Ragin 1989; Miethe and Drass 1999; Ragin 1987, 2000). The QCA method

13:34

P1: FCH/FFX 0521832993c01

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

INTRODUCTION

November 1, 2003

11

allows the logic of intensive, case-oriented research to be applied to studies involving substantial sample sizes (Ragin 2000). Instead of estimating to what degree a set of independent variables “accounts for” or “contributes to” variation in a particular dependent variable (for the problems with such an approach, see Becker 1998; Ragin 1987), QCA permits an examination of the configurations or combinations of attributes leading to that outcome variable. Variation in QCA refers to the diversity of unique conditions, measured case by case using the different combinations of variables within cases, that produce the outcome. QCA assumes that outcomes are produced by variables acting together, where the effect of any particular variable may be different from one case to another, depending on the values of the other attributes of a case. Hence, context and causal heterogeneity are given preeminent consideration in QCA. When applied to the study of homicide, QCA is an ideal approach because it allows us to model the diversity of common and unique combinations of attributes underlying homicide situations. By comparing these structures over time and across groups, QCA provides a logical and systematic method for the analysis of qualitative change in homicide situations.

Structural Elements Underlying Homicide Situations There are an enormous number of possible offender, victim, and offense characteristics that can form the structural basis of homicide situations. Offenders and victims are often classified according to their demographic features (i.e., sex, race, age, education, income, marital status), but measures of their physical composition (e.g., height, weight, hair color, dexterity, physical attractiveness, testosterone level, head size, skull abnormalities, and other physical stigmata) are also potentially important risk factors for various types of conventional and deviant behavior. In fact, some of the earliest criminological research (from Cesare Lombroso, Charles Goring, and Ernest Hooton to William Shelden) focused on similar biological and physical differences between criminals and noncriminals (see, for review, Vold, Bernard, and Snipes 1998). The earliest research on developing victim typologies also emphasized their social and physical features (e.g., “the wanton,” “the feeble-minded,” “the sickly,” “the helpless or defenseless”) as the primary basis for classification categories (see Fattah 1991; Von Hentig 1948). Of the possible offender attributes relevant to a situational analysis of homicide, those that are most widely investigated in previous empirical and theoretical studies are the offender’s sex, race, age, and income. This focus reflects the substantial differences in offending risks based on these particular characteristics. Homicide offenders in the United States are about nine times more likely to be male than female, and African Americans,

13:34

P1: FCH/FFX 0521832993c01

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

12

November 1, 2003

RETHINKING HOMICIDE

persons under 25 years old, and the poor are vastly overrepresented as offenders compared with their respective population distributions (see FBI 2001; Smith and Zahn 1999). There are also sound theoretical reasons for treating these four demographic characteristics as major structural elements of homicide situations. Specifically, each of these status attributes is connected to particular role expectations, behavioral opportunities, and life experiences that structure people’s exposure to risky situations and pattern their anticipated responses in these settings. Males in most Western cultures, for example, are often faced with qualitatively different life experiences and behavioral expectations than females, which ultimately shape their own actions and the reactions of others. Similarly, lower educational and economic opportunities, decreasing confidence in social institutions, a historical legacy of oppression, and greater levels of frustration and despair characterize the life experiences of many African Americans and subsequently structure their perceived and actual responses to anger-provoking situations (see Anderson 1994; LaFree 1998; Rose and McClain 1990; Wilson 1987). It is within this context that basic status characteristics of individuals become indicative of fundamental structural features of homicide situations. The same arguments are relevant for the inclusion of the victim’s sex, age, race, and income as major structural aspects involving victim characteristics. A wide assortment of offense characteristics have been identified as important features of violent criminal acts. These offense attributes include the motive for the crime, the victim–offender relationship, the type of weapon used, evidence of victim precipitation, the number of offenders and victims, the presence of alcohol or drugs, the number of bystanders, the time of the crime, and the properties of the physical location (e.g., home, street). Similar to particular offender and victim attributes, these offense circumstances are also considered structural features of homicide situations because they enable and constrain (i.e., structure) the likelihood of particular types of behavior within a situational context.

Data Sources for Studying Homicide A wide variety of types of data have been used in previous homicide research to study crime trends and their situational elements. The primary data sources include newspaper accounts, court documents, mortality statistics, and police reports. Court records and newspaper accounts have been the major source for deriving homicide patterns in historical analyses (see Brearley 1932; Dykstra 1968; Lane 1997; McKanna 1997; Redfield 1880). Police reports have become the most common data source for studying homicide in the last half of the twentieth century.

13:34

P1: FCH/FFX 0521832993c01

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

INTRODUCTION

T1: FCH November 1, 2003

13

Two general types of homicide data are used in the current study. Questions about the structure of homicide situations are examined primarily through the analysis of the U.S. Supplementary Homicide Reports (SHR), and state- and city-level police data (e.g., California and Chicago homicide files). Narrative accounts of homicides from several U.S. cities (Los Angeles, Miami, St. Louis, Las Vegas) are used to explore and describe the episodal nature and interpersonal dynamics underlying homicide transactions. The integration of both types of data is designed to further our understanding of both the structure and process underlying homicide situations. These data are used to investigate questions about both change and stability in the structure of homicide situations over time.

Outline of the Current Study The current study examines the situational context of homicide across subgroups and over time. Similarly to previous research, we conceptualize the homicide situation as a microenvironment that involves combinations of offender, victim, and offense elements. Changes in homicide situations and subgroup differences in them are examined through the analysis of narrative accounts and SHR data over the last three decades. The SHR data are highlighted as the most comprehensive data on homicide because of their national scope and inclusion of measures of most of the basic elements that underlie homicide situations. Both SHR data and narrative accounts of homicides in Los Angeles, Las Vegas, Miami, and St. Louis are identified as the major data sources to be used in the current study. Compared to previous studies of homicide, our situational approach is distinct in several respects. First, the homicide situation itself is subject to analysis by examining change and stability in these situations over historical periods, their relative prevalence, and the personal and offense characteristics that underlie them. Differences in homicide situations are empirically evaluated according to variation in offense elements (e.g., type of weapon, motive, number of offenders, location of the crime), victim characteristics, and offender attributes (e.g., age, gender, race). Second, we treat the homicide situation holistically as a configuration of offender, victim, and offense attributes rather than as a set of isolated and disparate elements. Third, the method of QCA is used to assess both similarities and differences in the structure of homicide situations across different subgroups and over time periods. The remaining chapters of this book are organized in the following way. Chapter 2 provides an overview of theoretical perspectives on the analysis of situations and illustrates how particular macro- and micro-level theories of crime may account for change and stability in the structure of homicide

13:34

P1: FCH/FFX 0521832993c01

14

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 1, 2003

RETHINKING HOMICIDE

situations over time. Data sources and analytic approaches for studying homicide situations are presented in Chapter 3, including a description of QCA as an alternative to the standard main-effects statistical model for analyzing homicide situations. Chapter 4 summarizes our current knowledge about the empirical distribution of homicide over time, across countries, and the nature of the offender, victim, and offense characteristics associated with them. This chapter also applies QCA to the SHR data to identify empirically the nature and diversity of the dominant structures of homicide situations in the United States over the last three decades. The next five chapters examine subgroup variations in homicide situations, including differences by the offense motive (Chapter 5), offender’s gender (Chapter 6), age (Chapter 7), race (Chapter 8), and prior relationship between the offender and victim (Chapter 9). The final chapter summarizes these results and discusses their implications for criminological theory and public policy on homicide.

13:34

P1: JMT 0521832993c02

CB617-Miethe-v1

November 6, 2003

C H A P T E R

T W O

Theoretical Development

theories of homicide and other criminal behavior are directed at three levels of explanation.1 Individual-level explanations focus on the characteristics of individuals that influence their differential risks of homicide offending or victimization. Macro-level theories explain differential rates of homicide across geographical units such as countries, states, cities, and neighborhoods. Most situational-level theories examine the context of homicide incidents, focusing on circumstances of the act and its spatial and temporal location. While an incredible volume of research exists on homicide correlates and its theoretical explanation,2 we propose that there are two related voids in this research tradition. First, there is wide ambiguity in both the definition of the homicide situation and what constitutes situational analysis. Second, most previous research on homicide is not integrative either across levels of explanation or in terms of its focus on all three basic elements of crime situations (i.e., offender, victim, and offense characteristics). These limitations, in turn, have prevented previous studies from fully documenting the empirical nature of homicide and the adequacies of existing macroand micro-level theories in accounting for change and stability in homicide situations over time and across subgroups. After reviewing previous applications of the “situation” in criminological research, this chapter provides the theoretical framework for examining homicide situations in the current study. We demonstrate how basic structural features of homicide situations can be identified and measured 1 2

The importance of level of explanation has been clearly discussed in previous research (see Messner and Rosenfeld 2001; Sampson and Lauritsen 1993; Short 1985, 1998). For the most recent reviews of the literature on homicide, see Martinez (2002) and Smith and Zahn (1999).

15

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

16

November 6, 2003

RETHINKING HOMICIDE

empirically. Several macro- and micro-level theories are used to illustrate how particular homicide situations may have changed their basic structural features over time. Subsequent chapters describe data sources and analytic procedures to study these homicide situations and evaluate how well change and stability in the nature of particular types of homicide situations are accounted for by current macro- and micro-level theories of crime.

Definitions of the Situation One of the central concepts in sociology and psychology is the situation. It is a central concern in the theoretical writings of Talcott Parsons, George Herbert Mead, and Erving Goffman. The situation is also the primary link between the fundamental disciplines underlying social psychology and continues to be a major source of controversy in traditional personality theory and experimental psychology. Although his major writings focus on dispositional theories (i.e., criminality), Edwin Sutherland in the 1940s was one of the first criminologists to recognize the importance of situational explanations for crime’s occurrence (Sutherland 1947). Situational-level explanations, theories, and variables are now becoming a mainstay in modern criminological studies (see Birkbeck and LaFree 1993; Miethe and Meier 1994; Sampson and Lauritsen 1993; Short 1997). Given the salience of the situation across disciplinary traditions, it is surprising that the concept lacks an uniform definition. For example, some ecological psychologists assert that a situation always includes an organism, a place, and action; different situations are defined by who is involved, where the action is taking place, and what type of activity is occurring (Pervin 1978). Changes in any of these components result in a change of the situation. The term “behavioral settings” is similar to this objective view of situations, treating them as having defined boundaries and physical properties that are associated with ongoing patterns of extra-individual behavior (see Barker 1965; 1968; Gump 1971). “Situations” for social ecologists and geographers also emphasize the physical properties of space and are defined by the language of microenvironments, locations, niches, and physical settings (see Davidson 1989; Hawley 1950). Defining situational factors as external causes, the person–situation controversy and its concern about the consistency of personality and behavior across situations have been a source of major debate within general psychology (see Furnham and Argyle 1981; Mischel 1968). It is within the field of symbolic interactionism, however, that the subjective meaning attached to action and social conditions becomes paramount in the definition of the situation (see, for review, Blumer 1969; Cooley 1964 [1922]; Goffman 1959; Lofland 1969; Mead 1934; Shibutani 1961). From this perspective, all human action takes place in situations, and differences in

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

THEORETICAL DEVELOPMENT

November 6, 2003

17

actors’ subjective interpretation of objective situations (and the actor’s particular goals, needs, and experiences) are why the same social conditions do not result in identical behavior. While social situations have objective properties (e.g., they involve individuals and are defined by time and space), for interactionists criminal behavior is produced largely by how the actor receives, interprets, and transforms previous and newly acquired meanings of that situation. According to Birkbeck and LaFree (1993:120), symbolic interaction theory contributes to the situational analysis of crime “by focusing attention on the meaning of situations for actors, by linking offender motivation and situational opportunities, and by conceptualizing crime and deviance as situationally precarious outcomes.” Consistent with the diverse definitions of the situation across disciplinary traditions, the meaning of situational analysis within criminological research is also equally varied. However, three general approaches to the situation have been used in previous homicide studies. First, some situational analyses of homicide examine the objective physical and temporal characteristics of dangerous places or environments. Second, other situational approaches focus attention on the homicide motive, circumstances, or syndrome as the dominant situational elements and develop subclassifications of homicides based on them. Third, other studies are more comprehensive and integrative in their approach by incorporating offender, victim, and offense elements in the definition and analysis of the homicide situation. The particular foci of these three types of situational analysis are described below.

Physical Properties of Dangerous Times and Places One of the oldest types of situational approaches in criminological research involves the identification of dangerous physical environments for criminal acts. Geographers and the early Chicago ecologists in the first half of the twentieth century initiated this tradition by mapping natural areas for crime (see Shaw and McKay 1942).3 More recent studies using the language of hot spots, microenvironments, behavioral settings, and domains follow the same tradition (see Davidson 1989; Lynch 1987; Sherman et al. 1989). Consistent with social disorganization theory, a large number of studies have found that high crime areas are characterized by higher rates of population turnover, ethnic and racial diversity, socioeconomic deprivation, 3

Ecological approaches are most often associated with macro-level analysis, but we think their treatment within a situational or micro-level framework is defensible for two reasons. First, the unit of explanation is often a crime rate which is derived from the aggregation of crime acts within a particular physical space (e.g., census tract, neighborhood, city). Second, the focus of ecological research is to identify the physical and social properties of these natural areas for criminal acts. These same characteristics are also the focal point of most situational analyses of crime.

14:10

P1: JMT 0521832993c02

18

CB617-Miethe-v1

November 6, 2003

RETHINKING HOMICIDE

and disruptive family relations (see Miethe, Hughes, and McDowall 1991; Sampson and Groves 1989; Sampson and Lauritsen 1993). Studies of communities over time reveal patterns of both stability and change in the criminal careers of neighborhoods (see Bursik and Webb 1982; Schuerman and Kobrin 1986). Structural changes in neighborhoods over time that are reflective of higher social disorganization (e.g., decreasing income, higher population mobility, increasing ethnic diversity) are also associated with increasing crime rates over time. Physical signs of low income and incivilities (e.g., abandoned buildings, multiunit dwellings, garbage and trash on the street) are indicative of lower levels of informal social control in the neighborhood, a commonly identified correlate of violent crime (see Skogan and Maxfield 1981; Taylor 1997). Research on the physical features of dangerous places has increasingly examined the behavioral setting or domains for crime (see Davidson 1989; Lynch 1987). From this perspective, the most common behavioral settings or domains include (1) the home environment, (2) school environment, (3) work environment, and (4) leisure environment. It is often assumed that crimes committed in these different domains are qualitatively distinct; have different frequencies of occurrence, different modi operandi, and different motivations; and are explainable by different predisposing and precipitating factors (Miethe and Meier 1994). For example, the home environment is a common location for violent acts, and crimes in these settings are more often expressive acts of anger by intimates and family members. Data on police calls for service are the common basis for identifying “hot spots” for crime (see Sherman et al. 1989). The nature of these frequent places of criminal activity varies by type of crime. Shopping malls, for example, may be considered prime locations for shoplifting and pickpocketing (because of the high density of expensive and portable property available in these settings), but they are not common environments for violent crime. Parking lots surrounding shopping malls, however, are sometimes the location for personal assaults and robberies. Bars, adult bookstores, and shelters for the homeless are more frequent locations for violence (see Hindelang, Gottfredson, and Garofalo 1978; Roncek 1981; Roncek and Maier 1991; Sherman et al. 1989). The specific physical characteristics that underlie dangerous “hot spots” involve basic properties of the physical space that enhance the likelihood of crime’s occurrence. For example, the amount and nature of physical space becomes important in accounting for the onset of crime because closed and confined spaces may (1) impede the ability of potential victims to flee from an attack, (2) decrease criminal motivations by limiting escape routes after crime commission, or (3) increase impulsive or spontaneous acts of violence because crowding in this particular public place may fuel tension, anxiety, and frustration (see Miethe and Meier 1994:67). By providing anonymity

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

THEORETICAL DEVELOPMENT

November 6, 2003

19

and cover for predatory acts, street lighting and patterns of traffic flow (both automobile and foot traffic) also represent physical attributes of crime “hot spots.” Seasonality effects and the dangerousness of public nighttime activity have also been identified as situational correlates in previous work (see FBI 2001; Miethe and Meier 1994). The higher violence in summer months is often attributed to factors such as greater public activity away from home and increased visibility and accessibility of persons during this season. Higher nighttime risks are said to be related to decreased guardianship and increased participation in dangerous activities (e.g., drug and alcohol use). As an approach for studying homicide situations, however, the focus on the physical properties of dangerous places, in and of itself, is limited in several related ways. First, it ignores other situational elements (e.g., victim– offender relationship, victim’s age and gender, presence of firearms) that may dramatically alter the objective dangerousness of any particular situation.4 Second, as illustrated in “domain” studies, the exclusive reliance upon physical properties to define situations often requires gross inferences about the underlying offender, victim, and offense attributes to explain the dangerousness of the particular physical environment. Rather than inferring these characteristics, a better strategy is to incorporate them directly into the measurement of the situational context. Third, as recognized within the symbolic interactionism tradition, this objective approach to situations makes the dubious assumption of identical behavioral responses to social conditions. While similar responses to structural conditions may occur because of shared life experiences and similar typifications of the situation, interactionists assert that the subjective assessment of any particular physical environment as conducive to violent crime is precarious and restrictive in its extraindividual generalizability.5

Circumstances, Motives, and Homicide “Syndromes” Another type of situational approach in homicide research involves the identification of subgroups or subtypes of homicide and the subsequent 4

5

For example, the danger of physical locations with particular objective conditions (e.g., dark with poor lighting, restrictive paths for fleeing by potential victims) may be easily neutralized by the presence of numerous bystanders or self-protection of potential victims through the possession of weapons. Situational approaches that incorporate elements of the offense and the characteristics of offenders and victims are discussed shortly as integrative approaches. It is interesting to note in this context, however, that there is some evidence of agreement in interpretations of situations conducive to crime. For example, interviews of burglars about target-selection processes often reveal similar assessments of attractive crime conditions. These risk factors for burglars include indicators of signs of occupancy, ease of access, and perceived attractiveness (see Miethe and Meier 1994).

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

20

November 6, 2003

RETHINKING HOMICIDE

analysis of the differences between them. These subgroups may derive from offender attributes (e.g., male vs. female killers, youth vs. adult offenders), victim characteristics (e.g., child vs. adult victims, single vs. multiple victim attacks), or offense circumstances (e.g., stranger vs. family vs. acquaintance slayings, gun vs. knife killings). Regardless of the particular subgroup that is investigated, the fundamental assumption of this approach is that these subgroups are qualitatively different from each other in their particular structural features. Consequently, situational analysis in this context involves the description and identification of the physical and social properties of the criminal act that give particular types of homicide their unique signatures. One of the earliest applications of this approach to situations is Marvin Wolfgang’s notion of victim-precipitated homicide. As a particular type of homicide, victim precipitation refers to situations in which the victim is “the first in the homicide drama to use physical force directed at his subsequent slayer” (Wolfgang 1958:252). Slightly over one-fourth of Philadelphia homicides from 1948 to 1952 were classified as victim-precipitated. Compared to other homicide situations, victim-precipitated homicides have higher proportions of each of the following characteristics: Black victims, Black offenders, male victims, female offenders, stabbings, victim– offender relationships involving male victims of female offenders, mate slayings, husbands who are victims of mate slayings, alcohol in the homicide situation, alcohol in the victim, victims with a previous arrest record, and victims with a previous arrest record of assault (Wolfgang 1958:264– 265). While operationally defined as the initiation of physical provocation by the victim, it is the combination of the definitional attributes and its empirical correlates that produce the uniqueness of these particular types of homicide situations.6 Of the types of subgroups examined in previous homicide research, most attention has been devoted to the study of differences in the motive and circumstances surrounding the homicide act. The language of motive and circumstances is what often underlies our conceptions of homicide situations and provides unique typifications of them. Such categories of motive and circumstances as domestic disputes, gang shootings, sexual assaults, and “thrill kills” provide immediate scripts and stereotypes to observers about the likely who, what, where, and why involved in the homicide situation. However, similar to studies of dangerous environments, previous research on homicide motives and circumstances has often failed to fully address 6

Wolfgang (1958:265) also indicates that victim-precipitated homicides tend to have higher proportions of each of the following characteristics: altercations and domestic quarrels, interracial slayings, and victims with primary relations to their slayers (i.e., close friends, relatives, acquaintances). However, when all of these offender, victim, and offense attributes are considered simultaneously, it should be clear that the subcategory “victim-precipitated” homicide occurs in an extremely diverse array of situational contexts. The importance of this observation and how we measure the level of diversity will be discussed later.

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

THEORETICAL DEVELOPMENT

November 6, 2003

21

the prevalence and nature of other crime elements (i.e., offender, victim, and offense attributes) that in combination define these particular types of homicide situations. One primary subclassification of motive in homicide situations distinguishes between instrumental and expressive acts. Instrumental crimes are those conducted for explicit, future goals (such as to acquire money or improve one’s social position), whereas expressive offenses are often unplanned acts of anger, rage, or frustration (see Block 1976; Block and Christakos 1995; Block and Zimring 1973; Decker 1993, 1996; Miethe and Drass 1999). In previous homicide research, killings that occur during the commission of another felony (e.g., robbery, rape, burglary) have usually been classified as instrumental motives because the violence is the presumed means to achieve the primary goal (e.g., money, sexual conquest). Arguments, brawls, romantic triangles, youth gang killings, and other interpersonal disputes are typically classified as expressive acts because their dominant motivation is the violence itself (see Block and Christakos 1995; Decker 1996; Polk 1994). However, situations of gang activity and drug activity are somewhat ambiguous in this expressive–instrumental typology because they may involve both types of circumstances (e.g., street gang violence may result from instrumental, entrepreneurial activities like drug trafficking or these violent situations may just involve impulsive, emotional, expressive outbursts).7 An alternative conception of motive and circumstances involves Kenneth Polk’s classification of violent situations in terms of general scenarios. According to Polk (1994:188), a theme of masculine competitiveness underlies four specific scenarios of lethal violence by men. First, some violent situations involve efforts to control the behavior of sexual partners, where perceived threats to a presumed proprietary right underlie violence directed at partners and their sexual competition. Second, “confrontational” homicides refer to situations of spontaneous exchanges between males over largely trivial matters that reflect some type of character or honor contest. Both parties in this situation are relatively willing participants in the escalation of violence. Third parties play a central role in these homicides either as a supportive social audience or by becoming directly involved in them. Third, masculine competitiveness is the basis of lethal violence that arises in the course of other criminal acts. The essential feature of these felony homicide situations is the “exceptional marginality, which leads some males to take extraordinary risks with life” (Polk 1994:190). Fourth, lethal violence is used in some circumstances as a means of conflict resolution, especially when 7

Alternative views have been expressed about the classification of motives. For example, Richard Felson (1993) argues that this distinction between instrumental and expressive motives is of limited value because even expressive acts of violence done in anger reflect an instrumental reaction to perceived wrongdoing.

14:10

P1: JMT 0521832993c02

22

CB617-Miethe-v1

November 6, 2003

RETHINKING HOMICIDE

other remedies are considered inadequate or unavailable. Polk considers these types of homicide scenarios to be distinctly masculine and primarily a working- or under-class phenomenon. Erving Goffman’s thoughts about “character contests” are also reflected in the situational approaches used by other researchers (see Athens 1985; Felson and Steadman 1983; Hepburn 1973; Horowitz and Schwartz 1974; Luckenbill 1977). Character contests in homicide involve a process of escalating confrontations between parties who are attempting to establish or save “face” at the other’s expense. While there is some disagreement on whether these confrontations involve mutual consent across stages of the violent transaction (see Athens 1985), the value of the notion of character contests for our purposes is that it focuses attention on the interplay between perceptions and actions of both parties in defining and generating homicide situations. Another approach to the identification of subtypes of homicide situations involves the notion of homicide syndromes (see Block and Christakos 1995). This classification scheme derives primarily from the combination of the offender’s motive (falling on an expressive–instrumental continuum), the victim–offender relationship, and “sibling offenses” (i.e., similar incidents that were not lethal). Examples of syndromes include spousal confrontations, child abuse by caretakers, expressive confrontations by acquaintances, instrumental crimes, and street gang violence. Chicago homicide data have been used to examine trends in these syndromes over time and some of the offender and victim characteristics associated with them (see Block and Christakos 1995). However, detailed, multiple comparisons of the differences in offender, victim, and offense attributes within and between these situational representations have not been conducted. Finally, a different orientation toward the classification of homicide situations by motive is illustrated by Jack Katz’s (1988) research on the “seductions of crime.” The central assumption in his work is that crime events cannot be explained without a close examination of the subjective meaning of crime for the offender. These meanings, in turn, derive from the actor’s subjective experience with criminal situations and the recognition of their seductive qualities. Beyond the obvious instrumental motives for crime, particular seductive features of criminal acts may include the “sneaky thrills” that property offenses like vandalism and shoplifting provide to some adolescents and the way robbery may enhance an offender’s self-image as a “hardman” (see Birkbeck and LaFree 1993). It is within the context of the “righteous slaughter,” the typical homicide identified by Katz (1988), that the actor’s interpretations of the seductive qualities of the homicide act are manifested. He describes these homicides “as an impassioned attempt to perform a sacrifice to embody one or another version of the ‘Good’” (Katz 1988:12). While many aspects of Katz’s

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

THEORETICAL DEVELOPMENT

November 6, 2003

23

work have been widely criticized (see Birkbeck and LaFree 1993; Polk 1994), his study nonetheless illustrates that the motives and goals of criminal offenders are more complex than typically viewed in past studies of homicide situations.

Integrated or Combinatorial Approaches Alternative conceptualizations of the situation directly incorporate offender, victim, and offense characteristics into their analytic scheme. These approaches make explicit their focus on the criminal act as the unit of analysis and recognize that the absence of any of these three basic elements is sufficient to thwart them. Major contributors to these integrated or combinatorial approaches to the study of homicide situations are outlined below.

Lofland’s Theory of Closure on Deviant Acts Drawing upon symbolic interactionism, John Lofland (1969) explores the situational cues and the interpretive processes that underlie an individual’s decision to commit deviant acts. From this perspective, actors have the capacity to commit a wide range of both conventional and deviant acts, but the conditions of threat and then encapsulation enhance their likelihood of commission of a particular defensive deviant act especially when accompanied by the presence of facilitating places, facilitating hardware, and facilitating others.8 True to symbolic interactionism, however, Lofland (1969) maintains that it is the actor’s subjective experiences and interpretations of these necessary conditions that ultimately determine the likelihood of closure on the deviant act. According to Lofland (1969:42), the initial step in the production of a defensive deviant act involves an object or action that is perceived by the actor as a clear and present threat. These threats may involve basic defenses of the physical body (as evident in victim-precipitated homicides) and types of social defacement or disgrace (e.g., threats to masculine identity). Under conditions of perceived threat and the reduction of possible alternative actions, the presence of facilitating places, hardware, and others is said to increase the likelihood of a deviant act like homicide. 8

Lofland (1969:40) makes a distinction between defensive and adventurous deviant acts. Defensive deviant acts are those involving a sense of threat that becomes a key experience around which actors organize their actions, whereas the key subjective experience in adventurous deviant acts is a sense of positive excitement and the perception of enchanting possibilities. Lofland (1969:39) also asserts that different phases in the production of deviance are involved among novice actors (whose deviant acts are relatively unpracticed and isolated) and among repeat or seasoned offenders. From our perspective, the most salient feature of this theory is the identification of places, hardware, and others as the major facilitators that underlie deviant acts.

14:10

P1: JMT 0521832993c02

24

CB617-Miethe-v1

November 6, 2003

RETHINKING HOMICIDE

While deviant acts can be committed in all social settings, it is also the case that some types of places are more facilitative of deviance than others. Lofland (1969:63) considers facilitating places to be bounded physical sites that provide protection from others “who might have an interest in interfering with the performance of the act.” Examples of facilitative, protective places include those with physical barriers (like walls), limited public access and traffic flow (like some alleys and streets in urban areas), and places at nighttime and with dim lighting. From this perspective, the home is considered a premier location for deviant acts because it is bounded by walls, legal rules concerning privacy, and interdependencies among the inhabitants that reduce the public visibility and surveillance of deviant activity within this setting. Other physical sites are attractive locations for deviant acts because they are accessible and render easier pathways for escape. Contrary to strict objective conceptions of dangerous places, Lofland’s approach emphasizes the actor’s subjective interpretation of these social conditions that ultimately determines the appraisal of situations as dangerous or facilitative of violence. Lofland (1969:69) defines hardware as “reasonably transportable nonhuman physical objects” necessary for the commission of a deviant act. Everyday household objects like knives, bottles, hammers, and bricks are widely accessible and available instruments whose mere presence facilitates the commission of a violent act. Also, these objects do not require specialized knowledge for their use as lethal instruments. Firearms are facilitative hardware for homicide in that they are widely available, do not require physical strength differentials or even direct bodily contact with the victim, and are relatively more likely to result in lethal injuries than attacks with other objects. Homicides by means of beatings or strangulations may involve the use of physical objects (e.g., baseball bats, rope, scarves), but external hardware is not required for the commission of these types of homicides. Given the wide availability of various types of facilitating hardware in everyday life, it would seem rather futile to attempt to reduce the likelihood of homicide through the restriction of particular hardware. In fact, under the adage “where there’s a will, there’s a way,” potential offenders display versatility and may easily substitute methods for any available hardware. However, the particular properties of firearms as hardware (e.g., they are the “great equalizer,” not requiring physical prowess or direct victim contact) make these physical objects less attractive for substitution. Under these conditions, restrictive access and supply of firearms may have meaningful crime-reduction benefits in homicide through decreasing the availability of a particularly unique type of facilitative hardware. If other methods are used as substitutes, reducing the supply of firearms may reduce only the lethality, but not the prevalence, of violence. The notion of facilitating others in Lofland’s scheme applies to situations in which the actions of others provide the course of least resistance

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

THEORETICAL DEVELOPMENT

November 6, 2003

25

for engaging in deviant conduct or offer encouragement for the offender to commit this act. Facilitative others who are victims are those who leave doors unlocked, keys in the ignition of cars, and walk alone at night in risky and dangerous situations. The actions of these persons themselves provide an opportunity structure for illegal activity. Persons who offer direct encouragement or tacit approval for deviance are often considered “significant others” (e.g., family members, friends), but even the action or inaction of strangers and bystanders may be perceived by offenders as fostering deviant acts. For example, inaction by bystanders may be interpreted as indifference, providing offenders with less resistance if they were to carry out a deviant act. The role of third parties in encouraging or discouraging deviance has also been recognized in other studies using a similar theoretical perspective (see Felson 1981; Felson and Tedeschi 1993). According to Lofland’s theory, facilitating places, hardware, and others are necessary conditions for closure on deviant acts. In addition, there also must be some level of motivation for action (caused by threat and encapsulation) and a modicum of knowledge and skill to commit the deviant act. The absence of any of these conditions is sufficient to thwart the selection of deviance from a range of alternative actions. Under conditions of severe threat and encapsulation, individuals may be especially prone to interpreting particular places, hardware, and others as conducive to crime commission. The fact that defensive deviant acts occur when people are experiencing severe threat may help explain why these offenses sometimes occur in situations that outside observers would not generally perceive as facilitative for their successful completion (e.g., in open public places, in the presence of numerous witnesses). Nonetheless, the value of Lofland’s (1969) theory is that it directly identifies three basic situational factors that underlie deviant acts and emphasizes the actor’s interpretation of the salience of these factors in selecting deviant acts from a range of alternative courses of action.

Criminal Opportunity Theories Criminal opportunity theories are explanations of crime events that cover both micro- and macro-levels of analysis. Lawrence Cohen and Marcus Felson’s (1979) “routine activity theory” was originally used to explain temporal changes in U.S. crime rates since World War II. A variety of empirical tests of this theory by the original authors and others, however, involve predictions about individuals’ risks of victimization from a property or violent crime (see Cohen and Cantor 1980, 1981; Cohen, Kluegel, and Land 1981; Miethe, Stafford, and Long 1987; Sampson and Wooldredge 1987). Michael Hindelang and associates’ (1978) “lifestyle-exposure theory” also focuses on the role of routine activities and lifestyles in explaining individuals’ risks of victimization. Terance Miethe and Robert Meier’s (1990) “structural-choice

14:10

P1: JMT 0521832993c02

26

CB617-Miethe-v1

November 6, 2003

RETHINKING HOMICIDE

model” and Michael Hough’s (1987) “model of target selection” derive from the same theoretical basis but place greater emphasis on offender motivation and target-selection processes. Each of these approaches falls within an opportunity framework because they are based on the premise that some situations provide more opportunities conducive to crime than others. From a routine activity perspective, macrostructural changes in routine activity patterns provide an opportunity structure for crime by affecting the convergence in time and space of three minimal elements of direct-contact predatory crimes: (1) motivated offenders, (2) suitable targets, and (3) the absence of capable guardianship. As necessary conditions, the absence of any of these elements is sufficient to deter criminal acts, in a manner similar to the inability to reach closure on deviant acts under Lofland’s (1969) theory without facilitating places, hardware, and others. At a macro-level of analysis, Cohen and Felson (1979) assert that large increases in crime rates could occur without any increase in the structural conditions that motivate offenders as long as there have been increases in the supply of attractive and unguarded targets for victimization. It is within this context that changes in the size and resale value of consumer goods, living arrangements, work and leisure activities, and the level of nighttime activity outside of the home create opportunities for criminal acts. Ironically, this perspective assumes that illegal activities feed upon the routine activities of everyday life. The basic premise underlying the lifestyle-exposure theory is that demographic differences in the likelihood of victimization are attributed to differences in lifestyles (see Hindelang et al. 1978). Variations in lifestyles are important because they are ultimately related to differential exposure to dangerous places, times, and others. In other words, particular lifestyles are directly linked to situations in which there are higher risks of victimization. According to the lifestyle-exposure theory, both ascribed and achieved status characteristics (e.g., age, gender, race, income, marital status, education) are major correlates of victimization risks because they carry with them shared expectations about appropriate behavior and structural obstacles that both enable and constrain one’s behavioral choices (Miethe and Meier 1994). For example, males are more prone than women to homicide victimization because they are traditionally socialized to be active in the public domain, are assertive and aggressive in social situations, have fewer restrictions on their daily life, and typically spend more time away from a protective home environment. These same factors associated with male socialization and weakened attachments to social institutions also explain their higher risk of homicide offending. Different lifestyles influence the nature of homicide situations by altering the likelihood that potential victims will be “in particular places, at particular times, under particular

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

THEORETICAL DEVELOPMENT

November 6, 2003

27

circumstances, and interacting with particular types of people” (see Miethe and Meier 1994:35). The structural-choice model of victimization is designed to integrate the routine activity and lifestyle-exposures theories. Miethe and Meier (1990:245) employ in their integrative model two assumptions derived from the base theories. First, routine activity patterns and lifestyles are assumed to create a criminal opportunity structure by enhancing one’s proximity to motivated offenders and exposure to risky situations. Second, the subjective value of a person or object and its level of guardianship determine the choice of the particular crime target. Under this revised theoretical approach, proximity and exposure are considered structural features (because they pattern the nature of social interaction and predispose individuals to riskier situations), whereas attractiveness and guardianship represent the choice component (because they determine the selection of the particular crime target within a sociospatial context). Hough (1987) has developed a similar model of target-selection processes that also rests on the assumption of a reasoning criminal. Under this formulation, exposure to motivated offenders, potential yield to the offender, and the accessibility of the target are the key factors that influence individuals’ risks of selection as a victim. The relative importance of particular target-selection factors and situational cues, however, differs by the type of offender (e.g., novices vs. seasoned offenders) and across different types of criminal acts (see Cornish and Clarke 1986).

The Crime Event Perspective Another integrated approach to the study of homicide situations involves the “crime event” perspective developed by Leslie Kennedy and Vincent Sacco (see Kennedy and Sacco 1996; Meier, Kennedy, and Sacco 2001; Sacco and Kennedy 1994, 2002). Similar to other integrative approaches, a basic assumption of this perspective is that crime events cannot be separated from the physical and social settings in which they occur. As social events, crimes are said to be intricately linked to the routine activities of subsequent victims and offenders, the places in which these activities occur, the behavior of witnesses and bystanders, and the particular circumstances of the situation. Like many other aspects of life, criminal events “are made more or less likely by the choices people make about how and where they spend their time, energy, and money” (Sacco and Kennedy 1994:116). This conceptualization of crimes as events also means that these situations have a temporal and episodic quality. Specifically, crime events have a beginning and end. Sacco and Kennedy (1994:137) assert that there are three stages of crime events: (1) the precursors of the event (i.e., the locational and situational factors that bring people together in time and space), (2) the

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

28

November 6, 2003

RETHINKING HOMICIDE

transactions (i.e., the interactions among participants that define the outcomes of their actions), and (3) the aftermath of the event (e.g., the report to the police, their response, the harm done and redress required). While consideration of the aftermath of the event is obviously important from the perspective of victims and the criminal justice system, the conceptions of precursors and transactions in this theoretical approach are the most similar to the treatment of “situations” in previous criminological research. However, it is the focus on the union of the participants and the physical setting that unites the crime event perspective with other situational approaches.

The Current Approach to the Study of Homicide Situations Drawing from previous research, our approach to the study of homicide situations assumes that these crime events are complex social phenomena involving an interplay between places, people, time, and action. These situations have structural properties that promote and restrict certain actions and provide a forum in which shared expectations of particular behavior are interpreted as appropriate or ambiguous. We further assume that a comprehensive understanding of homicide situations requires a holistic treatment of them as compound integrations of offender, victim, and offense circumstances. These combinations of the basic elements of crime are what define the underlying structures of homicide situations.9 Homicide situations are considered to have unique signatures when the particular combination of attributes that underlie them is observed among one subtype of homicide (e.g., spouse slayings) but not among another (e.g., stranger killings).10 To illustrate our operational definitions of homicide situations, subtypes, structures, and unique signatures, consider the following four narrative accounts of homicides in Miami, Florida:11 9

10

11

Our use of the word “structure” to define these combinations of offender, victim, and offense characteristics requires some clarification. Specifically, we assume that particular offender, victim, and offense characteristics in combination “structure” the nature of the homicide transaction. Particular role expectations and behavioral patterns are often associated with particular ascribed and achieved status characteristics (e.g., age, gender, race, income). Particular typifications and shared images are also connected with particular offense circumstances (e.g., guns elicit fear or defensive responses, alcohol use implies greater situational impulsivity, bystanders provide a feeling of external protection). It is within this context of enabling and constraining behavioral patterns and the nature of the interactions among parties that we use the term “structures” to define these combinations of personal characteristics and offense circumstances. The notion of unique signatures also applies to cases in which particular combinations of person and offense attributes are observed for a subtype of homicide at one point in time but not at some other time period. Within this context, we also refer to these homicide situations as emergent forms. These narratives are reported by Wilbanks (1984) and were derived from police reports of homicide in Miami in 1981. These particular narratives were selected for illustrative purposes only and are not intended to be representative of homicides in this city. The wording of

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

THEORETICAL DEVELOPMENT

November 6, 2003

29

Case #1: 25-year-old, Anglo male victim was killed by a 20-year-old, Anglo male offender. The offender came to the victim’s residence to resolve a drug dispute. The victim let him in, they argued, and the offender shot the victim and another Anglo male in the house. The offender was the former roommate of the two victims and they were all involved in drug trafficking. (Miami, case #298) Case #2: 25-year-old, Anglo male victim and the 19-year-old, Anglo male offender were patrons in a bar and an argument ensued. They went outside of the bar. The offender killed the victim by hitting him in the face with a large rock. (Miami, case #366) Case #3: 24-year-old, Black male victim and the 21-year-old, Black male offender were drinking at a bar when they became involved in an argument over a dice game. They went outside the bar and continued arguing. The victim pushed the offender twice and the offender then pulled out a gun and shot the victim once in the chest. The offender then walked up to the victim lying on the ground and fired another shot into the victim’s head and fled. (Miami, case #041) Case #4: 26-year-old, Black male victim and the 25-year-old, Black male offender were sitting on bleachers at a public park when they got into an argument over money from the sale of drugs. A fight ensued. The victim attempted to flee but the offender beat him with a pipe and then stabbed him several times.(Miami, case #248) From our perspective on the study of homicide, these lethal accounts are representative of four different homicide situations and numerous subtypes. Different subtypes of homicide situations in these narratives include intraracial slayings, Black homicides, Anglo homicides, male-on-male killings, young adult killers and victims, dispute-related violence, acquaintance killings, single victims, multiple-victim slayings, public location killings, gun homicides, direct-contact homicides, and combinations of these and other characteristics. The structure of these four homicide situations involves the particular combinations of offender, victim, and offense circumstances that underlie them. For example, all of these situations involve lethal arguments between young adult males of the same race who were acquaintances. Structural elements that are not shared across all situations, however, include the location of the killing (i.e., home in #1 vs. a public setting in #2, #3, and #4), the lethal weapon (i.e., gun, rock, gun, pipe/knife, respectively), the number of victims (e.g., multiple in #1 vs. single in the others), alcohol use (present these narratives has been changed somewhat from the original source to increase their readability.

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

30

November 6, 2003

RETHINKING HOMICIDE

in #2 and #3 vs. absent in the others), and drug sales (present in #1 and #4 vs. absent in the others). The unique signatures of homicide situations involve combinations of structural elements that are observed among one subtype of homicide but not in others. For example, the unique signature of drug disputes among Anglo males (case #1) compared to other subtypes of homicide situations is their location in the home and the killing of multiple victims. This type of drug dispute is not unique in any other way because other particular attributes that define its structure (i.e., male, gun use, nonalcohol use) are found in at least one of the other types of homicide situations. A similar logic would be used to examine the qualitatively unique signatures of other subtypes of homicide situations (e.g., gun vs. nongun slayings). Under our conceptual framework, situational analysis becomes an investigation of the unique and common signatures of different subsets of homicide. It also involves an assessment of change and stability in the nature of their particular structures over time. For example, this approach allows us to ask whether confrontational homicides are qualitatively distinct from instrumental homicides in their underlying offender, victim, and offense circumstances. Also, are some subtypes of homicide situations (e.g., the killing of intimates) more homogenous than other subtypes (e.g., stranger slayings) in their particular structures and have these patterns changed over time? Have homicides, in general, become more or less diverse in their structures over time? If so, what are the unique features of the emergent forms of homicide situations? Similar questions often form the basis for previous qualitative and quantitative studies of homicide (see, for review, Short 1997; Smith and Zahn 1999), but the homicide situation itself is rarely investigated in past work in terms of the structures of its composite properties. In short, our analysis of homicide situations is guided by two basic assumptions. First, we assume that homicide situations are diverse and complex, involving a potentially infinite number of offender, victim, and offense attributes that define their structures. Second, the diversity of these structures of homicide situations across subtypes and over time can be quantified through the coding of particular attributes associated with the criminal act.

Macro- and Micro-Theories of Change in Homicide Situations Given our focus on homicide situations and changes in them over time, there are several questions involving how various macro- and micro-level theories of crime can be used to predict change and stability in homicide situations over time. In other words, what changes in criminogenic social conditions and individuals’ behavioral patterns and lifestyles that are identified as risk factors in existing theories have occurred over time that may change the

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

THEORETICAL DEVELOPMENT

November 6, 2003

31

structure of particular types of homicide situations? Several criminological theories that would suggest changes in the structures of homicide situations are outlined below.

Crime and the American Dream Robert Merton’s theory of anomie is one of the most fundamental theories in the history of criminology. It attributes the high rate of crime and deviance in the United States to the discrepancy between societal goals, the institutionalized means of attaining them, and how the social structure distributes access to the legitimate means. From this perspective, high crime rates are intrinsic to the basic cultural commitments and institutional arrangements of American society. Using anomie-strain theory as the primary conceptual framework, Steven Messner and Richard Rosenfeld (2001) propose in their book “Crime and the American Dream” that high crimes rates in the United States are due to an unrestrained pursuit of monetary success and a deemphasis on the use of normative means of achieving this goal. These authors contend that the consequence of this lack of articulation is a “dark side” of the American dream, characterized by extreme and persistent economic inequality and the adoption of economic norms and the devaluation of noneconomic functions in other social institutions. Given their differential access to educational and occupational opportunities, particular social groups are often assumed to be more prone to illegal adaptations or innovations as a means of achieving the material goals of the American dream. These strains are especially onerous among the “truly disadvantaged” (Wilson 1987). Young, urban, African-American males are sent the same cultural message and experience the same focus on economic standards in other institutions, but their relatively low educational attainment and high unemployment places them at an enormous disadvantage in their pursuit of the primary economic goal.12 When applied to the nature of homicide situations, several empirical predictions are derivative from this theoretical orientation. First, the most prevalent homicide situations should involve young, urban, African-American male offenders because this group’s life experiences make them (a) more prone to engage in instrumental acts of lethal violence (e.g., burglary-, robbery-homicides) to achieve economic goals and (b) more susceptible to displaced forms of aggression due to frustrations from the failure to achieve 12

An important extension of Merton’s theory by Messner and Rosenfeld (2001) is their contention that increasing economic opportunities for the disadvantaged will not reduce their crime rates because these groups will remain disadvantaged whenever a social system has a “competitive allocation of monetary rewards [that] requires winners and losers” (p. 9).

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

32

November 6, 2003

RETHINKING HOMICIDE

these goals. Second, improvements in the economic wellbeing and life experiences of this particular social group in the 1990s are external social conditions that should decrease the prevalence and nature of their involvement in instrumental homicide situations. The accuracy of these specific predictions is evaluated later in our analysis of racial differences in homicide situations (see Chapter 8).

Declines in the Legitimacy of Social Institutions Gary LaFree’s (1998) book “Losing Legitimacy” explores traditional theories of crime and their ability to explain changes in street crime rates in the United States in the post–World War II time period. Three major post-war periods and crime trends are identified in his work: (1) an early period (1946–60) with stable, low crime rates, (2) a middle period (1961–73) with rapidly accelerating crime rates, and (3) a late period (after 1973) with stable, high crime rates. Data on street crime also indicate that African Americans, and in particular young, male African Americans, were disproportionately involved in producing these observed crime patterns (LaFree 1998:5). When current theories are applied to these crime trends, they are found to be deficient in their explanatory power. For example, LaFree (1998) argues that biological and psychological theories based on slowly evolving biological characteristics or deep-seated psychological traits are unable to explain the rapid changes in street crime rates over a short time period (e.g., the doubling of crime rates from 1960 to 1970). Merton’s theory of anomie is also limited in its ability to explain period-specific changes in crime rates because it assumes that strain is a constant and permanent feature of American society. Other economic theories are deemed deficient because of the paradox of “crime amidst plenty” in which crime rates were booming at a time of flourishing economic opportunities (see Wilson 1975). Furthermore, cultural deviance theories are unable to explain why the amount of criminal learning varies over time and how crime rates among young, male African Americans increased rapidly and then stabilized if there is a persistent “underclass” within this community. LaFree (1998) proposes that the best theoretical explanation for the changing rates of street crime involves the notion of declining institutional legitimacy. From his perspective, political, economic, and family institutions regulate crime by reducing individual’s criminal motivations, enhancing both informal and formal social control mechanisms to deter criminal behavior, and providing individuals with protection against the criminal acts of others. Social changes such as growing distrust in political institutions, collective actions and protests, economic inequality and inflation, and economic and ideological challenges in the American family are serious threats

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

THEORETICAL DEVELOPMENT

November 6, 2003

33

to institutional legitimacy and therefore thought to underlie changes in crime rates over time (LaFree 1998). From our perspective on the situational analysis of homicide, LaFree’s theory is important because it helps identify external forces that may alter the nature of homicide situations over time. For example, compared to social conditions in the 1970s and 1980s, the greater threats to the legitimacy of the family institution in the 1990s (as indicated by increases in divorce, single-parent families, and female labor force participation outside the home) are likely to result in multiple changes in the nature of homicide situations through their impact on increasing public activity and personal contact with others who are nonfamily members. These changes may include an increase in homicide situations involving nonfamily slayings by female offenders and a rise in stranger assaults across a wide range of persons and situations. The theory of declining legitimacy in social institutions also predicts a relatively higher frequency of homicide among young, male, African-American offenders over time and across a variety of situational contexts.

Changes in Routine Activities and Lifestyles Routine activity and lifestyle theories (Cohen and Felson 1979; Hindelang, Gottredson, and Garofalo 1978) focus on how changes in the routine activities of everyday life enhance criminal opportunities by increasing the convergence in time and space of motivated offenders, suitable targets, and the absence of capable guardians for protection. Specific changes over the last three decades in contemporary American society that may alter the nature of homicide situations include changes in leisure and work activities, youth behavioral patterns, residency patterns and interracial relations, weapon possession and distribution in urban areas, and basic production activities. Changes in work and leisure activities have created a wider set of structural circumstances that may be conducive to lethal acts of violence. Both work and leisure activities have consistently moved away from primary household activity into public domains, requiring greater contact with a wider array of people, places, and situations on a daily basis.13 The criminogenic risks associated with these routine activities include increased physical proximity to a wider pool of motivated offenders, increased exposure to risky and dangerous situations, increased attractiveness of the person as a visible and accessible target for victimization, and decreased guardianship or protection by other family members. 13

Time-utilization studies support this characterization of greater public activity in the last decade, especially among women (see Robinson, Andreyenkov, and Patrushev 1988).

14:10

P1: JMT 0521832993c02

34

CB617-Miethe-v1

November 6, 2003

RETHINKING HOMICIDE

From our perspective, the consequence of this greater public activity is an increased likelihood of homicides occurring in a wider diversity of situations. The balance of work and family activities may make young persons in urban areas especially prone to homicide situations that involve diverse structural elements. Older persons, on the other hand, may have more restrictive daily activity patterns and contact with particular others and settings, decreasing the diversity of structural elements involved in their homicide situations. In short, these arguments suggest that young and older persons kill and are killed under some types of situations that are qualitatively different. Several authors over the last decade have noted potential changes in the composition of youth offenders that are reflective of different criminal motivations and changing behavioral patterns. In fact, the most recent generation of juvenile offenders has been called “superpredators” who are radically impulsive and kill indiscriminately for various instrumental and expressive purposes (Bennett, DiIulio, and Walters 1996). Compared to their counterparts in earlier decades who have been described as homogeneous pools of victims and offenders (see Sampson and Lauritsen 1990; Singer 1981), this image of the superpredator suggests that youth homicide is becoming increasingly diverse in terms of its particular combination of offender, victim, and offense characteristics. This presumed increase in the diversity in youth violence in the 1990s is often assumed to include a rise in intergroup assaults, involving participants who vary in their age group, gender, race, and interpersonal relations. Unfortunately, the relative frequency of changes in these particular combinations of individual and offense attributes that underlie youth homicide situations has not been empirically documented. While many historical legacies of racial separation in the United States are slowly eroding, racial segregation still widely exists in housing patterns and interpersonal relations. From a routine activity or lifestyle perspective, increases in inter-racial residence patterns, marriage, and friendship networks are the same social forces that should influence the prevalence of interracial killings and alter the nature of interracial and intraracial homicide situations (see O’Brien 1987; Blau and Blau 1982). Another major change in routine activities involves the concentration and distribution of guns. Nearly one-half of American households have a least one firearm, and gun ownership varies by gender, urban and rural residence, and region of the country. While sport usage of guns is most common among older White males from small towns and rural areas, the use of guns in criminal acts is concentrated among young urban males, especially minorities (Cook and Moore 1999). Private handgun ownership has increased over the last three decades and these particular firearms are now, by far, the most common weapon in homicides (see FBI 2000; Kleck 1991). While various changes in gun attitudes and behavior have occurred over time, there are three major changes in basic activities surrounding guns that

14:10

P1: JMT 0521832993c02

CB617-Miethe-v1

THEORETICAL DEVELOPMENT

November 6, 2003

35

are especially important from our perspective on homicide situations. First, there is a growing circulation of both legal and illegal handguns in U.S. urban areas. Second, handgun ownership for either self-defense or offensive purposes has grown increasingly diverse over the last three decades, involving a wide range of demographic groups (e.g., men and women, young and old, Black and White). Third, recent changes in concealed weapon laws (i.e., from “may” to “shall” issue standards) increase the likelihood of persons carrying handguns in a public place. Under these conditions, the proportion of homicides involving handguns should be increasing over time, and these types of homicide situations in urban areas should be characterized by an increasingly diverse array of person and offense attributes that define their structures. Finally, changes in manufacturing and production activities have also affected the likelihood and nature of various kinds of predatory acts. For example, the emergence of lethal carjacking has been produced in a large part by changes in manufacturing technology (e.g., car alarms, tracking systems) which makes it more difficult to steal an unoccupied vehicle. The dramatic rise in ATM machines in open public places makes it easier for consumers to conduct financial transactions, but it also increases their visibility, decreases their protection, and subsequently increases their attractiveness as potential robbery victims. Both an increase in the prevalence of such robbery–homicide situations and changes in their structural features may be produced by these changes in production activities and modifications in the types of persons who are the victims and offenders.

Summary This chapter has covered various types of theoretical developments in the study of homicide situations. We reviewed various conceptualizations of the “situation” and developed our theoretical orientation to the study of homicide situations. We proposed that homicide situations are complex phenomena whose structure is defined by particular combinations of offender, victim, and offense circumstances. Several current macrotheories and microtheories of crime are used to make particular predictions about possible changes in the structure of homicide situations over time. Many of the basic themes outlined here will be further refined and developed in subsequent chapters. The important question of how we study homicide situations empirically and identify their common and unique structures is addressed in the next chapter.

14:10

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

C H A P T E R

October 15, 2003

T H R E E

Data and Methodology for Studying Homicide Situations

the current chapter examines the various data sources and different analytic approaches for studying homicide situations. We review previously used data sources for studying homicide and describe the data for our analysis of homicide situations. The examination of methodological approaches for studying homicide begins with the “main effect” statistical model and is followed by a discussion of the “interactive effect” model specification. Techniques of spatial analysis for developing homicide typologies and traditional qualitative methods are also discussed. After highlighting the conceptual and analytical problems with these traditional approaches, we present Qualitative Comparative Analysis (QCA) as an alternative approach for studying homicide situations. The discussion of QCA includes an overview of the logic of this approach, its previous applications, and the methodological issues underlying its extension to the study of homicide situations in large samples. The chapter concludes with a brief application of QCA to the study of gender differences in Chicago homicides to illustrate the construction of truth tables, the use of probabilistic rules for defining unique and common profiles, and the interpretation of QCA results.

Data Sources for Studying Homicide Various data sources have been used in previous studies of homicide to describe crime trends and their situational elements. These sources include newspaper accounts, court documents, mortality statistics, and police reports. While court records and newspaper accounts are the major source of information on homicide patterns in early Western history, police reports and national mortality statistics are the most typical choices for studying homicide trends in the twentieth century. Police reports are often better 36

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

DATA AND METHODOLOGY FOR STUDYING HOMICIDE SITUATIONS

October 15, 2003

37

suited for situational analysis than other sources because of their coverage of all types of homicide situations and their more extensive enumeration of offender, victim, and offense circumstances. Compared to other crimes, homicide is often thought to be the easiest to trace over time because it is “always taken seriously, almost always subject to law, never common enough to be completely tolerated or overlooked, and the legal disposition of murder cases usually is a matter of record” (Lane 1997:5). Unfortunately, the absence of comprehensive and uniform data on murder and its characteristics across time periods is a basic problem in all historical and comparative analyses. Depending upon the particular historical period, homicide accounts may be collected and disseminated in a variety of forms (see Monkkonen 2001). While national and international homicide data have increased dramatically in both comprehensiveness and specificity over the last century (see Brantingham and Brantingham 1984; LaFree 1999; Riedel 1999), numerous problems with the quality of the collected information continue to plague even recent data sources. The primary data sources for homicide research are described below.

Newspaper Accounts When other sources are unavailable or considered of dubious value, historical analysts have relied upon newspaper articles for a chronology of crime events and the frequency and pattern of homicides. Newspaper accounts have been used by several authors to construct estimates of homicides rates and trends. Redfield (1880) used killings listed in statewide newspapers in Kentucky, South Carolina, and Texas to reach conclusions about the high rates of Southern violence. Dykstra (1968) relied entirely upon newspaper accounts to study homicide in five cattle towns in Kansas in the 1870s. Wilbanks (1984) examined crime stories in the Miami Herald in the mid 1920s to document characteristics of offenders and victims (e.g., age, sex, race) and the circumstances of the homicide. Other authors (see Lane 1997; McKanna 1997) have used newspaper accounts to supplement official reports of homicide. Newspaper reporting of homicide has had a long history. The life and times of the Wild West outlaws and gunfighters were chronicled, if only with a modicum of accuracy, in the dime novel and penny press (see Inciardi, Block, and Hallowell 1977). Sensationalized accounts of violence became a common theme in the American press in the early twentieth century. Brearley (1932) reports that one third of the total space in the New York tabloid Daily News in a month in 1926 was devoted to crime news. The exploits of gangsters like Al Capone, John Dillinger, and Bonnie Parker and Clyde Barrow were chronicled in the headlines of newspapers across the country. These gangsters have been replaced as major media figures in the

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

38

T1: FCH October 15, 2003

RETHINKING HOMICIDE

last three decades by serial murderers (like David Berkowitz, Ted Bundy, Andrei Chikatilo, John Wayne Gacy, Theodore Kaczynski), mass murderers (like Charles Whitman and Timothy McVeigh), and political assassins (like Lee Harvey Oswald, Sirhan Sirhan, James Earl Ray and Carlos “the Jackal”). As a source of information on homicide situations, however, newspaper accounts have long been questioned on the grounds of sensationalism and their restrictive scope. For example, Brearley (1932) estimated that only about one-half of the homicides in the registrars of vital statistics for South Carolina were reported in a state daily newspaper. Newspaper coverage of homicide is far more common than for any other crime, and this coverage is clearly not proportional to the relative frequency of these criminal acts. Homicides involving multiple victims, high-status victims or offenders, and bizarre details are most likely to be covered in media accounts (see Wilbanks 1984). Under these conditions, newspaper reports seem especially problematic as an accurate data source on the prevalence and social characteristics of homicide situations generally.

Court Records Several types of court records (including indictments, trial transcripts, committals on convictions, and execution reports) have been the primary data source in many studies of homicide. Erickson (1966) used Essex County court records and the records of the Court of Assistants to study the level of crime in the Massachusetts Bay colonies in the last half of the seventeenth century. Criticizing other data sources for historical analysis, Lane (1979) employed the number of indictments prepared for grand juries as the only satisfactory measure of homicide rates in urban settings in the 19th century. Hindus (1980) examined criminal prosecutions for murder and other offenses for Massachusetts and South Carolina in the late eighteenth and nineteenth century. McKanna (1997) combined court data (e.g., criminal appearance docket, indictments from the Registers of Criminal Action) and coroner reports to study homicide from 1880 to 1920 in three diverse counties in the American West, covering the cities of Omaha (Nebraska), Trinidad (Colorado), and Globe (Arizona). Court data and observations have also been the primary source for a variety of studies of the criminal processing of murderers (see Sudnow 1965; Swigert and Farrell 1976; Wolfgang 1958). Court records provide the most systematic data that are available on homicide and other crimes in early American and Western European history (see Gurr 1981; Plint 1851; Quetelet 1842), but these data have several inherent problems as a measure of the prevalence and characteristics of homicide. First, the level of community and judicial tolerance for particular types of deaths (e.g., infanticide, abortion, lynchings) varies widely over time and within particular jurisdictions at the same time periods (Lane 1997). The

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

DATA AND METHODOLOGY FOR STUDYING HOMICIDE SITUATIONS

October 15, 2003

39

sole reliance upon court records in this case would seriously distort estimates of the prevalence and most common types of homicide circumstances. Second, there is major slippage in the purported number of homicides as one moves from coroner reports and death certificates to successive stages of criminal processing. Estimates of homicide’s prevalence from court data are limited by the fact that not all deaths result in the filing of formal charges or an indictment, and an increasingly smaller number of cases result in a guilty verdict and death sentence. Third, court records distort the sociodemographic profile of homicide offenders and victims because criminal cases become increasingly homogeneous as they move through the criminal justice system (see Miethe and Moore 1986). The general lack of power, status, and economic resources to successfully fight criminal processing also helps explain the overrepresentation of poor and minority defendants in court records (Reiman 1990; Walker, Spohn, and DeLone 2000).

Mortality Records Reports of death filed by coroners’ offices and medical examiners are widely used in homicide research over the last two centuries. Coroners’ inquests have been a primary data source in studies of homicide in the U.S. “Wild West” from 1880 to 1920, providing information on the victim’s race, gender, age, occupation, location, and cause of death (McKanna 1997). Similar information was gathered from Miami’s chief medical examiner from 1953 to 1983 to study homicide trends in this city (Wilbanks 1984). Reports from city and state vital statistics registration bureaus have been used to estimate national homicide victimization rates in the early 1900s (see Brearley 1932). Recent time series analyses have employed vital statistics data published by the National Center for Health Statistics (NCHS) to examine change and stability in homicide victimization patterns over the last century and the comparability of these trends with police reports (see Chen 1996; Eckberg 1995; Smith and Zahn 1999). Coroner reports have also formed the basis for major studies of homicide in other countries (see Polk 1994). Although coroner records and death registrations should be closer to the “true” incident of homicide than other sources (see Gurr 1981), mortality data have also been widely criticized on several grounds. First, the fee-for-service system among coroners in the early twentieth century contributed to the underreporting of deaths by homicide, especially when the fee was to be paid by the offender and apprehension of the perpetrator was unlikely (Zahn and McCall 1999). This system enables a slit throat to be ruled a ruptured aorta and a dead transient found with hands tied behind his back and wallet missing to be labeled a suicide (Lane 1997). Second, states in the United States entered the national reporting system at different times and with different homicide traditions, hampering the development

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

40

T1: FCH October 15, 2003

RETHINKING HOMICIDE

of national homicide victimization trends (Zahn and McCall 1999). Third, coroner records include no data on the characteristics of the offender and limited information on the victim and situational elements (Riedel 1999). The scarce information on offender and offense elements severely restricts the utility of mortality data for the analysis of homicide situations.

Police Records The most common source of data for homicide research over the last half century has been police incident and arrest reports. Starting in 1930, the vast majority of jurisdictions in the United States have reported their crime offenses and arrests to the Federal Bureau of Investigation, which compiles this information and disseminates it in their annual Uniform Crime Reports (UCR). A similar national uniform reporting system was first implemented in Canada in 1962 (see Silverman and Kennedy 1993). UCR data contain information on offenders and victims of murder and nonnegligent manslaughter, negligent manslaughter, and justifiable homicides. Murder and nonnegligent manslaughter involve the “willful (nonnegligent) killing of one human being by another,” and include deaths from a fight, argument, quarrel, assault, or the commission of a crime (FBI 1996:13). Gross negligence by the offender is required to be considered a negligent manslaughter, whereas killings by police officers in the line of duty and killings by private citizens in self-defense are examples of justifiable homicides. Starting in 1961, UCR data on homicide were augmented by the introduction of the Supplementary Homicide Reports (SHR). SHR data provide information on some offense elements and the sociodemographic characteristics of homicide offenders and victims. The utility of the SHR data for homicide research was greatly enhanced in 1976 through the inclusion of offender characteristics and a major revision in the coding of the offense elements and circumstances (Riedel 1999). Compared to other sources, UCR data and the SHR files provide the most comprehensive national information on the offender, victim, and offense characteristics underlying homicide in the last three decades. SHR data have been widely employed in recent homicide research (see Decker 1993; Dobrin, Wiersema, Loftin, and McDowall 1996; Flewelling 1996; Harries 1997; Maxfield 1989; Riedel 1993; Tennenbaum 1993; Williams and Flewelling 1987). Police reports of homicide for particular U.S. cities have also been a major source of data. Ferdinand (1967) examined arrest records in Boston for murder and other major offenses from 1849 to 1951. Wolfgang (1958) used police investigation reports in his study of Philadelphia homicides in 1948 to 1952. Luckenbill (1977) analyzed murder “rap sheets” from

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

DATA AND METHODOLOGY FOR STUDYING HOMICIDE SITUATIONS

October 15, 2003

41

a police department in a mid-sized California county to detect stages in homicide events. Several studies (Brewer, Damphousse, and Adkinson 1998; Lundsgaarde 1977) examined police reports in Houston over different time periods to develop profiles of homicide victim, offender, and offense elements. Police data have also been used extensively to study homicide in Chicago (see Block and Block 1992; Block and Christakos 1995; Block and Zimring 1973; Voss and Hepburn 1968). Specific studies of gang homicides in Chicago, Los Angeles and Pittsburgh have been conducted through the use of police data (Block and Block 1993; Maxson 1999; Maxson, Gordon, and Klein 1985; Tita, Cohen, and Blumstein 1996). Police data have also been the major source of information for homicide studies in other U.S. locations (see Chilton 1996; Rosenfeld 1997; Rosenfeld, Bray and Egley 1999; Wilbanks 1984; Zahn and Jamieson 1996) as well as in other countries (see Brantingham and Brantingham 1984; LaFree 1999; Polk 1994; Silverman and Kennedy 1993). Although often considered a more accurate measure of homicide than other crimes, UCR data have many well-known limitations for measuring the prevalence of crime and its distribution (see Gove, Hughes, and Geerken 1985; Mosher, Miethe, and Phillips 2002; Riedel 1999). For example, state and local agencies submit UCR data on a voluntary basis, and not all actual offenses are known to the police. Clearance rates for homicide have also decreased dramatically over the last three decades, resulting in higher proportions of cases in which offender characteristics are unknown (Riedel 1999). UCR and SHR data have been further criticized because no narrative information is provided on the homicide event, the official disposition of the case is not included in these files, and the SHR includes data on offenders only who are identified shortly after the incident (Wilbanks 1984). Regardless of their particular problems, there are two fundamental reasons why UCR and SHR data are the primary sources for recent statistical analyses of homicide. First, UCR data on homicide trends converge with NCHS data since the 1960s (see Cantor and Cohen 1980; Riedel 1999), providing some assurance of the validity of police data on homicide. Second, SHR data have no rivals in terms of their breadth and depth. No other source provides comprehensive, national data on the major components of crime: offender attributes, victim characteristics, and offense characteristics. Under these conditions, police reports provide unique data for the study of homicide situations.

Data Sources in the Present Study Two general types of homicide data are used in the present study. Questions about the structure of homicide situations are examined primarily through the analysis of the U.S. Supplementary Homicide Reports (SHR)

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

42

October 15, 2003

RETHINKING HOMICIDE

and police reports of homicide from the Chicago Homicide Project (see Block et al. 2001).1 These data are available through ICPSR for secondary analysis. A small section of the current analyses are carried out using the California Homicides Data File, which contains all homicides reported to the California Criminal Justice Statistics Center between 1987 and 1999.2 Narrative accounts of homicides from several U.S. cities (Los Angeles, Miami, St. Louis, Las Vegas) are used to assess and describe the episodal nature and interpersonal dynamics of homicide. The integration of both types of data is designed to further our understanding of both the process and structure underlying homicide situations.

Supplementary Homicide Reports (SHR) The analysis of the structure of homicide situations focuses on SHR data over the 1976 to 1998 time period. These data involve nearly 440,000 homicide incidents and provide measures of the following characteristics of their situational context: the sex, age, and race of the offender and the victim; the motive, the victim–offender relationship, the number of offenders, the number of victims, the type of weapon used, urban or rural location of the offense, and region of the country. Collectively, this grouping of offender, victim, and offense characteristics covers the basic structural elements of homicide situations.3 While the number of variables included in this list of structural elements may appear limited to some observers, it is important to recognize that the primary focus of our situational analysis involves the prevalence and uniqueness of particular structures of homicide situations defined by all possible combinations of these offender, victim, and offense elements. If each of these 13 variables from the SHR data were dummy coded with only two categories, our unit of analysis would involve the empirical distribution of the 8,192 possible distinct combinations of these attributes among the 440,000 1 2

3

The Chicago homicide data are described in detail in Chapter 6. The California data were made available by the Criminal Justice Statistics Center, California Department of Justice. Data were formatted for SPSS and codebooks were prepared by Marc Riedel, Center for the Study of Crime, Delinquency, and Corrections, Southern Illinois University. The content of the data file is similar to that found in the SHR with some additions such as the location of the homicide. The two major omissions in these data are the income and marital status of both offenders and victims. Unfortunately, these measures are not included in the SHR data and are rarely included in a systematic way in citywide studies of homicide. Measures of alcohol and drug use in the SHR data can be derived only through categories of the motive circumstance (e.g., brawl due to the influence of alcohol, brawl due to narcotics). More direct measures of substance use are contained in some citywide studies (e.g., Chicago Homicide Project), but data on these variables are often missing in police files. Other important criticisms of SHR data (e.g., the coding of the characteristics of only one victim in multiple-victim homicides) are discussed in great detail elsewhere (e.g., Maltz 1999; Maxfield 1989; Miethe and Drass 1999; Williams and Flewelling 1987).

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

DATA AND METHODOLOGY FOR STUDYING HOMICIDE SITUATIONS

October 15, 2003

43

homicides in this time period.4 Under these conditions, we do not consider our focus on only 13 variables to define the multiple structures of homicide situations to be restrictive when compared to other studies of homicide. To correct for the problem of missing data in the SHR, the method of Expectation-Maximization (EM) was used for imputing values for all cases with missing data. Expectation-maximization produces maximum likelihood estimates using an iterative 2-step procedure. In the E-step, the expected value of the log likelihood is computed. The M-Step computes new values for the means and covariance matrix using both imputed and nonmissing data. The process is repeated until convergence is reached. The EM procedure was carried out using a SAS macro developed by Paul Allison (Allison 1999), which uses the EM algorithm to carry out maximum likelihood estimation of the mean and covariance matrix of the multivariate normal distribution for incomplete data. The algorithms in the macro are modeled on those described by Schafer (1997). EM is among the preferred approaches for imputation when data are missing at random (see Allison 2000, 2002), conditions that are likely to hold in homicide data (Regoeczi and Riedel 2003).5

Narrative Accounts Narrative accounts of homicide incidents from several large cities (e.g., Los Angeles, Miami, St. Louis, Las Vegas) are used to describe the nature of these interpersonal transactions and their precipitating factors. Descriptive accounts of 405 homicides in Los Angeles were drawn from the files of the Los Angeles Police Department for the period of 1900 to 1976. A disproportionate stratified sampling design was used to select random subsets 4

5

These 8,192 possible distinct combinations of attributes are computed as 2 raised to the 13th power. If more than two categories are used for representing particular variables, which we will do for many of them, the number of possible configuration will increase accordingly. The following variables were used as predictors of missing data in the imputation model: (1) total number of victims; (2) total number of offenders; (3) victim age; (4) victim gender; (5) victim race (White [reference category], Black, and other); (6) offender age; (7) offender gender; (8) offender race (White [reference category], Black, and other); (9) weapon (gun, knife [reference category], other weapon); (10) victim–offender relationship (intimate partner [reference category], other family, friend/acquaintance, stranger); (11) motive (expressive [reference category], instrumental, other motive). When values are imputed for categorical variables such as race, motive, weapon, and victim–offender relationship, the initial values will be relatively meaningless since they frequently produce values between 0 and 1. In such situations a basic set of rules are applied to assign values of 0 and 1. In the case of a dichotomous variable, values of 0 and 1 are assigned on the basis of which is closer to the imputed value. For a four-category variable such as victim–offender relationship, the variable is represented by three dummy variables (hypothetically labeled X1 , X2 , and X3 ). All three dummy variables should be set to zero if 1 − X1 − X2 − X3 is greater than either X1 , X2 , or X3 . Otherwise, if X1 is greater than X2 and X3 , X1 should be assigned a value of 1 and X2 and X3 should be assigned values of 0, and so on (Allison 2002).

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

44

T1: FCH October 15, 2003

RETHINKING HOMICIDE

of approximately 50 homicide incidents from each decade over this time period. The narrative descriptions of homicides in Miami, Florida were originally developed and analyzed by Wilbanks (1984). A total of 569 homicides in 1980 are contained in this source, which is based on records from the medical examiner’s office, the police, as well as court records. All of these narratives are included in an appendix of the original source (Wilbanks 1984:187– 374). Homicide narratives for St. Louis are based on data drawn from the Records Division of the St. Louis Metropolitan Police Department. They include a total of 1365 homicides from 1990 through the end of 1995.6 A total of 350 descriptions of homicide incidents in Las Vegas during 1998 and 1999 were obtained from the Las Vegas Metropolitan Police Department. These narratives were in the form of short descriptive accounts of homicides that were released to the local media. Similar to the narratives in other cities, they provide a brief description of the offender, victim, and the circumstances surrounding the crime. Although the narrative accounts of homicide in each city are relatively short in length (most are between 1 and 2 paragraphs), they are of sufficient depth to identify general aspects of the homicide transactions and the nature of the interpersonal dynamics between the victim and offender. These narratives are used primarily to explore major themes underlying homicide transactions and the nature of within-group variability in subtypes of homicide situations. Our numerous readings of these narratives provided key insights about the nature of predisposing and precipitating factors underlying homicide situations. It is within this context that the narrative accounts also became important for guiding our examination of the structure of homicide situations.

Analytic Approaches for Studying Homicide Situations Homicide has been studied using both aggregate and individual units of analysis. Aggregate studies examine variation in homicide rates for various geographical units, including crossnational comparisons, U.S. trends, and the analysis of differences across states, SMSAs, cities, census tracts, and neighborhoods. Individual-level studies, in contrast, explore differences across a wide array of homicide subgroups. These homicide subgroups include comparisons by offender characteristics (e.g., males vs. females, Whites vs. Blacks, teenagers vs. adults), victim attributes (e.g., single vs. multiple victims), and offense circumstances (e.g., instrumental vs. expressive motives, gun vs. nongun killings, strangers vs. intimates vs. acquaintances). 6

We are extremely grateful to Rick Rosenfeld for providing us with these narratives.

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

DATA AND METHODOLOGY FOR STUDYING HOMICIDE SITUATIONS

October 15, 2003

45

While often overlooked in previous research, many studies of homicide subgroups are actually investigations of group differences in homicide situations. In other words, studies of homicide subgroups are not just concerned with differences on the basis of the particular attribute, but they are also concerned with variability between and within these subgroups across combinations of other offender, victim, and offense elements. For example, studies of gender differences in offending often assume that male and female homicides are qualitatively distinct in their particular offender, victim, and offense attributes that define the homicide situation. Similarly, stranger homicides are often assumed to be different in their combined situational elements from homicides involving intimates and acquaintances. Unfortunately, the statistical models used to examine such subgroup differences are typically not designed to examine the pattern and complexity of the variability across these subgroups.

The “Main Effects” Statistical Model The most common specification of the functional form of the relationship between homicide and its correlates is the “main effects” model. Under this model specification, the impact of an independent variable on a dependent variable is assumed to be constant across levels of all other independent variables. This main effects specification has both theoretical and statistical implications. Statistically, coefficients derived from this model are calculated in such a way that the values are constrained to be identical across levels of all other variables. Theoretically, the main effects model considers each variable to operate like a “master status,” having identical effects across all combinations of other attributes (see Miethe and Moore 1986). As a master status, this specification assumes, for example, that gender differences in risks of homicide offending are the same for both the young and old, Whites and Blacks, gun users and nongun users, stranger assaults and the slaying of intimates, and killings involving both instrumental and expressive motivations. It also assumes that the effect of gender is identical across all combinations of these offender, victim, and offense attributes. The main effects model underlies most aggregate- and individual-level studies of homicide. Previous research that examines variation in homicide rates across geographic units or over time typically relies upon this specification. Tests of social disorganization theory, for example, often assume that the impact of separate measures of low socioeconomic status, population mobility, ethnic heterogeneity, and low supervision of youth are the same across levels of the other variables (see Bursik 1988; Kornhauser 1978; Miethe, Hughes, and McDowall 1991; Miethe and Meier 1994; Sampson and Groves 1989; for an important exception see Krivo and Peterson 2000). Under this research tradition, the detrimental impact on homicide rates of high

16:41

P1: FCH/FFX 0521832993c03

46

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH October 15, 2003

RETHINKING HOMICIDE

population mobility (through its weakening of social control mechanisms in the community) is usually considered to be constant across other structural properties of communities. Tests of other theoretical perspectives (e.g., routine activities/lifestyle theory, relative deprivation, subcultural theories) with aggregate data often employ the same model specification (see Cohen, Felson, and Land 1980; Kovandzic, Vieraitis, and Yeisley 1998; Messner 1983; Messner and Blau 1987; Miethe and Meier 1994; Stahura and Sloan 1988). The vast majority of research on the determinants of individuals’ risk of homicide victimization and differences between groups in their offending also relies upon the main effects model. This specification underlies most research on homicide risks that uses the following statistical procedures: bivariate and multivariate contingency table analysis, simple and multiple regression, logistic regression, and discriminant function analysis. The topic of gender differences in risks of violent victimization illustrates the application of the main effects model to individual-level data. Men are far more likely to be victims of homicide than women, and this fact would be revealed in the estimation of a main effects model of the likelihood of homicide victimization by gender. If this logistic regression model were extended to include other measures of the individual’s demographic attributes (e.g., race, age, social class) and lifestyles (e.g., frequency of nighttime entertainment, type of job), the logit coefficient for gender represents the partial or net impact of this variable on the risks of victimization. The change in the magnitude of the unique effects of gender in models with and without controls for other variables indicates the extent to which the effect of gender is mediated by these other variables. Even in the multivariate model, however, the net impact of gender on victimization risks is considered constant across all levels and combinations of these other variables. This means that the actual magnitude of the higher risks of victimization for men is assumed to be identical within each race, age, income category, and lifestyle grouping. All other variables in the regression model are interpreted and constrained in the same way. Most statistical analyses of homicide that involve direct subgroup comparisons are also commonly based on a main effects specification. Research on the differences in the prevalence and characteristics of stranger vs. acquaintance homicide, for example, involves comparisons between these groups on various offender, victim, and offense characteristics (see Decker 1993; Riedel 1993; Williams and Flewelling 1987). These comparisons within each type of victim–offender relationship, however, are essentially main effect contrasts in that they (1) indicate whether categories of particular variables (e.g., weapon, motive) are more common among some types of victim– offender relationships than others and (2) assume that the impact of these relationship-specific variables is constant within categories of other variables (e.g., that the prevalence of a particular type of weapon within a relationship

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

DATA AND METHODOLOGY FOR STUDYING HOMICIDE SITUATIONS

October 15, 2003

47

category is the same regardless of the motive). Bivariate and multivariate analyses of other types of subgroup comparisons (e.g., differences between teenagers vs. adult offenders, White vs. Black, deaths by firearms vs. other weapons) are often based on the same main effects specification. The advantage of the main effects model is its simple functional form and its ease of interpretation. Unfortunately, this specification is often an oversimplification of social reality and fails to consider whether particular variables have context-specific effects. In fact, the underlying assumption of these techniques is that the context of a given variable is irrelevant to its effect (Ragin 2000). When applied to the study of homicide situations, the main effects model is both theoretically and analytically inconsistent with the assumption that the nature of subgroup differences in homicide risks varies widely across particular combinations of offender, victim, and offense elements.

Statistical Models with Interaction Effects In response to the limitations of the main effects model, researchers have increasingly turned to alternative specifications of the functional relationship between variables. The primary alternative functional form involves contextual or interaction effects. These interaction effects are often treated as contextual effects because they indicate that the impact of a variable is context-specific. Various types of homicide studies have employed interaction effects in their analysis. Racially disaggregated models of crime rates, for example, use a type of interaction effect when they test whether particular combinations of socioeconomic factors (e.g., high population mobility and low income) are equally important predictors of homicide and other crimes across racial groups (see LaFree and Drass 1996; Krivo and Peterson 2000; Messner and Sampson 1991; Sampson 1985a). Aggregate models of homicide rates that attempt to explain changes over time or variation across geographical units (e.g., cities, census tracts) sometimes include interactive terms directly to explore the nature of the variability within subsets of variables (see LaFree and Drass 1996). Individual-level studies of homicide risks and subgroup differences incorporate interactive effects by examining the joint impact of combinations of variables (e.g., combinations of age, sex, and racial groupings). Although statistical models with interaction terms may be better able to model the complexity of homicide situations, there are several problems with this general modeling approach. First, all empirically plausible interactions need to be hypothesized in advance (Ragin 1987). Second, the complexity of substantive interpretations increases in direct proportion to the number of variables included in the interaction terms. Researchers rarely interpret more than three-way interaction terms (i.e., combinations of three

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

48

October 15, 2003

RETHINKING HOMICIDE

or more variables considered simultaneously) even though an accurate portrayal of the data may require a more complex pattern of interaction effects. Third, the statistical problem of multicolinearity often plagues homicide studies (see Land, McCall, and Cohen 1990). This problem is compounded by the presence of multiple interaction terms, preventing researchers from isolating the unique impact of particular variables or combinations of variables. For example, an equation containing 64 interaction terms would be needed to examine combinations of six different causal conditions, generating high levels of colinearity among the terms and substantial obstacles to interpretation of the coefficients (Ragin 1987). Under these conditions, statistical models with interaction terms are not necessarily the ideal approach for modeling variability in homicide situations.

Spatial Analytic Procedures and Typologies Another alternative approach for studying homicide involves various forms of spatial analysis. Both cluster analysis and factor analysis are appropriate procedures for identifying unique homicide types and factors that may underlie them. Although cluster analysis has been used to identify and examine crime typologies for sex offenders and other offenders (see Knight and Prentky 1993; Rosenberg and Knight 1988), this approach has rarely been applied to examine homicide subgroups (for an exception see Riedel and Regoeczi 2001). Instead, homicide classifications have been constructed on a largely haphazard basis. Miethe and McCorkle (2001), for example, rely upon particular characteristics of the offender and offense to develop their classification scheme. Their homicide subtypes include (1) interpersonal disputes, (2) instrumental felony offenses, (3) youth group offending, (4) chronic violent offending, and (5) politically-motivated violence. Block and Block (1992) categorize homicides according to the offender’s primary motive and the victim–offender relationship. The following circumstances underlie their homicide syndromes: (1) acquaintance expressive, (2) instrumental, (3) intimate partner, (4) stranger expressive, (5) street gang, (6) other family expressive, (7) child abuse, and (8) sexual assaults. The major problem with these homicide typologies is that they violate the two basic assumptions of any classification system. These assumptions are that categories must be mutually exclusive and collectively exhaustive. Further, the nature of the variability within each of these homicide subtypes has not been subjected to systematic investigation. Given the limitations of previous classification systems, a preferred approach for the construction of homicide typologies would involve comprehensive comparisons of the structure of homicide situations across the full range of combinations of offender, victim, and offense elements. The method of Qualitative Comparative Analysis is well-suited for identifying

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

DATA AND METHODOLOGY FOR STUDYING HOMICIDE SITUATIONS

October 15, 2003

49

the nature of this complex structure that underlies homicide situations and subgroup variations in them.

Qualitative Approaches A variety of qualitative methodologies have been used to study homicide. These methods include case studies and ethnographies of dangerous persons and places, and content analyses of various narrative accounts of homicides (e.g., police arrest reports, media accounts, court transcripts). Several studies have used traditional qualitative methods to assess characteristics of homicide situations (see Jurik and Gregware 1992; Luckenbill 1977; Polk 1994). Although there are various ways to conduct qualitative research (e.g., by using grounded theory or analytic induction), a qualitative analysis is often conducted by sifting through volumes of data records (e.g., interviews, field notes, historical documents) to identify patterns, major themes, and elements of commonality and uniqueness that may be used to describe sets of observations or form the basis for developing theoretical explanations (see Coverdill, Finlay, and Martin 1994). Both the use of multiple coders for assessing interrater reliability and the triangulation of the observed findings with other methods are standard ways of improving the reliability of qualitative studies. Several limitations often underlie the use of qualitative approaches to the study of homicide and other types of violence. First, qualitative researchers often emphasize cases over variables and causal complexity over causal simplification, an extremely difficult task even for a skilled analyst when coding rules are not clearly articulated and more than a very small number of cases are investigated (Coverdill et al. 1994). Second, the representativeness of the cases is often a source of criticism because they are selected on grounds other than whether they are comprehensive or representative of a population (e.g., they are novel or serious, theoretically interesting, historically important). Third, it is impossible for other researchers to replicate the observed findings without a complete reproduction of the field notes, interview transcripts, or historical documents. Without this ability to reanalyze results, it is difficult to ascertain whether valid inferences have been drawn, contradictory or ambiguous evidence omitted, or alternative explanations overlooked (Coverdill et al. 1994:56). Compared to quantitative research approaches, qualitative methods involve case-oriented modes of research that are intensive, holistic, idiographic, nonstatistical, and often geared toward the illumination of particulars (Ragin 1987). The interpretive paradigm that underlies much qualitative research is also rooted in nonformal devices such as “thick” or rich description, narrative unfolding, and strong sensitivity to context, specificity, and culturally embedded meanings (see Griffin and Ragin 1994; Skocpol

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

50

October 15, 2003

RETHINKING HOMICIDE

1984). Formal methods of qualitative analysis like QCA are bridging approaches that operate within the “middle ground” between (1) the quantitative emphasis on explanation and causal generalizations and (2) the qualitative focus on interpretation and description.

Qualitative Comparative Analysis Current typologies of homicide situations and standard analytic approaches for examining sources of variation in them have generally been inadequate for modeling the full structure and complexity of these situations. As an alternative approach, the method of Qualitative Comparative Analysis (QCA) is applied in this study to examine the structure of homicide situations and the similarities and differences in these structures for different subgroups. The logic of QCA, its previous applications, and methodological issues underlying this analytic approach are described below.

The Logic of QCA QCA is a relatively new analytic technique that uses set theory and Boolean algebra to implement a case-oriented, comparative approach to data analysis (Drass and Miethe 2001; Drass and Ragin 1989; Ragin 1987, 2000). Traditional statistical approaches, in contrast, tend to be variable-oriented. This distinction is the key for understanding QCA, and especially its potential for the study of homicide situations. As an analytical method for case comparisons, QCA considers cases holistically, as complex configurations of attributes measured by a set of variables. It also assumes that events or outcomes are produced by variables acting together in combination. Under this specification, the effect of any particular variable may be different from one case to another, depending upon the values of other attributes of a case. QCA places importance on the total context of elements to assess whether the same outcome can be produced by different combinations of variables. Variation between an outcome variable and a set of explanatory variables in QCA is not measured variable by variable across situations. Rather, it refers to the diversity of unique conditions, measured case by case using the different combinations of variables within cases, that produce the outcome. In this manner, QCA allows for heterogeneity in causal conditions – different configurations of independent variables can lead to the same outcome – and produces a summary of these conditions through the application of a minimization algorithm based upon the rules of formal logic. Contrary to QCA, most traditional statistical techniques are variableoriented approaches that function to minimize the effect of context. Variables are treated as analytically distinct dimensions of phenomena that can be observed to a greater or lesser extent in each empirical case. Information

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

DATA AND METHODOLOGY FOR STUDYING HOMICIDE SITUATIONS

October 15, 2003

51

about a variable is abstracted from a sample of cases and used to estimate the average effect of each independent variable upon the outcome variable. These effects are assumed to hold across all cases, regardless of the values of other attributes. Under most statistical approaches, the explanation for why cases belong to a certain category is the same for all members of the category and usually is expressed in terms of an abstract model of the additive effects of the explanatory variables. The focus of the statistical model is to estimate the effects of individual variables averaged across all cases in order to determine to what degree a set of independent variables “account for” variation in a particular dependent variable. But to talk of independent variables accounting for X % of the variation in dependent variable is only meaningful on a statistical level and often lacks any direct translation into meaningful actions by real people (see Becker 1998). Furthermore, it does not permit an investigation of configurations of phenomena, instead examining the average effect of a variable across all contexts (Ragin 1987). These traditional approaches are premised on the notion that events can be accounted for by a single explanation as opposed to there being multiple paths to the same outcome. QCA, in contrast, is designed to examine multiple configurations or combinations of attributes leading to a particular outcome. It also summarizes variation in terms of the combinations of variables observed within individual cases, thereby preserving the complexity of situations while seeking to simplify them as much as possible (Becker 1998). In this respect, QCA embraces two fundamental aspects of social phenomena: (1) its causes are typically conjunctural and must come together in time and place to generate the particular phenomenon; and (2) the same outcome may be produced from several different combinations of causes (Ragin 1987, 2000). Similar to a statistical analysis, QCA requires the specification of a model relating a set of independent variables to an outcome variable. The independent variables are assumed to measure important structural features of the context within which the outcome occurs. There are various approaches for variable selection within QCA but it is often recommended that combinatorial theories that emphasize causal diversity form the basis for the inclusion of particular variables (see Amenta and Poulsen 1994; Ragin 2000). Because QCA is based upon Boolean algebra, and Boolean algebra is a binary mathematical system, all variables must be measured as dichotomies using the values 0 and 1.7 Independent variables that are not naturally dichotomous can be measured using sets of dummy variables (similar to regression analysis), where each dummy represents a category of the variable and one category is excluded. 7

This requirement of binary coding is no longer necessary with the application of fuzzy-set logic to QCA (see Ragin 2000).

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

52

T1: FCH October 15, 2003

RETHINKING HOMICIDE

An analysis using QCA begins with the construction of a truth table. A truth table lists all of the unique configurations of offender, victim, and offense variables appearing in the data, along with the corresponding category (or categories) of homicide observed for each configuration. The truth table provides information about which configurations are unique to a category of the outcome variable (we call these “unique signatures”) and which are observed within both categories (called “contradictions” or “common” configurations/profiles). By comparing the number of unique and common configurations for each category of the outcome variable, QCA provides an estimate of the extent to which types of homicide are similar or unique in their situational profiles. These unique and shared configurations also define subsets of conditions that can be compared and simplified in the next step of the analysis. Once a truth table has been constructed, QCA compares the configurations within a subset, looking for commonalities that allow configurations to be combined into simpler representations. QCA can simplify the configurations unique to each category of the outcome variable, those common to both categories of the outcome variable, or some combination of the two. The process of simplification in QCA is accomplished by a minimization algorithm, which identifies and eliminates unnecessary variables from configurations. Several different algorithms are available through QCA programs to derive these minimized solutions.8 QCA considers a variable unnecessary if its presence or absence within a configuration has no impact on the outcome associated with the configuration. For example, suppose configurations ABC and ABc (where upper case indicates presence or 1, and lower case indicates absence or 0) both produce the same type of violent crime (Y). Since both configurations produce the same outcome, they would be assigned to the same category of the outcome variable. QCA would consider variable C to be irrelevant when combined with A and B (i.e., within the context of A and B) since Y occurs regardless of the value of C. Thus, QCA would replace these two configurations (ABC, ABc) with the single, simpler configuration AB. However, variable C is removed only from these two configurations; it may be necessary to produce Y when combined with other variables. It is in this manner that QCA produces case-based, rather than variable-based, results. QCA repeats these comparisons until no further reductions are possible. Redundancies among the remaining reduced configurations are eliminated, producing the final solution: a statement of the conditions 8

The original QCA program was developed by Kriss Drass and Charles Ragin and is available for distribution through the Center of Urban Affairs and Policy Research at Northwestern University. For recent additions to QCA and the fuzzy-set version of QCA (called FSQCA), the reader should consult Charles Ragin at the University of Arizona.

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

DATA AND METHODOLOGY FOR STUDYING HOMICIDE SITUATIONS

October 15, 2003

53

unique to the subset of configurations minimized. In the present study, these results can be interpreted as subprofiles of the basic unique and common structures that underlie homicide situations.

Previous Applications of QCA As a methodological approach for developing empirically based typologies and evaluating combinatorial theories, QCA has been increasingly used to examine similarities and differences in the combinations of attributes that underlie a classification variable (see Amenta and Halfmann 2000; Amenta and Poulsen 1994; Coverdill et al. 1994; Drass and Miethe 2001; Miethe and Drass 1999; Ragin 1987, 2000). QCA is often viewed as a “middle ground” methodology between the traditions of qualitative and quantitative analysis. In fact, Coverdill et al. (1994) contend that QCA offers the prospect for better theory than quantitative approaches (through its logic of multiple conjunctural causation) and better data analysis than traditional qualitative approaches (through its requirement that all cases be considered in the analysis). Previous applications of QCA have been based primarily on “small-N ” samples in the areas of political, organizational, and historical sociology. For example, these applications include studies of labor management practices (Coverdill et al. 1994), early welfare state formation (Hicks 1994), ethnic political mobilization (see Becker 1998; Ragin 1987), social movement organizations (Ragin 1987), peasant rebellions (Ragin 1987), social spending on the aged (Amenta and Poulsen 1994), Latin American revolutions (Wickham-Crowley 1991), assessing discrimination (Ragin, Mayer, and Drass 1984), the structure of metropolitan juvenile courts (Becker 1998), and case congregation in aids litigation (Musheno, Gregware, and Drass 1991). Most of these applications involve sample sizes of fewer than 30 cases. Applications of QCA in “large-N ” studies are rare within and outside of criminology. The criminological studies include research on differences between robbery and other violent crime (Drass and Miethe 2001), probation officers’ recommendations (Drass and Spencer 1987), and the social context of instrumental and expressive homicides (Miethe and Drass 1999). The sample sizes in these studies ranged from a minimum of 200 cases to a maximum of 34,329 cases.

Methodological Issues in QCA As a methodological synthesis or hybrid approach for addressing quantitative researchers’ concerns with explanation and qualitative researchers’ emphases on interpretation, QCA offers a diverse perspective that has both strengths and limitations. The primary strengths of QCA are that it directs

16:41

P1: FCH/FFX 0521832993c03

54

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH October 15, 2003

RETHINKING HOMICIDE

attention to (a) cases as wholes and as configurations of potential causes, (b) contingent causes and combinations of causes, (c) causal heterogeneity across subsets of cases, (d) fusion of concerns for deviant cases and generalizations, (e) fusion of idiographic and nomothetic modes of explanation, (f) systematizing small-N studies and bringing case-oriented virtues to moderate-to-large-N studies, and (g) versatility as an inductively intensive manner for building theory and as a deductively grounded method for testing the fit of a well-articulated theory to the facts (Hicks 1994:97). The methodological limitations often associated with QCA involve (a) its absolute or deterministic conception of necessary and sufficient conditions and (b) the adverse impact of measurement error on these deterministic solutions. Most applications of QCA have used a deterministic solution to the classification of combinations of elements in a truth table. Under this formulation, a particular combination of attributes is considered unique only when it has a nonzero frequency for one category of the outcome variable and a zero frequency for the other category of the outcome variable. Applied to gender differences in homicide, this absolute decision rule for unique profiles would require that a particular combination of attributes that is found in male-perpetrated homicides (e.g., expressive motives, killings with guns, night-time slayings, public places, intragroup attacks) is never observed among homicides committed by women. Combinations of attributes that have a nonzero frequency for both categories of the outcome variable are considered a contradiction or common configuration. Even if a configuration was found to disproportionately represent one group (e.g., 15,000 males vs. 2 females), this particular combination would be classified as a contradiction under this deterministic solution. Unfortunately, the discovery of unique profiles becomes increasingly unlikely as the sample size increases because of the possibility of various sources of measurement error (e.g., coding error, low interrater reliability) and the mere idiosyncracies of humans, placing severe limits on the utility of QCA for comparative research in large-N designs. As a means of overcoming this limitation of QCA with large samples, previous research has developed frequency-based rules for establishing “unique” and “common” configurations. The selection of various relative rules for defining unique and common configurations is conceptually similar to the choice of different significance levels (e.g., .01 vs. .05 vs. .10) in conventional statistical analyses (see Miethe and Drass 1999). We have used a “10% difference” rule in our previous work for classifying configurations as unique or common because this standard is often viewed as the lower limit for claims of substantive differences across categories within contingency table analyses (see Fox 1998; Regoeczi, Miethe, and Drass 2001). Under this decision

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

DATA AND METHODOLOGY FOR STUDYING HOMICIDE SITUATIONS

October 15, 2003

55

rule, a configuration is unique to a particular outcome variable if its relative frequency within that combination exceeds its overall marginal distribution by 10 percentage points. Tests of statistical significance for sample proportions can also be used to define relatively unique configurations. The use of a relative difference rule for defining unique and common configurations offers a probabilistic alternative to the deterministic solution underlying applications of QCA in small-N designs. These probabilistic techniques permit an investigation of causal complexity in the face of the randomness and error that are built into social science research and social life more generally (Ragin 1987). To further reduce the impact of measurement error in configurations with small cell frequencies, general rules for establishing minimum cell frequencies for each configuration have also been used in past studies (e.g., 5 cases per configuration because this is a general rule of thumb for identifying minimum cell frequencies in the analysis of contingency tables [Miethe and Drass 1999]). Different minimum frequencies are also used across categories to maintain proportionality across groups. For example, we used minimum cell frequencies of 5 for female offenders and 35 for male offenders in our study of Chicago homicides because males were overrepresented as offenders by roughly a 7:1 ratio (Regoeczi, Miethe, and Drass 2001). Minimum rules based on cell proportions (e.g., include a configuration if it accounts for more than .01% of the cases within each group) are an equivalent approach for standardization for groups of disparate sizes. Rules for minimum cell frequencies or proportions are not used in small sample applications of QCA with their emphasis on deterministic solutions.

An Illustration of QCA To illustrate the application of QCA in a large-N design, we will conduct a brief comparative analysis of gender differences in homicide situations. This analysis is based on the Chicago homicide data for the 1960s (1965–69) and includes only 4 independent variables: (1) motive [(1) = expressive, (0) = instrumental], (2) victim–offender relationship [coded (1) family/intimate, (0) = other], (3) location of the crime [coded (1) = outdoor, 0 = indoor], and (4) weapon [coded (1) gun, (0) = other]. A more complete list of variables will be included in subsequent analyses of gender differences in homicide situations using these Chicago data (see Chapter 6). Here, we use only 4 variables to illustrate the QCA technique. The method of QCA begins with a truth table, a tabular representation of all possible combinations of the included attributes. As shown in Table 3.1, all 16 of the possible configurations of the 4 variables are observed within this sample. The cell frequencies for male offenders (outcome = 1) and

16:41

56 Gun =Gun Gun

436 251 138

Outdoor Outdoor Indoor

1. Non-Fam/Int 2. Non-Fam/Int 3. Non-Fam/Int

Expressive Expressive Instrument

Total N

Male Offenders:

Panel C: Unique Profiles of Homicide Situations

414 234 137

Male

22 17 1

Female

0% 18.7% 81.3%

16 0 3 13

25.0% 56.2% 18.8%

16 4 9 3

95% 93% 99%

% Male

% N

Configurations

%

N

Configurations

0% .2% 99.8%

2,810 295 1,096 1,419

N

Absolute Rule

% 10.5% 39.0% 50.5%

Male Male Male

Relative Rule

QCA Classification

Common Common Common

%

Homicide Cases

2,810 0 5 2,805

N

Homicide Cases

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

Number of Situations Observed Situations Unique to Female Offenders Situations Unique to Male Offenders Contradictions (Common Situations/Cases) for both Genders

Panel B: Configurations and Cases (using Relative Rule of 10% Difference)

Number of Situations Observed Situations Unique to Female Offenders Situations Unique to Male Offenders Contradictions (Common Situations/Cases) for both Genders

Panel A: Configurations and Cases (Absolute or Deterministic Rule)

Table 3.1. Qualitative Comparative Analysis of Gender Differences in Homicide Situations (Chicago 1965–69)

P1: FCH/FFX 0521832993c03 T1: FCH October 15, 2003 16:41

Non-Fam/Int Non-Fam/Int Non-Fam/Int Family/Intim Family/Intim Family/Intim

57

Expressive Expressive Expressive Expressive

Indoor Indoor Outdoor Outdoor

Outdoor Indoor Outdoor Outdoor Outdoor Indoor

=Gun Gun Gun =Gun

Gun =Gun =Gun =Gun Gun Gun

340 303 39 52

Total N

112 107 99 2 2 1

Expressive Expressive Expressive

Indoor Outdoor Indoor

Gun Gun =Gun

542 376 10

Total N

MODEL: Offmale = Family/Intimate + Expressive + Outdoor + Gun

1. Non-Fam/Int 2. Non-Fam/Int 3. Non-Fam/Int

Both Male and Female Offenders:

Panel D: Common Profiles of Homicide Situations for Each Gender

Family/Intim Family/Intim Family/Intim Family/Intim

Instrument Instrument Instrument Instrument Instrument Instrument

497 302 9

Male

187 199 15 38

Male

111 102 93 2 2 1

45 74 1

Female

153 104 24 14

Female

1 5 6 0 0 0

92% 80% 90%

% Male

55% 66% 38% 73%

% Male

99% 95% 94% 100% 100% 100%

Male Male Male Male Male Male

Female Female Female Female

Common Common Common

Common Common Common

Absolute Rule Relative Rule

QCA Classification

Common Common Common Common

Absolute Rule Relative Rule

QCA Classification

Common Common Common Male Male Male

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

1. 2. 3. 4.

Female Offenders:

4. 5. 6. 7. 8. 9.

P1: FCH/FFX 0521832993c03 T1: FCH October 15, 2003 16:41

P1: FCH/FFX 0521832993c03

58

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH October 15, 2003

RETHINKING HOMICIDE

female offenders (outcome = 0) are used to determine whether particular configurations are unique or common to both genders. Under the absolute or deterministic solution used in small-N designs, there are 3 situational profiles (i.e., combination of attributes) that are found among male offenders and not observed among female homicide offenders. These exclusively unique male profiles, however, account for only 5 of the 2,810 homicides in this decade (see Panel A of Table 3.1). No situational profiles were exclusive to female offenders, meaning that there were no combinations of attributes that had a nonzero frequency for female offenders and a zero frequency among male offenders. The majority of configurations (81%) involved situational contexts that were contradictions (i.e., included both male and female offenders) and these common profiles represented nearly all Chicago homicides in the 1960s (2,805 of 2,810). In short, the QCA results under a deterministic rule of uniqueness indicate that both male and female offenders almost always kill under similar situational contexts. A small number of unique signatures (n = 3 configurations) are found among male offenders, but they represent only a trivial number of homicides (n = 5 cases). A 10% difference rule for defining unique signatures requires that a particular cell frequency exceed its marginal frequency by 10% for that configuration to be considered relatively unique to that particular group. The marginal distribution for male offenders is 83% (i.e., 83% of the Chicago homicides in the 1960s were committed by males), so a particular situational profile would be considered unique to males under this decision rule if males accounted for more than 93% of the offenders in this situational profile (i.e., 83% + 10% rule = 93%). In a similar vein, relatively unique female profiles are defined by a percentage greater than 27% female (i.e., 17% [the marginal percentage of female offenders] + 10% rule = 27%). Profiles with cell proportions outside these standards (i.e., cell proportions from 74% to 92% male or from 8% to 26% female) are considered situational contexts that are common to both male and female offenders. The pattern of unique and common profiles is changed dramatically once the 10% rule is used to define unique situational contexts (see Panel B of Table 3.1). Under this rule, the vast majority of configurations (13 of 16) are now unique to either male or female offenders, and nearly half of all homicides are accounted for by these unique profiles. These results indicate that both qualitatively unique and common combinations of attributes underlie gender differences in the situational contexts for homicide. The truth table in QCA presents the full listing of all possible combinations of attributes in the data and provides the relative cell frequencies necessary for identifying the unique and common signatures. When more than 5 variables are included in a large-N design, the size of the truth table becomes increasingly large, requiring for practical purposes that interpretations

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

DATA AND METHODOLOGY FOR STUDYING HOMICIDE SITUATIONS

October 15, 2003

59

and presentations be based on a relatively small number of “dominant” (i.e., frequently occurring) profiles, since the truth table may include literally hundreds of configurations. The nature and relative frequency of each observed situational context for Chicago homicides in the 1960s are presented in Panel C of Table 3.1. For comparative purposes, we also present how each configuration is classified under both the absolute (i.e., deterministic) and relative difference (i.e., probabilistic) rules for defining unique signatures. The nature of the particular combinations of attributes in a truth table provides the basis for determining how homicide situations are qualitatively unique and similar across each gender. For example, the unique signatures for male homicide offenders (n = 9 configurations) involve primarily nonfamily members as victims who are killed in outdoor locations for both instrumental and expressive motives using either guns or other weapons. Two situational contexts are especially dominant among these uniquely male profiles (representing 414 and 234 of the homicide cases), and both of them involve expressively motivated killing of non-family members in outdoor locations regardless of weapon. In contrast, all unique situations for female offenders (n = 4) involve expressively motivated killings of family members, and indoor locations are the primary setting for the two most dominant female profiles. When males and females kill in common situational contexts, the two most dominant situations involve expressively motivated killings of nonfamily members with firearms (see Panel D of Table 3.1).9 The amount of situational clustering (LaFree and Birkbeck 1991) in a QCA study involves an examination of the relative frequencies across profiles and their representativeness of all homicides within a particular group. For example, female-perpetrated homicides that occur in situations unique to females are extremely concentrated (i.e., have high situational clustering) because the two dominant unique profiles account for 87% (257 of the 295) of all homicides that occurred in uniquely female situations. In contrast, male offenders exhibit less situational clustering because the two most dominant unique profiles for them account for only 59% (648 of the 1096) of the homicides in uniquely male situations. If a minimum frequency rule were applied in this example, the conclusions would change very little. This is the case because only 3 configurations would have failed the minimum frequency rule of at least 5 cases per cell 9

It should be obvious that this pattern of unique and common signatures would not be detected in a standard linear regression analysis of gender differences in homicide. This additive model assumes the effects of each variable are similar across contexts, an assumption whose violation is easily seen in the truth table for QCA. A “main effects” regression analysis, by definition, would not properly specify the pattern of complex interactions between the types of victim– offender relationship, motives, locations, and weapon use in homicide that is the basis of the combinatorial principles that underlie QCA.

16:41

P1: FCH/FFX 0521832993c03

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

60

T1: FCH October 15, 2003

RETHINKING HOMICIDE

and these configurations in total accounted for only 5 homicides in Chicago from 1965 to 1969. Our subsequent analyses of the Chicago homicides and the national SHR data will include more configurations that are excluded under minimum frequency rules because our QCA models will include more causal factors (e.g., often more than 10 variables) and thereby increase the number of possible configurations. The language of absolute and relative decision rules, unique and common signatures, dominant situations, situational clustering or concentration, and minimum frequencies will be used throughout the remainder of this study to describe our QCA results. The basic description and interpretation of QCA results in our subsequent applications will follow the logic of the presentation used in this illustration.

Summary Both quantitative and qualitative approaches to the study of homicide have been reviewed in this chapter. Neither of these approaches by themselves can offer a variable- and case-oriented focus (i.e., treat cases as wholes and as configurations of variables that represent potential causes), address causal heterogeneity across subsets of cases, integrate concerns for deviant cases and generalizations, cover both idiographic and nomothetic modes of explanation, or provide versatility in methods for building and testing theories on the basis of small and large samples. As a middle ground or bridge between these competing paradigms, QCA offers a strong alternative that builds upon the strengths of both quantitative and qualitative traditions. A brief illustration of QCA to study gender differences in homicide situations was used in this chapter to further illustrate the logic of QCA, its particular language, and the various decision rules that guide the analysis. Our applications of QCA in the remaining chapters apply minimum frequency rules and a relative difference or probabilistic standard of unique and common profiles to “large-N ” samples of homicides. The specific rules are specified and comparisons are also conducted to examine the consequences of our particular decision rules on substantive conclusions about the nature of homicide situations. The ideas of unique and common signatures, dominant configurations, and situational clustering form the basis for much of our empirical analyses of homicide situations involving various subgroups (e.g., instrumental vs. expressive motives, male- vs. femaleperpetrated homicides, Black vs. White homicides, youth vs. adult offenders, stranger vs. acquaintance vs. family/intimate homicides).

16:41

P1: JMT 0521832993c04

CB617-Miethe-v1

November 3, 2003

C H A P T E R

F O U R

The Empirical Distribution of Homicide

homicide is not randomly distributed over persons, places, or time. Some groups of people are more likely than others to be involved in lethal violence as its offenders and victims. There are also fundamental differences within and across geographical areas in their rates of homicide. Like individual offenders, countries, cities, neighborhoods, and other physical environments have criminal “careers” of violence that vacillate over time. The current chapter summarizes the empirical distribution of homicide across countries, over time, and across various social groupings. Trends in England and the United States are emphasized because of the greater availability of longitudinal data in these countries. After presenting brief histories of homicide in these countries, we use Supplementary Homicide Reports (SHR) to identify the most prevalent structures of homicide situations in the United States over the last three decades. QCA is used to empirically document change and stability in the structure of homicide situations over time. The observed results are then discussed in terms of their implications for further research and public policy.

Historical Trends in Homicide Based on previous studies and a reanalysis of existing data, it is possible to describe trends in the prevalence of homicide and its characteristics over time. These historical trends and changes in homicide characteristics are provided below.

Stability and Change in the Prevalence of Homicides Data limitations and changing definitions of homicide preclude the development of a comprehensive time series for homicide rates across Western 61

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

62

November 3, 2003

RETHINKING HOMICIDE

history. Historical data prior to the early twentieth century are especially problematic, but definitional issues and disparate counting of incidents also plague current homicide data. Nonetheless, a general portrayal of homicide trends and a determination of the most dangerous times are possible from various sources. Contrary to contemporary political rhetoric on crime and its prevalence, most authoritative sources conclude that homicide and other forms of violence have generally declined throughout Western history. Records from the thirteenth century suggest that murder and other types of interpersonal violence were a far more common occurrence in medieval England than current times (Given 1977). In contrast, Lane (1997:297) describes homicide trends in American history as representing a long distended U-curve, or perhaps a reverse J-curve, with “high rates from the earliest available nineteenth century statistics dropping to lows somewhere in the middle twentieth century . . . and shooting up sharply in the 1960s.” Gurr (1981:326–327) concludes that the dominant crime pattern in the United States is “the occurrence of three pronounced upsurges of interpersonal violence which began roughly fifty years apart: ca. 1860, 1900, and 1960.” The substantial drop in crime rates throughout the 1990s followed by increasing rates again at the turn of the new century is the current U.S. historical pattern (see FBI 2002).

Homicide Trends in England Court records on homicide in England can be traced back nearly 800 years for particular cities and counties. These judicial records are less comprehensive for some time periods (e.g., the mid-fourteenth to the mid-sixteenth centuries) due to massive political and social unrest in the aftermath of the “Black Death” (i.e., bubonic plague). Nonetheless, previous researchers (e.g., Beattie 1974; Cockburn 1977; Given 1977; Hanawalt 1979) have been able to construct profiles of crime trends and offense characteristics in England over a remarkably long historical period. Based on these accounts, medieval England was an incredibly violent society compared to the levels of interpersonal violence in the last half of the twentieth century. The merging of homicide court records and population estimates indicates that homicide rates in particular cities and counties in England in the thirteenth century were far higher than contemporary patterns. For example, the estimated homicide rates for the counties of Bedford and Kent in this early century are over twenty times higher than patterns in England in the 1970s (Given 1977). Homicide rates in the first half of the fourteenth century remained relatively high compared to contemporary standards. Homicides in medieval times typically involved persons with strong social ties (i.e., primarily acquaintances and neighbors), the poor, and males as victims and offenders.

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

November 3, 2003

63

As evidence of somewhat distinct historical trends, homicides in the earlier centuries were relatively rare within the context of the nuclear family and far more likely than recent times to be a group activity and to occur in rural areas.1 Robbery also accounted for a substantially higher proportion of homicides in the Medieval era than in contemporary England (Lane 1997). A more complete historical record of homicide in England begins again around the mid-1500s. From this time period to the early 1700s, crime rates exhibited a series of waves and spurts. Homicide rates are estimated to have declined substantially during this time period, and lethal killings became less common in robbery-homicides and other felony situations. Cockburn (1977:57) describes the vast majority of murders in this period as occurring “during acts of sudden, unpremeditated aggression and resulted from attacks with a variety of knives and blunt instruments.” Men and the poor remained the primary offenders and victims of homicide. Homicide in England in the eighteenth and nineteenth century followed a vacillating pattern but declined noticeably starting in the 1850s (see Gurr 1979). Lethal violence represented only a small proportion of the “crime problem” at the end of the nineteenth century. There were no obvious changes in the characteristics of homicide offenders and the situations in which people were killed during this time period. The level of homicide in the twentieth century also followed various waves similar to prior historical periods in England. Serious crime in London reached its lowest recorded level in the late 1920s, waffled up and down around the mid-century, and increased sharply from 1950 to 1974. The post– World War II increase was especially dramatic, with murders in London growing by 300% in the 1950–74 time period (Gurr 1979:356).2 Rates of homicide in England remained fairly stable at 1.2 per 100,000 in 1979 to about 1.5 per 100,000 in the late 1990s (Barclay, Tavares, and Siddique 2001; Brantingham and Brantingham 1984). During the twentieth century, homicide became primarily an urban problem and an increasing proportion of property offenses involved violence. The act of homicide in England at the end of that century, however, has 1

2

Roger Lane (1997) attributes the group nature of these homicides to the collective nature of medieval economic and social life. He also characterized most homicides in this era as spontaneous acts from simple arguments among neighbors. This group and the spontaneous nature of medieval homicides are illustrated by the following transaction: “if John raised his hoe against William, brothers or friends, always nearby, were likely to join in, and the result was often a fracas that ended with a single victim and several killers” (Lane 1997:19). Gurr (1979) also reviewed patterns of violence over time in other Western societies, including other English-speaking countries and regions (e.g., New South Wales, Canada, Republic of Ireland) and Scandinavian countries (Denmark, Finland, Norway, and Sweden). An upsurge in homicide and assault is observed in each of the countries from the 1950s to the mid-1970s. Japan is a major exception to this pattern, reducing by half its rate of violent crime over this same period. LaFree and Drass (2002) provide a more detailed comparison of homicide victimization rates across various countries in the post–World War II period.

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

64

November 3, 2003

RETHINKING HOMICIDE

Figure 4.1. U.S. Homicide Victimization Rates over Time. Source: NCHS (2000)

retained the following structural characteristics: males and young persons as victims and offenders, primary relationships between parties, interpersonal disputes as the motive, and knives and blunt objects as the primary lethal weapon (Lane 1997).

Homicide Trends in American History Compared to England, where more complete historical data are found, the absence of systematic records until the twentieth century makes it difficult to judge the accuracy of inferences and attributions about long-term changes in American homicides. Figure 4.1 graphically represents the homicide victimization rates per 100,000 in the United States over the twentieth century. The source of these data is the U.S. Vital Statistics from the National Center for Health Statistics (NCHS series). As shown in Figure 4.1, homicide rates in the United States exhibit distinct patterns over time. A U-shaped trend is clearly evident in the mid-century period. Homicide rates increased over the first third of the century, dropped in the mid-century before rising sharply in the 1960s, and fluctuated but remained relatively high during the last quarter of the century. Homicide

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

November 3, 2003

65

victimization rates have decreased dramatically throughout the 1990s. Current data indicates that homicide rates have turned upward again in the first couple years of the twenty-first century. The examination of homicide rates for particular U.S. cities over the last 150 years reveals patterns of both common and unique histories (see Monkkonen 2001, 1994).3 For example, many large cities (e.g., Boston, Chicago, Philadelphia, St. Louis) had relatively stable homicide rates in the 1860–1900 time period, followed by a sharp increase in the next two decades. These cities also exhibited a dramatic J-shaped trend in the last half of the century, with low rates in the 1930–50 time period followed by a dramatic increase throughout the 1960s and high but fluctuating rates in the 1970–96 time period. In contrast, other cities (e.g., Milwaukee) had low and remarkably stable rates from the 1870s until an abrupt rise in the 1960s. Washington, DC had a pattern similar to Milwaukee except for an incredibly high rate at the start of its series in 1860. Cities like San Francisco exhibited another pattern, with a less dramatic increase in the 1960s and generally higher homicide rates in the 1860–1920 period than found in the 1930–95 period. Conclusions about homicide that derive from these data, however, are limited because only a small part of American history is covered and the context for understanding homicide is ignored in these quantitative tallies. Drawing upon the extensive historical literature (see especially Brown [1975], Lane [1997] and Monkkonen [1991]), we present a brief description of the wider social context of homicide in America over the last three centuries.

Homicide in the Seventeenth and Eighteenth Century Instances of homicide are clearly found in narrative accounts and court records of Colonial America. However, definitions of homicide varied across the colonies, and different settlements may have been more or less susceptible to violence. The histories of violence in the early colonies varied dramatically for settlers of New England and the colonial South. Puritan beliefs and values greatly shaped the communal tolerance for deviance and its control in the Massachusetts Bay colonies (see Erickson 1966). Plymouth Colony and other New England settlements were largely closed societies. Few people were allowed to live outside communal boundaries and the settlers had little in terms of portable wealth that might encourage economically motivated predatory acts of violence. Both the absence of large plantations to produce crops like tobacco and the Puritan goal of maintaining Godly communities 3

The following summary trends reported here for select U.S. cities are based on a reanalysis of data provided by Monkkonen (1994).

17:23

P1: JMT 0521832993c04

66

CB617-Miethe-v1

November 3, 2003

RETHINKING HOMICIDE

made the New England colonies more reluctant to admit either “rowdy bachelors” or “Pagan Africans” into their settlements (Lane 1997:47–48). Farmers in this region were more likely to work alone compared to southern farmers, decreasing the chance of disputes and trivial arguments about coworkers’ payments and responsibilities. Puritan leaders viewed sin and religious dissent as the major threat to community tranquility. In fact, more women accused of being witches and Quakers who preached heretical doctrines were hanged than murderers in the Massachusetts colonies prior to 1660 (Lane 1997). The relatively unique historical pattern in the New England colonies of executing persons for such acts as infanticide, adultery, and bestiality is explained by this association of sin with crime. Servants, especially poor unmarried women, were the people most commonly accused, convicted, and hanged for infanticide in this time period. Given this profile of the New England colonies, it should not be surprising that homicide was a relatively rare occurrence. The lack of portable wealth and the closed communal boundaries provided little opportunity for bandits or outlaws. Strong and homogeneous religious beliefs served as an effective mechanism of social control for various types of deviance. Most homicides involved people of equal status (e.g., neighbors, companions, family members) and fellow servants, laborers, and small farmers. While court records are widely known to undercount homicide occurrences (especially those killings by high-status offenders against low-status victims), a homicide rate of less than 1 per 100,000 annually is produced from court convictions in Massachusetts (Lane 1997). Life in the Southern colonies differed in several fundamental ways from New England. First, there was greater exploitation of servants and slaves in the Southern colonies to farm the land, grow tobacco, and work semitropical crops like rice and indigo. Second, various marginalized individuals from England (e.g., rogues, beggars) came to these settlements as indentured servants who would receive parcels of land when their terms of servitude ended. Death rates from overwork, malnutrition, and general ill-treatment among these poor and exploited young men approached 80% annually, forcing landowners to constantly import new laborers (Lane 1997). The end result of these actions was greater population diversity and the development of a wide class of young bachelors with few bonds to conventional institutions. Even though their social and economic conditions were quite different, the Southern colonies had a history of violence which paralleled the pattern in New England. Most homicides in the South and North involved social equals (e.g., fellow servants, Blacks, Native Americans, family members). Cases of Native Americans attacking White settlers occurred and were followed by serious retaliation by both groups of settlers. Pocahontas’ uncle,

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

November 3, 2003

67

for example, broke peace in 1622 with a surprise attack against the Jamestown settlement, killing nearly 350 people, roughly one-third of the White population (Lane 1997:37). Seventeenth-century New Englanders like John Eliot, Roger Williams, and other Puritan leaders had established peaceful relations with the Native Americans, but greed and racial prejudice ultimately eroded these ties and subsequently lead to the King Philip’s War of 1675–76 (Brown 1975). Pontiac’s Rebellion in Pennsylvania in 1763 followed a similar process of friendly White–Native American relations turning deadly. As a measure of homicide’s prevalence, court data are often criticized because they exclude killings that are undetected, have no suspects, and those deemed justifiable through extra-legal and nonadjudicative processes. However, Lane (1997:48) argues that undetected homicides may be less common in colonial times, and especially in New England, for two reasons. First, tight neighborly vigilance assured that little deviance would go undetected. Second, the New England justice system was thought to have taken more seriously its responsibility for the lives of the disadvantaged, as indicated by masters being tried for murdering servants and Whites being tried for killing Native Americans. Consequently, available court data indicate that homicide was comparatively rare in the early American colonies and not a major threat to social order. Trends in prosecution rates for violent offenses varied dramatically across cities in the 1700s. New York City, for example, began this century with a relatively high crime rate, experienced a lower period in the 1720s, which was followed again by a rise in the 1730s and accelerated growth after 1760. The proportion of prosecutions for violent crimes more than doubled over this century (Bridenbaugh 1955). Philadelphia’s rates of prosecutions for serious crimes (e.g., robbery, rape, assault, and murder) have been estimated to be high and relatively flat for the first half of the eighteenth century, but also increased dramatically after 1760 (Brantingham and Brantingham 1984). While exhibiting the same upturn at the middle of the century, Charleston, South Carolina had lower crime rates than other cities throughout this time period. However, as reflective of the different religious traditions and economic bases, the majority of prosecutions in Charleston in the late 1700s were for violent offenses, whereas crimes of order and morality (e.g., bastardy, fornication, vagrancy) were the major source of prosecutions in the northern colonies (like Middlesex County, Massachusetts) in the same era (Hindus 1980). Prior to the American Revolution and shortly thereafter, violent acts often took the form of vigilantism directed at particular groups and policies. For example, armed bands known as the Paxton Boys murdered peaceful “praying Indians” in villages near Philadelphia in the 1760s, whereas surprise attacks of isolated settlements and cabins were a common tactic of the

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

68

November 3, 2003

RETHINKING HOMICIDE

Woodland Native Americans (Brown 1975; Jacobs 1972). In contrast, those who objected to new taxes and other imperial policies burned houses, trashed furniture, and laid on tar and feathers (Lane 1997:70). Not counting Native American attacks and deaths during the American Revolution, it seems clear that most disputes and conflict with the English authorities involved nonlethal violence. The Boston Massacre of 1770, resulting in the death of five Americans by British troops, was a clear exception to this pattern. Violence became a dominant theme in America at the end of the eighteenth century. Charges for assault and battery were, by far, the most common basis for indictments in a study of court records in Ohio County, Virginia in 1801–10, whereas charges for murder were rarely issued during this time period (Steel 1991). Acts of collective violence such as Shays’ Rebellion in Massachusetts (1786–87), the Whiskey Rebellion in western Pennsylvania (1794), and Fries’ Rebellion in eastern Pennsylvania (1798– 99) also dotted the American landscape in this period. The Harpe brothers, among the earliest serial murderers in American history, brutalized anywhere between 20 to 38 victims in the frontier states of Kentucky and Tennessee in 1798 and 1799 (Brown 1975).

Homicide in the Nineteenth Century Two critical historical events had a major impact on violence in America in the nineteenth century: the Civil War and expansion of the Western frontier. Relatively unique forms of violence (like assassinations, urban riots, and vigilante groups) also emerged during this century. By all authoritative accounts (see Lane 1997; Brown 1975), the era of the Civil War and Reconstruction was a tumultuous period of violence that had an overwhelming impact on subsequent decades. The last half of the nineteenth century was one of the most violent times in American history, “an era of Ku Kluxers, lynch mobs, White Caps, Bald Knobbers, night riders, feudists, and outlaws” (Brown 1979:22). The growth in violence and the activities of particular dissent groups are directly tied to the Civil War and the increased acceptance of violence as a method of conflict resolution. Major changes occurred in American society throughout the early to mid1800s that had profound implications on crime and its social control. Cities in the East and Midwest had unprecedented growth, fueled by enormous European immigration. Immigration increased dramatically population diversity in U.S. cities, but it also provided a new context to play out unresolved conflicts based on ethnicity and religion. The Catholic Irish migrants, in particular, were especially notorious for their violence and hostility toward official authority, anger linked to their persecution for centuries by AngloSaxon Protestant rulers (Lane 1997).

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

November 3, 2003

69

Other social changes and major reform efforts also took place in this time period. A temperance movement began in the 1820s to restrict the use of alcohol. The first public school system was created in this era, emphasizing discipline and the development of a responsible, orderly population (Lane 1997). Penitentiaries were developed on the assumption that human behavior was malleable through the forced teaching of virtue and hard work. Urban police forces were established in the mid-1800s to deal with mob riots and the enforcement of temperance practices. The rapid growth of an urban commercial economy resulted in more impersonal labor relations and a change in farm practices to better coordinate economic activities. Small handguns (like Colt’s revolver and Derringer’s pistol) began to be manufactured in quantity in the mid-1840s, providing an easily portable weapon for both self-protection and starting trouble. Each of these developments provided emergent forms of social control in an increasingly diverse and complex society. While the control of young males has been a persistent problem throughout history (McCorkle and Miethe 2001), the “bachelor” subculture was the major source of violence in the Antebellum and post–Civil War eras. Composed of adult men under 35 years old, this large group of unattached, low-income wage earners lived in boarding houses filled with others similarly situated who had essentially no supervision or responsibilities (Lane 1997). As “muscle” for local authorities, members of the bachelor subculture were often recruited to help maintain order or formed special alliances for vigilante activities. The lifestyles of these young men (e.g., gambling, drinking in saloons and whorehouses, fighting), however, also constantly placed them in dangerous situations. Coupled with their involvement in vigilante activities, it isn’t surprising that homicide in the mid-1800s has been considered “almost wholly the business of pugnacious young unmarried males” (Lane 1997:126). The relatively low rate of homicide among civilians during the Civil War period is often attributed to the wide recruitment of these men into military service. The pre–Civil War era was characterized by rising violence in both the North and South (see Brown 1975). During the frontier expansion, local citizens increasingly took matters in their own hands and formed vigilante committees when formal authority was absent or deemed incapable of resolving the problem. Organized attacks on outlawry was initially pioneered by the South Carolina Regulators and confined to a small number of states. Vigilante committees focused on issues such as crime, race, religion, political elections, and land rights. It is estimated that hundreds of vigilante groups sprang up in the pre– and post–Civil War period. Examples of these groups include the exploits of John Brown and the abolitionists, the Mollie Maguires, the Committee of Vigilance in San Francisco, the Ku Klux Klan (KKK), and less organized

17:23

P1: JMT 0521832993c04

70

CB617-Miethe-v1

November 3, 2003

RETHINKING HOMICIDE

lynch mobs. Illustrating the diversity of views toward group violence, John Brown’s raid at Harper’s Ferry was alternatively viewed as either an inspiration for a holy war against slavery or proof that the South could never coexist in an union that included free states and harbored abolitionists (Brown 1979:22). Violence in the nineteenth century was expressed in different forms. The Civil War wreaked mass destruction and death across the country and was a precipitating factor in many subsequent conflicts. William Clarke Quantrill used the war as an opportunity for seeking revenge and profiteering. His followers (e.g., the James brothers, the Youngers, the Daltons) achieved their own notoriety in the Wild West. The tradition of feuding in the Appalachian Mountains (e.g., the Hatfield–McCoy feud, the Martin–Tolliver and Hargis–Cockrell feuds of Eastern Kentucky) and the Southwest (e.g., the Sutton–Taylor feud, Graham–Tewksbury feud in Arizona) was fueled in many cases by the animosities invoked by the Civil War when local residents were divided in their Confederate and Union loyalties (Brown 1979). Lynch mobs were evident prior to the Civil War, but urban labor riots and vigilante groups such as the KKK received far greater prominence in the postwar period. Formal dueling was virtually eliminated by the end of nineteenth century, and assassination emerged as a new form of ideologically based killing (Brown 1975). Regionally distinct rates and patterns of violence resulted from racial tensions in the South, labor problems in the North, and Native American warfare and unsettled conditions in the West (Lane 1997:146). The expansion of the Western frontier was associated with armed confrontations with various Native American tribes and growing interpersonal disputes among settlers. Western migration involved vicious attacks on both sides. The Black Hawk War of 1832 nearly annihilated the entire Sac and Fox Native Americans of Illinois. Under the leadership of Chief Little Crow, a major Sioux uprising in 1862 in the new state of Minnesota resulted in the massive slaughter of local traders and farmers. The official records indicated 644 civilian deaths, with the amount of “rape, mutilation, and butchery of children exceptional even by the fearsome standards in that tradition” (Lane 1997:167). Colonel J.M. Chivington lead an attack by local militia on a peaceful Cheyenne village of mostly women and children at Sand Creek, Colorado, that killed nearly three hundred people. With few exceptions (e.g., Custer’s demise at the Little Bighorn River in 1876), large-scale cavalry charges and extensive battles were not common in the Western territory. Isolated farmers and wagon freighters were the prime targets for Native American attacks. Most Native Americans were disarmed and largely banished to reservations by the late 1880s. The “cowboy” tradition began after the Civil War and involved rounding up cattle in South Texas and then herding them hundreds of miles across the open plains to the railheads of Kansas (Lane 1997:170). Cattle towns

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

November 3, 2003

71

like Abilene and Dodge City have often been characterized as drunk and disorderly, and some authors conclude that these towns were especially violent and had relatively high rates of homicide in their cattle-trading period. McKanna (1997), for example, estimates that Abilene in the early 1870s had a homicide rate of about 80 per 100,000 and Dodge City reached a rate of 160 per 100,000 in 1878. Mining towns like Bodie, California, in the late 1870s and early 1880s also had enormously high homicide rates, estimated at 116 per 100,000 annually. The high level of violence in these frontier towns was attributed to a dominant male culture characterized as “young, single, intemperate and armed” (Lane 1997:171). Due to the limitations of available data, the actual level of violence on the Western Frontier is subject to debate (see Drago 1970; Friedman and Percival 1981; McGrath 1984; Prassel 1972; Rosa 1969). Nonetheless, a common trend based on the historical records is that most killings in this period involved drunken brawls and fights among willing parties, often motivated by the steadfast refusal to back down when challenged. The high public tolerance for these character contests among men is indicated by the wide acceptance of self-defense as a legitimate justification for the offense. Collective acts of violence through the Civil War, vigilante groups, and the Native American attacks characterized much of the nineteenth century. The absence of law enforcement authority was a major precipitating factor in the emergence of vigilante justice, while the typical homicide on the Western frontier involved intraracial disputes, neighborhood quarrels, and tavern brawls among young unattached males. The Wild West bandits and gunfighters received wide notoriety after the Civil War, but these infamous characters were relegated to the status of historical legends by the turn of the century. Most authoritative accounts suggest that rates of “ordinary,” interpersonal homicide decreased after the mid-century.

Homicide in the Twentieth Century Our ability to describe homicide trends and patterns in the United States has been greatly enhanced in the twentieth century. Multijurisdiction mortality statistics have been compiled as early as 1900, although national trends based on these data prior to 1930 are considered unreliable because of changes in the size of the area of coverage (see Gurr 1981; Riedel 1999). National estimates of the number of homicides known to the police have been available since 1933 through the Uniform Crime Reports compiled and distributed through the FBI and other law enforcement agencies (see Mosher, Miethe, and Phillips 2002). As noted earlier in this chapter, the U.S. homicide rate has gone through various waves and spikes in the twentieth century. Homicides tended to increase after 1900 to a peak in the early 1930s, dropped until the mid-1940s,

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

72

November 3, 2003

RETHINKING HOMICIDE

temporarily increased after World War II, dropped again in the 1950s, dramatically increased from the early 1960s until 1980, decreased to the mid1980s, generally increased again until the early 1990s, and has consistently decreased over the last half of the 1990s (see FBI 2000; Zahn and McCall 1999). Several historical events and circumstances have contributed to the changing volume of homicide in this century. The prohibition era from 1919 to 1933 provided the context for deadly battles between bootleggers and law enforcers, and bloody skirmishes between rival suppliers of alcohol to “speak easy” establishments. Competing organized crime families waged vicious attacks on each other in efforts to control illegitimate and legitimate business markets. Bank robbers and bandits like John Dillinger, Pretty Boy Floyd, and Bonnie and Clyde received some admiration for their Robin Hood–like image, but they were dubbed by law enforcement as “public enemies” of the 1930s. Civil unrest reached a peak in the mid-1960s, with multiple deaths occurring in the Watts (Los Angeles) riots of 1965, the Newark riot in 1967, and the Detroit riot in 1967. Assassinations of President John F. Kennedy, Robert Kennedy, and Martin Luther King cut short dreams of peace, prosperity, and racial equality (Lane 1997). Both mass murders and serial murders by lone offenders emerged in the last three decades as relatively new forms of homicide. Terrorist attacks by disgruntled Americans and international terrorist groups now pose an increasingly dangerous threat to the quality of life. Public attention in the last half of the century has been directed at other forms of homicide. Marital violence and the battered child syndrome are now recognized as major public health problems. Sibling abuse, teen killers, gang violence, school shootings, drug-related homicides, and elderly offenders involved in acts of mercy killings are common themes in contemporary media depictions of homicide. The methods (e.g., bombings, type of gun use) and motives (e.g., “thrill kill,” robbery, expressive violence) of modern killers have also been well documented (see Block and Christakos 1995; Katz 1988; Smith and Zahn 1999).

Offender, Victim, and Offense Elements in Homicide over Time All homicides involve the convergence of offender, victim, and offense elements in time and space. Common and distinct historical patterns in these offender, victim, and offense attributes are described below.

Offender Characteristics According to SHR data for the 1990s, homicide offenders are disproportionately male, young, and African American. Spatial analysis of crime locations

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

November 3, 2003

73

further suggests that homicide offenders are also more likely to be poor and economically disadvantaged (see Harries 1997; Martinez and Lee 1999; Short 1997; Wolfgang 1958). Australian and Canadian data reveal a similar pattern of disproportionate offending risks for males, the young, and particular ethnic or racial minorities (see Mouzos 2000; Polk 1994; Silverman and Kennedy 1993). Although female offending has received a great deal of recent attention, the vast majority of known homicides involve male perpetrators, and this pattern has been stable over time. The historical record is filled with accounts of female murderers, especially involving cases of infanticide and spouse slayings.4 However, killings by women are clearly the exceptional case even back to the Middle Ages (Lane 1997). Women murderers were sparse in Colonial America, rarely acknowledged in the 1800s, and accounted for only a small proportion of homicides in the 1900s. Approximately 90% of the murderers committed to prison in the 1920s were male (Brearley 1932). Males also represent about ninety percent of homicide arrestees in recent UCR data (FBI 2001). At least 85% of the homicide offenders in Canada and England are male (Brantingham and Brantingham 1984: Silverman and Kennedy 1993). As is true of other crimes, persons under 30 years old are the most common offenders in homicide. Based on our reanalysis of McKanna’s data (1997), persons in their twenties were the modal category for those accused of homicide in California in the mid- to late 1800s. Brearley (1932:82) estimates that the average murderer in the 1920s is about 30 years old. Over half of homicide arrestees in 1998 were under 25 years old (FBI 1998). Historical narratives in the eighteenth and nineteenth century make reference to homicide being nearly the exclusive domain of young males (see Lane 1997). Over the last three decades, however, increased attention has been given to the rise in homicide by teenagers. Persons under 18 accounted for less than 10% of the total homicide arrestees in 1974, but this proportion increased to 15% by 1998 (FBI 1998; FBI 1974). SHR data indicate persons under 20 years old represent about one-quarter of homicide offenders in the 1990s (see Figure 4.2). While accounting for only about 1% of all homicide arrestees, killings by juveniles under 15 years 4

The enormous undercounting of infanticide throughout early American history may also dramatically minimize the prevalence of female offenders. However, at particular times (e.g., since the 1900s) and in particular places (e.g., Massachusetts colonies), better records are available that suggest that infanticide was relatively uncommon. Lane (1997) notes that less than two percent of U.S. homicides in the 1920s were infanticides, attributing this low rate to the growth of hospital births (which provided greater health care and witnesses) and the invention of pasteurized milk (which made it easier for young women to leave infants at the poorhouse or foundling home). When infanticide is excluded, women are clearly underrepresented as homicide offenders throughout history.

17:23

P1: JMT 0521832993c04

74

CB617-Miethe-v1

November 3, 2003

RETHINKING HOMICIDE

Figure 4.2. Offender’s Sex and Age

old have also increased over the last three decades. Murders committed by juveniles, however, have occurred throughout American history. Unfortunately, these deaths have increased in frequency in modern times and are being perpetrated by an increasingly younger offender population (Fox 1996). Minority overrepresentation, especially among African Americans, has characterized U.S. homicide arrestees over the last century. Brearley (1932) found that about half of the persons sent to prison for homicide in the 1920s were African Americans. This racial group accounted for nearly 56% of homicide arrestees in 1998 (FBI 1998). Rates of homicide by African Americans were lower in the nineteenth century and earlier time periods. Other ethnic groups like the Irish and Italians have been identified in the late nineteenth and early twentieth century as being especially prone to violence (Lane 1997). SHR data indicate that the proportion of homicides involving Black offenders has remained relatively stable at around 50% over the last three decades (see Figure 4.3). Another enduring characteristic of homicide offenders is that they come disproportionately from lower and working class backgrounds. The low social class of homicide arrestees and persons in prison for homicide has been widely noted across various historical periods from Colonial America onward

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

November 3, 2003

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

75

Figure 4.3. Offender’s Race

(see Brearley 1932; Lane 1997; McKanna 1997; Monkkonen 1991). Class biases in the definition of homicide and legal treatment of offenders may account for some of these differences. However, other theories (e.g, strain theory) place blame on the anger and frustration associated with being economically disadvantaged as the explanation for the predominance of lower income persons as homicide offenders. Aside from these univariate distributions, it is also the case that particular combinations of offender attributes exhibit stability and change over time. For example, the bachelor subculture, comprised of young, single, low-income males, has remained the most notorious category of homicide offenders in Colonial America, the Antebellum period, the post–Civil War and period of Western expansion, the times of riots and demonstrations in the 1960s, and the era of gang violence in central cities from the 1960s to the present. However, the racial and ethnic composition of the members of this male subculture (e.g., from Irish to African American and Hispanic), their motivations for offending (e.g., expressive or instrumental goals), and their relative prevalence in different regions of the United States (e.g., from the South to Northern cities) have changed over time.

Victim Characteristics It is widely assumed in previous studies that criminal offenders and victims possess many of the same characteristics (see Brearley 1932; Sampson and

17:23

P1: JMT 0521832993c04

76

CB617-Miethe-v1

November 3, 2003

RETHINKING HOMICIDE

Figure 4.4. Victim’s Sex and Age

Lauritsen 1990, 1993; Singer 1981). Available data also support this image of homogeneous pools of victims and offenders in homicide. Similar to homicide offenders, SHR data for the 1990s indicate that homicide victims are disproportionately male, young, and African American. Over three-fourths of the homicide victims are male, about one-half are under 30 years old, and slightly under one-half of them are African Americans (see Figures 4.4 and 4.5). These demographic attributes of homicide victims are similar across the twentieth century. For example, Brearley (1932) found that in the mid1920s about 80% of homicide victims were male, about 43% were African Americans, and that individuals 25 to 30 years old were the most frequent victims. Similar gender and age differences in homicide victims are also derived from historical accounts of earlier centuries, but the data are less reliable and comprehensive (see Lane 1997; Monkkonen 1991). The harsh treatment of slaves in Colonial America and the Antebellum era, as well as the racially motivated lynch mobs in the post–Civil War period, suggest a pattern of disproportionate representation by African Americans as homicide victims throughout most of American history. The principle of homogeneity between victims and offenders also appears to characterize social class differences in risks of homicide victimization. Given their close physical proximity to homicide offenders and their

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

November 3, 2003

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

77

Figure 4.5. Victim’s Race

social marginality, persons who are poor, immigrants, or members of the underclass should be especially vulnerable to homicide. The overrepresentation of low-income persons and recent immigrants as victims of violence has been widely acknowledge in historical accounts across centuries (see Brown 1975; Lane 1997; Monkkonen, 1991, 1975). Although victimization of the “rich and famous” grabs the media spotlight, most homicide victims in the past and present come from inner city areas characterized by low income and social disorganization.

Offense Elements Elements of the offense involve the characteristics and activities surrounding the criminal act. This category includes the motivation for the crime, interpersonal dynamics (e.g., the victim–offender relationship, intraracial versus interracial crimes), the type of weapon used, the number of offenders, the presence of alcohol and drugs, and the spatial and temporal location of the offense. The motivation for the criminal act is an important element in understanding homicide. Previous research often dichotomizes motive into “instrumental” and “expressive” categories. Instrumental homicides are those conducted for explicit, future goals (as to acquire money or improve one’s social position), whereas expressive homicides are often unplanned acts of anger, rage, or frustration (see Block 1976; Block and Christakos 1995; Block

17:23

P1: JMT 0521832993c04

78

CB617-Miethe-v1

November 3, 2003

RETHINKING HOMICIDE

Figure 4.6. Expressive and Instrumental Motives

and Zimring 1973; Decker 1993). Although they may have elements of anger and rage associated with them, most homicides that occur during the commission of other felonies like robbery, rape, burglary, drug sales, and “car jackings” are usually considered instrumentally motivated homicides. Based on SHR data for the 1990s and the classification above, the majority of U.S. homicides involve expressive motivations (see Figure 4.6). Arguments were the single most common motive for expressive homicides. Among the instrumentally motivated homicides, robbery was by far the most common motive, followed by narcotic law violations, burglary, and rape (see Figure 4.7). The prevalence of arguments and brawls as the primary motive for homicide has a long history. Altercations and brawls were the most frequent classification for 760 murders in Chicago during the mid-1920s (Brearley 1932). Mutual quarrels and brawls are widely noted in the literature on cattle towns and boom towns on the Western frontier (see Lane 1997; McKanna 1997). Descriptions of murder in medieval England and Colonial America also emphasize arguments and quarrels as the primary motivating factors (see Brantingham and Brantingham 1984; Lane 1997). The frequency of robbery and holdups as the precipitating motive for homicide appears to vary dramatically over time. The absence of portable wealth contributes to the apparently low rate of robbery-homicide in Colonial America (especially the Massachusetts colonies). Contrary to popular images of the murderous sprees by Western bandits and outlaws, our reanalysis of McKanna’s data on California homicides from 1850 to 1899 reveals

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

November 3, 2003

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

79

Figure 4.7. Types of Instrumental Motives

robbery as a motive in only about 13% of cases. Over one-third of the homicide victims in Chicago during 1920 were slain while resisting a burglary or robbery (see Brearley 1932), whereas robbery and burglary were the precipitating circumstance in less than twenty percent of murders in the 1990s. Thus, in contrast to the relatively stable trend for arguments and brawls, the prevalence of instrumental homicides has varied noticeably over the past three centuries. The victim–offender relationship has long been recognized as a distinguishing element in homicide. Comparisons are often made about the relative frequency of homicides by family members, acquaintances, and strangers (see Decker 1993, 1996; Riedel 1993; Williams and Flewelling 1987). After the exclusion of cases in which the victim–offender relationship is unknown, slightly more than one-half of the homicides in the SHR files in the 1990s involved acquaintances. Family members were involved in approximately one-quarter of the homicides in this time period, whereas the victim and offender were strangers in slightly less than one-quarter of the cases (see Figure 4.8). Historical narratives and court records suggest that the relative frequency of different types of victim–offender relationships in homicide has changed over time. Court data on Philadelphia homicides from 1839 to 1901 reveal patterns similar to those in the 1990s, with nearly 50% committed by

17:23

P1: JMT 0521832993c04

80

CB617-Miethe-v1

November 3, 2003

RETHINKING HOMICIDE

Figure 4.8. Victim–Offender Relationship

acquaintances, roughly twenty percent by family members, and the remaining thirty percent involving strangers (Lane 1979). Wolfgang’s (1958) study of Philadelphia homicides for the period 1948 to 1952 revealed a similar pattern for family homicides (25%), but a substantially higher proportion by acquaintances (61%) and a lower proportion by strangers (14%). Instances of domestic homicide (e.g., infanticide, spouse slayings) and stranger attacks (e.g., robberies) are clearly contained in historical accounts of the Colonial and Antebellum eras. Unfortunately, data for these time periods are insufficient for estimating the relative prevalence of particular types of victim–offender relationships. The available data on homicide over the last century suggest the following trends in victim–offender relationships. First, homicides by friends and acquaintances remain the modal category throughout this century. Second, homicides involving family members have generally declined over time, with SHR data showing a decrease from 29% to 25% from the 1970s to the 1990s. Third, lethal violence by strangers has become more common over time, increasing from about one-eighth of Philadelphia homicides between 1948 and 1952 to about one-fifth of all U.S. homicides in the 1970s and to slightly less than one-fourth of U.S. homicides in the 1990s. These proportions of stranger homicides are slightly higher than observed when missing data are imputed, supporting the claim of some previous researchers that homicides with an “unknown” victim–offender relationship are disproportionately stranger homicides (see Messner, Deane,

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

November 3, 2003

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

81

Figure 4.9. Weapon Used

and Beaulieu 2002; Pampel and Williams 2000; Regoeczi and Riedel 2003; Williams and Flewelling 1987). SHR data indicate that homicide is primarily an intraracial crime, a pattern which has remained fairly stable over time. In fact, nearly 90% of U.S. homicides in the 1970s, 1980s, and 1990s are intraracial. Over 90% of the homicides in Philadelphia between 1948 and 1952 were intraracial (Wolfgang 1958). Over 80% of the African-American victims in Washington, DC and Chicago in the late 1910s and early 1920s were slain by African Americans (Brearley 1932). However, homicides during race riots, the Antebellum period, and the over 3,700 lynching incidents by vigilante groups from 1880 to 1930 typically were interracial in nature (see Brown 1975; Lane 1997). The vast majority of contemporary U.S. homicides involve the use of firearms, far outdistancing knives, other sharp objects, and other weapons (see Figure 4.9). This represents a clear historical change. Lane (1997) notes that only about 15% of Philadelphia homicides in the 1840s were committed with a firearm, a product of the expense of firearms, the wide availability of other weapons (like sticks, clubs, bricks), and the spontaneous quarrels, brawls, and arguments that characterized most homicidal events at this time. Studies also suggest that firearms are less often the weapon of choice for homicides in other Western cultures (see Brantingham and Brantingham 1984; Polk 1994; Silverman and Kennedy 1993). The typical homicide occurs in the context of a single offender and single victim. Less than ten percent of homicides in the 1970s, 1980s, and 1990s involved multiple victims (see Figure 4.10). Instances of gang violence and other collective acts of homicide are widely publicized, but these killings

17:23

P1: JMT 0521832993c04

82

CB617-Miethe-v1

November 3, 2003

RETHINKING HOMICIDE

Figure 4.10. Multiple Offenders and Multiple Victims

in contemporary American society represent only a minority of homicides. However, the prevalence of killings by multiple offenders has changed somewhat over time, increasing from about eight percent in the 1970s to about thirteen percent of homicides in the 1990s. Killings in riots and by lynch mobs, by definition, involve multiple offenders, and have had a long history in this country. Nonetheless, most acts of interpersonal violence involve disputes between one offender and one victim. A majority of homicides involve the presence of drugs or alcohol in either the victim, offender, or both parties (see Auerhahn and Parker 1999; Parker 1995; Parker and Rebhun 1995; Wolfgang 1958). While narcotic law violations are a relatively infrequent motive in homicides in the SHR data, disputes over drug transactions and crimes committed under the influence of drugs are widely recognized as criminogenic conditions. Information on drug and alcohol use by offenders is not usually contained in either court or police records of homicide, making it difficult to draw inferences about the role of substance abuse over time. Nonetheless, frequent historical references to the “drunken brawls” in the Wild West and the finding that more than one-third of the California homicide victims and offenders in the years 1849 to 1899 were drinking at the time of the offense (McKanna 1997) gives some indication of the longstanding involvement of alcohol in homicide situations. Homicide does not occur equally over time and space. The disproportionate occurrence of homicides and other violent crimes in evening hours and on weekends has been well established in previous research (see

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

November 3, 2003

83

Miethe and McCorkle 2001; NCHS 1993; Perkins and Klaus 1996; Wolfgang 1958). These patterns also hold over time. For example, over one-half of the California homicides in the latter half of the 1800s occurred between 6:00 p.m. and midnight. Saturday and Sunday were also the most lethal days in the mid-1800s (see Lane 1997). The street was the most common location for homicide in Philadelphia and California in the mid-1800s. However, the victim’s home is the most common location for homicide in more recent studies (Miethe and McCorkle 2001; Wilbanks 1984; Wolfgang 1958). The high rate of homicide offending in contemporary urban centers represents a pattern of both stability and change over time. Over half of the homicides in the 1990s occurred in urban areas with a population of greater than 100,000. Groups of young men have terrorized urban residents with threats of homicide and other violent crimes since the late 1880s, and the greater risk of homicide in modern cities is an extension of that historical legacy. Prior to this time period, however, rural settlements and people that lived on the Western frontier were at far greater risk of becoming homicide offenders and victims than the residents of the “civilized” eastern cities (Lane 1997). The lawlessness of these settlements contributed to the use of homicide as a method of dispute resolution.

Homicide Situations in SHR Data, 1976–985 As described in previous chapters (see Chapters 1 and 2), the structure of homicide situations is defined by combinations of offender, victim, and offense elements. Measures of these structural attributes are contained in the SHR data for the late 1970s (1976–79), the 1980s, and most of the 1990s (1990–98). A total of 439,954 homicides are contained in these SHR data from 1976 to 1998 (FBI 1976–98). The number of homicides for each time period includes 73,942 for the late 1970s, 189,014 in the 1980s, and 176,998 in the 1990s. Missing data on several variables (e.g, circumstance underlying the crime, the victim–offender relationship) dramatically reduce the number of homicides with complete information. The results presented below include imputed data, based on the expectation–maximization algorithm described in Chapter 3. For this examination of homicide situations over time, offender and victim characteristics include their gender, age, and race. Age is coded in three categories to compare youth (persons under 20) with two groups of adult 5

Detailed justifications for the use of SHR data and the coding of particular variables are included in other chapters. Specifically, Chapter 3 described the SHR data, whereas Chapters 5 through 9 provide justifications for the coding of particular offender, victim, and offense characteristics.

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

84

November 3, 2003

RETHINKING HOMICIDE

offenders/victims (i.e., 20–29 year olds vs. 30 and older). Victims and offenders are defined as the same age if they are within one age category from each other. Race of the victim and offender are coded as White, African American, and other (e.g., Asian, Native American). Hispanics are excluded in this analysis because data on ethnicity were not included throughout the entire SHR series. Offense elements involve the circumstances surrounding the homicide. These include the victim–offender relationship (coded as strangers, acquaintances, or family/intimates), number of co-offenders (coded as single or multiple offenders), type of weapon (coded as firearm or other method), the location of the offense (coded as large cities [population >100,000] or less urban areas), and motive (coded instrumental, expressive, or “other”). Instrumental motives involve all felony-type circumstances in the SHR data, including robbery, rape, burglary, drug offenses, motor vehicle theft, and other felony situations. Expressive homicides include arguments, brawls, romantic triangles, and youth gang slayings. The “other” category includes a variety of situations such as killings by babysitters, sniper attacks, killings within institutions, gangland killings, and additional circumstances not covered by the instrumental and expressive categories. As indicated earlier, the univariate frequency distributions for each offender, victim, and offense characteristics are similar for each decade. Males and African Americans are clearly overrepresented as both offenders and victims across each decade. The modal age category for both offenders and victims involves persons in their 30s and older. Over three-fourths of the parties are nonstrangers across all decades, with acquaintances representing the dominant category. For other offense variables, the modal categories across all decades are single offenders (89%), expressive motives (54%), firearms as the lethal weapon (66%), and urban location (66%). Only small changes in these univariate distributions have occurred over time. The first step in the analysis of the structure of homicide situations involves documenting the observed number and relative frequency of combinations of offender, victim, and offense characteristics. Accordingly, we have identified (1) the number of different homicide situations per decade, (2) the number of different situations that involve an average of at least 5 homicides per year (applying a minimum frequency rule), and (3) the 25 most prevalent homicide situations that are characterized by unique and common signatures over time.6 6

As discussed in Chapter 3, several different decision rules can be used to count observations in QCA. In the current analysis, a configuration must have a minimum frequency of 5 per time period to be included. This decision rule is used to minimize the effect of random fluctuations and coding errors in the enumeration of observed configurations.

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

November 3, 2003

85

Figure 4.11. Number of Homicide Situations over Time

The Diversity of Homicide Situations One measure of whether homicide situations are becoming more diverse is the relative frequency of their structural profiles over time. Of the maximum 32,768 combinations of structural elements that are theoretically possible with the dummy coding of the 15 variables, only 5,591 of these situational profiles are represented by at least one U.S. homicide between 1976 and 1998 (see Figure 4.11).7 The number of distinct homicide situations per decade increased from 3,457 in the 1970s to 4,609 in the 1980s and 4,705 in the 1990s. When a minimum frequency of 5 homicides per year is used to define prevalent situations, the number of different situational contexts drops dramatically for each decade. In fact, less than 700 different situations are observed each decade under this minimum frequency rule (623, 623, and 656, respectively). Four important conclusions derive from this pattern of homicide situations over time. First, the large number of different homicide situations observed in each decade suggests that previous research has grossly oversimplified the structural diversity of homicide. It is hard to imagine how any existing homicide typology (see Block and Block 1992; Miethe and McCorkle 7

The specific categories of the variables included in the QCA model are offender characteristics (male, <20, 20–29, White, Black), victim characteristics (same sex, same race, same age group), and offense attributes (family/intimates, acquaintances; expressive, instrumental motives; gun usage, multiple offenders, urban area over 100,000).

17:23

P1: JMT 0521832993c04

86

CB617-Miethe-v1

November 3, 2003

RETHINKING HOMICIDE

Figure 4.12. Number of Homicide Cases in Situations over Time

2001) can adequately account for the over 3,400 distinct combinations of offender, victim, and offense attributes that underlie homicides in each decade. Second, homicide situations have become more diversified in their structures over time. This is indicated by the growing number of different structural profiles in the SHR data from the late 1970s to the 1990s. Third, most homicide situations are actually representative of very few homicides. About 82% of the different configurations in the late 1970s represent less than 5 homicides per year, and about 86% of the different situational contexts in the 1980s and 1990s include fewer than 5 homicides per year. However, situational contexts that satisfy our criterion of minimal occurrence account for at least 81% of homicides for each decade (see Figure 4.12). Thus, while most homicide situations are infrequent in their occurrence, the clear majority of homicide cases in each decade are concentrated within a relatively small subset of prevalent homicide situations. Fourth, the complexity of the task of describing and understanding homicide has been greatly reduced as the unit of analysis shifts from individual homicides to “prevalent” homicide situations. For example, for the 1990s this reduced complexity is indicated by the difference between (1) attempting to explain the occurrence of nearly 177,000 homicides and (2) trying to account for the structure of 656 prevalent homicide situations that account for the vast majority of these homicides.

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

November 3, 2003

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

87

Figure 4.13. Profile of Unique and Common Homicide Situations

QCA and the Analysis of Homicide Situations Qualitative Comparative Analysis (QCA) provides a systematic way of determining whether unique or common profiles best represent the structure of homicide situations over the last three decades. We use a 10% difference rule to define unique profiles and apply a minimum frequency rule of 5 homicides per year to focus on the nontrivial and less idiosyncratic situational contexts (see Chapter 3 for a discussion of these rules). Application of the minimum frequency rule across all decades reduces the total number of configurations to 618 prevalent homicide situations. The comparative analysis of these 618 situational contexts is summarized in Figure 4.13 and described below.

The Prevalence of Unique and Common Homicide Situations over Time Analyzing the 618 prevalent homicide situations that satisfy the minimum frequency requirement reveals that over two-thirds of them (n = 448 configurations) involve combinations of offender, victim, and offense attributes that are observed across all three decades. These “common” or “stable” profiles over time represent about 78% of homicides that happen within prevalent situations (see Figure 4.13).

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

88

November 3, 2003

RETHINKING HOMICIDE

Only a small minority of situational contexts are unique for any given decade and these unique profiles represent only a small proportion of homicide cases. It is interesting to note, however, that the proportion of unique profiles (and the proportion of homicides represented by them) has also increased over time (see Figure 4.13). The number of unique profiles and the number of cases explained by them have increased dramatically from the late 1970s through the 1990s (i.e., n = 22 vs. 102 unique profiles and n = 7,426 vs. 54,718 homicide cases). Nonetheless, this observation should not detract from the dominant conclusion that the situational contexts for homicide have largely remained stable over the last three decades.

The Nature of the Unique and Common Homicide Situations over Time Examination of the truth table in a QCA analysis reveals the particular patterns of unique and common signatures that underlie homicide situations. The most prevalent situational contexts that are common over time and unique to each decade are summarized below.

Common Situational Contexts over Time We have just shown that the majority of U.S. homicides occur in situational contexts that have not changed their structural features over the last three decades. A basic question surrounding these stable structural profiles, however, involves the nature of the particular combinations of offender, victim, and offense attributes that define them. The 25 most dominant situational contexts that are represented by stable structures over time are presented in Table 4.1.8 Three major types of homicides contained among these stable structures are killings within the context of (1) interpersonal disputes, (2) domestic violence, and (3) instrumental, felony-type circumstances.9 For each 8

9

These 25 profiles represent far less than 1% of all possible configurations observed in the last three decades (25/5,591), but the descriptive value of this small number of common situational contexts is indicated by the fact that they account for over one-fifth of all U.S. homicides from 1976 to 1998. The three major subtypes of homicide situations that underlie these stable structures over time are derived from the entire set of nontrivial common profiles that represent an average of at least 5 homicides per year (n = 418 common profiles). While our conclusions about common and unique profiles are based on the nontrivial situational contexts that satisfy the minimum frequency rule of 115 cases (i.e., an average of 5 homicides for each year in the 1976–98 period), these conclusions are largely unchanged when the analysis is based on all observed configurations (not just those that satisfy the minimum frequency rule). There are a wide variety of particular types of homicide situations with common signatures over time. We focus here on these three types of homicide situations for illustrative purposes and their general interest to criminologists. Each of these types of homicide situations

17:23

Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale

O3098 O2029 O3098 O3098 O2029 O3098 O3098 O2029 O2029 O2029 O3098 O3098 O3098 O2029 O3098 O3098 O3098 O3098 O3098 O3098 O2029 O3098 O2029 O3098 O3098

Oblack Oblack Owhite Owhite Owhite Owhite Oblack Owhite Oblack Oblack Owhite Oblack Owhite Owhite Oblack Owhite Oblack Oblack Oblack Owhite Owhite Owhite Owhite Owhite Oblack

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex

=Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race Acquaint Acquaint Acquaint Acquaint Acquaint Family/Intim Acquaint Acquaint Acquaint Stranger Acquaint Acquaint Family/Intim Acquaint Family/Intim Family/Intim Stranger Family/Intim Stranger Stranger Acquaint Family/Intim Stranger Acquaint Family/Intim

Expressive Expressive Expressive Expressive Expressive Expressive Expressive Expressive Expressive Instrumental Expressive Instrumental Expressive Expressive Expressive Othermotive Instrumental Expressive Expressive Instrumental Expressive Expressive Instrumental Expressive Expressive

Gun Gun Gun Gun Gun Gun NonGun Gun Gun Gun NonGun Gun Gun NonGun Gun Gun Gun Gun Gun Gun NonGun Gun Gun NonGun NonGun

Characteristics Underlying the Dominant Homicide Situations Urban Urban Urban Rural Urban Rural Urban Rural Rural Urban Urban Urban Urban Urban Urban Rural Urban Urban Urban Urban Rural Rural Urban Rural Urban

SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff SingOff

10,825 10,508 6,792 6,611 5,309 4,305 4,164 3,942 3,559 3,395 3,131 2,977 2,854 2,845 2,711 2,584 2,492 2,235 2,162 2,152 2,101 2,051 2,050 1,972 1,951

N

CB617-Miethe-v1

89

Percent of all homicides in common situations over time accounted for by these 25 dominant profiles: 30.1% (95,678/317,345) Percent of all homicides in all three decades accounted for by these 25 dominant “common” profiles: 21.7% (95,678/439,954)

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

Ranking

Table 4.1. Profiles of the 25 Most Dominant Homicide Situations Common across All Three Decades (SHR Data, 1976–98)

P1: JMT 0521832993c04 November 3, 2003 17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

90

November 3, 2003

RETHINKING HOMICIDE

of these types of homicide, the vast majority of killings happen in situational contexts that have the same structural profiles over time. Some individual elements (e.g., male offenders, intraracial violence) are major structural features across and within all three types of homicide situations, whereas other attributes (e.g., age of offenders, urban locations, gun usage) are better represented as context-specific effects. The nature of the stable structural profiles for these three major types of homicide is described below. Interpersonal disputes cover a wide range of arguments over money, status, or property. These homicide situations include what the literature refers to in various ways as either honor contests, character contests, or confrontational homicides (see Daly and Wilson 1988a; Luckenbill 1977; Polk 1994; Wolfgang 1958). Disputes are the most prevalent circumstance surrounding homicide, and nearly three-fourths of lethal disputes occur in stable situational contexts over time. In other words, the structural contexts of lethal disputes in the U.S. have been largely invariant across the late 1970s, 1980s, and 1990s. Lethal disputes that have stable structures “almost always” involve male offenders, offenders and victims of the same race and age groups, single offenders, nonstrangers, and combinations of these attributes.10 In addition, these intragroup disputes usually include victims and offenders of the same sex, guns as the lethal weapon, and large cities as the crime location. More often than not these stable profiles also consist of adult offenders in their 30s or older and White offenders. When compared to lethal disputes with structures unique to a particular decade, situational contexts for dispute-related homicides, which have persisted over time are clearly overrepresented by conjunctions of the following attributes: male offenders, adult offenders, White offenders, nonstrangers as victims, lethal weapons other than firearms, and/or single offenders. For example, the combination of adult males who kill others of their same age group and race accounts for 83% of the lethal disputes that occur within stable situational contexts over time, but this profile represents only 33% of dispute-related homicides in situations unique to each decade. Although lethal disputes vary widely in their structural profiles (see Table 4.1 for an illustration of this diversity), the stable situational contexts for these homicides are consistent with the notion of scenarios of masculine violence (see Daly and Wilson 1988a; Polk 1994). In particular, the

10

and others defined by motive, gender, age, race, and the victim–offender relationship are addressed more fully in the remaining chapters. To describe the relative prevalence of particular attributes or combinations of attributes, we will use the linguistic qualifiers of “almost never” (1% to 10% of the time), “rarely” (10% to 20%), “sometimes” (20% to 40%), “more often than not” (50% to 65%), “usually” (65% to 80%), and “almost always” (80% to 99%). Ragin (2000) uses similar language in his applications of QCA and a fuzzy logic extension of these comparative methods.

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

November 3, 2003

91

predominance of male offenders within these common profiles suggests that masculinity is a persistent theme that underlies lethal disputes over each of the last three decades. We explore the issue of masculinity and its unique manifestations in particular situational contexts in greater detail in subsequent chapters (see Chapters 6, 7, 8). Another major type of homicide situation with stable structures involves the killing of family members and intimate partners. About two-thirds of these homicides are found in situations with common structural profiles over time. These homicide situations collectively can be viewed as contexts for domestic violence because the victims are usually either spouses (including common law and ex’s) and other intimate partners (including ex- and current boy/girlfriends, lovers). What are the individual elements and combinations of attributes that differentiate between historically unique and stable structures for these homicides? Killings of family members and intimate partners that occur within stable structural contexts almost always involve male offenders, adult offenders, persons of the same age and race, and combinations of these factors. These persistent contexts for domestic violence also usually involve intersex attacks and expressive motives, and more often than not they are represented by White offenders and occur in large urban settings. The particular profile that best differentiates between stable and unique situational contexts for these homicides involves White male offenders over 30 years old. This profile is about 30 times more frequent in domestic violence situations with stable structures than those with unique signatures over time (39.4% vs. 1.3%). A particularly interesting feature of the stable homicide structures is that none of them involve particular variables that have uniform effects across all contexts. Take, for example, gender differences in the killing of family members and intimate partners. Male offenders are almost always found in the stable contexts for domestic violence (i.e., 90% of these homicides involve male offenders), but male-perpetrated homicide is actually either more or less prevalent than female offending in this context depending upon the particular combinations of attributes that define the entire structural profile.11 Our more general point is that the assumption of constant effects across all contexts (i.e., the “main effects” model) does not represent any of the structural elements used to defined historically unique and common configurations. Instead, the dominant profiles of homicide situations presented in Table 4.1 provide clear evidence of the value of a conjunctive approach that allows each variable to have a full range of context-specific effects. 11

Specific situational contexts for domestic violence in which women predominate as offenders and the more general question about gender differences in situational profiles involving intimate partners, family members, and other victim–offender relationships are addressed in Chapters 6 and 9.

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

92

November 3, 2003

RETHINKING HOMICIDE

A third major type of homicide situation is killings during the commission of other felonies (e.g., robbery homicides, rape homicides, drug-related homicides). These instrumental-type homicides account for more than onefourth of U.S. homicides in the last three decades, typically occurring in situational contexts with stable structures over time. Instrumental homicides almost always involve intragroup attacks by male offenders and these attributes are equally prevalent among felony homicides within historically unique and stable situational contexts. In both contexts, these homicides also usually involve firearms and occur in large cities, and rarely are committed by multiple offenders or against family members and intimates. The common signatures for instrumental homicides over time are based primarily on the nexus of adult male offenders, guns, strangers, and urban locations. When treated as “main effects” aggregated across contexts, many of these particular attributes (e.g., males, strangers, guns, large-city locations) do not distinguish between stable and unique situational contexts. However, some combinations of these attributes (e.g., the shooting of strangers during felonies in large cities) are more than twice as prevalent within stable situational contexts, other profiles (e.g., shootings by males over 30) are over 15 times more representative, and still others (e.g., shootings by males in large cities) are only slightly more descriptive of stable than decade-specific contexts for instrumental homicides. Although most instrumental homicides occur in situational contexts stable over the last three decades, it is difficult to identify the primary sources of stability in their structures because most of the variables do not have constant main effects across situational contexts. Instead, the sources of change and stability are rooted in a complex interplay among particular offender, victim, and offense characteristics. We can say that stable contexts for instrumental homicides are clearly overrepresented by adult offenders and rarely involve either females or juvenile offenders, but even among these attributes their effects are context-specific, depending on their particular joint occurrence with other structural factors.

Unique Homicide Situations in the 1970s and 1980s Only a small minority (7%) of U.S. homicides occur in situational contexts that are unique to either the 1970s or the 1980s. Nonetheless, we briefly explore the unique profiles in these decades because they illustrate situational contexts that have greatly diminished in their prevalence over time and they further inform our understanding of the stable structures over time. Several combinations of structural elements predominate in these unique profiles. Unique structural profiles for homicides in the 1970s are disproportionately represented by (1) female offenders who kill male intimate partners

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

November 3, 2003

93

and (2) male juvenile offenders who commit instrumental homicides with weapons other than firearms. These two profiles account for 72% of the homicides that occurred in situational contexts unique to the 1970s. Both of these profiles are about 12 times more likely to occur in homicide situations unique to the 1970s than either those unique to other decades or in stable structures over time. These unique profiles in the 1970s are not unexpected given the literature on homicide and its temporal patterns. For example, the decline in spousal homicides committed by female offenders is well documented (see Browne and Williams 1989; 1993; Browne, Williams, and Dutton 1999), and the current findings about the relative frequency of these killings in situational contexts unique to the 1970s are consistent with these studies. Similarly, the unique structural context for juvenile homicides may stem from the widely noted proliferation of lethal violence by juveniles throughout the 1980s and early 1990s that has been strongly linked to the interplay of guns and drugs (see Blumstein 1995). From this perspective, the unique structure for young offenders may simply reflect a change in the availability of firearms to juveniles and the subsequent substitution of guns for knives and blunt objects as the lethal weapon in homicides associated with illegal drug transactions, robberies, and other felony situations. A decline in arson-related homicides by male juveniles may also explain this particular unique signature involving instrumental homicides in the 1970s. The primary source of the unique signatures for U.S. homicides in the 1980s also involves lethal weapons other than firearms. These situational contexts involving nonfirearms are overrepresented by male offenders, offenders in their 20s, instrumental motives, and particular combinations of these factors. Another type of unique structure for this time period involves intragroup attacks (e.g. victims and offenders of the same age group and race), multiple offenders, White adult male offenders, and combinations of these attributes. It is unclear what particular circumstances are at the basis of these structures. However, it bears repeating that these unique situational contexts for both the 1970s and 1980s account for only a small proportion of U.S. homicides.

Unique Homicide Situations in the 1990s From the standpoint of homicide situations, the 1990s are a qualitatively distinct decade because a substantially larger proportion of homicides occur in situations unique to this particular decade compared to previous decades. While most homicides occur in stable structures over time, about 15% of them are found in situational contexts unique to the 1990s. The 25 most prevalent situational contexts unique to this decade are shown in Table 4.2.

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

94

November 3, 2003

RETHINKING HOMICIDE

Unique homicide situations in the 1990s are overwhelmingly represented by male offenders and guns as the lethal weapons. Male offenders are found in 99% of these unique situations in the 1990s and guns are used in more than 99% of them. In addition, these unique situations for this decade almost always involve intragroup members (i.e., victims/offenders of the same sex, same race, and same age group), Black offenders, and large city locations. They also usually involve single offenders and acquaintances, and rarely include intimate partners or adult offenders over 30 years old. Compared to other decades and stable contexts over time, the major distinguishing features of unique structures in the 1990s, in order of their discriminatory power, include the following: juvenile offenders (58% of the homicides in unique 1990 profiles vs. 8% in the “other” profiles), Black offenders (84% vs. 47%), guns (99.7% vs. 64%), intrasex (97% vs. 68%), large-city locations (88% vs. 64%), nonintimates and nonfamily members (94% vs. 71%), multiple offenders (21% vs. 5%), and instrumental motives (36% vs. 25%). Particular combinations of these structural attributes contribute to extremely large percentage differences between unique and nonunique homicide situations in the 1990s. For example, young male offenders in urban areas are ten times more prevalent within structural contexts unique to the 1990s than in other time periods (50.4% vs. 4.5%). Differences across unique and nonunique contexts in this particular profile are even greater among Black offenders (40% vs. 1.2%) and less dramatic but still noteworthy among White offenders (10.4% vs. 2.7%). As indicated by its presence in 12 of the 25 most prevalent homicide situations unique to this decade (see Table 4.2), lethal shootings by juvenile Black males of other Black males within large U.S. cities are major structural features that underlie emergent homicide situations in the 1990s. The predominance of urban youth violence as an emergent context for homicide is widely noted in previous research (see Anderson 1998; Cook and Laub 1998; Fagan and Wilkinson 1998; Heide 1999). The literature has also emphasized the roles of gender, race, and firearms in urban violence. What is less consistent with the popular image of youth violence, however, is that these emergent contexts for young Black male homicides also involve primarily (1) intragroup members (i.e., they are not usually predatory acts against victims of a different age, sex, or race) and (2) single rather than multiple offenders. The extent and nature of the emergent forms of urban youth violence are addressed more fully in Chapter 7.

Theoretical Accounts of the Findings Using SHR data over the last three decades, this analysis of homicide situations reveals several patterns about the nature of homicide. These major

17:23

Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale

O < 20 O2029 O < 20 O2029 O < 20 O2029 O < 20 O2029 O < 20 O < 20 O < 20 O < 20 O3098 O2029 O < 20 O < 20 O < 20 O < 20 O < 20 O2029 O2029 O3098 O < 20 O < 20 O < 20 Oblack Oblack Oblack Oblack Owhite Oblack Oblack Oblack Oblack Oblack Oblack Oblack Oblack Oblack Oblack Oblack Oblack Owhite Owhite Oblack Oblack Owhite Oblack Oblack Oblack

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex

=Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race =Race Acquaint Acquaint Acquaint Stranger Acquaint Acquaint Stranger Acquaint Acquaint Stranger Acquaint Acquaint Acquaint Acquaint Acquaint Stranger Acquaint Stranger Acquaint Acquaint Stranger Acquaint Acquaint Stranger Family/Intim

Expressive Instrumental Instrumental Expressive Expressive OtherMotive Instrumental Expressive Expressive Expressive OtherMotive Expressive Expressive Instrumental Expressive Instrumental Instrumental Expressive Expressive Instrumental OtherMotive Expressive Instrumental Instrumental Expressive

Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun

Characteristics Underlying the Dominant Homicide Situations Urban Urban Urban Urban Urban Urban Urban Urban Rural Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Rural Urban Urban Urban Urban Urban

SingOff SingOff SingOff SingOff SingOff SingOff SingOff MultOff SingOff SingOff SingOff MultOff SingOff MultOff SingOff SingOff SingOff SingOff MultOff SingOff SingOff SingOff MultOff MultOff SingOff

2,652 2,360 1,421 1,333 1,142 1,057 1,033 796 792 753 710 633 612 558 546 534 490 485 483 475 444 358 349 341 322

N

CB617-Miethe-v1

95

Percent of all homicides in situations unique to the 1990s accounted for by these 25 dominant profiles: 28.0% (20,679/73,960) Percent of all homicides in the 1990s accounted for by these 25 dominant “unique” profiles: 11.7% (20,679/176,998)

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

Ranking

Table 4.2. Profiles of the 25 Most Dominant Homicide Situations Unique to the 1990s (SHR Data, 1990–98)

P1: JMT 0521832993c04 November 3, 2003 17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

96

November 3, 2003

RETHINKING HOMICIDE

findings and how well they are accounted for by existing theoretical perspectives are summarized below.

Major Findings Our application of QCA to the study of homicide situations focuses on their structural features. Our primary findings are related to the diversity of these situations and their unique and stable structural profiles over time. First, our analysis illustrates that homicide has a complex and diverse structure. Using 15 offender, victim, and offense attributes to define structural contexts, about 5,600 distinct structural profiles are found among all U.S. homicides in the SHR data from 1976 to 1998. The number of distinct structural profiles reduces to 618 when we restrict our focus to situations that represent an average of at least 5 homicides per year over the last three decades. The number of prevalent homicide situations has increased across decades, suggesting greater diversity in the structure of homicide over time. These findings indicate that homicide situations have a far more diverse and complex structure than often assumed in previous research and public discussions of homicide. Second, most homicides occur in situational contexts that have not changed their structures over time. The pattern of stable structures also holds for three major types of homicide situations: (1) interpersonal disputes, (2) the killing of family members and intimate partners,12 and (3) instrumental homicides during the commission of other crimes. Stable homicide situations in general and within these three types of homicides are usually represented by adult male offenders who kill persons of the same age and race. Third, homicide situations have become increasingly characterized by decade-specific structural profiles. Unique contexts for homicides in the 1970s are linked to two particular themes (i.e., intimate partner killings by women and juvenile offenders in instrumental homicides with weapons other than firearms), whereas homicide situations unique to the 1990s almost always involve male offenders and firearms. Black offenders and urban settings are also major structural features of emergent forms of homicide in the 1990s.

Theoretical Accounts As described in Chapter 2, several theoretical perspectives have been widely used to explain individual risks and aggregate rates of homicide across 12

As will become evident in Chapter 9, further disaggregation by victim–offender relationship indicates that changes have occurred over time in the structure of intimate partner homicides. The stability found here may thus be a reflection of a lack of change in the structure of killings involving family members and not necessarily for all types of domestic violence situations.

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

November 3, 2003

97

groups. How well do current theoretical perspectives account for the three major findings about homicide situations observed in this chapter? No current theoretical formulation is able to easily explain the diversity of homicide situations identified in this analysis. This is the case for two basic reasons. First, most theories offer explanations of level differences in individuals’ propensity toward violence or aggregate crime rates rather than level differences in the situational context of violent acts. In other words, they have been applied to questions of why some groups, places or times have higher homicide rates, but do not propose interpretations of why particular situational contexts of homicides are more prevalent than others. This “neglected situation” in criminological research (see Birkbeck and LaFree 1993) is indicated by less theoretical attention to situational analysis and the failure to include measures of the situational context in most theories of offender motivation and aggregate analyses of crime rates. Second, most theories and empirical tests of them are based on a main effects specification that further ignores the situational context because it assumes that the effect of each variable is constant across contexts. When the homicide situation is rarely the unit of analysis and empirical treatments of homicide often greatly simplify or even ignore the situational context, it shouldn’t be surprising that most existing theories are unable to explain the diversity of homicide situations. This applies to both theories of offender motivation and aggregate crime rates. For example, most theories of offender motivation (e.g., learning theories, social control theories, various psychological and biological theories) neglect the situational context because they often do not consider the victim attributes and offense characteristics that provide a general opportunity structure for the occurrence of homicides. Most macrotheories of homicide rates over time (e.g., anomie/strain theory, social disorganization, differential opportunity) are also limited in explaining the diverse contexts for homicide because they typically assume that particular causal factors (e.g., economic inequality, value conflict) have constant effects across contexts or among particular subgroups when crime rates are disaggregated (e.g., that the effects of joblessness and economic inequality are the same for all Black males). The only theoretical perspectives that have some predictive utility in the understanding of the diversity of homicide situations are criminal opportunity theories (e.g., Cohen and Felson’s routine activity theory (1979)) and those that derive from symbolic interactionism (e.g., Lofland’s facilitating conditions (1969)). Criminal opportunity theories focus on the convergence of offender motivation, suitable targets, and the absence of capable guardians as necessary conditions for crime. Lofland’s theory (1969) offers the notions of “facilitating places, hardware, and others” as similar requirements for criminal acts. By emphasizing the union of offender, victim and offense elements in time and space, both theories have potential for explaining the diversity of homicide’s situational contexts.

17:23

P1: JMT 0521832993c04

98

CB617-Miethe-v1

November 3, 2003

RETHINKING HOMICIDE

From the offender’s perspective in both of these theories, any particular set of structural features may be perceived as attractive or facilitating conditions, thereby accounting for the wide diversity of homicide situations observed in this study. For example, the observed pattern of high levels of situational clustering (i.e., a small number of situational contexts account for almost all homicide incidents) is logically consistent with these theories. In particular, the large number of idiosyncratic structures illustrates the subjectivity in an offender’s interpretation of attractive or facilitating conditions, whereas the small number of dominant structures would indicate that some of these perceptions are shared across groups of offenders. Unfortunately, by emphasizing the offender’s interpretation of situational elements, these theories are basically unfalsifiable because any combination of attributes may be perceived as producing attractive or facilitating conditions. Nonetheless, we will use these criminal opportunity theories in subsequent chapters as a general conceptual framework to more fully explore differences in homicide situations over time and across particular subgroups (see especially Chapter 7). It is possible to interpret the patterns of stable and unique homicide situations over time from a variety of macrosociological perspectives. Structural strain theory, for example, is most widely used to explain why crime rates are higher in the United States than other countries and why some subgroups have higher offending rates than others (e.g., some ethnic minorities and the poor experience more frustration and strain because of blocked access to legitimate means of achieving success). When applied to the question of stability and change in the structures of homicide situations, this theoretical approach would argue that (1) stability in situations derives from the relative persistence of strain in American life over the last three decades and (2) the emergence of unique situations involving young Black urban males is due to recent historical circumstances that have increased their particular exposure to strain (e.g., high joblessness, the exodus of service jobs and their replacement by high tech jobs in which they are less able to compete, growing despair due to these particular life experiences). The particular manifestation of this strain could take the form of both instrumental homicides (e.g., as a means of achieving economic success) or expressive homicides (e.g., arguments and disputes precipitated by frustration and anger). As we will describe more fully in subsequent chapters, any theoretical perspective that is designed to explain U.S. homicide and its situational context must recognize the central role of masculinity within lethal violence. Males are and always have been overrepresented as offenders, and all major situational contexts that have remained stable over time involve male offenders as a primary structural feature. Emergent situational contexts in the 1990s almost always involve male offenders as well. Whether these killings are motivated by masculine competitiveness over honor,

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

THE EMPIRICAL DISTRIBUTION OF HOMICIDE

November 3, 2003

99

social status, or sexual partners, or are committed for dispute resolution, the fact remains that homicide has a distinctly male character across a variety of situational contexts. The persistence of male offenders across contexts has not been fully explored in previous research, in spite of the large volume of literature on gender differences in homicide and other types of crime (see Daly and Wilson 1988a; Messerschmidt 1986; Polk 1994). Much of the popular literature on homicide and current public policy about homicide is based on relatively low-frequency situational contexts for these crimes. For example, alleged epidemics of both serial killing and random violence have been largely socially constructed by the wide media coverage of specific instances that fit this image (see Best 1990, 1999; Jenkins 1994). However, our analysis of SHR data indicates that stranger homicides involving expressive motives are rare events in the last two decades and especially when they are considered within the context of intergroup attacks (i.e., the typical image of these predatory acts by strangers). Similarly, images of youth violence in a group context (e.g., gang violence) have often been emphasized as dominant themes in the last two decades, but our analysis of homicide situations indicates that such contexts for violence are relatively rare for each decade and do not have particularly unique signatures in the 1990s (e.g., they have not become more intergroup in nature). The attention devoted to female offenders and media portrayals of rising rates of instrumental violence by women are also disproportionate to the empirical frequency of these situational contexts of violence. Our general point is that all too often theoretical, empirical, and policy research has focused on situational contexts for homicide that represent only a small minority of these violent acts. The particular attributes of female offenders, instrumental motivations, stranger killings, interracial slayings, multiple offenders, and teenage offenders are rarely, if ever, found among the top 25 situational profiles for each decade. Contrary to claims of its changing character and the emergence of particular unique situational contexts, homicide remains in the 1990s primarily an expressive act of violence carried out with guns in large cities by single, adult, male perpetrators who victimize acquaintances of a similar age, sex, and race during arguments and disputes. Rather than focusing disproportionately on the exceptional case and situations (e.g., teen offenders, child victims, female killers), public policy initiatives would best be served by also addressing the predisposing and precipitating events that underlie these 25 most dominant homicide situations.

Conclusions The remaining chapters of this book explore in greater detail these questions and the particular nature of subgroup variation in homicide situations.

17:23

P1: JMT 0521832993c04

CB617-Miethe-v1

100

November 3, 2003

RETHINKING HOMICIDE

The analyses will explore differences in the structure and prevalence of homicide situations based on the motive for the crime, the offender’s sex, the offender’s age, the offender’s race, and the victim–offender relationship. By doing so, we hope to further illustrate the value of QCA as a method for describing the sources of stability and change in the structure of homicide situations over time.

17:23

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

C H A P T E R

F I V E

Instrumental and Expressive Motives in Homicide Situations1

one common basis for distinguishing types of homicide involves their motive. Typical motives and circumstances surrounding homicides include trivial altercations, jealousy, revenge, romantic triangles, robbery, sexual assault, burglary, and disputes in drug transactions. These motives are often subclassified to differentiate between two general types of homicides: “expressive” and “instrumental” homicides. Using narrative accounts of homicides in Los Angeles over the twentieth century and the national Supplementary Homicide Reports (SHR), this chapter examines the episodal nature and structure of expressive and instrumental homicides. Descriptive case accounts are analyzed to identify patterns in the transactions that underlie the most common types of expressive homicides (i.e., confrontational disputes among male offenders) and instrumental killings (i.e., robbery-homicides). Qualitative Comparative Analysis (QCA) is used on the SHR data to determine whether instrumental and expressive homicides are qualitatively different in their structures (i.e., combinations of offender, victim, and situational characteristics). We do this empirically by identifying the most prevalent combinations of individual and situational elements unique to each type of homicide, as well as those common to both, through QCA’s systematic process of holistic comparison.

Instrumental and Expressive Crimes Although not without its critics (e.g., Felson 1993; Polk 1994), the distinction between instrumental and expressive crimes has been widely used in criminological research. Instrumental crimes are those conducted for explicit, future goals (such as to acquire money or improve one’s social position), 1

Portions of this chapter were originally published in Miethe and Drass (1999).

101

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

102

November 3, 2003

RETHINKING HOMICIDE

whereas expressive offenses are often unplanned acts of anger, rage, or frustration (see Block 1976; Block and Christakos 1995; Block and Zimring 1973; Decker 1993; 1996; Siegel 1998). The instrumental/expressive distinction often parallels the differences between planned (premeditated) and spontaneous (“heat of passion”) offenses. Across a variety of different crimes, it has been common practice among criminologists to consider instrumental acts as qualitatively different from expressive acts. Taxonomic systems for rapists, for example, employ the instrumental/expressive distinction as a fundamental dimension (Prentky et al. 1985; Rosenberg et al. 1988). The instrumental/expressive distinction is also a critical dimension in typologies of motor vehicle theft, vandalism, political terrorism, workplace violence, and intrafamily assaults (Amir 1988; Gelles 1987; Miethe and McCorkle 2001; Swanton 1989; Whittingham 1981). Violent crimes are often distinguished from other offenses (like corporate crime) based on their relative frequency of instrumental and expressive motives. The presumed difference in the relative frequency of expressive and instrumental motivations for violence in the North and South has been used as an explanation for changes in regional variation in homicide rates among African Americans (Rose 1984). Furthermore, studies of criminal careers have employed this distinction, asserting that early criminal careers begin with the commission of expressive crimes and serious criminal careers are often continued for instrumental reasons (Petersilia 1980). From the perspective of crime control, the relative prevalence and nature of instrumental and expressive crimes have important policy implications. Under the deterrence doctrine, the threat of legal sanction is considered the most effective for instrumental crimes by persons with low commitment to a criminal lifestyle (see Chambliss 1967; Parker and Smith 1979; Thomas and Williams 1977). Expressive crimes, in contrast, are often viewed as undeterrable by legal sanctions. Other crime prevention and intervention strategies are also tied directly to the instrumental/expressive distinction. Treatment programs for impulsivity and anger management, for example, are directed at the control of expressive acts, whereas an assortment of social betterment measures (like job training, family counseling, and educational enhancement) are often advocated as corrective actions to abate the conditions that motivate instrumental crimes. Crime control efforts through environmental design and situational crime prevention techniques may be equally effective for both instrumental and expressive crimes, by increasing the “costs” for the reasoning instrumental offender and decreasing the opportunities for spontaneous or expressive criminality (see Clarke 1983; 1997; Cornish and Clarke 1986). Regardless of the particular crime control strategy, however, it is often assumed that instrumental and expressive crimes are unique in the characteristics of their offenders, victims, and situational elements.

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

103

When applied to the study of homicide, particular circumstances and motivations are often categorized as either instrumental or expressive crimes. Arguments, brawls, romantic triangles, youth gang killings, and other interpersonal disputes are typically classified as expressive acts because their dominant motivation is the violence itself (see Block and Christakos 1995; Decker 1996; Polk 1994). Killings that occur in the commission of another felony are the most commonly classified instrumental homicides. While many homicides in these felony-type circumstances are often a “side effect” of another criminal act (Maxfield 1989), these killings are usually classified as instrumental crimes because the death of the victim is a potentially expected outcome in the pursuit of the primary goal. This same rationale is used to support the legal classification of felony homicides as first-degree murders. Previous research provides empirical support for this subclassification of homicide circumstances as instrumental and expressive crimes. Interviews with convicted robbers, for example, indicate that a majority of these offenders were motivated by instrumental reasons such as getting money and purchasing drugs (Feeney 1986; Gabor et al. 1987; MacDonald 1975). Most muggers do not intend to kill their victims (Maxfield 1989), but the precipitating actions that result in the robbery are clearly motivated by instrumental pursuits. Youth crime, especially when it occurs in a group context, is often described as impulsive, situational and opportunistic, spontaneous, and nonutilitarian. Street fights are the major triggering event in gang-related homicides (Block and Christakos 1995), providing additional support for the classification of youth gang homicides as expressive offenses. However, for other homicide situations (e.g., revenge or “pay back” killings), it is difficult to classify motive unequivocally on an instrumental–expressive dichotomy (see Polk 1994). Although most criminological research treats instrumental and expressive acts as distinct entities, several alternative conceptualizations have been employed in past studies. Block and Block (1993), for example, consider expressive and instrumental motives as different points along a continuum. These authors also note that both types of motives may be present among the same types of crimes. For instance, street gang violence may result from instrumental, entrepreneurial activities (like drug trafficking), whereas other violent gang situations involve impulsive, emotional, expressive outbursts (Block and Block 1993).2 In contrast, Felson (1993:104) interprets all 2

While Block and Block (1993) emphasize the instrumental–expressive continuum, this theoretical conceptualization often reduces to a simply dichotomy in statistical models of crime types and empirically based taxonomies. For example, Block and Christakos (1995) acknowledge the instrumental–expressive continuum, but then proceed to identify homicide syndromes that derive from dichotomizing instrumental and expressive motives as one of their classification criteria. As discussed shortly, our analysis permits an empirical evaluation of the proper conceptualization of these two general types of homicides by determining whether

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

104

November 3, 2003

RETHINKING HOMICIDE

aggressive behavior as goal-oriented or instrumental because “it is an attempt to achieve what people value.” From this perspective, even expressive acts of violence done in anger reflect an instrumental reaction to perceived wrongdoing. Given the widespread use of the instrumental–expressive distinction in criminological research, an important question involves whether these types of homicides are fundamentally or qualitatively distinct. In other words, are instrumental and expressive homicides best characterized as unique or common in their particular combinations of offender, victim, and offense elements that define the structure of homicide situations?

The Structure of Motive-Based Homicide Situations The method of Qualitative Comparative Analysis (QCA) provides a formal approach for systematizing our analysis of instrumental and expressive homicide situations. Four basic questions underlie these comparative analyses. First, do instrumental and expressive homicides have similar or different structures in terms of their particular offender, victim, and situational attributes? Second, what are the common and unique characteristics of these homicide types? Third, have these common and unique characteristics changed over the last three decades? Fourth, do widely assumed subsets of instrumental (e.g., rape- and robbery-murders) and expressive homicides (e.g., lovers’ triangles, brawls, and arguments) have structural profiles that are internally consistent with all other subtypes within these general categories? In other words, are the unique structural profiles or signatures of expressive and instrumental homicides also observed among the various subsets of circumstances that are often included in these two general motive categories? Previous research provides mixed predictions about whether different individual and situational attributes underlie different homicide motives. Under Felson’s (1993) assertion that all aggressive actions derive from the same basic instrumental values, both expressive and instrumental homicide should occur in the same contexts and share common explanatory factors because uniquely expressive motives do not exist. A finding of different contexts for instrumental and expressive homicides would question the characterization of a singular motivational structure for all homicides. In contrast, other researchers (e.g., Decker 1996; Maxfield 1989) observe that expressive violence is more concentrated among persons in primary or intimate relations, whereas instrumental violence is more common among strangers. Decker (1996:434) also contends that persons excluded from mainstream urban life (e.g., young, Black males) are more prone to deviate from this instrumental and expressive crimes are fundamentally unique or share the same combination of offender, victim, and situational factors that underlie their structure.

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

105

pattern of primary-expressive and stranger-instrumental homicides. These latter observations suggest that unique combinations of individual and situational factors may underlie different homicide motives.

Data and Measurement Our comparative analysis of expressive and instrumental homicides is based on the FBI’s Supplementary Homicide Reports (SHR) from 1976 to 1998.3 The specific offender characteristics used to differentiate types of homicide include their gender, race (White vs. Black), and age (<20, 20–29, and 30 or older). Similar variables are included as victim characteristics. The joint attributes of the victim and offender were used to classify homicides as involving persons of the same sex, same race, and same age group. Situational and offense elements involve the victim–offender relationship (coded family/ intimates, acquaintances, or strangers), gun use, the number of offenders (one vs. multiple), and the location of the homicide (large city more than 100,000 vs. “others”). Combined, this particular set of offender, victim, and situational elements captures the major structural elements of crime events that have been examined in past research (see Sacco and Kennedy 2002; LaFree and Birkbeck 1991; Miethe and Meier 1994). Looking at the modal categories for the variables, homicides in the total sample most often involve offenders who are male, African Americans, and over 30 years old. The typical victim in this sample also possesses these same characteristics. Victims and offenders are acquaintances in the majority of homicides, two-thirds involve a firearm, and over half occur in large cities. Most of these patterns hold for both instrumental and expressive homicides. The major exception is that stranger assaults are far more common among instrumental crimes, whereas slayings of family members or intimates (like girlfriends) are more common among expressive homicides. Over three-fourths (77%) of the homicides involve expressive motives and the remaining one-fourth (23%) are instrumental homicides.

Results of Qualitative Comparative Analysis (QCA) Our application of QCA begins by looking at the distribution of configurations of individual and offense elements by type of homicide. This information, obtained from the truth table, allows us to identify how much 3

This analysis is based on homicides that have nonmissing data on each of the included variables. The total sample size is 197,178 cases for the 1976–98 period. We have not applied imputation methods for missing data here because one of our goals in this particular study is to see how well our derived structures for expressive and instrumental homicide actually correspond with the subcategories of motives included within these two general categories. If we used the imputed data here, the imputation method would be another possible explanation for the lack of convergence in the results across subsets of motives. Accordingly, only cases with complete information on a variable are included in this analysis.

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

106

November 3, 2003

RETHINKING HOMICIDE

heterogeneity exists within each motive category and how well the instrumental–expressive dichotomy differentiates homicide events. For this analysis, a minimum cell proportion of .1% is used to identify nontrivial configurations of offender, victim, and offense attributes.4 The terms “prevalent” or “dominant” structures describe each distinct combination of individual and offense characteristics that satisfies the minimum proportion rule. As described in Chapter 3, this minimum cell proportion rule eliminates combinations that are observed infrequently. In fact, about 85% of the distinct structural configurations are eliminated by applying this rule. While the minimum rule eliminates most configurations, the remaining dominant structures still represent approximately 86% of all homicides in the 1976–98 period. Thus, focusing on the dominant configurations does not seriously limit our inferences about the nature of instrumental and expressive homicides.5 Combinations of attributes are defined as “unique” structures in this analysis when their observed cell proportions among expressive or instrumental homicides exceed by at least 10 percentage points their respective marginal distributions. “Common” structures are those in which a difference of less than 10% exists between their cell proportions and their marginal proportions. Thus, each dominant structure can be classified as uniquely expressive, uniquely instrumental, and common among both expressive and instrumental homicides.6

Comparisons between Expressive and Instrumental Homicides Our analysis of the structure of homicide situations involves a comparison of the relative frequencies of dominant configurations of individual and offense attributes across types of homicide (instrumental vs. expressive). These comparisons of expressive and instrumental homicides are conducted 4

5

6

Given our sample sizes for expressive (n = 152,322) and instrumental (n = 44,856) homicides, the minimum proportion rule of .1% translates into a minimum frequency of 152 and 45, respectively, for each group. Under this decision rule, a configuration must contain at least 152 expressive homicides or at least 45 instrumental homicides to be included in this analysis. Application of this rule reduces dramatically the number of configurations (from 1934 to 282) but retains the vast majority of homicide cases (86% [170,021 of the 197,178]). We have conducted all analyses without using the minimum frequency rule and the results are similar. Thus, our conclusions about unique and common configurations for expressive and instrumental homicides are not affected by this decision rule. As described in Chapter 3, minimum frequency rules are often used in QCA applications to minimize measurement error on configurations with a small number of observations. A minimum frequency rule is used here without loss of generality. Given that 77% of our sample involves expressive homicides, uniquely expressive homicide situations are those in which at least 87% (77% + 10%) of the homicides in these structures have expressive motives. Uniquely instrumental homicide situations are those in which instrumental homicides are found in at least 33% of these structures (23% + 10%), whereas “common” structures are those in which the proportion of expressive homicides included in them ranges from 67% to 87%.

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

107

Figure 5.1. Homicide Situations and Cases with Unique and Common Motives

to identify their unique and common signatures. The overall pattern of these unique and common profiles is summarized in Figure 5.1. Of the 282 dominant structures of homicide situations in the 1976–98 period, the clear majority (78%) are profiles that are unique to either expressive or instrumental homicides. These unique situational profiles also account for over two-thirds of the homicides in these three decades (see Figure 5.1). Unique signatures for each motive also best represent the majority of homicide situations and cases within each decade.7 Based on these results, it is clear that the structures of the situational contexts for expressive and instrumental homicides are qualitatively distinct in most cases. Expressive and instrumental homicides are often different types of homicides with unique structural profiles.

The Nature of Unique and Common Profiles We have just shown that expressive and instrumental homicides are best represented by distinct structures. What are the particular offender, victim, and 7

Specifically, the proportion of situational contexts that were unique to each decade was 81.5%, 79.7%, and 75.4%, respectively. These unique profiles for expressive and instrumental homicides represented at least 65% of the homicides in each decade (73.1% of the homicides in the 1970s, 68.6% of the homicides in the 1980s, and 65.2% of the homicides in the 1990s). These findings indicate that homicide situations have become increasingly common to both expressive and instrumental motives over time, but unique profiles are still the dominant pattern even in the 1990s.

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

108

RETHINKING HOMICIDE

Table 5.1. Particular Offender, Victim, and Offense Characteristics Underlying Uniquely Expressive, Uniquely Instrumental, and Common Homicide Situations

Characteristics Male Offender <20 Yr Old Offender 20–29 Yr Old Offender 30 and Older Offender White Offender Black Offender Same Age of V–O Same Race of V–O Same Sex of V–O Family and Intimates Acquaintances Strangers Guns as Lethal Weapon Multiple Offenders Large City Locations

Uniquely Expressive

Uniquely Instrumental

Both Motives

77.3% 5.7% 17.3% 77.0% 47.3% 52.7% 98.4% 99.8% 51.3% 53.2% 46.8% 0% 70.9% .8% 58.3%

100% 23.0% 59.5% 17.5% 39.4% 60.6% 85.2% 73.9% 80.0% 0% 34.6% 65.4% 66.6% 43.9% 68.8%

99.2% 18.8% 63.6% 17.6% 48.9% 51.1% 96.0% 94.9% 91.6% 6.0% 81.6% 12.4% 65.4% 5.5% 59.2%

offense attributes that in isolation and in combination distinguish between unique and common profiles for expressive and instrumental homicides? An initial way to assess the nature of unique profiles for expressive and instrumental homicides involves an assessment of the relative importance of particular offender, victim, and offense characteristics that underlie them. This question is addressed by comparing the relative frequencies of each characteristic among exclusively expressive homicide situations, exclusively instrumental situations, and homicide situations with structures similar for both expressive and instrumental motives. The results of these “main effects” comparisons are summarized in Table 5.1. The single most powerful attributes that best distinguish situations unique to expressive homicides from instrumental homicide situations involve, in order of relative differences, the victim–offender relationship, the offender’s age, the number of offenders, intrasex relations, intraracial relations, and the offender’s gender (see Table 5.1). Some of these differences across groups are substantial. For example, strangers represent about two-thirds of the victims within situational contexts unique to instrumental homicides, but strangers are never found in situations unique to expressive homicides. Uniquely instrumental homicide situations are also strongly over-represented by multiple offenders (44% vs. 1%), intrasex assaults (80% vs. 51%), male offenders (100% vs. 77%), and offenders under 20 years old

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

109

(23% vs. 6%). In contrast, uniquely expressive homicide situations are greatly overrepresented by offenders in their 30s or older (77% vs. 18%), intraracial attacks (99.8% vs. 74%), and assaults on nonstrangers (100% vs. 35%). The offender’s race, type of lethal weapon (coded gun or nongun), and location (coded large city vs. less urban area) do not strongly differentiate between these types of homicide situations. Although the magnitude of these “main effects” is large in many cases, they ignore the basic fact that the unique signatures for instrumental and expressive homicides are derived from the combination of attributes that underlie them. What this means is that the effect of most of these variables is context-specific, with particular combinations of other attributes making these situational contexts more or less likely for each type of motive. Even in the case of male offenders (who are found in 100% of the situations unique to instrumental homicides), the classification of situational contexts as unique or common for each type of motive ultimately rests on the conjunction of male offenders with other individual and offense attributes. To illustrate the diversity of particular situational contexts that are unique and common to each type of motive, Table 5.2 summarizes the 10 most dominant profiles found within uniquely expressive homicides, uniquely instrumental homicides, and both motives. These 10 most dominant combinations represent only a small proportion of their respective situational contexts and the homicides found within them, but they nonetheless serve to highlight the fundamental differences and similarities of the structures that underlie these general motives. The major theme found within the unique signatures for expressive homicides involves their occurrence in situations of intragroup conflict among older adults who victimize primary group members (e.g., family, intimates, acquaintances). These situations are almost always attacks by single offenders on single victims, and they usually involve male offenders, guns as the lethal weapon, and urban locations.8 This dominant profile is similar to the characterization of uniquely expressive homicide situations that is derived from the summary of “main effects” for particular variables. Intragroup conflicts (e.g., same age, same race and nonstrangers) account for nearly all (98%) of the homicides in situations unique to expressive motives compared to only slightly more than one-quarter (26%) of homicides in uniquely instrumental situations. 8

The variable “number of victims” is not included in most of our analysis because single victims represent over 95% of all homicides across a wide variety of situations. The only possible exception is family slayings in which the proportion of multiple victims (including situations of murder–suicide) is slightly higher. Even in these killings, multiple victims are rare. An inspection of the unique profiles, however, revealed that they were almost always single-victim situations. This is why we included this variable in our general description of the characteristics of uniquely expressive homicide situations.

17:26

110

Omale Omale Omale Omale Omale Omale Omale Omale Ofema Omale

O3098 O3098 O3098 O3098 O3098 O3098 O3098 O3098 O3098 O3098

Oblack Owhite Owhite Owhite Oblack Oblack Owhite Oblack Owhite Owhite

=Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale

O2029 O2029 O2029 O2029 O2029 O2029 O3098 O2029 O < 20 O2029

Oblack Oblack Oblack Oblack Oblack Owhite Oblack Owhite Oblack Owhite

Acquaint Acquaint Acquaint Fam/Intim Acquaint Acquaint Fam/Intim Fam/Intim Fam/Intim Acquaint

Gun Gun Gun Gun Gun =Gun Gun Gun Gun =Gun

=Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex

Stranger Acquaint Stranger Stranger Stranger Stranger Stranger Stranger Stranger Acquaint

Gun Gun Gun Gun Gun Gun Gun Gun Gun  Gun =

Attributes Underlying the Dominant Homicide Situations

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex

Urban Urban Urban Urban Urban Urban Urban =Urban Urban =Urban

Urban =Urban Urban =Urban =Urban Urban Urban Urban =Urban Urban

Singoff Multoff Multoff Singoff Multoff Singoff Singoff Singoff Multoff Singoff

Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff

47% 40% 70% 77% 91% 35% 33% 43% 96% 63%

% Instrum

89% 91% 90% 96% 93% 91% 97% 99% 98% 88%

% Express

990 683 657 650 642 502 399 392 387 357

Nins

5,728 5,072 3,756 3,654 3,450 2,939 2,295 2,077 1,922 1,848

Nexp

CB617-Miethe-v1

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Ranking

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race

Attributes Underlying the Dominant Homicide Situations

Panel B: Unique Instrumental Motives

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Ranking

Panel A: Unique Expressive Motives

Table 5.2. Profiles of the 10 Most Dominant Situations Unique and Common to Expressive and Instrumental Homicide (SHR 1976–98)

P1: JMV 0521832993c05 November 3, 2003 17:26

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale

Ranking O2029 O2029 O2029 O2029 O < 20 O2029 O2029 O2029 O3098 O3098

Oblack Owhite Owhite Oblack Oblack Oblack Owhite Owhite Owhite Owhite

Panel C: Common to Both Motives

=Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex Acquaint Acquaint Acquaint Acquaint Acquaint Acquaint Acquaint Acquaint Acquaint Acquaint

Gun Gun Gun Gun Gun =Gun =Gun =Gun =Gun Gun

Attributes Underlying the Dominant Homicide Situations Urban Urban =Urban =Urban Urban Urban Urban =Urban =Urban Urban

Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff

80% 87% 84% 86% 78% 86% 82% 81% 86% 70%

% Express

7,938 3,751 3,544 3,256 3,187 2,543 2,357 2,058 1,639 1,201

Nboth

P1: JMV 0521832993c05 CB617-Miethe-v1 November 3, 2003

111

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

112

November 3, 2003

RETHINKING HOMICIDE

The unique signatures for instrumental homicides are more often represented by intergroup attacks by younger male offenders within a group context and in urban settings. The use of guns, Black offenders, and intrasex attacks of strangers are also major elements among the 10 most dominant profiles unique to instrumental homicides. Guns, urban locations, and Black offenders are particular combinations of attributes whose representativeness among instrumental homicide situations is not derivable from the “main effects” of each variable separately. This particular combination of attributes represents about 38% of the homicides in uniquely instrumental situations, but this profile is found in only about one-fifth (21%) of the homicides within situational contexts unique to expressive motives. Differences between instrumental and expressive situational contexts are even more dramatic when male offenders and multiple offenders are added to this base profile. These acts of urban gun violence by Black males in a group context are found in over one-fifth of the uniquely instrumental situations, and they are never observed within situational contexts unique to expressive motives. Situational contexts common to both expressive and instrumental homicides also involve intragroup attacks but in this case the predominant profile involves offenders in their 20s who victimize acquaintances. These intragroup attacks among the 10 most dominant situational contexts are equally represented by guns and other weapons, urban and less urban locations, and White and Black offenders. The single most dominant profile in this category (i.e., Black male offenders in their 20s who victimize acquaintances with guns in urban areas who are of the same age group, sex, and race) is over twice as prevalent as the next most common profile (see Table 5.2). It is important to emphasize, however, that common situational profiles for expressive and instrumental homicides represent only about one-third of the U.S. homicides during the 1976–98 time period.

Changes over Time in the Nature of Unique Profiles When separate analyses are done for each decade, the majority of expressive and instrumental homicides continue to occur in situational contexts that are unique to each type of motive. However, there have been some specific changes over time, especially when patterns for the 1990s are compared with earlier decades. The nature of these decade-specific changes are summarized below. Similar to other decades, expressive homicides in the 1990s are far more likely than instrumental homicides to be an intragroup phenomena, involving persons of the same age group and race. These expressive homicide situations also occur disproportionately among primary group members (e.g., family members, acquaintances). In contrast, the unique signatures of

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

113

instrumental homicides focus on attacks by intergroup members (e.g., persons of a different age category, different races, strangers), Black offenders, and multiple offenders. As an indication of the magnitude of differences in particular profiles, homicides involving young urban males with guns are more than 22 times more likely in situational contexts unique to instrumental homicides than uniquely expressive situations. When the profiles of expressive homicides are compared over time, the vast majority of configurations and cases of expressive homicides in the 1990s occur in structural contexts that are common to other decades.9 The common structures for expressive homicides over time almost always include adult offenders (96%), victims and offenders of the same age group (98%) or race (97%), and single offenders (98%). The primary contributors to the “emergent” profiles for expressive homicides in the 1990s are lethal shootings by urban teenage males in a variety of contexts. In fact, nearly one-half of the expressive homicides that occur in situations unique to the 1990s involve lethal shootings by urban teenage males, but this profile covers less than 1% of the homicides in situational contexts not unique to this time period. Although an increasingly prevalent profile over time, it is important to emphasize that this particular situational context is relatively rare for homicides with expressive motives, accounting for only 6% of these homicides between 1976 and 1998. The structural context for instrumental homicides is also best represented by a pattern of stability over time, although less so than is true among expressive homicides. Slightly over one-half of the configurations of attributes and the instrumental homicides within them are represented by stable situational contexts over time. However, this also means that a sizeable minority of instrumental homicides are found within historically distinct situational contexts. What is the nature of these stable and changing structures for instrumental homicides over time? Stable structural contexts for instrumental homicides over time almost always involve adult male offenders who kill other adults who are nonfamily members. The profile that derives from the combination of these attributes accounts for 83% of the instrumental homicides that occur in stable structures compared to only 55% of the instrumental homicides that are found with historically specific structures. Compared to felony-homicides in other decades, instrumental killings in the 1990s are far more likely to involve teenage offenders, Black offenders, guns as the lethal weapon, and large urban locations. Multiple offenders have also become more prevalent over time in some particular situations 9

Specifically, 64% of the situational contexts involving expressive motives did not change over time. These stable contexts accounted for 70% of homicides with expressive motives over the 1976–98 time period.

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

114

November 3, 2003

RETHINKING HOMICIDE

involving instrumental homicides. Felony-type situations involving young urban males with guns represent almost one-half (46%) of the homicides in uniquely instrumental contexts in the 1990s, but they are found in only 4% of the situations not unique to instrumental homicides in this decade. These large differences are found only among Black offenders (42% vs. 3%, respectively). Among White offenders, lethal shootings by young urban males during other felonies is not a major “emergent” context for instrumental homicides unique to the 1990s (5% vs. 2%). This context-specific pattern would not have been revealed by the simple application of the additive effects of each variable in a main effects model. Although some decade-specific patterns underlie the structures of expressive and instrumental homicides, it bears repeating that most structures of both types of homicide have not changed over the last three decades. Regardless of the decade, our QCA results indicate that intragroup attacks by nonstrangers are the situations that are most unique to expressive homicides. Attacks by adult male strangers and multiple offenders provide the unique signatures of the structural context of instrumental homicides.

Internal Consistency within Subsets of Instrumental and Expressive Homicides Our final analysis focuses on whether there is internal consistency or homogeneity in the structural profiles within subtypes of instrumental and expressive homicides. In other words, do the different kinds of instrumental homicides (e.g., robbery homicides, rape homicides) and different kinds of expressive homicides (e.g., lovers’ triangles, arguments over money, youth gang slayings) share the same common structural profiles as found within each general category of motives? The internal consistency of the structures of subtypes of instrumental and expressive motives is summarized in Table 5.3. As shown in the top panel of Table 5.3, subsets of homicide circumstances that are often classified within instrumental homicides vary widely in terms of whether they share the same structural profiles unique to other felonytype situations. Over 90% of killings in circumstances of rape and prostitution occur in situations with the same structural profiles unique to other instrumental homicides. The vast majority of homicides during robberies (89%), burglaries (88%), and auto thefts (86%) also occur in structural contexts similar to other instrumental homicides. In contrast, fewer homicides involving narcotic law violations (65%) and gambling (58%) happen in structural contexts that are unique to other instrumental homicides. The results for subtypes of expressive motives exhibit a similar pattern of internal consistency. Nearly 90% of homicides involving lovers’ triangles, arguments over money, and other arguments occur in situations with

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

115

Table 5.3. Comparability of Structural Profiles within Subtypes of Expressive and Instrumental Homicides

Percent of Profiles Unique to Homicides

Percent of Profiles Common to Both Expressive and Instrumental Homicides

Instrumental Homicides Rape Prostitution Robbery Burglary Auto Theft Larceny Other Sex Offenses Suspected Felony Other Felony Arson Narcotics Laws Gambling

92.1% 90.5% 88.6% 87.8% 86.0% 78.9% 78.1% 75.1% 71.4% 70.3% 65.0% 58.5%

7.9% 9.5% 11.4% 12.2% 14.0% 21.1% 21.9% 24.9% 28.6% 29.7% 35.0% 41.5%

Expressive Homicides Lovers’ Triangles Other Arguments Arguments over Money Brawl under Alcohol Brawl under Drugs Youth Gang Killings

88.7% 87.8% 85.3% 84.3% 75.4% 66.8%

11.3% 12.2% 14.7% 15.7% 24.6% 33.2%

Circumstance

structural profiles that are common to other expressive homicides. Brawls under alcohol and drugs occur somewhat less frequently in situations unique to other expressive homicides. Only about two-thirds of youth gang slayings occur in a structural context similar to expressive homicides; the remaining one-third involve structural profiles found among both expressive and instrumental homicides, a finding consistent with previous research suggesting that these homicides have both instrumental and expressive dimensions (see Block and Christakos 1995). The dominant conclusion from Table 5.3 is that structural profiles for each subtype of motive are internally consistent with other motive circumstances within the general category of instrumental and expressive homicides. As least in terms of their structural features, these results support the treatment of particular circumstances as types of expressive homicides and other circumstances as types of instrumental homicides. In short, the structural contexts of instrumental and expressive homicides are usually

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

116

November 3, 2003

RETHINKING HOMICIDE

qualitatively distinct, and this is true even among subsets of particular circumstances within these general motive categories.

Narrative Accounts of Expressive and Instrumental Homicides The analysis thus far has focused on the structure of homicide situations, examining whether instrumental and expressive homicides have unique structures and whether these structural contexts have changed over time. Our general conclusion is that these structures are in most cases qualitatively different for each type of motive. Now, we want to examine more closely the homicide context by exploring the nature of the situational dynamics found within homicide transactions. Narrative accounts of homicide events are used to focus primarily on the processual elements underlying these lethal encounters. By uniting these narrative accounts with our previous analyses, the remaining part of this chapter seeks to increase our understanding of both the process and structure that underlie homicide situations. Our QCA results and previous research have identified major subtypes of homicide situations within the general categories of expressive and instrumental homicides. Confrontational disputes among males and arguments among primary group members are the most common situations often linked to expressive homicides, whereas killings during the commission of a robbery are considered the most prevalent type of instrumental homicides. The nature of the interpersonal conflict and episodal transaction underlying these homicide situations is identified through the analysis of homicide narrative accounts from U.S. cities. The primary narrative accounts used here are based on homicides in Los Angeles throughout the twentieth century.10

Confrontational Disputes and Character Contests Killings that occur in the context of arguments and disputes are often considered “confrontational homicides” that resemble in many cases a form of “character contest” (see Athens 1985; Daly and Wilson 1988a; Luckenbill 1977; Polk 1994). These particular types of lethal disputes are usually characterized in previous research as male-on-male assaults among low-income persons in public locations that involve mutual combat and “face-saving” defenses to one’s social self. These confrontational disputes are generally found to represent a sizeable minority, if not majority, of all homicides (see Daly and Wilson 1999; Polk 1994; Savitz et al. 1991). 10

The characteristics of these narrative data are presented in Chapter 3.

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

117

Although arguments and quarrels among males are the dominant circumstances for Los Angeles homicides for each decade, the specific nature of these lethal disputes, the characteristics of the offender and victim, and the situational context vary widely and are not necessarily consistent with the popular image of character contests. For example, notice the variation in the context and triggering events in the following male-on-male arguments that resulted in homicides:11 Victim was killed with a blow to his head. Both the male victim and male offender were teamsters employed at a transfer company. A quarrel started over the merits of their respective teams when the victim called the offender a “bastard” and the offender retaliated by striking the victim over the left eye with his fist. The victim put his hands over his face and sank to his knees. The offender struck him twice more, once on each side of the head. The victim sank to the floor and was dead when the officers arrived a few minutes later. (Los Angeles, April 1910, 6:25 p.m.) Male victim was killed by veterinary surgeon at his “hospital.” The victim’s brother stated that the two brothers went into the place together, and while the doctor was working on a horse the victim threw a peanut shell at the doctor. The doctor, possibly believing that it was horse manure from a nearby pile, became enraged and struck the victim with a hard blow with his fist, throwing him to the stone pavement and causing a fracture of the skull. The victim, unconscious, was picked up and brought to the receiving hospital and died the following day. (Los Angeles, December 1910) The male victim was shot and killed by the male offender. The offender had insulted the victim’s sister-in-law. The victim went to the offender for an explanation. The offender in return asked the victim what he was going to do about it, and then the offender shot the victim through the stomach. (Los Angeles, May 1923: 8:50 p.m.) The 40-year-old Negro male victim was shot to death as the result of a previous quarrel over money involved in a gambling deal. The alleged offender had lost money in a craps game at the home of the victim. After the game, the suspect threatened the victim. (Los Angeles, January 1945) The 48-year-old Caucasian male victim was in a bar and went to the restroom. At that time the male offender took the victim’s seat and picked up the victim’s change and beer. An argument started and the owner of the bar 11

Our descriptions of the homicide incidents differ from the original police narratives in several ways. First, we have excluded personal identifiers (e.g., names of the victim, offender, witnesses, police officers) and replaced specific names of locations with generic names (e.g., Bob’s grocery store is described as a “grocery store”). Second, sentences and some wording in the original narrative accounts were rearranged to improve the readability of the accounts. None of these changes, however, substantially altered the sequence or context of the crime event. Pejorative language and slang comments are not changed from the original narrative because these words and phrases are a part of the historical context.

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

118

November 3, 2003

RETHINKING HOMICIDE

put the offender out the back door. The offender then came back to the front door and had a fight with another patron who knocked the offender down. The offender then left and returned in a few minutes with a butcher knife. The victim attempted to pacify the offender and offered to shake hands but the offender grabbed the victim by the lapel and pulled the victim toward him, stabbing the victim several times with the butcher knife. The victim was taken to the hospital where he expired. (Los Angeles, June 1951)

As illustrated by these male confrontations, lethal disputes have some common features (e.g., they often involve a previous dispute, the victim plays a contributory role, many involved types of “status threats” or public “dissing”), but they are also quite diverse in their particular triggering event, the lethal weapon, and other offense characteristics. Regardless of the historical period, the full sample of lethal disputes in Los Angeles indicates that arguments over gambling debts, girlfriends, rip-offs in drug transactions, and verbal insults about one’s masculinity, race, and family background are major circumstances underlying these killings. While the particular precipitating action may be viewed in many of these homicides as a “trivial altercation” (see also Wolfgang 1958), they are clearly not trivial in terms of their consequences. The importance of alcohol as a major precipitating factor is indicated by the occurrence of many of these confrontational homicides in or outside bars and the frequent mention of drinking as a primary activity when the crime occurred. Although character contests are often portrayed as public defenses of one’s social self in the presence of others, narrative descriptions of homicides in Los Angeles also illustrate that lethal disputes often occur in private settings (e.g., the victim’s home) and witnesses are not necessarily present in either public or private attacks. Slightly more than half (51%) of the dispute-related homicides in Los Angeles over the century occurred in either the victim’s home or other nonpublic setting (e.g., hotel room, boarding house). Third parties and public locations, however, are more common characteristics of expressive homicides among acquaintances and strangers than family members and intimates. Under these conditions, the nature of the situational context in which lethal disputes are resolved provides at best mixed support for their classification as character contests or other types of social defenses in public places. The most notable change in the nature and prevalence of lethal disputes in Los Angeles over the century is the rather dramatic increase in homicides involving multiple offenders since the 1960s. These situations have changed from drunken brawls to gang fights among young males. Multiple offenders were present in less than 5% of lethal disputes prior to 1960s but increased to nearly 20% of lethal disputes in subsequent decades. Our analysis of

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

119

SHR data for Los Angeles in the early 1990s indicates that over one-third of multiple offender homicides were classified as gang-related compared to only 6% of these homicides in the 1970s.

Defenses of “Significant Others” The full sample of lethal disputes in Los Angeles provides evidence that many expressively motivated homicides involve the defense of “significant” others (e.g., girlfriends, daughters, brothers, sisters-in-law, drinking buddies). Most lethal disputes involving the defense of others are also characterized by some level of either verbal or physical provocation by the victim. Offender’s claims of self-defense are often substantiated by witnesses, resulting in legal charges being dropped against the defendant. Within this context, the narratives of lethal disputes support Wolfgang’s (1958) assertion that it is often only a matter of chance who ends up being labeled the victim and offender in a homicide. However, homicides that are precipitated by a defense of others exhibit wide diversity, as illustrated in the following narrative accounts from various cities: The victim, witness, and several others had been drinking on a front porch. Actually, the group consisted of mostly family members and the victim simply joined the group. The victim snatched a wine bottle from the witness and drank from it. The witness told the victim to put the bottle down. The witness said the victim had sores on his mouth and the witness didn’t want to catch anything from the victim. The victim punched the witness in the face. The two fought and the victim banged the witness’ head into a car several times. Others broke up the fight and the witness’ brother chased the victim down the street. The victim said he would be back with his ‘thumper.’ The suspect had been in his auto. He saw the fight, but didn’t know the cause. The witness told the suspect what had happened. The suspect retrieved his shotgun and then heard screaming outside. The suspect heard his sister yell to someone not to start shooting with babies around. The suspect guessed she was yelling at the victim. The suspect went outside and saw the victim with a pistol. Those present said the victim waved the pistol at everyone. The suspect saw the victim chasing the witness, threatening to shoot the witness, so he (the suspect) shot the victim with a shotgun. Afterwards, the witness hid the weapon for the suspect. The witness said he believed the suspect had saved his life. (St. Louis, case #91054) A group of people was gathered in the street shooting dice or watching, socializing. Witness #4 and #8 began to fight. Associates of the victim and witness #4 said they began by ‘play fighting’ but that witness #8 broke his necklace and then he started fighting seriously. Witness #8 claimed the fight was serious from beginning, that they were fighting over $20 that witness #4 owed

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

120

November 3, 2003

RETHINKING HOMICIDE

to witness #8 and refused to pay. In any case, when the victim, witness #1, and witness #2 realized the fight was serious, they tried to break it up. Witness #8 then called for his brother, who punched the victim. This led to a kind of free for all in street, with onlookers trying to break up the fight. The suspect then hollered to witness #8 and his brother to ‘move out of the way’ and opened fire. Crowd dispersed, running in different directions. The victim collapsed. Witness #1 and witness #2 were also shot. (St. Louis, case #91196) The victim was living with his mother and stepfather. At approximately 3 p.m., the victim and his mother had an argument over a bottle of wine. Victim slapped his mother’s face. His mother then called her husband, who began arguing with the victim. The offender (husband/stepfather) went into the kitchen and got a knife and returned to where the victim was standing and stabbed him in the chest with the knife. The offender then called the police. The victim was taken to the hospital where he died. (Los Angeles, September, 1962) At approximately 10:48 p.m., patrol officers responded to a call that a female was at a door screaming. The caller then heard three shots being fired and a vehicle leaving at a high rate of speed. The patrol officers located a male and female victim at the address. They advised the officers that they were walking down the street when they were approached by three men in their late 20s to early 30s. The men tried to “pick up” the female. Words were exchanged. The male and female ran from the area and were chased by the three men. Two were on foot and the third was driving a small pickup truck. The male and female tried to hide when they were discovered by the three men. The male tried to defend himself from the assault by two of the pursuing men. After fighting for several minutes, the male pulled a handgun and fired several rounds at the three men, hitting one in the upper chest. The three ran to the pickup truck and fled the area. Police were later called to nearby location and discovered one of the attackers dead in the back of the pickup from the gunshot wound. After interviewing all people involved and recovering the gun, it appears that the shooting was in self defense. (Las Vegas, June 1998) The 35-year-old male victim and the 29-year-old male offender were friends who met almost every morning to talk and have coffee. On the morning of 6–25–75, they became involved in an argument when the victim made a remark about the suspect’s mother and aunt. The suspect demanded an apology which the victim would not give. Offender went home, got a knife and returned. The victim got up and tried to placate the offender and the offender stabbed the victim in the upper abdomen piercing the heart causing the victim’s death. The offender fled and was later arrested and filed on for one count of murder. (Los Angeles, June 1975)

As illustrated by these accounts, lethal violence is sometimes the outcome of an effort to intervene on behalf of a friend or family member. Under these conditions, violence may function as a form of social control in terms of

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

121

achieving functional goals. Violent behavior within these contexts may also be fused with issues of respect and identity, particularly when the individual intervening on behalf of another is shown disrespect by the antagonist (Oliver 1994). As Anderson (1998) notes, the ability to look after oneself and one’s loved ones grants an individual a level of respect and deference. As the last narrative exemplifies, defenses of other may be both physical and symbolic. Throughout the narratives we have examined, it is not uncommon to find that such disparaging comments about significant others are often the basis of “trivial” altercations that have deadly consequences. Both physical and symbolic defenses of others have also been noted by other researchers as a major circumstance in violent confrontations and disputerelated homicides (see Katz 1988; Luckenbill 1979; Oliver 1994; Polk 1994). The defense of children seems to be an especially important precipitating factor in female-perpetrated killings of intimate partners (see Browne 1987; Goetting 1995; Jurik and Gregware 1992).

Instrumental Violence and Felony-Type Homicide Situations Killings during robberies are the most prevalent type of instrumental homicide situations in Los Angeles and elsewhere (see UCR 2000). The narrative accounts, however, indicate that robbery homicides in Los Angeles occur in a variety of contexts and that these homicides differ widely in terms of whether the killing was an intended consequence or an unfortunate “side effect” of the robbery (see Maxfield 1989). As illustrated in the following narratives, there are substantial differences within and between street muggings and robberies of institutions in their modus operandi and precipitating factors:

Homicides in Street Muggings and Personal Robberies Victim was found in a park about 8:40 p.m. suffering from a gun shot wound in the lungs. Officers responded to the call and brought the victim to the hospital where he passed away. The victim made a statement to the police to the effect that yesterday he met a man in a saloon who took the victim to a park to meet some girls. After being in the park for some time, the offender pulled a gun on the victim, snapped it three times, and then shot him. The offender then robbed the victim of $12. The offender was arrested by the police, brought to the hospital, and positively identified by the victim. (Los Angeles, August 1911) The male victim was a driver of an auto containing two bank employees. He was shot with a shotgun and instantly killed by three bandits in a 7 passenger machine. The bank men were driving from the post office to their bank at 8:30 a.m., when the bandits’ car drove along side their car and without warning shot the victim. Covering the other passengers, the bandits took the

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

122

November 3, 2003

RETHINKING HOMICIDE

valuables, a large suitcase, etc., containing liberty bonds, valuable paper of private corporations, some currency, to the value of several hundred thousand dollars. The bandits escaped in the dense traffic. (Los Angeles, February 1923) The suspect had met the victim a short time previously in a beer parlor, and on the night of the killing, they had been to several bar rooms and then bought a bottle of liquor and went to the victim’s room. After consuming this bottle, they became involved in an argument, whereupon the suspect struck the victim, breaking her nose and cutting her mouth. She started to scream, then the suspect choked the victim and put a gag in her mouth, strangling her to death. Before leaving the apartment, the suspect ransacked the place, taking whatever money he could find. The suspect was picked up by the police detectives after residents had identified him as being seen with the victim. Fingerprints were found in the apartment, which were the suspect’s, and after several hours of questioning he confessed to the murder and told the story. (Los Angeles, February 1939) The 25-year-old Caucasian victim and a companion were picked up while hitching by two Negroes in a dark colored late model sedan. The victim was robbed and assaulted by the two suspects and when he resisted he was shot. On the orders of the suspects, he was removed from the holdup car and the holdup men left the scene. The companion called an ambulance and the victim died en route to the hospital. (Los Angeles, July 1945) The 60-year-old male Caucasian victim called the night clerk on the house phone and stated “Charlie, I’m shot.” Witnesses went to the room and found him seated on the bed bleeding from the mouth. When an ambulance crew got there, he was pronounced DOA. The victim was in a habit of carrying large sums of money and leaving his door open. Apparently, someone attempted to rob him and when he resisted, shot him. The following morning a revolver was found in the alley with indications that it had been dropped from a considerable height. It was identified as the murder weapon. (Los Angeles, February 1954: 7:20 p.m.) The 31-year-old Caucasian victim operates a catering lunch wagon and had just stopped at 61st street when one male Negro (suspect) walked up and bought a cup of coffee. The victim was talking with another customer. The suspect then pulled a sawed-off shotgun, 12-gauge double barrel, from under his coat and shot the victim in the left side. The suspect took the victim’s changer and removed the victim’s wallet from his pocket while he was on the ground. The wallet was later found in the vicinity of the catering truck. The suspect also took the wallet from the customer to whom the victim was talking at the time of the shooting. (Los Angeles, February 1962) While returning home from a laundromat the 15-year-old male victim was accosted by 3 males. He was ordered to hand over his letterman’s jacket. When he refused, he was stabbed fatally. Witnesses identified the killers. Two 18-year-old, Negro males were charged with murder. (Los Angeles, January 1973)

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

123

Robbery-Homicides of Store Owners and Businesses The male victim, the proprietor of a saloon, was closing his place when a male suspect in a black mask stepped out from behind a curtain in the back of the saloon and ordered the victim to throw up his hands. He threw them up and the man fired a gun. The bullet struck the victim in the forehead and went back over the skull. The suspect was described as about 5 ft. 6 in. and weighing about 150 pounds. (Los Angeles, October 1900) Victim was assaulted in his grocery and hardware store by two men who undoubtedly entered the store for the purpose of robbery. Witnesses in front of the store state that they saw the two men close in on the victim, one carrying a hand axe. They heard the victim say “don’t hurt me! You can have my money!” and that one of the men struck the victim on the head with the axe knocking him down. The men then dragged the victim to the cellar door, struck him twice more with the ax, took him down the cellar and left him there, fleeing out of the store. The victim was brought to the receiving hospital where he died shortly after. Both male offenders are in their 20s and are said to live in Las Cruces, New Mexico. The father of one of the offenders is a wealthy man of that town. (Los Angeles, October 1910: 4:00 p.m.) The male victim was shot and killed in his store by an unknown man accompanied by another. Both suspects escaped, leaving in a Ford. A third male is believed to have been waiting outside the store. The motive appeared to have been an attempted store holdup. There were two witnesses. (Los Angeles, December 1922: 6:50 p.m.) The 35-year-old Caucasian victim was closing his liquor store when he was held up. Indications were that the victim struggled with the suspect and was shot by the suspect. The victim ran from the store yelling “I’m shot.” Witnesses described the suspect as a 34–40 year old Caucasian. (Los Angeles, November 1951) The 73-year-old Negro victim was a bookmaker and gambler who had a shoe shine parlor. When he did not return as usual, his wife started looking for him. She found the victim in the rear of his parlor lying on his back with his clothing disarranged and his pockets turned inside out. An autopsy revealed that the victim had been strangled with his own neck tie. (Los Angeles, February 1976)

A closer examination of these robbery-homicide narratives across the decades reveals several common elements in the nature of these offenses. First, multiple offenders and greater levels of offense planning appear to be more salient factors in institutional robberies, whereas street muggings are more indicative of spontaneous acts by single offenders. Second, when the victim and offender are acquainted, the robbery appears to be primarily an afterthought that follows the fatal attack. Third, based on witness statements it appears the methods used by offenders to make initial contact with the victim (e.g., driving up in a car, asking the victim for change or directions,

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

124

November 3, 2003

RETHINKING HOMICIDE

accosting attractive targets on the street) have remained similar across the entire century. Witness statements further indicate that some degree of victim resistance is a persistent situational element that precedes these deadly robberies in each decade.

Drug-Related Robberies Killings within the context of illegal drug transactions are common circumstances for robbery homicides in many large U.S. cities. Our examination of homicide narratives in the last two decades in select cities (e.g., Las Vegas, Miami, St. Louis) reveals that many of these lethal drug-related robberies are precipitated by the offender’s perception of a “rip-off” or other type of “dissing.” However, the distinction between instrumental and expressive motivations becomes blurred in many drug-related robberies, as illustrated by the following narratives from Miami in the 1980s (see Wilbanks 1984): Two victims (a 35-year-old Black male and an approximately 25-year-old Black female) were found shot to death on a roadway in far south Dade. The two victims came to Miami from Detroit to buy heroin. The sellers, two Latin males and a Latin female decided to rip-off the victims for the $29,000 they had with them instead of making the sale. The offenders handcuffed the two victims and took them to far southeast Dade. The male victim was taken from the car (left female victim with female offender in the car) and killed him. They came back and debated whether they should kill the female (while she begged for her life). The female offender gave orders for her to be killed and so she was killed. Later the wife of one of the offenders came forward (after numerous beatings by her husband) and told the whole story to the police (she was with the offenders when the killings took place). She is now in protective custody. (Miami, case #504 and #573) The 27-year-old Latin male victim set up a drug deal between groups of 5 or 6 Blacks. The Blacks took the money and fled, then the Latins shot the victim in anger after the rip-off and then took the victim to the hospital where he died. (Miami, case #011) The 41-year-old victim was a Columbian who had arrived in Florida only one week before. The victim was involved in the drug traffic and was killed (shot) in a dispute over the selling of drugs. The offender was caught before he could take the drugs and money (which he stole in this killing/robbery) and flee. The offender was tied to the crime partly by teeth marks on the body. The victim was killed in a brutal fashion. (Miami, case #103) . . . It appears that the 25-year-old Latin male victim and the 21-year-old Latin male offender had been involved in a drug deal. The victim called the offender and said he had one gram of TCP he would sell for $90. The offender was not willing to pay this price (he suggested $75) and so told another he would rip-off the victim (he also later told others he killed the victim). (Miami, case #107)

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

125

For many of the homicides that occur within the context of drug transactions, an instrumental motivation can be reasonably inferred because of the planning involved in the offense and the specific objective was to rip off the drugs (see case #504). However, in other homicides within drug transactions, the motive appears to be a more spontaneous or expressive response to a perceived wrong (see case #11). The nature of the criminal event (e.g., teeth marks and the brutal killing noted in case #103) also suggests a blend of instrumental and expressive motives in these homicide situations. The variation in the nature of these killings may help explain the earlier noted QCA finding of a lower percentage of narcotic-related homicides occurring in structural contexts unique to other instrumental homicides than was the case for many other types of felonies.

Summary of Findings Our analyses of narrative accounts of homicides over the last century and SHR data from 1976–98 reveal several general patterns about expressive and instrumental homicides. These patterns are summarized below. First, expressive homicides are far more prevalent than instrumental homicides across the entire twentieth century. Based on trends in Los Angeles, homicides in the context of arguments, brawls, and other interpersonal disputes increased over the last half century, from 40% in the 1900–19 period to 65% in the 1960–77 period. SHR data indicate greater stability in the relative proportion of expressive and instrumental homicides but some changes within these categories as well. For example, the proportions of homicides involving youth gang slayings and narcotic law violations have both increased substantially between the 1970s and 1990s. However, each of these types of homicide situations represents less than 10% of all homicides in the last decade. Robberies continue to be the most common circumstance underlying instrumental homicides, while arguments are the most common basis for expressive homicides in each decade. The predominance of expressive motivations in homicide is also revealed in archival accounts of the full span of American history (see Lane 1997). Second, expressive and instrumental homicides are unique in their nature and situational context. Based on our examination of narrative accounts, expressive homicides often involve confrontations among men, trivial altercations, victim provocation, and the presence of alcohol. When the parties have a prior relationship, these lethal disputes are often the termination of a previous and longstanding dispute. However, the image of a “character contest” does not accurately describe all of these lethal disputes because these homicide situations occur in both public and private settings, and are not restricted to lower class members. In contrast, instrumental homicides are more often attacks by strangers and multiple offenders in a variety of different situational contexts. Elements of planning are also more

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

126

November 3, 2003

RETHINKING HOMICIDE

commonly observed in some types of instrumental homicides (e.g., those involving institutional robberies) than other types (e.g., street mugging). Third, the vast majority of homicides in the United States occurred within a relatively small number of structural contexts. In fact, less than 300 distinct combinations of offender, victim, and offense attributes form the structural basis for about 86% of all U.S. homicides in the last three decades. The vast majority of these structural profiles underlying homicides are unique to either instrumental or expressive homicides. Fourth, instrumental and expressive homicides are often qualitatively different in terms of their structural bases. At least two-thirds of homicides in each decade occur in situational contexts that are unique to either instrumental or expressive homicides. Intragroup disputes among nonstrangers that involve only one offender are the most common structural contexts unique to expressive homicides, whereas intergroup killings by male strangers in a group context within large cities are the distinct attributes often underlying lethal situations unique to instrumental homicides. Fifth, although the situational context for both types of motives has remained relatively stable over time, some “emergent” situational contexts for homicide have occurred in the 1990s. For both instrumental and expressive homicides, this changing structural context has involved an increased proportion of homicides by young Black urban males with guns in a group context. This particular situational context for homicide accounts for only a small proportion of all homicides over the last three decades, but it represents an emergent trend for both instrumental and expressive homicides. Sixth, the unique structural profiles for instrumental and expressive homicides accurately represent the majority of homicides within each of these general categories. Internal consistency is highest among instrumental homicides that occur within the context of rape and prostitution. Lover’s triangles, arguments over money, and other arguments are the types of expressive motives that represent the highest proportion of homicides sharing structural features with other expressive motives. In contrast, homicides occurring within the context of gambling and drug activity are the least similar to other instrumental motives in terms of their underlying structure. Nearly one-third of youth gang homicides (often considered an expressively motivated crime) occur in structural contexts similar to both expressive and instrumental homicides.

Conclusions and Implications The current research was motivated by an ongoing debate about the nature and usefulness of the instrumental–expressive distinction. Most researchers have explicitly assumed that instrumental and expressive crimes are qualitatively distinct, while others characterize these general types as polar ends of

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

127

a continuum. The work by Felson (1993), in particular, suggests that this distinction is of little practical value because all offenses ultimately are thought to involve a pursuit of instrumental goals. Our results provide empirical support for both characterizations of expressive–instrumental crimes. For example, the homicide narratives and QCA results provide strong evidence for treating instrumental and expressive homicides as distinct phenomena. Narrative descriptions indicate clear differences in the nature of the situational dynamics across motive circumstances. QCA reveals that the vast majority of homicides occur in structural contexts that are qualitatively different for expressive and instrumental homicides. However, the finding that about one-third of homicides occur in situational contexts that are common to both instrumental and expressive homicides provides some evidence of comparability across motive categories. The most important conclusion that derives from our analysis, however, is that characterizing homicides as either instrumental, expressive, or a combination of the two is a complex issue (see also Polk 1994). Both within and across these general crime categories, there is variation in the nature and prevalence of particular combinations of offender, victim, and situational factors. Under these conditions, the treatment of instrumental–expressive crimes as a simple dichotomy may result in a misrepresentation of both their common and unique signatures. This is especially true of homicides that occur in the context of youth gang violence and drug transactions, where the distinction between instrumental and expressive motives is the least clear. Over the last two decades, there has been a growing debate about the relative merits of general theories of criminality and situational-based theories of crime events. General theories like Gottfredson and Hirschi’s (1990) self-control approach purport to account for all forms of crime across situational contexts, whereas most theories of victimization (like routine activities and lifestyle/exposure theories) focus on the interplay between target exposure, attractiveness, and guardianship in establishing a particular situational context facilitating criminal acts (see Miethe and Meier 1994). Within the context of theory construction and evaluation, the results of the current study have several implications. For those interested in general theoretical development, our results indicate that some of the distinctions made in past research between different types of homicide (e.g., instrumental vs. expressive, gang vs. non-gang, drug-motivated vs. other instrumental killings) are based on a somewhat weak foundation because these particular subsets are not always unique in their particular configurations of offender, victim, and situational characteristics. However, the finding of unique structures in a majority of cases also suggests that motivespecific explanations, rather than a general theory, provide a better account

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

128

November 3, 2003

RETHINKING HOMICIDE

of the structure of homicide situations. The narrative data provide further support for the value of motive-specific explanations. For advocates of situational approaches, the discovery of both common and unique characteristics underlying homicide situations also tells us much about the nature and explanation of homicide. For example, our QCA models of unique combinations of attributes reduce dramatically the complexity of the full array of situational elements without over-simplification of the situational dynamics in homicide. Particular characteristics of these unique configurations, in turn, may be used for more rigorous theoretical development to account for this variation in homicides across various situational contexts. By examining more closely unique configurations of instrumental and expressive homicides, we may better understand the social processes and triggering mechanisms that are involved when particular individuals and situational factors converge in time and space. Our qualitative analysis of homicide narratives greatly supplements the QCA results by illustrating aspects of the situational dynamics and processual elements that are commonly found among subsets of instrumental and expressive homicides. Several major themes are present within each of these situations. For example, confrontation homicides do not necessarily follow the image of character contests because they happen in both public and private locations, with and without witnesses, and are often the result of long-term conflicts rather than situationally emergent affronts to one’s social identity. The narratives from Miami homicides provide clear verification of the QCA results that drug-related homicides have structural elements and situational dynamics that are indicative of both instrumental and expressive motivations. Under these conditions, the integration of traditional qualitative methods to study process and QCA to identify unique and common structures provides a more comprehensive approach to the study of homicide situations than either method used separately. From the perspective of public policy, our results have several immediate implications in relation to the distinct structural profiles for both instrumental and expressive homicides which emerged in the 1990s. Lethal shootings by urban teenage males became an increasingly prevalent and distinct profile for both expressive and instrumental homicides over time. Homicides within this emergent context also disproportionately involve Black youth and multiple offenders. However, teenage offenders are relatively rare in both instrumental and expressive homicides. Rather than focusing on these isolated and rare forms of homicide, public policy would be better informed by focusing on the dominant characteristics of homicide situations. For both instrumental and expressive homicides, the dominant elements of homicide are that the offenders and victims are disproportionately Black adult males and they occur in the context of urban areas and with guns. The primarily unique aspect among the general motives

17:26

P1: JMV 0521832993c05

CB617-Miethe-v1

November 3, 2003

INSTRUMENTAL AND EXPRESSIVE MOTIVES IN HOMICIDE SITUATIONS

129

is that instrumental killings are more likely to involve intergroup assaults by multiple offenders who are strangers and expressive homicides are more typically an intragroup phenomena involving adult acquaintances or family members. Persistent in their effects over time, it is these basic attributes underlying homicide situations that should occupy the attention of public policy.12 Finally, increasing attention over the last decade has been given to the topic of crime profiling (see Godwin 2001). Profiling in the study of homicide often involves the comparison of different types of homicide, exploring their common and unique features. For theoretical work on the general topic of profiling, our results suggest that general typologies of homicide based on a few underlying categories or “syndromes” may be of limited value because even among categories of instrumental and expressive homicides there are approximately 300 subprofiles within each decade that are either distinct or common across crime categories. For law enforcement personnel interested in the creation of crime profiles, our results illustrate how to empirically identify particular configurations of offender, victim, and offense attributes that have relatively unique occurrences within particular types of crime. However, as in general research on homicide subtypes, it is important to recognize that these profiles of homicide structures are complex, involving distinct intersections of particular sets of attributes. By providing an analytic tool for examining differences within and between subgroups, we think that QCA has enormous potential for developing and validating crime profiles across a variety of contexts. 12

Specific policy implications of this research will be addressed in Chapter 10.

17:26

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

C H A P T E R

S I X

Gender Differences in the Structure of Homicide Situations

one of the most consistent criminological findings over time and across cultures is the substantial overrepresentation of males as violent offenders. Males are the offenders in about ninety percent of the U.S. homicides known to the police in the 1990s. This proportion has been remarkably stable throughout American history (see Brearley 1932; Lane 1997; McKanna 1997; Simon 1975; FBI 2000). Women murderers were sparse in Colonial America, rarely acknowledged in the 1800s, and accounted for only a small proportion of homicides in the 1900s. Killings by women are clearly the exceptional case in other countries, and this trend dates back to the Middle Ages (see Gartner, Baker, and Pampel 1990; Hartnagel 1982; Lane 1997; Silverman and Kennedy 1993). Rates of homicide victimization are also far higher among men than women (see Gartner 1990). Although the gender gap in rates of homicide offending exhibits little change over time, a large volume of literature suggests that there are major differences between men and women in the types of circumstances and situations in which they kill. These differences are most prominent with respect to the killing of intimate partners (see Adinkrah 1999; Campbell 1992; Daly and Wilson 1988a; Goetting 1995; Jones 1980; Jurik and Winn 1990; Wilson and Daly 1992; Wolfgang 1958). Unfortunately, most of these studies examine gender differences with respect to particular variables (e.g., victim–offender relationship, motive, or victim age), rather than in the totality of combinations of offender, victim, and offense elements that define the basic structure of homicide situations. Using the method of Qualitative Comparative Analysis (QCA) and narrative accounts of homicide situations, the present chapter examines stability and change in the structure of homicide situations in Chicago and

130

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

131

other cities over various time periods.1 We define the structure of homicide situations by their combination of offender, victim, and offense attributes. Gender-specific comparisons of homicide situations over the last four decades are made to evaluate the unique and common elements of male and female homicide and potential changes in their structure over time. The findings are then discussed in terms of their implications for future research.

Patterns of Homicide by Male and Female Offenders A major assumption underlying previous research on gender and homicide is that male and female offenders are qualitatively distinct, killing under different circumstances and situations. Historically, female offenders have been associated with higher rates of infanticide and killings that occur within intimate partner relations (see Block and Christakos 1995; Browne, Williams, and Dutton 1999; Jensen 2001; Kaplun and Reich 1976; Lane 1997; Wilson and Daly 1993). Male-perpetrated homicides have long been linked with notions of masculine competitiveness and a subculture of young, single, poor, and transient males (see Daly and Wilson 1988a; Lane 1997; Polk 1994). Studies of intimate partner homicides by women and men indicate differences in circumstances prior to and at the time of the homicide, the motivations for the lethal assault, and the location of the crime (see Browne, Williams, and Dutton 1999; Jurik and Winn 1990; Riedel and Best 1998; Silverman and Kennedy 1993; Wilson and Daly 1993; Wilson, Daly, and Wright 1993). Differences in weapon use by gender have also been observed in previous research, with a greater use of knives occurring among femaleperpetrated homicides (Riedel and Best 1998; Silverman and Kennedy 1993; Wolfgang 1958). Aggregate studies of gender-specific homicide rates report mixed findings concerning differences in the structural correlates of male and female homicide rates (see Jensen 2001; Steffensmeier and Allan 1992; Steffensmeier and Haynie 2000; Weisheit 1993). Homicides committed by women are argued to be predominantly motivated by self-protection against aggressive and threatening behavior by 1

Chicago homicide data are used in this chapter to explore the structure of homicide situations for two reasons. First, these data cover a longer historical period than the SHR data, providing a greater opportunity to study how macro-level changes in routine activities over time affect the situational context in homicide. Second, the Chicago homicide data provide several measures of the situational context that are not available in the SHR data (e.g., time of the offense, whether it occurred in a private or public location). Both of these variables are important contextual factors for understanding differences in male- and female-perpetrated homicide. The nature of this particular sample of homicide cases will be described shortly.

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

132

T1: FCH November 3, 2003

RETHINKING HOMICIDE

males (Daly and Wilson 1988a; Goetting 1987; Polk 1994; Rosenfeld 1997; Wilbanks 1983). Browne (1987) asserts that women kill in response to physical threats from their male victims – a defensive reaction by women, because of harm or potential harm to them, their children, or other family members. Jurik and Gregware (1984) also find that histories of violence underlie cases in which women kill someone they know. Other researchers (Browne 1987; Ward, Jackson, and Ward 1979; Wolfgang 1958) contend that men and women commit homicide in manners reflective of traditional gender roles. Although the notion of “character contests” is widely applied to explain the motivations for male homicide (see Luckenbill 1977; Polk 1994), this metaphor does not appear to represent most women offenders of homicide (Jurik and Gregware 1992). In their review of studies of women murderers, Silverman, Vega, and Danner (1993:187–188) conclude that the following characteristics and patterns are associated with homicides by females: median age of over 30, African-American offenders, married, low education, low socioeconomic status, substance abusers, history of physical abuse, intact family relations, few prior offenses, victims who were almost exclusively male intimates, killings in the home, firearm usage, intoxication at the time of the offense, arguments and abusive situations as precipitating motives, absence of codefendants, and few felony murders. Although gender differences in homicide offending and its correlates have been identified in previous research, conclusions from these studies about the gender-specific nature of the situational context of homicide are limited in several fundamental respects. First, most studies have used cross-sectional designs, prohibiting an investigation of whether the nature of homicide situations for each group has remained stable or changed over time. Second, gender differences with respect to particular variables (e.g., victim–offender relationship, motive, weapon use) are often examined in statistical models as main or additive effects. Unfortunately, when homicide situations are defined by the conjunction of offender, victim, and offense characteristics, the main effects model is incapable of evaluating whether men and women kill under qualitatively different situations because it assumes that each variable has a constant impact across all contexts. In short, the main effects model is theoretically and analytically incompatible with the study of conjunctive or combinatorial causes that underlie homicide situations.

Theoretical Perspectives Theories of homicide can be distinguished in terms of their emphasis on the offender or the situational context of the crime. Offender-based theories focus on the sources of criminal motivation across contexts, whereas theories

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

133

of homicide situations explore the interplay of particular offender, victim, and offense attributes that underlie violent acts. Various theoretical perspectives have been used in previous research to explain gender differences in the prevalence of homicide, its situational context, and both change and stability in its prevalence and context over time.

Theories of Homicide Offending The disproportionate involvement of males as homicide offenders has been explained by a number of biological, psychological, and sociological theories. For example, some research links male violence to testosterone and other biological factors (see Archer and McDaniel 1995). Sociological theories emphasize differential opportunity, social learning, subcultural influences and value conflict, lifestyles, and gender roles or masculine socialization as contributory factors for higher levels of male homicide (see Cohen and Felson 1979; Daly and Wilson 1988a; Hindelang, Gottfredson, and Garofalo 1978; Polk 1994; Wolfgang and Ferracuti 1967). Evolutionary perspectives have also been developed to account for the overrepresentation of males in homicide offending (see Daly and Wilson 1988a; 1988b; 1999; Wilson and Daly 1995). While the stability of gender differences in homicide offending is explained by some theories (e.g., biological predispositions, historical legacies of masculine competitiveness), changes over time are typically accounted for by causal factors identified by alternative theoretical perspectives (e.g., changing economic opportunities, declining legitimacy of social institutions, changes in socialization practices). Regardless of whether theories of offender motivation are offered as general or gender-specific approaches, they tend to be nomothetic explanations for differences in prevalence rates across groups and across contexts. However, these theories are usually limited because they either ignore or cannot account for variation in these patterns across different situational contexts. For example, some general theories of crime (e.g., Gottfredson and Hirschi’s “Self-Control Theory” (1990)) attribute the high aggregate rates of homicide offending by males to lower self-control, but this theory cannot easily explain higher proportions of female-perpetrated homicides within particular situational contexts (e.g., knife killings of family members or intimates in indoor locations). To explain such context-specific differences, self-control theory must argue that differences in criminal opportunity within this particular context negate the impact of self-control. While opportunity factors have been increasingly recognized as an important feature of self-control theory, no current formulation of this general theory contends that opportunity differences will overpower and replace low self-control as the predominant causal factor in homicide or other crimes.

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH

134

November 3, 2003

RETHINKING HOMICIDE

Other theoretical approaches that focus exclusively on offender motivation are also unable to readily account for the possibility of both unique and common situational contexts for homicides within and across genders. For example, current theoretical formulations that are gender-specific (e.g., theories about masculine competitiveness, character contests) may account for some of the uniqueness of particular victim and offense characteristics underlying male- and female-perpetrated homicides. However, these gender-specific theories seem limited in their explanation of homicides that occur within situational contexts common to both genders. For example, masculine competitiveness may be ineffective as an explanation of homicides involving expressive killings of acquaintances with guns if both men and women are found to kill in such situations. Under these conditions, neither general theories nor gender-specific theories of offender motivation are adequate for explaining the wide diversity of possible structures that underlie homicide situations. This is the case because most theories of offender motivation largely neglect both the victim and offense attributes that help define the structure of homicide situations.

Theories of the Structure of Homicide Situations The question of whether male- and female-perpetrated homicides are distinct really focuses on whether men and women kill in qualitatively unique situations. If homicide situations are defined by combinations of offender, victim, and offense characteristics, it is easy to see that most theories of offender motivation are limited for answering questions about unique or common signatures because they often exclude two of the three basic structural elements of homicide situations. As a general framework for studying homicide situations and changes in them over time, there are several combinatorial or conjunctive theories of crime that are based on the integration of offender, victim, and offense elements. The most relevant theoretical bases for studying homicide situations involve Lofland’s (1969) work on facilitating conditions and criminal opportunity formulations of crime and victimization (e.g., routine activities, lifestyle theories). These theories have been described more fully in Chapter 2 and therefore are only briefly summarized here. John Lofland (1969) uses symbolic interactionism to explore the situational cues and the interpretive processes that underlie an individual’s decision to commit deviant acts. From this perspective, necessary conditions for the completion of deviant acts are the presence of facilitating places, facilitating hardware, and facilitating others.2 2

Other necessary conditions include some level of motivation for action (caused by threat and encapsulation) and a modicum of knowledge and skill to commit the deviant act. Under this

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

135

When applied to the study of homicide, several structural elements of the situational context would contribute to these facilitating conditions for lethal violence. For example, most outdoor locations at nighttime are considered facilitative places because dim lighting reduces public visibility, whereas the home may be a prime location for violence because it is bounded by walls, legal rules about privacy, and interdependencies among the inhabitants that reduce the public visibility and surveillance of deviant activity within this setting (Lofland 1969:63). Everyday household objects like knives, bottles, hammers, and bricks are widely accessible and available instruments whose mere presence facilitates the commission of a violent act. Firearms are especially facilitative hardware for homicide in that they are widely available, do not require physical strength differentials or even direct bodily contact with the victim, and are more lethal than other hardware. “Facilitating others” includes any person (e.g., “significant others,” co-offenders, third parties) who provides little resistance against the lethal violence or may be perceived as encouraging it. Criminal opportunity theories (e.g., routine activities, lifestyle/exposure theories) are explanations of crime events that cover both micro- and macrolevels of analysis. From this perspective, various macrostructural changes in routine activity patterns and lifestyles provide an opportunity structure for crime by affecting the convergence in time and space of three minimal elements of direct-contact predatory crimes: (1) motivated offenders, (2) suitable targets, and (3) the absence of capable guardianship. Individuals’ differences in routine activities and lifestyles are important under this perspective because they are ultimately related to differential exposure to dangerous places, times, and others. Based on a criminal opportunity framework, several social trends may contribute to particular changes in the nature of male and female offending in homicide. Specifically, increases in public activity (through the rise in nondomestic labor), wider population mobility, decreases in marriage, and the rise in single-person (primary) households are fundamental changes that may increase one’s chances of contact and conflict with acquaintances and strangers in public settings, thereby increasing rates of expressive acts against nonintimates and persons of different ages, races, and genders in outdoor locations. These potential changes in the situational context of homicide, however, may be especially relevant for particular types of offenders (e.g., women, persons under 30 years old) and less predictive of changes among other combinations of victim, offender, and offense characteristics. For example, changes in marriage patterns over the last several decades have been linked to declining rates of intimate partner homicides (see Rosenfeld theory of facilitating conditions, the absence of any of these factors will thwart the commission of the criminal act.

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

136

T1: FCH November 3, 2003

RETHINKING HOMICIDE

1997), and these societal changes may potentially result in a displacement of homicides by particular offenders (e.g., young White women) to other targets and locations (e.g., acquaintances or stranger victims in outdoor locations). Criminal opportunity theory suggests that differences in lifestyles and routine activities may also influence the nature of gender differences in homicide situations in other ways. For example, due to lifestyle differences that enhance male’s exposure to a wider range of facilitating places, hardware and others, homicide situations involving male perpetrators may be far more diversified in the range of offender, victim, and offense characteristics that underlie them. This may include a greater level of intergroup attacks (including intersex, interracial, and interage), stranger homicides, and more diversity in the temporal and spatial locations of the crime. In contrast, traditional routine activities and lifestyles for women may place constraints on their opportunities for homicide, limit the range of motives under which they kill, and restrict the pool of potential targets, hardware, and locations. Under these conditions, female-perpetrated homicides should be characterized by less diversity in the combination of offender, victim, and offense elements that define the structure of homicide situations in earlier decades (e.g., 1960s), but these homicide situations should become more diversified in later decades (e.g., 1990s) as structural constraints on women’s routine activities and criminal opportunities are weakened over time. From the joint perspectives of criminal opportunity theory and Lofland’s facilitating conditions, several structural changes in American society over the last four decades should influence the nature of gender differences in homicide situations over time. Specifically, increases in single-person households, increased nonfamilial activity, population mobility, greater density of gun ownership, and increased public activity at night should alter the structure of homicides for each gender (because these factors are representative of changes in facilitating places, hardware, and others, and decreased guardianship). However, most of these structural changes have had a far greater impact on the routine activities of women than men, suggesting that homicide situations among female offenders should be especially susceptible to these changes over time. If these theoretical orientations are useful for understanding homicide situations, increases in intergroup conflict, nonfamily victims, guns as the lethal weapon, outdoor locations, and nighttime offending should produce substantial changes in the structure of homicide situations for both men and women between the 1960s and the 1990s.

Research Questions Previous research has clearly demonstrated the predominance of males as homicide offenders. However, what is less obvious is whether males and

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

137

females kill under qualitatively different situations, involving relatively distinct combinations of particular offender, victim, and offense attributes. Much research has explored either offender characteristics, victim attributes, or offense characteristics of homicide, but previous studies have not employed a combinatorial analytic method to identify the particular combinations of offender, victim, and offense attributes that may provide unique signatures for homicide within and between genders. Using QCA and data on homicides in Chicago and other U.S. cities, we examine three related questions about the nature of gender differences in homicide situations. First, do homicide incidents involving male and female offenders exhibit common or unique patterns in their combination of offender, victim, and offense elements? Second, have gender differences in the basic structure of homicide situations (i.e, configurations of offender, victim, and offense factors) changed over time? Third, what are the dominant profiles of homicide situations that uniquely occur for each gender? Narrative accounts of homicides in several U.S. cities will be used to explore more fully the episodal nature of these gender-specific homicide situations. After documenting the dominant profiles underlying homicide situations, the observed results will be compared with expectations based on existing theories of offender motivation and criminal opportunity. The findings are then discussed in terms of their implications for future research.

Methods and Analysis The primary data source for the current study is the Chicago Homicide Project that is available through ICPSR. This data source covers all homicides in Chicago police records from 1965 to 1994. These data have been widely used by homicide researchers over the last three decades (see, for example, Block 1976; Block et al. 2001; Block and Christakos 1995; Block and Block 1992; 1993; Zimring, Mukherjee, and Van Winkle 1983). A total of 22,704 homicides are included in this data file over the study period.3 Although the analysis is restricted primarily to a single city, Chicago homicides are comparable to other large U.S. cities in terms of their gender and other offender, victim, and offense attributes. The high concentration of homicides among Black offenders in this city is the most disparate with national 3

To correct for the problem of missing data in the Chicago homicide files, the method of Expectation Maximization (EM) was used for imputing values for all cases with missing data. A comparison of results with and without imputation of missing data revealed some differences in the prevalence of particular homicide situations. The single largest increase after imputation was the rise in instrumental homicides, which increased by 6% among female offenders and by 8% for male offenders, whereas the marginal distributions for most variables changes by only 2% to 4% after imputing missing data. The analyses based on the EM imputation are presented here because they maximize the sample size and they are thought to best represent the distribution of all homicides in Chicago in the study period.

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

138

T1: FCH November 3, 2003

RETHINKING HOMICIDE

trends. Our selection of the Chicago data source was due to its comprehensive longitudinal design and its inclusion of many key variables that are not available in other sources (e.g., SHR data). Narrative accounts of homicide cases in various cities are also used to further explore gender differences in the nature of homicide transactions. The major variables include offender characteristics, victim characteristics, and offense elements. The variables included in this analysis represent those offender, victim, and offense attributes widely used in past studies and particular offense attributes (like type of lethal weapon, time of day, public or private location, multiple offenders) that are indicative of the presence of either facilitating places, facilitating hardware, or facilitating others (see Lofland 1969). A listing of these variables, their coding, and frequency distributions over time periods are presented in Table 6.1. Measures of offender and victim characteristics include their gender, race, and age.4 Race is coded as either White, African-American, or Hispanic. Victims and offenders of other racial/ethnic groupings were excluded. Three categories are used to classify the offender’s and victim’s age, allowing us to compare young persons (<20 year olds) with two adult groups (20–29 year olds, 30 and older). Homicides are considered “same age” (i.e., intra-age) when the victims and offenders differ by less than one category. “Different age” homicides involve differences of greater than one category (e.g., a 30 or older offender and a victim less than 20 years old). Dummy variables are also constructed for “same race” and “same sex” groupings. This coding scheme was used to evaluate directly whether there are gender differences in the magnitude and nature of homicide as an intragroup phenomenon. Similar to other studies (see Decker 1993, 1996; Hewitt 1988; Maxfield 1989; Petee et al. 2001; Polk 1994; Williams and Flewelling 1987), the victim–offender relationship is coded into the categories of family members/intimates, acquaintances, and strangers. Intimates and family members are persons involved in primary group relations and include spouses (both common-law, marital, and ex-spouses), parents, children, boy/ girlfriends, and homosexual lovers. Analyses of gender differences within 4

Rather than exclude cases of multiple offenders or multiple victims, we used the response categories of the “first” offender and victim in the coding of these variables. This decision affects the formal interpretation of particular variables in situations of multiple offenders (29% of homicides) or multiple victims (7% of homicides). For example, in the case of the first offender being under 20 years old, the formal interpretation in multiple offender homicides would be of the form “situations that involve at least one teenage offender are more common among male than female offenders.” Due to the relative infrequency of multiple offenders or multiple victims in Chicago homicides and for sake of brevity, we will not make repeated reference to the “at least one . . .” interpretations in the text. It is important to note that the general patterns of our results about gender differences are similar when multiple offender and victim situations are either included or excluded.

15:40

Year of Homicide 1965–69 1970–89 1990–94 Type of Homicide Female Offender∗ Male Offender Offender Characteristics African American Hispanic White∗ <20 yrs. old age 20 to 29 age 30 to 98∗ Victim–Offender Characteristics V–O Same Age Group V–O Different Age Group∗ V–O Same Race V–O Different Race∗ V–O Same Sex V–O Different Sex∗ V–O Family/Intimate V–O Acquaint V–O Strangers∗

Variables 2,810 15,457 4,437 2,735 19,969 17,468 2,956 2,280 5,787 9,714 7,203 20,354 2,350 20,112 2,592 16,696 6,008 4,322 12,273 6,109

Ofem Omale Oblk Ohsp Owht Olt20 O2029 O3098 =Age =Age =Race =Race =Sex =Sex Fami Acqu Stran

Total N

Yr6569 Yr7089 Yr9094

Value Labels#

139 89.1 10.9 87.9 12.1 81.4 18.6 13.5 57.4 29.1

75.3 14.3 10.4 27.1 42.7 30.2

% Male

93.4 6.6 94.9 5.1 12.2 87.8 63.4 31.5 5.1

86.5 3.0 10.5 7.3 38.5 54.2

% Female

90.0 10.0 87.9 12.1 77.9 22.1 19.3 54.9 25.8

77.1 5.9 17.0 21.8 36.7 41.6

% Male

Yr 1965–69

90.2 9.8 93.7 6.3 19.8 80.2 51.5 33.2 15.4

91.2 5.9 2.9 23.7 37.1 39.3

% Female

88.7 11.3 90.1 9.9 84.5 15.5 10.5 63.5 25.9 (cont.)

79.1 14.9 6.0 36.8 40.2 23.0

% Male

Yr 1990–94

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

93.5 6.5 93.4 6.6 16.3 83.7 59.4 29.6 11.0

88.7 3.9 7.3 13.7 43.3 43.0

% Female

Yr 1965–94

Table 6.1. Coding, Value Labels, and Frequencies for All Variables (Chicago Homicides, 1965–94)

P1: FCH/FFX 0521832993c06 T1: FCH November 3, 2003 15:40

Instr Expr Gun Knife Oweap Mult Sing Outd Indoor Nite Day

Value Labels# 6,322 16,382 14,474 4,572 3,658 6,494 16,210 9,582 13,122 15,666 7,038 22,704

Total N 11.1 88.9 41.2 44.2 14.6 12.7 87.3 21.4 78.6 65.7 34.3 2735

% Female 30.1 69.9 66.8 16.8 16.3 30.8 69.2 45.1 54.9 69.4 30.6 15661

% Male 3.0 97.0 40.0 49.3 10.7 5.6 94.4 18.0 82.0 68.1 31.9 467

% Female 19.5 80.5 59.7 21.5 18.8 19.6 80.4 38.8 61.2 71.5 28.5 2343

% Male

Yr 1965–69

18.3 81.7 37.8 47.1 15.1 20.0 80.0 25.9 74.1 69.0 31.0 410

% Female

32.5 67.5 75.0 10.9 14.2 37.0 63.0 53.8 46.2 70.9 29.1 4027

% Male

Yr 1990–94

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

140

Notes: # descriptive labels used in QCA results. ∗ excluded categories. For all variables, 1 = Yes and 0 = No.

Offense Characteristics Instrumental Motive Expressive/other Motive∗ Firearm Knife/Stabbing Other Weapon∗ Multiple Offenders Single Offender∗ Outdoor Location Private/Indoor Location∗ Offense at Night Offense During Day∗ Sample Sizes (N ):

Variables

Yr 1965–94

Table 6.1. (cont.)

P1: FCH/FFX 0521832993c06 T1: FCH November 3, 2003 15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

141

the category of family members and intimates showed that female offenders were disproportionately involved in higher rates of killings of spouses/ lovers, and their children and stepchildren.5 These findings provide empirical support for our decision to combine various subcategories into the general category of “family/intimates.” Offense characteristics included those used in most previous studies of homicide. Primary offense attributes are those like motive and weapon which are included in all major homicide data reports (e.g., SHR data). Motive is coded to contrast instrumental acts with expressive and other motives. Instrumental motives include felony-type situations (e.g., robbery, sexual assault, burglary, drug offenses), whereas expressive homicides include arguments, brawls, and romantic triangles.6 Weapon use includes three categories: firearms, knives and sharp objects, and “others” (e.g., blunt objects, hands/feet, arson, strangulations). Other offense characteristics cover additional aspects of the situational context and include the number of offenders, the location of the offense (i.e., outdoor vs. indoor location), its time of day (i.e., evening vs. daytime), and the year of the offense. The year of the offense is coded to contrast homicides in the 1960s (1965–69) and in the 1990s (1990–94). These particular time periods were chosen to provide the maximum contrasts over time and to cover time frames in which structural changes in the routine activities and lifestyles in American society would be most pronounced (e.g., temporal changes in household size, number of one-person households, mobility, nighttime activity, nonfamilial activity outside the home, greater interracial contact).7 Two analytic approaches are used in the current study of gender differences in homicide situations. First, the method of Qualitative Comparative Analysis (QCA) is used to identify the structure of homicide situations that are (1) unique to male offenders, (2) unique to female offenders, and

5

6

7

For example, female offenders were involved in about 12% of all homicides in the sample, but they accounted for 41% of child killings, 49% of the killing of intimates, and 15% of the patricides. It is within this context that the category of family/intimates is said to be disproportionately represented by female offenders. Nearly all (over 99%) of the homicides within the category “expressive/other” involve various types of arguments or altercations, fights, or brawls. Our categorization of particular motives as being either instrumental or expressive is empirically justified by Miethe and Drass’ (1999) findings of high internal consistency among the specific motives that fall within the general categories of instrumental and expressive homicide. Changes over time were also examined when the full data were coded to contrast time periods before and after 1980. As one might expect, changes over time were less dramatic with this coding of time period. The contrast between the 1960s and 1990s is the “best test” condition for examining the consistency of our results with expectations based on criminal opportunity theories.

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

142

T1: FCH November 3, 2003

RETHINKING HOMICIDE

(3) common to both male and female offenders.8 Second, a qualitative assessment of homicide narratives for various cities is conducted to identify possible gender differences in the dominant themes that underlie situational transactions in homicide.

Results Univariate Distributions The initial analysis involves differences in the univariate distributions of offender, victim, and offense attributes for homicides involving male and female offenders. These gender-specific comparisons over time are summarized in Table 6.1. Similar to national trends, both males and African Americans are strongly overrepresented as homicide offenders in Chicago for each time period (see Table 6.1). Approximately 88% of the offenders are male and nearly eighty percent are African Americans. Homicides are primarily intraracial, intrasex, and involve victims and offenders who are within similar age groups. The modal homicide also includes the following attributes: (a) offenders between 20 and 29 years old, (b) acquaintances, (c) expressive motives, (d) firearms as the lethal weapon, (e) single rather than multiple 8

As described in Chapter 3, our application of QCA involves the use of a “relative” or probabilistic rule for defining unique profiles and contradictions. Specifically, we define a configuration as unique when its relative frequency within that combination exceeds its overall marginal distribution. For example, given that 12% of the Chicago homicides in this sample involve female offenders, a particular combination of attributes would be deemed “relatively unique” to females if females accounted for over 22% of the offenders in this particular profile. A 10 percentage point difference over its marginal proportion is the rule used here to define relatively unique configurations. Such a decision rule is commonly used when QCA is applied to large samples (see Amenta et al. 1992; Miethe and Drass 1999; Ragin et al. 1984). As a way of reducing measurement error in configurations with small cell frequencies, several general rules were also employed to establish minimum cell frequencies for each configuration. For example, a minimum cell frequency of 30 cases was used to define “prevalent” configurations in the general analysis of gender differences. Under this rule, all configurations of offender, victim, and offense attributes that involved less than 30 cases over the 30 year period were excluded from the analysis. This minimum rule translates into the substantive decision that homicide situations that account for no more than one homicide per year are largely idiosyncratic and ignorable due to their low prevalence. Minimum cell frequency rules of 5 and 35 were used in the respective analyses of female- and male-perpetrated homicides in the 1960s and 1990s. The value of 5 cases per configuration was selected for female offenders because this is a general rule of thumb for identifying minimum cell frequencies in the analysis of contingency tables (see Miethe and Drass 1999), whereas a minimum of 35 cases per configuration was employed for male offenders to make our coding rule proportional to the roughly 7:1 marginal distribution in favor of male offenders in Chicago (i.e., 88% male vs. 12% female). As will be shown shortly, comparisons of the number of configurations and cases before and after the application of the minimum frequency rule provide a direct measure of the relative diversity and “situational clustering” (Birkbeck and LaFree 1993) of the structures of homicide situations across genders and over time.

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

143

offenders, (f) private indoor locations, and (g) evening hours. These modal categories are similar for each time period. Male and female homicide offenders differ on several of these offender, victim, and offense characteristics (see Table 6.1). Compared to male perpetrators, female offenders have higher proportions of each of the following characteristics: African-American offenders, offenders age 30 or older, victims and offenders within the same age groupings, intraracial attacks, victims and offenders of different sexes, slayings of intimate partners or family members, expressive motives, knives as the lethal weapon, sole offenders, indoor locations, and daytime offending. The nature and magnitude of these gender differences are generally similar over each time period. Most of these bivariate trends have been observed in other studies of gender and homicide (see Block and Zimring 1973; Jurik and Winn 1990; Mann 1996; Wilbanks 1983; Wolfgang 1958).

Qualitative Comparative Analysis Three distinct types of qualitative comparative analysis are conducted to address the research questions in this chapter. First, gender differences in the structure of homicide situations are examined across the entire time frame (1965–94) to assess the nature of their unique and common signatures. Second, change and stability in the structure of male homicide situations are explored by comparing patterns in the 1960s and 1990s. Third, change and stability in the structure of female homicide situations are examined for the same time periods. The results of these analyses are summarized below.

Gender Differences in the Structure of Homicide Situations Do male and female offenders kill in qualitatively different homicide situations? Table 6.2 summarizes the pattern of unique and common profiles across both genders and the specific nature of the most dominant configurations that underlie their distinct signatures. As shown in Panel A of Table 6.2, the 22,704 Chicago homicides from 1965 to 1994 are contained within 2,827 different combinations of offender, victim, and offense characteristics. After excluding configurations that contain less than 30 cases (i.e., idiosyncratic situations that represent at most one homicide per year), a total of 138 prevalent homicide situations (5%) remain in the analysis. This relatively small proportion of prevalent homicide situations, however, still represent more than half (54%) of all Chicago homicides in this 30-year period. This pattern of prevalent configurations and cases indicates that homicides have enormous diversity in their underlying structures (n = 2,827), but there are also relatively high levels of situational clustering because more than half of all Chicago homicides are

15:40

144

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

=Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex Acqu Acqu Acqu Acqu Acqu Acqu Acqu Acqu Acqu Acqu

Expr Expr Expr Expr Expr Expr Expr Expr Expr Expr

Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun

O2029 O2029 O3098 Olt20 O3098 Olt20 O2029 O2029 Olt20 O2029

Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk Ohsp

Sing Sing Sing Sing Sing Mult Mult Sing Sing Sing

Nite Nite Nite Nite Nite Nite Nite Day Nite Nite

Outd Indoor Indoor Outd Outd Outd Outd Outd Indoor Outd

597 of 604 460 of 467 415 of 424 335 of 336 271 of 276 226 of 228 223 of 223 160 of 163 155 of 158 151 of 151

Profile Ns

22,704 12,267 1,346 7,349 3,572

Percentage of all male homicides explained by these dominant male profiles: 15.0% (2,993 of 19,969)

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race

Male Offenders:

(4.9%) 19.6% 56.5% 23.9%

2,827 138 27 78 33

99% 98% 98% 100% 98% 99% 100% 98% 98% 100%

% Male∗

(54.0%) 11.0% 59.9% 29.1%

%

%

N

N

Homicide Cases

Configurations

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

Panel B: Unique Profiles within the Most Dominant Homicide Situations

Number of Situations Observed Number of Homicide Situations after N > 30 Cell Frequency Rule Situations Unique to Female Offenders Situations Unique to Male Offenders Contradictions (Common Situations/Cases) for both Genders

Panel A: Configurations and Cases (using 10% Difference Rule)

Table 6.2. Qualitative Comparative Analysis of Homicide Situations Unique to Male and Female Offenders, 1965–94

P1: FCH/FFX 0521832993c06 T1: FCH November 3, 2003 15:40

=Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex Fami Fami Fami Fami Fami Acqu Acqu Acqu Acqu Acqu

Expr Expr Expr Expr Expr Expr Expr Expr Expr Expr

Knife Knife Gun Gun Gun Knife Knife Gun Gun Knife

O3098 O2029 O3098 O2029 O3098 O3098 O2029 O2029 O3098 O2029

Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk

Sing Sing Sing Sing Sing Sing Sing Sing Sing Sing

145

=Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

Indoor Indoor Indoor Indoor Indoor Indoor Indoor Indoor Indoor Indoor

187 of 265 176 of 233 160 of 325 102 of 198 87 of 212 82 of 136 61 of 89 44 of 98 41 of 87 38 of 59

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex Acqu Acqu Acqu Acqu Acqu Acqu Fami Fami Acqu Acqu

Expr Expr Expr Expr Expr Expr Expr Expr Expr Expr

Knife Gun Knife Knife Knife Knife Oweap Gun Oweap Oweap

O3098 O2029 O2029 O2029 O3098 O3098 O3098 O3098 O2029 O2029

Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk

Sing Sing Sing Sing Sing Sing Sing Sing Sing Sing

Nite Day Nite Nite Nite Day Nite Nite Nite Day

Indoor Indoor Indoor Outd Outd Indoor Indoor Indoor Indoor Indoor

217 188 172 124 100 95 81 76 73 66

Profile Ns

71% 76% 49% 52% 41% 60% 69% 45% 47% 64%

% Female∗∗

MODEL: Offmale= Olt20 + O2029 + Oblk + Ohsp + =Age + =Race + =Sex + Fami + Acqu + Instr + Gun + Knife + Mult + Outd + Nite Notes: OWEAP refers to lethal weapons other than firearms and sharp objects like knives. ∗ This column represents the percentage of homicides in the particular profile that involve male offenders. ∗∗ This column represents the percentage of homicides in the particular profile that involve female offenders.

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Nite Nite Nite Nite Day Day Day Day Nite Nite

Percentage of all female homicides explained by these dominant female profiles: 35.8% (978 of 2,735)

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race

Profile Ns

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

Panel C: Most Dominant Homicide Profiles Common to both Males and Females:

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Female Offenders:

P1: FCH/FFX 0521832993c06 T1: FCH November 3, 2003 15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

146

T1: FCH November 3, 2003

RETHINKING HOMICIDE

concentrated within a rather small number (n = 138) of prevalent homicide situations. When the 10% difference rule is used to define uniqueness among the prevalent homicide situations, the majority of both configurations and homicide cases are represented by unique signatures. Over 75% of the 138 observed homicide situations are relatively unique to either female or male offenders and these unique profiles account for about 71% of the homicides (see Panel A of Table 6.2). These results indicate that most male and female homicide offenders kill in qualitatively different situations.9 Table 6.2 provides a summary of gender differences in the 10 most dominant homicide situations. Several clear patterns emerge from these genderspecific profiles of homicide situations. First, there is far more situational clustering among female offenders than male offenders. Over one-third of all female homicides in the entire 30-year period are concentrated within the 10 most dominant homicide situations that are relatively unique to women, whereas the 10 most dominant profiles account for only 15% of all male homicides. Second, the dominant male situations are exclusively intragroup, expressive attacks of acquaintances by non-White males with guns. These dominant male profiles also usually involve younger offenders (under 30 years old), occur in outdoor locations at night, and are committed by single rather than multiple offenders. Third, the dominant profiles for female offenders are exclusively expressive attacks within indoor locations by one Black offender who kills a male victim of similar age and race. Family members or intimate partners as victims, nighttime offenses, knives as the lethal weapons, and offenders aged 30 or older are additional characteristics often found within these 10 dominant female-perpetrated homicide situations. The overall profile of gender differences in homicide situations that derives from this analysis is consistent with general portrayals of homicide in previous research. Specifically, as interpersonal disputes among similarly positioned males in public places (and sometimes with multiple offenders as an audience), the 10 most common male homicide situations have the same structural characteristics that often underlie honor or character contests (see Luckenbill 1977; Polk 1994). These metaphors do not fit the pattern of dominant profiles for female offenders because these lethal disputes occur mostly among family members in private, indoor locations. Instead, the 9

This conclusion is less supported under an “absolute” or deterministic definition of unique configurations and when no minimum frequency rule is used. Under these conditions, almost half (47%) of all Chicago homicides (n = 22,704) occurred in situations that were exclusively unique to either males or females – i.e., they had a nonzero cell frequency for one group (e.g., males) and a zero cell frequency for that particular configuration in the other group (e.g., females). Nonetheless, unique situational profiles for each gender remain a dominant pattern among Chicago homicides even when this highly restrictive, “small-N ” standard is used to define unique configurations (see Ragin 2000).

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

147

image of domestic violence is the primary theme that best describes the nature of the 10 most dominant situations unique to female offenders. An examination of the full array of 138 prevalent profiles (not just the 10 most dominant for each gender) reveals several additional patterns about their unique signatures among particular homicide subtypes.10 First, instrumentally motivated homicides are almost always male-perpetrated (95%) so it isn’t surprising that the vast majority of these homicides occur in situations unique to male offenders. These killings are most often intragroup attacks of nonfamily members with guns at night in outdoor locations with multiple offenders. Second, lethal interracial attacks are rare in Chicago, and they also occur almost always in situational contexts unique to male offenders. These attacks are eclectic in their signatures, involving both instrumental and expressive motives and no particular concentration among types of lethal weapon, time of day, location, number of offenders, or age groups. Third, stranger assaults are another homicide subtype committed primarily by males (95%), and they often involve instrumental attacks on other males at night and in outdoor locations by multiple offenders with guns as major elements of their unique signatures. In contrast, women who kill strangers do so primarily in intersex attacks of an instrumental nature that occur in indoor locations and with a more diverse array of weapons. While our interpretations thus far have focused on gender-specific profiles, it is important to emphasize that in many cases the structures of homicide situations are similar across genders. As shown in Panel C of Table 6.2, the most prevalent profiles that are common to both men and women often involve expressive assaults by single, older, non-White offenders who kill people of a similar age and race during the evening hours at indoor locations using various lethal weapons (both guns and knives). The joint attributes of single offenders and expressive motives are found in all 33 of the prevalent homicide situations common to both male and female offenders.

Stability and Change over Time in Male and Female Homicide Situations Have the situational contexts for male and female homicide offenders changed over time? Tables 6.3 and 6.4 summarize for each gender the pattern of unique and common homicide situations in Chicago in the 1960s and 1990s. As shown in Panel A of Table 6.3, the 6,370 male-perpetrated homicides in the 1960s and 1990s are distributed within 1,433 different combinations of offender, victim, and offense characteristics. After applying a minimum 10

The full truth table containing these 138 specific profiles is not presented here but is available from the authors upon request.

15:40

148

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex Acqu Fami Acqu Acqu Acqu Acqu Acqu Acqu Fami Fami

Expr Expr Expr Expr Expr Expr Expr Expr Expr Expr

Gun Gun Gun Gun Knife Knife Knife Knife Gun Gun

O3098 O3098 O3098 O3098 O3098 O2029 O2029 O3098 O3098 O2029

Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk

Sing Sing Sing Sing Sing Sing Sing Sing Sing Sing

Nite Nite Nite Day Nite Nite Nite Nite Day Nite

Indoor Indoor Outd Indoor Indoor Outd Indoor Outd Indoor Indoor

97 of 134 43 of 56 43 of 80 40 of 50 38 of 58 26 of 36 25 of 35 25 of 44 22 of 35 20 of 35

Profile Ns

72% 77% 54% 80% 66% 72% 71% 57% 63% 57%

% 1960s∗

Percentage of all male homicides in the 1960s explained by the 10 unique situations in this decade: 16.2% (379 of 2,343)

=Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

Male Offenders in the 1960s:

(31.3%) 19.0% 29.6% 51.4%

(2.2%) 32.3% 41.9% 25.8%

1,433 31 10 13 8

6,370 1,995 379 591 1,025

%

%

N

N

Homicide Cases

Configurations

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

Panel B: Unique Homicide Situations for Male Offenders in Each Decade

Number of Situations Observed in the 1960s and 1990s Number of Homicide Situations after N > 35 Cell Frequency Rule Situations Unique to the 1960s Situations Unique to the 1990s Contradictions (Common Situations/Cases) for both Decades

Panel A: Configurations and Cases (using 10% Difference Rule)

Table 6.3. Qualitative Comparative Analysis of Stability and Change in Male Homicide Situations over Time

P1: FCH/FFX 0521832993c06 T1: FCH November 3, 2003 15:40

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex Acqu Acqu Acqu Acqu Stran Acqu Acqu Acqu Acqu Acqu

Expr Expr Expr Expr Inst Expr Inst Expr Expr Expr

Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun

Olt20 O2029 Olt20 O2029 Olt20 Olt20 O2029 Olt20 O2029 Olt20

Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk Ohsp Ohsp

Mult Mult Sing Sing Mult Mult Sing Mult Sing Sing

149

Outd Outd Outd Outd Outd Indoor Outd Outd Outd Outd

98 of 118 65 of 84 49 of 55 48 of 60 47 of 50 43 of 48 42 of 45 38 of 47 36 of 41 35 of 38

83% 77% 89% 80% 94% 90% 93% 81% 88% 88%

=Age =Age =Age =Age =Age =Age =Age =Age

=Race =Race =Race =Race =Race =Race =Race =Race

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex Acqu Acqu Acqu Acqu Stran Acqu Stran Acqu

Expr Expr Expr Expr Expr Expr Inst Expr

Gun Gun Gun Gun Gun Gun Gun Gun

Olt20 O2029 O2029 Olt20 O2029 O2029 O2029 Olt20

Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk

Sing Sing Sing Sing Sing Sing Mult Sing

Nite Nite Nite Nite Nite Day Nite Day

Outd Outd Indoor Indoor Outd Indoor Outd Indoor

171 169 129 71 62 50 38 37

Profile Ns

MODEL: Yr6090= Olt20 + O2029 + Oblk + Ohsp + =Age + =Race + =Sex + Fami + Acqu + Instr + Gun + Knife + Mult + Outd + Nite Notes: ∗ This column represents the percentage of male homicides in this particular profile that occurred in the 1960s. ∗∗ This column represents the percentage of male homicides in this particular profile that occurred in the 1990s.

1. 2. 3. 4. 5. 6. 7. 8.

Nite Nite Day Day Nite Nite Nite Day Nite Nite

% 1990s∗∗

Percentage of all male homicides in the 1990s explained by these 10 unique situations in this decade: 12.4% (501 of 4,027)

=Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

Profile Ns

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

Panel C: Dominant Homicide Profiles for Male Offenders Common to Both Decades:

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Male Offenders in the 1990s:

P1: FCH/FFX 0521832993c06 T1: FCH November 3, 2003 15:40

150

1. 2. 3. 4. 5. 6. 7. 7. 7. 10.

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex Fami Fami Fami Fami Acqu Acqu Acqu Fami Acqu Fami

Expr Expr Expr Expr Expr Expr Expr Expr Expr Expr

Gun Knife Gun Knife Knife Knife Gun Knife Gun Gun

O3098 O3098 O2029 O2029 O3098 O2029 O3098 O3098 O3098 O3098

Oblk Oblk Oblk Oblk Oblk Oblk Owht Oblk Oblk Oblk

Sing Sing Sing Sing Sing Sing Sing Sing Sing Sing

Nite Day Nite Day Nite Nite Day Nite Nite Nite

Indoor Indoor Indoor Indoor Indoor Indoor Indoor Indoor Indoor Outd

37 of 49 24 of 36 19 of 24 12 of 18 11 of 16 10 of 16 8 of 8 8 of 8 8 of 10 7 of 8

Profile Ns

76% 67% 79% 67% 69% 67% 100% 100% 80% 88%

% 1960s∗

Percentage of all female homicides in the 1960s explained by these 10 unique situations in this decade: 30.8% (144 of 467)

=Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

Female Offenders in the 1960s:

(56.1%) 36.6% 7.7% 55.7%

(10.8%) 47.0% 26.5% 26.5%

314 34 16 9 9

877 492 180 38 274

%

%

N

N

Homicide Cases

Configurations

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

Panel B: Unique Homicide Situations for Female Offenders in Each Decade

Number of Situations Observed in the 1960s and 1990s Number of Homicide Situations after N > 5 Cell Frequency Rule Situations Unique to the 1960s Situations Unique to the 1990s Contradictions (Common Situations/Cases) for both Decades

Panel A: Configurations and Cases (using 10% Difference Rule)

Table 6.4. Qualitative Comparative Analysis of Stability and Change in Female Homicide Situations over Time

P1: FCH/FFX 0521832993c06 T1: FCH November 3, 2003 15:40

=Race =Race =Race =Race =Race =Race =Race =Race =Race

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex Acqu Fami Fami Fami Acqu Acqu Fami Fami Fami

Expr Expr Expr Expr Expr Expr Expr Expr Expr

Knife Oweap Oweap Oweap Knife Knife Knife Oweap Knife

O3098 O2029 O2029 O2029 O3098 Olt20 Olt20 O2029 Olt20

Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk

Sing Sing Sing Sing Sing Sing Sing Sing Sing

Nite Day Nite Nite Nite Nite Nite Day Day

151

9 of 15 6 of 10 4 of 5 4 of 6 3 of 5 3 of 5 3 of 5 3 of 5 3 of 5

60% 60% 80% 80% 60% 60% 60% 60% 60%

=Age =Age =Age =Age =Age =Age =Age =Age =Age

=Race =Race =Race =Race =Race =Race =Race =Race =Race

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex Fami Fami Fami Fami Fami Acqu Acqu Fami Fami

Expr Expr Expr Expr Expr Expr Expr Expr Expr

Knife Knife Gun Gun Knife Knife Knife Knfie Gun

O3098 O2029 O3098 O2029 O3098 O3098 O2029 Olt20 O2029

Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk Oblk

Sing Sing Sing Sing Sing Sing Sing Sing Sing

Nite Nite Day Day Nite Day Nite Nite Nite

Indoor Indoor Indoor Indoor Outd Indoor Indoor Indoor Outd

65 53 25 13 10 9 8 5 5

Profile Ns

MODEL: Yr6090= Olt20 + O2029 + Oblk + Ohsp + =Age + =Race + =Sex + Fami + Acqu + Instr + Gun + Knife + Mult + Outd + Nite Notes: ∗ This column represents the percentage of female homicides in this particular profile that occurred in the 1960s. ∗∗ This column represents the percentage of female homicides in this particular profile that occurred in the 1990s.

1. 2. 3. 4. 5. 6. 7. 8. 9.

Indoor Indoor Indoor Indoor Outd Indoor Indoor Indoor Indoor

% 1990s∗∗

Percentage of all female homicides in the 1990s explained by these unique situations in this decade: 9.3% (38 of 410)

=Age =Age =Age =Age =Age =Age =Age =Age =Age

Profile Ns

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

Panel C: Dominant Homicide Profiles for Female Offenders Common to Both Decades:

1. 2. 3. 3. 5. 5. 5. 5. 5.

Female Offenders in the 1990s:

P1: FCH/FFX 0521832993c06 T1: FCH November 3, 2003 15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

152

T1: FCH November 3, 2003

RETHINKING HOMICIDE

frequency rule that excludes configurations with less than 35 cases, only 31 prevalent configurations involving 1,995 male homicides remain in the analysis. Expressed as percentages, this pattern means that 2% of the observed configurations account for 31% of all homicides by male offenders in the two decades. For female offenders, 34 (11%) of the configurations satisfy the minimum frequency rule of 5 cases and these prevalent situational profiles account for 492 (56%) of all female homicides in these two decades (see Panel A of Table 6.4). While the number of prevalent situations are comparable for each gender (suggesting no major differences in the overall diversity of situations in which each group kills), female homicides exhibit greater situational clustering because a higher proportion of all female homicides (56% vs. 31%) are concentrated within these prevalent homicide situations. Under the 10% difference rule, patterns of both stability and change over time are found among the prevalent homicide situations for each gender. Almost three-fourths of the 31 prevalent homicide situations involving male offenders are unique to either the 1960s or the 1990s and these temporally unique profiles account for nearly half of the male-perpetrated homicides in these two decades (see Panel A of Table 6.3). Similarly, about 74% of the 34 prevalent homicide situations involving female offenders are temporally unique and these profiles represent about 44% of the female perpetrators in these two decades (see Panel A of Table 6.4). For both male and female offenders, a slight majority of homicide cases occur in situations that are common over time. The specific profiles of the most dominant situations that have changed over time for male and female offenders are presented in Panel B of Tables 6.3 and 6.4. Emergent forms of male-perpetrated homicides in the 1990s retained many of the same structural features as homicides in the 1960s (e.g., intra-age and intraracial assaults by non-Whites) but other attributes changed over time. The most obvious changes in particular offender and offense characteristics that contribute to the 10 most dominant situations unique to male offenders in the 1990s include a rise in outdoor locations (9/10 in the 1990s vs. 3/10 in the 1960s), teenage offenders (6/10 vs. 0/10), multiple offenders (5/10 vs. 0/10), and guns as the lethal weapon (10/10 vs. 6/10). It is the change in both the marginal and joint distributions of these variables that provide the unique signatures for male-perpetrated homicides in the 1990s.11 Similar to male offenders over time, emergent forms of femaleperpetrated homicides in the 1990s retained some of the same structural 11

The emergence of unique signatures in the 1990s involving young males with guns was also observed in the SHR data (see Chapter 4 ). Thus, this pattern is consistent across different aggregate units, providing some evidence for the generalizability of our findings across contexts.

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

153

features of homicides in the 1960s (e.g., expressively motivated intraage and intraracial assaults by single offenders) but other attributes were different over time. The clearest changes in particular offender and offense attributes unique to female offenders in the 1990s include a rise in lethal weapons other than firearms (9/9 vs. 5/10) and teenage offenders (3/9 vs. 0/10). For the other offender and offense characteristics that define homicide situations, changes in the joint distribution of these variables (rather than their marginal distributions) are the primary source for the unique signatures. Most of the unique configurations among female offenders occurred within the 1960s rather than the 1990s, whereas an opposite temporal pattern is found in the unique profiles for male offenders. The profiles of homicide situations that exhibit stability in their structures over time are presented in Panel C of Tables 6.3 and 6.4. Temporally stable profiles among male offenders typically involve intragroup, expressively motivated shootings by young (less than 30 years old), single, Black offenders at nighttime. Outdoor locations and attacks on acquaintances are other characteristics that contribute to these common homicide situations at both time periods (see Panel C of Table 6.3). For female offenders, the most dominant situations common in both time periods involve particular combinations of expressive killings of opposite sex family members or intimates by single, Black, and adult offenders of the same age and race as their victim. Other particular offense elements that are often found in these stable homicide situations for female offenders include nighttime hours, indoor locations, and knives as the lethal weapon (see Panel C of Table 6.4). The persistence of killings by women in the context of domestic violence is the major source of their stability in homicide situations over time.

Narrative Descriptions of Homicide Situations The conclusion that men and women usually kill in qualitatively different situations is also revealed when the nature of the homicide transaction is more fully investigated. The QCA results indicate that females kill primarily within the context of domestic violence and men kill in a variety of instrumentally and expressively motivated situations. Narrative accounts of homicides that represent these major themes and illustrate the situational dynamics in these violent events are summarized below.

Felony-Homicide Situations While women are rarely involved in felony homicide situations, their participation in these instrumental homicides is usually quite different than their male counterparts. Women who kill in this context are often involved in

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

154

T1: FCH November 3, 2003

RETHINKING HOMICIDE

“rip-offs” that stem from various types of drug or sexual transactions (e.g., prostitution) and they work with male offenders in an accessory role to gain initial access to potential victims or entice them into illegal activities (see Jurik and Gregware 1992). These themes are clearly revealed in the following homicide narratives: A Latin female knocked at the 23-year-old Anglo male victim’s door asking for the victim. As he came to the door the woman stepped back and at least two Latin males opened fire on the victim. The victim ran to his bedroom and grabbed his handgun. More gunshots were exchanged and the victim was fatally wounded. The victim was involved in the drug trade and the police believed the killing was a dispute over drugs. (Miami, case #093, Wilbanks 1984:201) An 80-year-old, Anglo male was robbed in his hotel room by a 25-yearold Anglo female and a 24-year-old Black male. The female offender was a prostitute and got the victim to open the door by offering sex. The victim was beaten and choked to death. They got $37 from the robbery-murder. (Miami, case #356, Wilbanks 1984:224) The 70-year-old Black male victim had been drinking all day and was often in the company of prostitutes. On this occasion it appears that he invited a prostitute up to his room and that she then robbed him and stabbed him with a butcher knife. The 28-year-old Black female prostitute has been charged with murder before but beat the charge. She is now at large. (Miami, case #10, Wilbanks 1984:240)

Instrumentally motivated homicides by male offenders are mostly committed within the context of robbery and many of these homicide situations involve illegal drug transactions. While these illegal drug activities are similar in some respects (e.g., parties trying to rip off each other through insufficient payment or selling “bad” dope), greater planning is often found among street-level drug-related homicides that involve large-scale transactions. Consistent with Maxfield’s (1989) views that robbery is often an unplanned consequence of homicide, narrative accounts also indicate that many personal robberies begin as interpersonal disputes and the thefts ultimately occur at the conclusion of these lethal disputes. Elements of planning and calculation are more commonly observed in male-perpetrated robbery-homicides involving business establishments (see also Chapter 5).

Domestic Violence Situations The situational dynamics involving homicides within the context of domestic relationships also exhibits some gender differences. Substance abuse (both alcohol and drug use) and jealousy are major elements of the situational dynamics in the killing of intimate partners by male offenders. Suicide

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

155

following the homicide is also far more common in intimate slayings by male offenders (Dawson and Gartner 1998). In contrast, threats of physical violence immediately before the fatal attack and histories of abusive relations between the parties are strong situational dynamics in female-perpetrated homicides within the context of intimate partners. The similarity of the situational dynamics within each gender and their differences across groups are illustrated in the following sample of murder cases from large urban counties in the United States (see Langan and Dawson 1995):

Husband Killing Wife Three weeks before the murder the 32-year-old husband beats his 35-yearold wife (the victim) severely. She tells others that next time he is going to kill her. She wants a divorce, but he wants a reconciliation. On the day of the murder, the two are at a bar, and he is drinking. Later a witness sees her outside a car in the middle of the street. The wife screams, then slumps to the ground, dying from multiple stab wounds to the chest. The husband flees the scene. Later he is found dead from a drug overdose. (San Diego, case #89: p. 5) The 29-year-old wife (the victim) accuses the 32-year-old husband of being interested in another woman. An argument turns into a fight. The wife is armed with a kitchen knife. The husband has scissors. Both sustain stab wounds, but her wounds are fatal. The husband pleaded guilty to negligent manslaughter and was sentenced to 5 years in prison. (Oklahoma City, case #30: p. 14) Both are unemployed. One week before the murder, the 22-year-old husband is released from prison and moves in with his 22-year-old wife (the victim), who is living at his grandmother’s house. At the time of the murder, the wife tells the husband that she wishes to leave him to be with a man she was seeing while the husband was in prison. She also informs him that she might be pregnant by the other man. The husband strangles her, telling the grandmother he killed his wife, and calls the police. Family members and friends say he has a history of assaulting his wife and threatening her life. He also has a long criminal record. He pleads guilty to nonnegligent manslaughter and was sentenced to 16 years in prison. (Albuquerque, case #9: p. 14) He, age 41, has been sadistically abusing his 41-year-old wife (the victim) for about 8 years. He frequently binds, gags, and blindfolds her and forces her to perform sex acts on him. He sometimes cuts and burns her while she is bound. A year prior to the murder she calls the police following a beating. He is arrested and convicted and then thrown out of the Navy, terminating a 22-year career. Soon after, they separate. On the night of the murder, he has drunk about 20 beers and calls his estranged wife, ordering her to take off her clothes and await his arrival. When he enters her apartment, she is

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

156

T1: FCH November 3, 2003

RETHINKING HOMICIDE

nude and holds out her wrists to him. He binds, gags and blindfolds her, cuts her with a knife, and wraps tape and rope around her neck, strangling her. Hours later he is arrested. He admits killing her and intending to do so for some time. He blames her for ruining his career, saying he was going to kill her sooner or later. The jury convicted him of first-degree murder. He was sentenced to 26 years in prison. (San Diego, case #15: p. 10)

These scenarios provide further insight into the events leading up to the killings. Consistent with other research on intimate partner homicide (see Daly and Wilson 1988a; Polk 1994; Wallace 1986), themes of jealousy and possessiveness emerge repeatedly in these case histories. Lethal violence during a failed attempt at reconciliation is also an important circumstance surrounding intimate partner homicides by male offenders, and underscores the danger faced by women seeking to terminate relationships with abusive men (see also Rasche 1993). In some of these scenarios, the woman is also armed or may even be the first to take up a weapon. Yet these actions were not sufficient to secure their safety. Unfortunately, whether this is a reflection of physical strength differentials or a greater determination to inflict harm on the part of the male cannot be deciphered from these narratives. While the precipitating event in many of these descriptions are disputes and arguments, prior histories of abuse are also common in cases of maleperpetrated intimate partner killings. Thus, these incidents of lethal violence are often the culmination of a pattern of abuse and violent behavior for both genders. The killing described in the last narrative shows clear indications of being premeditated, a pattern which other studies have demonstrated to be characteristic of intimate partner homicides committed by men (see Dawson and Gartner 1998; Polk 1994). The distinct situational contexts for male- and female-perpetrated intimate partner homicides are evident when comparing the above case descriptions with those in which the wife kills the husband:

Wife Killing Husband The 28-year-old husband (the victim) has a long history of assaulting his 25-year-old wife. At the time of the murder a witness sees the husband in the kitchen chasing the wife with a machete in his hand. The fight ends when the wife stabs the husband once. She is arrested the same day and claims self-defense. (New Orleans, case #95: p. 5) The 49-year-old husband (the victim) is drunk and gets into an argument with his 50-year-old wife. According to her, at some point he throws a fan at her. She gets a knife and lunges at him, cutting his abdomen. The jury acquits the wife of murder charges. (Chicago, case #28: p. 9) She, a 29-year-old artist, has lived with her 47-year-old common-law husband (the victim), a musician, for a few years. He frequently beats, starves,

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

157

and tortures her. He repeatedly tells her he is going to kill her. For some time leading up to the murder, he has not let her eat or shower. On the day of the murder they are going to a bar when she asks him to let her go home because she is tired. He says “no” and begins hitting her and calling her names. The beating ends when she stabs him. He has numerous past arrests and convictions. The jury acquitted her. (Manhattan, case #72: p. 9) The husband (the victim) and his common-law wife are arguing about a variety of things. Throughout the argument the husband beats her. When the common-law husband comes after her in the kitchen, she grabs a knife. She stabs him in the back as he is walking away. She pleads guilty to negligent manslaughter and was sentenced to 2 years in prison. (Dayton, OH, case #3: p. 13)

The structures of female-perpetrated homicides that emerged in the QCA analysis are further emphasized in these accounts. For example, knives are the lethal weapon in these cases of expressively motivated killings by women of their intimate partners. They take place indoors and do not involve additional offenders. However, the details provided in these narrative shows some additional patterns that cannot be gleaned from the quantitative analysis. In particular, these include evidence of past domestic abuse characterizing the relationship between the parties and the precipitating role of the victim in these lethal encounters. These scenarios offer strong evidence of male victims acting as initial aggressors in confrontations arising over arguments fueled by a variety of different topics. This greater level of victim provocation by male intimate partners may be the fundamental reason why there are more acquittals and fewer severe criminal sentences for women who kill in this situational context (see also Daly and Wilson 1988a; Langan and Dawson 1995; Lundsgaarde 1977). When narrative descriptions of intimate partner killings are compared over time, we find evidence of stability in the nature of the offenses and the sequential transactions underlying these expressive homicide situations across each decade. These similarities over time are illustrated in the following narratives of male- and female-perpetrated spouse killings from Los Angeles.

Husband Killing Wife Victim was murdered in her home at 10:30 a.m. by her husband. Husband was a bartender who had been living apart from his wife for a short time and returned to the house this morning. Husband said he came home to induce his wife to make up with him. They again quarreled and, losing his temper, he knocked her down with a blow of his fist. Then, he jumped on her head with the heels of his shoes, crushing her skull and killing her instantly. Husband was immediately arrested for the crime and later executed by hanging at San Quentin prison. (Los Angeles, April 1910)

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

158

T1: FCH November 3, 2003

RETHINKING HOMICIDE

Victim was shot and killed by her husband. Her husband committed suicide with the same gun immediately afterwards. They were not living together, but had quarreled considerably about property and money. There were no witnesses to the shooting, but neighbors heard the shots at about 11 p.m. (Los Angeles, June 1926) The 45-year-old, Caucasian victim and her husband had been separated for three weeks. On this particular night she went to a cabin with her husband for the supposed purpose of completing their income tax forms. He strangled her to death with his hands and then tied a towel around her neck. On his arrest, he readily admitted killing her, giving as his reason that she had been messing around with Mexicans. She had previously had to call police because of beatings he had given her. (Los Angeles, February 1954) The victim, her husband, and a male mutual friend had been out drinking in a local bar. All three returned to the victim’s apartment, where they had several more drinks. The friend left the apartment and after he left the victim and her husband argued. The suspect (the husband) choked his wife to death, then hid his keys and wallet to make it look like someone else murdered his wife. (Los Angeles, July 1962)

Wife Killing Husband Victim was killed by his wife at about 7:30 p.m. in their auto. The victim had been spending his money on a young woman and beat up his wife often. Victim was beating up his wife when the shooting occurred. (Los Angeles, December 1922) The victim was shot by his wife with a .38 caliber revolver. The offender stated that they had a family quarrel and he stayed away all night. Upon his arrival home, the quarrel started again and he went into the bathroom. She got the revolver and shot through the door, striking the victim in the right chest and through the intestines. The husband died from the wounds. (Los Angeles, July 1931) The 32-year-old Negro victim died as a result of gunshot wounds inflicted by his wife. She was arrested and the Coroner’s jury found this to be a case of self-defense. Deceased had previously beaten up the suspect. He had also stated that he was going to kill suspect and went to her room with a loaded and cocked .45-caliber gun. (Los Angeles, January 1945) The 59-year-old Caucasian victim and his 60-year-old wife engaged in an all night domestic dispute and fight. At about 8:00 a.m. the fight erupted again. The wife stabbed her husband in the heart with a 9-inch butcher knife. The wife went to a bar in a taxi where she reported the crime to the police. (Los Angeles, May 1966) The victim, the 35-year-old Negro husband of the suspect, came home late and a fight ensued. The victim struck his wife and then produced a 12gauge shotgun and a box of shells. She countered with a .22 revolver and

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

159

fired six times, hitting him twice. The murder complaint was dismissed on the grounds of self-defense. (Los Angeles, May 1974)

In spite of the span of decades from which these cases are drawn, we find continued evidence of variation according to the gender of the offender. Based on the narrative accounts, homicides by males who kill their wives, girlfriends, or ex-partners often fit the image of “rage” killings, characterized by anger and committed under the influence of alcohol with any available lethal weapon. Suicide by the male offender is not an uncommon consequence of these killings. In contrast, women who kill their husbands and boyfriends are often provoked by physical and verbal threats by the victim, they frequently occur during periods of separation, and they rarely commit suicide after the lethal attack. Other major aspects of these homicide disputes include residential locations and the absence of third parties who may have minimized the likelihood of violence in this setting. No obvious pattern of change in the nature of the situational transactions underlying homicides within the context of domestic disputes is observed in the examination of Los Angeles narratives over the twentieth century. In other words, the fundamental nature of themes of male sexual proprietariness and jealousy in intimate partner homicides is not unique to a particular decade or time frame. In this respect, changes in broader social conditions such as the economy, family composition, or the structure of communities may have little impact on the situational dynamics of killings between intimate partners. Where these changes may play a significant role, however, is in reducing the prevalence with which either women in general or certain subgroups of women are drawn into situational conditions where the escalation of a dispute into lethal violence is likely. These patterns also emphasize how homicides which are diverse in their structures may share similar underlying processes. Male sexual proprietariness appears to be the underlying processual dynamic in killings committed by both men and women within the context of intimate partner relationships. These narrative accounts also reveal that both types of lethal disputes are characterized by previous histories of conflict, a finding consistent with past research (see Jurik and Gregware 1992; Polk 1994). While men frequently play the role of the initial aggressor in both types of homicides, differences in attributes such as weapon use and location may produce distinct situational contexts for male- and female-perpetrated homicides. As is illustrated in the following narratives, the theme of male possessiveness and sexual jealousy is also evident in another type of homicide: lover’s triangles. The Negro victim was shot and killed at his home by the 60-year-old male Negro offender. The offender had gone to the victim’s house to see the offender’s wife who was living in the victim’s house. The offender had some

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

160

T1: FCH November 3, 2003

RETHINKING HOMICIDE

words with the victim and the shooting followed. The offender claimed selfdefense. (Los Angeles, July 1910) The male victim was shot and killed about 6:35 p.m. by the male offender who escaped. The victim had stolen the wife of the offender, thus incensing the latter. (Los Angeles, February 1923) The 30-year-old, Caucasian male victim was beaten and kicked to death by the offender and a companion. The attack occurred sometime during the night and the victim was left at the scene of the beating and not discovered until noon of the next day. Both suspects were arrest. The motive claimed by one of the offenders was that the victim had been out with his wife. (Los Angeles, May 1945) The 35-year-old, Negro male victim was found lying in the front seat of his automobile with a fatal gunshot wound to his head. Investigation revealed that he had been shot by a male suspect who after being arrested admitted he had fired a shot at the victim. He shot at the victim after he had seen his wife leave the victim’s car and the victim tried to pull away from the curb. (Los Angeles, February 1954) The 24-year-old, Negro male suspect observed his estranged wife with the victim, her 35-year-old Negro boyfriend, sitting in a vehicle. Suspect followed them to an alley and parked. Suspect requested to talk to his wife, and as the victim and suspect’s estranged wife walked to the suspect, he fired a shot at his wife with a .22 caliber revolver, hitting her in the left temporal area. The boyfriend tried to flee as the suspect fired several shots at him. One of these shots hit the boyfriend in the chest causing death. (Los Angeles, June 1964)

A substantial majority of fatal lover’s triangles involve males killing the current boyfriend or husband of their ex-lover (see also Daly and Wilson 1988a). The same underlying themes of sexual jealousy and proprietariness that lead men to engage in lethal violence toward their intimate partners result in male-on-male homicides in many of these lover’s triangles. Thus, while we have a different situational structure than in killings of women by male intimates, the processual elements are similar. As will become evident, that the processes underlying homicides may cross-cut what are typically treated as different types of killings is a pattern that reemerges throughout the chapters of this book.

Conclusions and Implications Our application of QCA to the study of gender differences in homicide situations reveals several major findings. First, homicides in Chicago occur in a diverse array of situational contexts but a relatively small number of these situations (n = 128) account for more than half of all homicides during the 1965–94 time period. Second, the majority of these most prevalent

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

161

homicide situations are qualitatively distinct for male and female offenders. Males kill predominantly in situations involving acquaintances of the same age, sex, and race who are shot by single offenders in disputes in public locations. The killing of family members and intimates of the opposite sex within private locations are key facets that often underlie female-perpetrated homicide situations. Third, both stability and change are found in the dominant homicide situations for each gender over time. Emergent contexts for male homicide often involve the growth of killings in outdoor locations, gun usage, multiple offenders, and teenage offenders, whereas female offenders in the 1990s have shifted away from guns and kill under more idiosyncratic circumstances. The implications of these findings are discussed below. Contrary to homicide typologies in past research that identify only a small number of subtypes (see Block and Christakos 1995; Miethe and McCorkle 2001; Polk 1994), the discovery of over 2,800 different homicide structures in the current study suggests that homicide is far more diverse than previously considered. This finding also runs counter to claims made that homicides take place in a very limited number of settings and social circumstances (e.g. Lundsgaarde 1977). For those concerned with typology construction, the pattern of unique and common structures provides an empirical basis for deriving general and gender-specific types of homicide situations. The task of typology construction, however, is less formidable when one considers the level of situational clustering in homicides. For example, over one-third of all female homicides in Chicago over the last four decades occurred within a well-defined and small subset of relatively unique situations. These deadly situations involve expressive acts of violence by Black women who attack nonstranger males of the same age and race. These nonstrangers are usually family members and their killings often take place in private locations, fitting the common image of domestic violence. In contrast, the 10 dominant male homicide situations resemble character contests in terms of their structural features (e.g., public disputes among similarly situated males) and these dominant profiles alone account for about one-eighth of all male-perpetrated homicides. Under these conditions, the presence of situational clustering (especially among female homicides) helps simplify the number and variation in homicide situations within a more comprehensive typology of homicides. Both criminal opportunity theory and Lofland’s (1969) notion of facilitating conditions were considered the best orientations for studying homicide situations because they are combinatorial theories that place causal importance on the union of particular offender, victim, and offense attributes for understanding criminal acts. How well do these theories account for the empirical results in the current study? The dominant patterns that underlie the stability in homicide structures over time are consistent with several principles derived from these

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

162

T1: FCH November 3, 2003

RETHINKING HOMICIDE

combinatorial theories (e.g., target accessibility, visibility, convenience, availability, low guardianship, facilitating places). From this perspective, gender differences in the physical location of homicides may be attributed to gender differences in what constitutes “facilitating places.” For example, major sources of offender motivation for males are character contests in public locations (see Luckenbill 1977; Polk 1994). Given the readily available pool of similarly motivated males in public settings, it would follow that public places are a facilitating context for male homicides. In contrast, women’s homicide is often a response to histories of abuse by male intimate partners that take place within the home (Browne 1987; Jurik and Gregware 1992). Such private locations thus become a facilitating context for two reasons: (1) they afford males a low visibility environment for domestic abuse and (2) they provide a physical environment for self-defensive actions by women. The home environment also provides a readily available supply of potential lethal weapons for women to carry out their defensive responses to violence. Most of the dominant attributes linked to changes in the situational profiles for male offenders are consistent with Lofland’s image of facilitating hardware (e.g., guns), facilitating places (e.g., public locations increase the risk of contact with similarly motivated males), and facilitating others (e.g., multiple offenders provide a social audience to encourage criminal acts). However, neither of these combinatorial theories can easily account for the rise in teenage offenders for both genders (unless they are linked to agespecific theories of offender motivation) or the shift away from guns as the preferred lethal weapon by women under particular circumstances. Although not offering a complete explanation for the observed results, combinatorial theories are far better able to describe these results than theories that focus exclusively on criminal motivation. In fact, no current formulation of either general or gender-specific theories of offending that ignores victim and offense attributes can account for the pattern of unique and common signatures found here. For example, the gender-specific differences in physical locations for homicide cannot be explained without looking jointly at aspects of both criminal motivation and opportunity. Under these conditions, future research would be better served by trying to advance the state of conjunctive theories rather than theories that focus solely on offender motivation. Evidence of both stability and change over time in the structure of male and female homicides is inconsistent with much of the theoretical work and public commentary on social change and crime rates. Our most plausible explanation for these disparities involves the fact that many social changes identified in past research (e.g., increases in interracial contacts, increased instrumental motivations, rising teen violence) have occurred within relatively low-frequency categories. For example, homicides involving teenage offenders in Chicago almost doubled from the 1960s to the 1990s, but

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

GENDER DIFFERENCES IN THE STRUCTURE OF HOMICIDE SITUATIONS

163

teenagers still accounted for only a minority (37%) of homicide offenders in the 1990s. Stranger killings (from 22% to 25%), interracial killings (from 11% to 10%), instrumentally motivated homicides (from 17% to 31%), and killings by Hispanic offenders (from 5% to 14%) also represent changes within relatively infrequent categories. However, by focusing on the most dominant profiles, our QCA results emphasize patterns that derive from the most prevalent categories within each variable (e.g., adult offenders, intraracial attacks, Black homicides, and expressive motives).12 Such an emphasis, by definition, may have overshadowed some of the changing temporal patterns within these low frequency situational profiles. From the perspective of crime control policy, the nature of the unique and common signatures for male and female homicides has several direct implications. First, as primarily interpersonal disputes among similarly situated persons in public places, homicide prevention efforts for males should be directed at the particular facilitating conditions (e.g., encouragement by a social audience, status enhancement, social empowerment through displays of masculinity) that underlie these uniquely male situations. Second, most female offenders still kill within the context of domestic violence. Because of the large concentration of female-perpetrated homicides in this context, prevention efforts should continue to focus on the enhancement of resources for women whose only viable solution to their own histories of domestic violence may be the killing of their abusive partners. Third, much of the public commentary on homicide (and the reform efforts that follow from them) focuses on homicide situations that are extremely rare for either or both genders. This includes the wide public attention given to (1) subtypes of female murderers (e.g., women who kill for instrumental purposes, juvenile offenders), (2) interracial slayings, and (3) homicides involving multiple offenders. Prevention efforts directed at these rather infrequent situations will net little positive return because of the rarity of their occurrence and the fact that many of them (e.g., female juvenile killers, interracial slayings) are extremely idiosyncratic in the particular combinations of offender, victim, and offense attributes that define their structures. The implications of the current study can be briefly restated. Male and female homicides are found to exhibit unique signatures, involving in most 12

Another possible explanation for the stability of homicide situations over time involves the operation of conflicting social forces. For example, the rise in public activity and subsequent increased contact with a more diverse pool of potential crime targets may not necessarily alter the nature of homicide situations for either female offenders or victims because the women that are involved in these riskier activities may either have less criminal motivation (e.g., because they have greater social resources or “capital” than less mobile women) or they have greater guardianship (e.g., they take greater precautions while in these public places). Under such conditions, various contravening forces may nullify the impact of particular social changes on the nature of homicide situations. Unfortunately, we are unable to address empirically the plausibility of this explanation because of the limitations of current data.

15:40

P1: FCH/FFX 0521832993c06

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

164

T1: FCH November 3, 2003

RETHINKING HOMICIDE

cases different combinations of offender, victim, and offense attributes. Female homicide is relatively rare and more concentrated than male homicide within a small number of dominant homicide situations. These dominant profiles that underlie male and female homicide exhibit patterns of both change and stability over time. Many of the particular factors that contribute to the unique profiles for male offenders in the 1990s (e.g., increases in gun usage, multiple offenders, outdoor locations) are congruent with principles of criminal opportunity theory, but the particular nature of the unique signatures for female homicides in this same period is not easily explained by this theoretical orientation.

15:40

P1: JMT 0521832993c07

CB617-Miethe-v1

November 1, 2003

C H A P T E R

S E V E N

Change and Stability in the Structure of Youth Homicide

youth violence has emerged as one of the most challenging social problems in contemporary American society.1 The increased prevalence of violence by youthful offenders over the last three decades has been referred to as an “epidemic” by some authors (see Cook and Laub 1998) and has generated a wide array of public policy interventions (e.g., increases in adult certifications, gang abatement programs, sentencing enhancements for crimes committed with firearms or in school zones). Past empirical research has focused on describing trends in youth homicide over time and particular differences in patterns within subgroups (e.g., male vs. female, White vs. nonWhite youth). Unfortunately, limitations of the analytic procedures used have often prevented previous research from isolating the particular nature of change and stability in the basic structure of youth homicide situations over time and their unique characteristics when compared to homicides by adult offenders. Using the Supplementary Homicide Reports (SHR) from 1976 to 1998 and narrative accounts of homicide situations, in this chapter we examine the nature of change and stability in the structure of youth homicide over time. The structure of youth homicide is defined by basic combinations of offender, victim, and offense attributes that underlie homicide situations. Through the method of qualitative comparative analysis (QCA), unique and common structures of youth homicide situations are empirically 1

Although the concept of “youth” has multiple meanings (e.g., is it a particular chronological age or state of mind), we use the age of 19 and under to designate “youth” offenders. The term “juveniles,” “teenagers,” and “young” will be used interchangeably because most youth offenders are in their teenage years (even though technically most persons under 18 are not juveniles). The contrast group in most of our analysis will be “adult” offenders, defined here as persons in their 20s or older. The specific meaning attached to the term “youth” has been addressed elsewhere (see Moore and Tonry 1998).

165

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

166

November 1, 2003

RETHINKING HOMICIDE

identified. The specific nature of these changes in youth homicide over time and their differences from adult homicides are then discussed in terms of their implications for future research and public policy on youth violence.

Trends and Correlates of Youth Homicide The tendency for teenagers and young adults to be disproportionately involved in violent crime has long been recognized. Historical accounts of violence in the United States often describe a “bachelor subculture” of young, unattached males who were widely involved in individual and collective acts of violence (see Lane 1997). In fact, young males have remained the most notorious category of homicide offenders throughout American history, including early Colonial America, the Antebellum period, the post–Civil War and period of Western expansion, the times of riots and demonstrations in the 1960s, and the era of gang violence in central cities from the 1960s to the present. Data from the FBI’s Uniform Crime Reports (UCR) and the Supplementary Homicide Reports (SHR) have been used increasingly over the last three decades to estimate age-specific rates of homicide offending and victimization (see Cook and Laub 1998; Fox 1996; Sickmund, Snyder, and Poe-Yamagata 1997; Zahn and McCall 1999). Teenage and adult homicide rates are often further dissaggregated by gender, race, or a particular offense attribute (e.g., gun vs. nongun killings). Trend analyses in past research, however, have been restricted to multiple comparisons of at most three combinations of variables (e.g., homicide rates for young, AfricanAmerican males). Data analysis and estimations from previous studies indicate several major trends in youth homicides over the last three decades. First, the proportion of homicide arrestees under 18 years old increased from less than 10% in 1974 to about 15% in 1995, but subsequently dropped down again to about 10% of all homicide arrestees in 1999 (FBI 1999; FBI 1974). Homicide rates for teenage offenders increased by an incredible 172% from 1985 to 1994 while actually decreasing for adults over 25 years old during this same period (Blumstein 1995; Cook and Laub 1998; Fox 1996). Second, the rise in youth homicide is not consistent across other demographic groups and offense attributes. For example, the homicide offending rate among white male teenagers has doubled since 1985 while more than tripling among Black male teens. Regardless of race, offending rates among female teenagers have remained relatively stable over time (Fox 1996). Similar to offender characteristics, victim and offense attributes in youth homicides also exhibit various patterns of change and stability over time

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

November 1, 2003

CHANGE AND STABILITY IN THE STRUCTURE OF YOUTH HOMICIDE

167

(see Cook and Laub 1998; Fox 1996; Sickmund, Snyder, and Poe-Yamagata 1997). For example, males are more than five times more likely than females to be victims of teen killers, and this pattern has persisted over time. The victim and offender in these homicides are typically acquaintances, followed in frequency by strangers and family members. Acquaintance and stranger killings have increased over time, while the proportion of teen homicides involving family members has decreased. Similar trends have been observed when homicide data are not disaggregated by age (see Zahn and McCall 1999). Other offense attributes include the motive or circumstances of the crime, the number of offenders, number of victims, and type of weapon. Arguments remain the most common motive or circumstance surrounding youth homicide, accounting for over one-third of these homicides in each decade (Fox 1996). The number of juvenile killings committed with a gun (especially a handgun) has quadrupled since 1985, whereas slayings involving other weapons have remained stable (see Cook and Laub 1998; Fox 1996; Sickmund et al. 1997). Our analysis of SHR data reveals a rise in the proportion of youth homicides involving multiple offenders (from 16% in the 1970s to 23% in the 1990s). The proportion of homicides involving multiple victims, however, has remained around 7% across all three decades. Over the last decade there has been a major change in the image of youth violence. The most recent generation of juvenile offenders has been called “superpredators” – youth who are radically impulsive and kill indiscriminately for various instrumental and expressive purposes (Bennett, Dilulio, and Walters 1996; Dilulio 1995, 1996). This image departs from the typical view of homicide as involving homogeneous pools of victims and offenders (see Sampson and Lauritsen 1990; Singer 1981), and suggests that youth homicide has become increasingly diverse in terms of its particular combinations of offender, victim, and offense characteristics. This alleged increase in the diversity of youth violence in the 1990s should include a rise in intergroup assaults whose participants vary in their age group, gender, race, and interpersonal relations. Although various claims have been made about the similarity and uniqueness of homicide across age groups, these conclusions are often based on bivariate trends and a main effects model. However, by presuming constant effects across context, this model specification does not allow for contextspecific differences in the impact of gun usage, race, urban location, or any other correlate of youth violence. Under these conditions, it remains unclear whether full combinations of offender, victim, and offense attributes that define the basic structure of youth homicide have changed or remained the same over time and whether the situational context for youth homicide is, in fact, qualitatively unique from adult homicide situations.

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

168

November 1, 2003

RETHINKING HOMICIDE

Research Questions Much has been written on youth violence, its causes, bivariate distribution over time, and possible solutions (see Anderson 1998; Cook and Laub 1998; Fagan and Wilkinson 1998; Farrington 1998; Heide 1999; Howell and Hawkins 1998; Moore and Tonry 1998; Zimring 1998). Our concern is to empirically document whether there have been fundamental changes in the structure of homicide situations over the last three decades that represent emergent forms of youth homicide. Using SHR data for the last three decades, we examine the relative prevalence of combinations of offender, victim, and offense attributes that underlie homicide situations involving youth offenders. Three research questions are addressed. First, what are the most prevalent combinations of offender, victim, and offense elements that define the basic structure of homicide situations involving youth offenders? Are these dominant situations similar for both youth and adult offenders? Have they changed over time? Second, what is the level of diversity and concentration of homicide situations among youthful offenders? Have youth offenders killed in an increasingly more diverse array of situations over time and are these situations more or less concentrated than homicides by adult offenders? Third, are there emergent forms of youth homicide in the 1990s whose combinations of offender, victim, and offense attributes are not commonly observed in earlier decades or among adult offenders? What are the relative frequencies and particular characteristics of these emergent forms of youth homicide?

Methods and Analysis SHR data for the 1976 to 1998 period are used to describe the basic structure of homicide situations involving youth offenders. The particular measures of offender, victim, and offense attributes included in this analysis of homicide situations are the offender’s race (White, Black, other); similar age, race, sex categories of the victim and offender (same age, same race, same sex); the victim–offender relationship (intimate/family, acquaintance, stranger); motive (instrumental, expressive/other); weapon (gun, other); multiple offenders (yes, no); and urban location of the homicide (city over 100,000, other). The specific coding of these variables has been discussed in earlier chapters (see Chapter 3). Several different types of analyses are conducted to examine the research questions underlying this study. First, we describe univariate trends in youth and adult homicide over time. Second, QCA is used to identify unique and common structures of homicide situations involving youth and adult offenders. Third, narrative accounts of homicides involving youth offenders are

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

November 1, 2003

CHANGE AND STABILITY IN THE STRUCTURE OF YOUTH HOMICIDE

169

examined to explore the basic processes underlying the situational dynamics in youth homicide. The results of these analyses are summarized below.

Results Univariate Trends over Time Table 7.1 presents the relative frequency of particular offender, victim, and offense characteristics over time among youth and adult offenders. For each time period, homicides by youth offenders are associated with a high and relatively stable proportion of each of the following main effects: male offenders (91%), same age (66%), same race (86%), same sex (79%), acquaintance slayings (51%), and expressive motives (61%). However, an increasing proportion of youth homicides over time involve either Black offenders (49% in the 1970s to 61% in the 1990s), guns as the lethal weapon (58% to 78%), multiple offenders (16% to 23%), and urban settings (65% to 74%). Comparisons of the univariate trends across age groups reveal both similarities and differences. Similar to their youth counterparts, homicides by adult offenders are also associated with high and relatively stable proportions of either male offenders, acquaintance killings, expressive motives, or victims and offenders of the same age or same race. The proportion of homicide offenders who killed within a large urban location also increased somewhat over time for both age groups. However, contrary to youth offenders, there was no appreciable increase over time in the proportion of adult homicides involving either Black offenders, guns as lethal weapon, or multiple offenders. These observed univariate trends within age groups are similar to those reported in other studies (see Cook and Laub 1998; Fox 1996; Sickmund et al. 1997). Based on these findings alone, one would conclude that youth homicide has changed somewhat in its aggregate trends over time, especially concerning the prevalence of guns, multiple offenders, and Black offenders. The bivariate profile of adult offenders has remained stable over time.

Dominant Homicide Situations for Juvenile and Adult Offenders The simultaneous classification of offender, victim, and offense attributes defines the basic structures of homicide situations. QCA is now used to examine (1) whether unique or common situational profiles best characterize the structure of youth and adult homicides and (2) what is the particular nature of the most dominant homicide situations unique and common to each age group.

15:21

89% 49 49 62 86 75 19 48 32 42 58 58 23 20 16 65 12,097

1970s 90% 46 52 63 86 77 18 49 33 42 58 60 22 18 19 70 31,136

1980s 93% 37 61 69 86 82 13 54 33 37 63 78 11 11 23 74 39,678

1990s 91% 42 56 66 86 79 16 51 33 39 61 68 17 15 20 71 82,911

Total 84% 48 50 96 91 54 31 48 20 24 76 66 19 15 7 60 61,845

1970s 86% 50 48 96 90 65 30 49 21 25 75 63 21 17 9 63 157,878

1980s

88% 46 52 94 89 68 28 50 22 26 74 68 16 16 10 67 137,320

1990s

Adult Offenders

87% 48 50 95 90 66 30 49 21 25 75 65 19 16 9 64 357,043

Total

CB617-Miethe-v1

170

Source: FBI SHR Data 1976–98.

Male Offenders White Offenders Black Offenders Same Age Group Same Race Same Sex Family/Intimates Acquaintances Strangers Instrumental Motive Expressive/Other Gun Knife Other Weapon Multiple Offenders Urban Location Sample Sizes

Variable

Youth Offenders

Table 7.1. Age-Specific Distribution of Homicide Cases across Variables over Time

P1: JMT 0521832993c07 November 1, 2003 15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

November 1, 2003

CHANGE AND STABILITY IN THE STRUCTURE OF YOUTH HOMICIDE

171

Based on the 12 offender, victim, and offense attributes examined here and a minimum frequency rule of 5 cases per year, there are 383 distinct homicide situations found in the SHR data from 1976 to 1998 (FBI 1976– 1998). About one-fourth (26%) of these situations are relatively unique to youth offenders (i.e., they involve combinations of attributes that are disproportionately found in killings involving teenage offenders), but these particular unique profiles account for only about 11% of all homicides. Slightly more than one-tenth (13%) of homicide situations are unique to adult offenders but these unique signatures represent only about 17% of all homicides. Nearly two-thirds (61%) of homicide situations are common for each age group, and these common situational contexts for both age groups account for the vast majority (72%) of homicides over these three decades. Although the bivariate findings suggest some age-specific differences (e.g., higher levels of multiple offenders, guns, and urban locations in youth homicide), most U.S. homicides occur in situations that are common to both juveniles and adult offenders. This finding on its face value has important implications for theory development because it suggests that age-specific explanations may not be necessary for generating an understanding of the structure of homicide situations.2 Homicides with common structures across age groups are best characterized by intragroup conflicts, involving victims and offenders with similar characteristics (see Panel A of Table 7.2 for the 10 most common profiles). Across the entire set of 235 distinct situational contexts common to both age groups, these shared profiles are almost always represented by male offenders (92%) who kill persons of their same age group (99%), race (90%), or sex (83%). Acquaintances as victims, instrumental motives, urban locations, and single offenders are other structural elements that are more frequently observed in situations common to both age groups. Conjunctively, male offenders who kill other male acquaintances of the same age and race are over 8 times more prevalent in situational contexts common to both age groups than in situations unique to either juvenile or adult offenders. Situational contexts unique to youth offenders are disproportionately represented by Black male offenders, guns, multiple offenders, stranger assaults, and urban locations. The two most prevalent situations unique to youth offenders are lethal shootings by urban males in a group context involving victims of a similar age, race, and gender (see Panel B of Table 7.2). This combination of attributes represents about one-third of all homicides that occur in unique situational contexts for young offenders, but is found 2

While the dominant pattern is common homicide structures across age groups, youth homicide has some unique signatures, especially as it relates to guns, urban setting, and its group context. The most prevalent profiles unique to young offenders are described shortly.

15:21

172

omale omale omale omale omale omale omale omale omale omale

oblack owhite owhite oblack oblack oblack owhite oblack oblack owhite

=age =age =age =age =age =age =age =age =age =age

=race =race =race =race =race =race =race =race =race =race

omale omale omale omale omale omale omale

oblack owhite oblack oblack oblack owhite oblack

=age =age =age =age =age =age =age

=race =race =race =race =race =race =race

=sex =sex =sex =sex =sex =sex =sex

=sex =sex =sex =sex =sex =sex =sex =sex =sex =sex

acquaint acquaint acquaint stranger stranger acquaint stranger

acquaint acquaint acquaint acquaint acquaint acquaint acquaint stranger stranger stranger

expressive expressive expressive instrumental expressive expressive instrumental

expressive expressive expressive expressive instrumental expressive expressive expressive instrumental expressive

gun gun gun gun gun gun gun

gun gun gun gun gun nongun nongun gun gun gun

multoff multoff singoff multoff multoff singoff singoff

singoff singoff singoff singoff singoff singoff singoff singoff singoff singoff

urban urban urban urban urban urban urban

urban urban =urban =urban urban urban urban urban urban urban

3,601 2,534 2,515 1,955 1,348 1,300 1,228

Nprof

30,955 17,171 14,478 10,996 9,547 9,435 8,411 7,850 7,683 6,278

Nprof

33% 34% 47% 29% 36% 40% 81%

% teen

18% 15% 9% 14% 24% 11% 13% 21% 23% 16%

% teen

CB617-Miethe-v1

1. 2. 3. 4. 5. 6. 7.

Panel B: Top 10 Unique Juvenile Profiles:

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Panel A: Top 10 Common Profiles:

Table 7.2. Qualitative Comparative Analysis of Homicide Situations Common to Both Age Groups, Unique to Juvenile Offenders, and Unique to Adult Offenders

P1: JMT 0521832993c07 November 1, 2003 15:21

owhite oblack owhite

=age =age =age

=race =race =race

173

omale omale omale ofemale omale omale omale omale ofemale ofemale

owhite owhite oblack oblack owhite owhite oblack owhite owhite oblack

=age =age =age =age =age =age =age =age =age =age

=race =race =race =race =race =race =race =race =race =race

=sex =sex =sex =sex =sex =sex =sex =sex =sex =sex

=sex =sex =sex

family/intimate family/intimate family/intimate family/intimate family/intimate family/intimate family/intimate family/intimate family/intimate family/intimate

stranger acquaint family/intimate

expressive expressive expressive expressive expressive expressive expressive expressive expressive expressive

gun gun gun

gun gun gun gun gun nongun nongun nongun gun nongun

expressive instrumental expressive

singoff singoff singoff singoff singoff singoff singoff singoff singoff singoff

multoff singoff singoff

=urban urban urban urban =urban =urban urban urban =urban urban

urban urban =urban

8,913 5,960 5,115 4,588 4,405 4,303 3,999 3,936 3,893 3,713

Nprof

1,203 1,155 1,034

98% 96% 95% 94% 92% 95% 95% 96% 97% 91%

% adult

36% 74% 58%

Notes: “Nprof ” = number of homicides within this profile; “% teen” = the percent of teenage offenders within this profile; “% adult” = the percent of adult offenders within this profile. Source: SHR Data, 1976–98.

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Panel C: Top 10 Unique Adult Profiles:

8. omale 9. omale 10. omale

P1: JMT 0521832993c07 CB617-Miethe-v1 November 1, 2003 15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

174

November 1, 2003

RETHINKING HOMICIDE

Figure 7.1. Proportion of Common and Unique Situations over Time by Age

in only 2% of the situational contexts not unique to youths. Male offenders, guns, urban environments, and multiple offenders are structural elements whose interplay forms the basis of the unique signatures for youth homicide. Black offenders are also a major component of these unique profiles. The structural context of homicides unique to adult offenders is dramatically overrepresented by lethal arguments involving family members or intimate partners of the opposite sex. This particular profile is found in 9 of the 10 most frequent situational contexts unique to adult offenders (see Panel C of Table 7.2), and accounts for 80% of the homicides in uniquely adult situations compared to only 2% in other situational contexts. Unique adult situations almost always involve expressive motives (95%) and female offenders are also a major distinguishing feature. In fact, about one-fourth of the homicides in situations unique to adults involve females who kill male family members or intimate partners, but they account for only about 2% of these situations of domestic violence that are not unique to adults.

Stability and Change in Age Differences in Homicide Situations over Time When age differences in homicide situations are compared within each decade, there is an interesting rise in the level of uniqueness of homicide situations over time. Specifically, the proportion of homicides that occur in situational contexts similar for each age group dropped from about 73% in the 1970s to about 50% in the 1990s (see Figure 7.1).

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

November 1, 2003

CHANGE AND STABILITY IN THE STRUCTURE OF YOUTH HOMICIDE

175

The primary contributors to these increasingly unique homicide situations for youth offenders in the 1990s include the emergence of larger proportions of urban attacks by multiple offenders, stranger violence, and instrumentally motivated crimes with guns by young Black males. This pattern is consistent with the types of homicides that have been attributed to the epidemic of drug market violence that began in the mid 1980s and started rapidly declining after peaking in 1993 (see Blumstein 1995). Given that about one-half of homicides in the 1990s occur in situations unique to each age group, our general conclusions are that (1) homicide situations have become increasingly unique for each age group over time and (2) youth homicide is characterized by structural profiles that are qualitatively both unique and common to adult offenders, with a movement toward increasing distinctness. Contrary to much of the popular imagery about the changing nature of youth violence, some offender, victim, and offense attributes are rarely part of their underlying structures. The two most obvious instances of this unwarranted attention have been juvenile female offenders and interracial killings. Female offenders were involved in less than 10% of the juvenile homicides over the three decades, and female offenders were found in only about 4% of the homicides that occurred in situations unique to juveniles in the 1990s. Similarly, there was no major insurgence in interracial attacks by juvenile offenders. Only about 16% of juvenile-perpetrated homicides that occurred in situations unique to the 1990s were interracial killings compared to about 10% for those in the 1970s. Based on these results, it is clear that juvenile homicide remains primarily a male offense against persons of the same race.

The Situational Context of Particular Types of Youth Homicide Aside from the general context of youth homicide, there are several particular types of homicide situations that are often included in discussions of youth violence. We briefly describe the structural context of two of these: (1) comparisons between younger and older juvenile offenders and (2) the juvenile “superpredator.”

Differences between Older and Younger Juvenile Offenders It is well known that in the United States homicide arrests increase through successive years of adolescence (see FBI 2000). However, less attention has been given to whether the nature of homicide situations is similar or different for younger and older adolescents. In other words, do younger juvenile offenders kill in different situational contexts than older juvenile offenders.

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

176

November 1, 2003

RETHINKING HOMICIDE

Consistent with many legal standards of immaturity, we use the age of 16 as the boundary for comparing younger (i.e., under 16) and older (i.e., 16 and older) adolescents. Four general findings are revealed from this comparative analysis. First, the vast majority (84%) of all youth homicides occur within situational contexts that have the same structural features for both younger and older juveniles. The unique structural contexts for each age group represent only a small proportion of juvenile homicides. Second, for both older and younger juveniles, the most dominant situational contexts for youth-perpetrated homicides involve intragroup attacks by male offenders. These common profiles for both age groups are also primarily killings motivated by arguments and other disputes in large cities by lone offenders who shoot acquaintances with a firearm. The main effects of each of these variables are shown in Figure 7.2. Third, situational contexts unique to younger juvenile offenders are almost always intraracial shootings of family members. The majority of these killings also involve lone male offenders who kill other males in lethal arguments. Similar to recent research on child homicide (see Alder and Polk 2001), the single best factor that distinguishes homicides by juveniles under 16 is the predominance of family members as their victims. Fourth, dominant profiles unique to juveniles 16 and older involve a greater prevalence of interracial attacks, strangers as victims, lethal weapons other than firearms, instrumental motives, small urban and rural locations, and particular combinations of these attributes. Even though these differences exist between younger and older adolescent offenders, it is important to emphasize that the vast majority of youth homicides occur in situational contexts that are similar for both younger and older offenders. This diverse pattern of killings involving younger and older juvenile offenders suggests that there are several distinct forms of youth homicide: (1) the killing of family members by younger adolescents, (2) instrumentally motivated intergroup killings by older adolescents, and (3) dispute-related killings of acquaintances in urban areas by adolescents generally. The existence of multiple types of youth homicide has important implications for both theoretical explanations of lethal violence by youth offenders as well as the development of initiatives seeking to prevent the occurrence of these killings.

The Juvenile “Superpredator” Although widely criticized from a variety of perspectives (see Blumstein 2002; Cook and Laub 1998; Elikann 1999), a new generation of juvenile criminals were deemed “superpredators” in the 1990s by the popular press

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

November 1, 2003

CHANGE AND STABILITY IN THE STRUCTURE OF YOUTH HOMICIDE

177

Figure 7.2. Main Effects across Context [Age Differences]

and some academic researchers (see Dilulio 1995, 1996; Fox 1996; Gest and Pope 1996). Claims were made that “today’s bad boys are far worse than yesterday’s and tomorrow’s will be even worse than today’s” (Bennett, Dilulio, and Walters 1996:26–27). Caused by “moral poverty” (i.e., a context in which children grow up without love, care, and guidance from responsible adults), the result of this new, inadequately socialized cohort of superpredators is that America is now home to thickening ranks of juvenile “super-predators”– radically impulsive, brutally remorseless youngsters, including even more pre-teenage boys, who murder, assault, rape, rob, burglarize, deal deadly drugs, join gun-toting gangs, and create serious communal disorders. They do not fear the stigma of arrest, the pains of imprisonment or the pangs of conscience. (Bennett, Dilulio, and Waters 1996:27)

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

178

November 1, 2003

RETHINKING HOMICIDE

From this description of the superpredator, it is possible to develop several profiles of their presumed unique signatures. First and foremost, these offenders have emerged in the 1990s and they are young males whose “radically impulsive” killing implies that they kill indiscriminately (i.e., across different types of victims) and in a variety of contexts (e.g., both impulsive/expressive crimes and felony-type situations).3 Guns and multiple offenders are also implied as emergent elements of these homicide situations. It is also reasonable to assume that these unique situations involving superpredators are overrepresented by urban locations (due to the reference to “communal disorder,” a property more typically linked to large urban areas) and Black youth (due to their higher exposure to deterioration in family life over time). Under these assumptions about the nature of juvenile superpredators, the implication is that the situational context for these juvenile homicides has shifted over time such that they constitute an emergent form of homicide in the 1990s. However, when QCA is applied to examine the unique and common structures of male juvenile homicides with firearms in a group context, we find little evidence that the situational context of these superpredator homicides has changed their structures over time. In fact, 80% of this small subset of youth homicides occur in situational contexts that have not changed over time and only 7% of these homicides happen in situational contexts unique to the 1990s. The other possibility is that these homicides haven’t changed in terms of their structure, but have increased in amount over time. Based on the analysis of profiles of juvenile offenders over time, however, there is also little evidence of a dramatic rise over the last two decades in the volume of juvenile superpredators. Male offenders under 20 years old account for a growing proportion of homicides over time, but even in the peak period of 1990–94 they represent only 21% of all homicides. When juvenile homicides are broken down further in terms of the characteristics consistent with the image of the superpredator, only 4% of the homicides in the 1990s involve young males who shoot others in the company of other offenders. This percentage reduces to 2% when we also include the additional elements of Black offenders and large urban areas. This profile is found within about 1% of U.S. homicides in the late 1970s. Thus, juvenile homicides have increasingly involved particular characteristics associated with the notion of a superpredator, but they represent essentially a trivial proportion of U.S. 3

In their critique of the notion of superpredator, Cook and Laub (1998) argue that changes in youth homicide in the 1990s are better represented as a period effect rather than a cohort effect as implied by Dilulio et al.’s explanation. John Dilulio has recently recanted on his position about superpredators.

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

November 1, 2003

CHANGE AND STABILITY IN THE STRUCTURE OF YOUTH HOMICIDE

179

homicides. Furthermore, the situational profile associated with the superpredator is not new to the 1990s. Similar to other research on this topic (see Cook and Laub 1998), these results provide essentially no support for the claim that an emergence of “thickening ranks” of juvenile superpredators has characterized the 1990s. Male juveniles who kill with guns in a group context are rare in the United States across the last three decades and the situational contexts for these killings have not changed dramatically over time.

Homicide Narratives for Juvenile Offenders Our analysis thus far has highlighted the structural aspects that underlie homicide situations for juvenile offenders, adult offenders, and subsets of younger and older juvenile offenders. Now we explore the processes underlying homicides committed by youths through a qualitative analysis of St. Louis homicide narratives of offenders in this age group. A careful inspection of these 267 homicide accounts reveals the following major themes in killings committed by youth offenders. To begin with, the group nature of youth offending is readily apparent in these narratives. There are two important points to be made about this. First, it demonstrates that the group context of youth crime extends beyond the less serious types of offending that is more characteristic of criminal activity committed by individuals within this age group (see Short and Strodtbeck 1965; Warr 2002). In other words, in spite of the gravity of the offense, youth homicide displays a strong group nature. Second, this group context is pervasive across a variety of types of killings by youths. It is probably to be expected, for example, that homicides resulting from gang disputes or feuds will have a group nature to them, as illustrated in the following narratives: The victim has been in a confectionery with friends when they were approached by a group of friends of the suspect (though the suspect wasn’t among them). The suspect’s friends apparently thought the victim’s group were Crips. It is not clear whether they were. But the suspect’s group tried to intimidate the victim’s group, who tried not to be intimidated. The store owner heard the words ‘Blood’ and ‘Cuz’ so told everyone to leave. The Bloods left first, followed shortly by the victim’s group. The suspect was approaching the area on foot and was told that his associates were being chased by Crips, so he opened fire. Afterwards, the suspect’s group claimed that the victim’s group had chased them into an auto and were armed. Police never found evidence that the victim and friends were armed. The victim’s friend said that, if they had been armed, they could have prevented the victim’s death. (St. Louis, case #91019)

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

180

November 1, 2003

RETHINKING HOMICIDE

The victim was walking towards a street corner, when suspects #2, #3, #4, and #5 drove up and opened fire on him. Suspect #1 approached the victim on foot. The suspects had set out to get a Crip, had recognized the victim as such, and shot him. Reportedly, the suspects then stood among the crowd that gathered and watched the police and EMS technicians. Since 1991, there had been a feud between the two gangs. The victim appears to have been a member of the 5900 Wabada Rolling 60’s Crips, formerly the B.O.D.s, or Boys of Destruction. The suspects were members of the Horseshoe Posse Bloods. (St. Louis, case #93050)

Both of these scenarios provide examples of how gang membership can facilitate violent activity in the form of gang-motivated violence. That is, these homicides resulted from assaults committed on behalf of the gang itself (see Rosenfeld, Bray, and Egley 1999). By belonging to a gang, youth commit themselves to a lifestyle imbued with violence that often reaches lethal levels. The particular mechanisms or forces that compel this type of violence by gang members have been long investigated in the literature on youths, gangs, and delinquency (see Short and Strodtbeck 1965; Fagan and Wilkinson 1998). Drug market violence (i.e., drug debts and disputes) is another common situational context for youth homicides evident in these narratives. As with gang feuds, these are also frequently group-oriented, as can be seen from the following accounts. The victim sold drugs for the suspect. Two weeks prior, the suspect had given the victim five ‘rocks’ to sell, for which the victim owed the suspect $80. The victim had not paid. The suspect went to the victim’s residence with witness #3 to collect. The suspect and victim argued over the money. The victim knocked on witness #1’s door (this witness served as ‘back-up’ if the victim needed help). Witness #1 and witness #2 went outside onto porch. They saw the suspect with pistol. Witness #1 said not to shoot in his front yard. Witness #2 said not to shoot up her house. Witness #3 said they would not shoot, would not disrespect the witness’ residence. The suspect and witness #3 were trying to goad the victim into fighting witness #3. The victim refused, kept trying to persuade the suspect to put down his gun. The suspect would not put down gun. The suspect wanted his money and the victim didn’t have it. When the victim would not fight witness #3, and the suspect would not put down gun, the victim told the suspect that he (the suspect) may as well go on and shoot him (the victim). It is possible that the victim thought the suspect wouldn’t shoot because people were watching. However, when the victim said the above, the suspect started shooting. The victim turned and ran into his residence, where he collapsed. (St. Louis, case #90122) The location was a place for selling drugs. The suspect and some half dozen others were selling drugs there. 20 minutes prior, the victim had purchased a rock from suspect #2. Suspect #2 sold him soap (‘ganked’ him). Apparently

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

November 1, 2003

CHANGE AND STABILITY IN THE STRUCTURE OF YOUTH HOMICIDE

181

this was done, in part, because the victim was unknown to them. The suspects then watched the victim drive down the street, make a U-turn, and return. So the suspects open fired on the victim’s vehicle. The victim tried to drive away, but hit several parked vehicles before his car stopped. The suspects fled scene. There seems to have been no argument over the soap. The suspects fired at the victim before he spoke. (St. Louis, case #92008)

These accounts exemplify the violent nature of the crack market, which, in combination with the proliferation of guns, has been offered as an explanation of the dramatic increase in youth homicides in the late 1980s and early 1990s (Blumstein 1995; Blumstein and Cork 1996; Cork 1997). The above narratives also emphasize another major underlying theme apparent in all of the above accounts: the use of firearms in lethal violence among youth. However, guns are not only pervasive in gang- and drugrelated killings by young offenders. Whether the homicide is the result of retaliation, gang feuds, drug disputes, robberies, arguments, or jealousy, guns are widely used across a variety of contexts in youth homicides. As will become evident, this is a theme which permeates the remaining accounts in this chapter. The prevalence of guns in these descriptions of youth homicides may reflect the symbolic value of gun ownership among adolescents in terms of being a source of masculinity, identity, power, and respect (Fagan and Wilkinson 1998; Gibbs and Merighi 1994). Youths also frequently engage in lethal violence in a group context in the course of arguments and disputes. As illustrated below, these can take various forms: The incident occurred in the suspect’s residence. The victim was in the house with witness #2 and the suspect. The suspect was angry with the victim for drinking beer in his (the suspect’s) father’s house, and told the victim to stop. The victim didn’t. The suspect took the beer bottle from the victim and threw it out the back door into the alley. The victim and the suspect continued to argue about the beer. Witness #1 and one other youth then returned from a fast food restaurant with hamburgers. The suspect told the victim to leave the house, and the victim said he would but talked to others, asking them why the suspect had a gun in his waistband. The suspect pointed the gun at the victim and fired. When the victim fell, the suspect stood over the victim and continued to fire the weapon until he was ‘dry firing.’ The suspect needed to prove that he was not a punk, not playing games (about the beer), etc. The suspect did not flee scene. (St. Louis, case #90045) The suspect had gone to above location to pick up the witness and their baby. The victim and his friends had been playing cards on the hood of the car. When they saw the suspect, they began to taunt him. The suspect accused them of disrespecting him. They surrounded the suspect. He told them he was neither Crip nor Blood, and wasn’t into ‘gang stuff.’ They claimed they weren’t either. The suspect pulled gun. The victim’s friends then fled but the

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

182

November 1, 2003

RETHINKING HOMICIDE

victim did not. The suspect shot the victim and drove away. (St. Louis, case #92063) Earlier this day, when the victim, witness #1, and witness #2 were leaving a girl’s house, witness #1 took offense that suspect #2 was ‘looking’ at him. And challenged suspect #2 with ‘what the hell are you looking at?’ Words were exchanged. Suspect #2 appeared to be getting a pistol from his glove compartment when a female cousin persuaded him to put it away. Still angry, suspect #2 told the three that he would come back and kill them. That evening, the victim, witnesses, and one or two others were leaving in witness #3’s car, eating a pizza, when the suspects drove up. Suspect #1 got out of the car and shot at group. The victim, witness #1, and witness #2 ran. When the shooting stopped, they discovered that the victim had collapsed in vacant lot. (St. Louis, case #95094)

The situational contexts of the youth homicides described in these accounts all contain elements of honor contests, masculine competitiveness, or issues of respect (see Anderson 1998; Luckenbill 1977; Polk 1994). They closely fit the image of what Wilson and Daly (1985) describe as the “young male syndrome” in their overriding concern with saving face. In all of these cases the confrontation occurred in the presence of others, a finding consistent with prior studies demonstrating the significance of third parties in the escalation of disputes (see Felson, Ribner, and Siegel 1984; Luckenbill 1977; Oliver 1994; Polk 1994; Tedeschi and Felson 1994). Displays of masculinity and the need to acquire status and respect may be particularly important themes underlying lethal violence committed by youth given the salience of these kinds of issues for individuals at this developmental stage generally (Fagan and Wilkinson 1998; Wilkinson and Fagan 2001). The resolution of conflicts through fighting and violence may reflect a constant demand for respect in combination with restricted avenues for attaining it (Wilkinson and Fagan 2001). The necessity of demonstrations of toughness or violence as a means of self-preservation among inner-city adolescent males has emerged as a significant theme in previous studies (see Anderson 1997; 1998; Canada 1995; Wilkinson and Fagan 1996; 2001). According to Anderson (1998), the rules for achieving and maintaining respect are learned early in life. The issue of confrontational homicides is taken up further in the following chapter. There are also times when these disputes turn into ongoing feuds, as is demonstrated by the following narrative. The feud began in 1989 between two young men (one age 20) over an 18-yr-old girl. One man shot the other at least twice. Both men were incarcerated at the time of the shooting. However, ever since that shooting, associates of the two men had continued shooting at one another on an average of once a week. The victim and suspects belonged to the group associated with the

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

November 1, 2003

CHANGE AND STABILITY IN THE STRUCTURE OF YOUTH HOMICIDE

183

victim of the original shooting in 1989. On the eve in question, the victim and suspect #2 had spotted a member of the other group and decided to ‘hit’ him. They picked up suspect #1 and weapons and headed back to location of intended target, who was standing on the street. The victim and suspects tried to take up a position in the gangway, but there wasn’t enough room for them all to shoot, so suspect #2 ran around back of building and took up a position in another gangway. The victim was to begin firing first, but target and his associates saw suspect #2, who then opened fire. The victim and suspects all fired weapons, the target and associates returned fire. The victim and suspects were all armed with .38s. The suspects said that target always carried a 9 mm, and the man he’d been standing with always carried a .44. When the shooting stopped, the victim was dead. Not clear whose shot(s) killed him. The suspects tried to flee scene in the victim’s auto but had an auto accident, from which they fled on foot. (St. Louis, case #91197)

In this case, a dispute between two males over a woman takes on a tremendous amount of significance as it escalates into a full-blown feud that no longer even involves the original participants. An important question to be addressed in future research concerns how and why initial disputes between two parties escalate into ongoing feud, that are kept alive by two much larger groups of individuals. For example, these processes may involve what others have referred to as the victimization of “generalized others” (see White, Rosenfeld, and Phillips 2002). The permeation of violence in the lives of many of these youths is evident in the way in which they resort to violence as a way of correcting past wrongs committed against them, their friends, family, or fellow gang members. This theme is exemplified in the following narratives: This shooting was supposedly in retaliation for someone (a group) shooting into the home of suspect #2’s mother. However, that shooting was supposedly in retaliation for the suspects having flourished and fired weapons prior to that. After suspect #2’s home was shot up, the suspects went to above location and fired into a crowd. The victim had been watching a basketball game when shot. The suspects were involved in several shooting incidents this day. Earlier, they committed a drive-by. They used a stolen vehicle with a sunroof. Two of the suspects stood up through the sunroof to shoot. They shot a 16-yr.-old youth in the leg. At the time of this shooting, they were using a stolen van. In addition to the victim, they shot and wounded an 11 yr. old. Suspect #2 said it was ‘about respect.’ The opposing gang members were jealous of him and wanted to be like him because he had money, women, and respect. The suspects were Crips. At the time of this shooting, the suspects ‘shot up the place.’ (St. Louis, case #93043) In 1992, the suspect’s cousin was the victim of homicide in the county. The suspect was extremely upset by it, and had threatened to kill everyone involved. Supposedly, this suspect was responsible for a 1994 homicide as a

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

184

November 1, 2003

RETHINKING HOMICIDE

result. The victim had a brother one year younger than him who had been arrested for the homicide of the suspect’s cousin. The victim’s brother had been released one month prior to this homicide. The victim was the same height as his brother, and there was some similarity of appearance. Apparently the suspect mistook the victim for his (the victim’s) brother. At time in question, the victim and his cousins (witnesses) were walking to their grandmother’s house when someone shot at them from an alley. Everyone ran, but the victim collapsed at the spot where the shooting began. (St. Louis, case #95085)

Within these contexts, violence appears to act as a form of self-help or social control (see Black 1983; Goffman 1967). In other words, violence functions as a means of dispute resolution or revenge. Either unwilling or unable to rely upon the systems of formal social control for retribution, these youths seek to make their own justice. That formal social control may be essentially nonexistent in particularly dangerous neighborhoods is reflected in the dependence on self-help of its adolescent residents (Wilkinson and Fagan 2001). According to Anderson (1998:81), the code of the street – rules which regulate interpersonal behavior in public, especially violence – is “actually a cultural adaptation to a profound lack of faith in the police and the judicial system, and in others who would champion one’s personal security.” These views translate into a willingness of residents to defend themselves and their loved ones by whatever means necessary, and the ability of individuals to “take care of themselves” becomes another source of status. The theme of respect is also central to the events leading up to the killing in the first account. However, it plays out somewhat differently than the more traditional confrontational homicide in which threats to identity and masculinity are explicit elements of a dispute that occurs within a more confined time frame, and often in the presence of others. Furthermore, the situations described here appear to be infused with both a level of instrumentality in the sense that the shooting was committed for a specific gain, as well as a strong expressive element that is evident in the spontaneity and carelessness with which the violence is carried out. Unfortunately, the existence of this latter element often leads to individuals becoming “unintended victims” of homicides when the killings go awry, as occurred in both of these cases. Perhaps what we are seeing here is evidence of Gottfredson and Hirschi’s (1990) low self-control criminal or the risk-taking behavior characteristic of young males (Alder and Polk 1996; Wilson and Daly 1985), where a hereand-now orientation and lack of foresight makes for poor offending skills, subsequently resulting in the wrong person being killed. All too often homicide is characterized either as methodically planned or totally spontaneous and impulsive. In contrast, our readings of homicide narratives suggest the presence of both types of circumstances in group settings that turn lethal. Both the group context of youth homicide offending,

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

November 1, 2003

CHANGE AND STABILITY IN THE STRUCTURE OF YOUTH HOMICIDE

185

as well as the chaotic and disorderly nature of its commission, are clearly evident in situations of robbery homicides. Witnesses #1, #3, and #4 were playing basketball with friends when the victim, the suspect, witness #2 and witness #5 walked by. The victim claimed witness #1 had shot at him in the past. He challenged the witness with “what’s up, cuz?”, told him he had shot at him, etc. Witness #1 claimed he didn’t know what the victim was talking about and tried to take off. The victim and suspect chased him and grabbed him. Witness #5 had been carrying a gun, which the suspect took and used to threaten witness #1. The victim and the suspect both punched witness #1 and robbed him of 2 rings at gun point. There are 2 versions of what happened next. Version 1: the victim told the suspect “go ahead and shoot him, man.” When the suspect fired the gun, witness #1 ducked and the victim, who had been holding witness #1, was struck. Version 2: the victim asked the suspect for the gun. Just as the suspect was handing it over, the victim tried to hit witness #1 one more time. His swing hit the gun, causing it to discharge. Two groups of youngsters then ran from the scene in different directions. (St. Louis, case #91068) Suspect #1 had a job but lost it. His mother said he began ‘running the streets’ (had prior arrests). A friend said that he had been talking for a while about ‘jacking’ someone, and talked about wanting to kill someone. He had a pregnant girlfriend in Cincinnati, whom he wanted to visit for Father’s Day. She had offered to send him a bus ticket, but he told her that he’d rather drive. He told his friends that he wanted to go ‘in style.’ The day before this incident, the suspects stole a Cadillac from a dealership. Suspect #2 then drove them to Cape Girardeau where suspect #1 and suspect #3 tried to rob a man of his Honda. They abandoned the idea because suspect #1 couldn’t drive a stick shift. They took the man’s wallet and let him go. The suspects then persuaded suspect #2 to drive them to St. Louis so suspect #1 could steal a Lexus. (Suspect #3 was going to accompany suspect #1 to Ohio.) When in St. Louis, the Lexus vehicles they saw were driven by women, so they didn’t want to steal one because suspect #1 “didn’t want to shoot a woman.” At Northwest Plaza, they saw the victim drive into a parking lot. He didn’t have a Lexus, but a jeep-like vehicle. Suspect #1 decided that he liked that. Suspects #1 and #3 then approached the victim as he exited his vehicle, and forced him to get back in. Suspect #1 was armed with pistol. Suspect #3 drove the victim’s vehicle. They then followed suspect #2. He drove looking for a ‘secluded’ spot where they could shoot the victim. Suspect #2 didn’t know his way around St. Louis that well. They drove downtown. As the suspects approached McKinley Bridge, the victim jumped and ran. Suspect #1 shot him. The victim fell, then got back up and ran. He ran to a bank where he did business, and where the president was a personal friend. He reached bank, said he’d been shot, and collapsed. Suspect #2 returned to southern MO. Suspects #1 and #3 drove on in the victim’s vehicle. They were pursued at high speed by Highway Patrol from IN to OH. They wrecked their vehicle. Suspect #3 was killed. Suspect #1 required hospitalization. (St. Louis, case #94104)

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

186

November 1, 2003

RETHINKING HOMICIDE

Both of these narratives demonstrate the haphazard nature in which many robbery-related homicides by youths are carried out. The spontaneous character of the decision to take the items by force or to select a particular target are exemplified in these accounts. The group nature of these offenses is also readily apparent. What eludes explanation is the unprovoked nature of the killing in the second case. The lack of provocation and the shooting of robbery victims who were essentially cooperating with the offenders is a relatively common theme underlying robbery homicides by youths. This may tie in with the earlier noted spontaneity and unplanned character of these types of killings among young offenders. Not having taken sufficient steps to prevent being identified by their victims, they eliminate them instead. Alternatively, it is possible that the violence in these situations provides a form of entertainment for those involved (see Tomsen 1997). It is clear that more research attention needs to be devoted to developing an understanding of what turns these robberies into homicides when no resistance is offered by the victim. At times, what begins as a robbery homicide becomes infused with notions of masculinity, and the events leading up to the killing subsequently bear a closer resemblance to an honor contest than a felony-related killing. The following narrative illustrates these underlying processes: The suspect was walking with the witnesses when they saw the victim get out of an auto to use a pay phone. They passed by and when they got to a vacant lot the suspect said he wanted to go back for the victim’s jacket. Witness #1 remained on the vacant lot. Witness #2 gave the suspect his gun and accompanied the suspect back to where the victim was on the phone. The suspect said “what’s up, Blood? Give me your jacket.” The victim responded “get for real.” The suspect then shot the victim and he and the witnesses fled. Afterwards, witness #2 asked the suspect why he shot the victim. The suspect said he hated for people to think he was a joke. (St. Louis, case #91031)

This example of a would-be robbery turned face-saving homicide demonstrates several important points. First, it shows how the use of violence as a way of accomplishing masculinity manifests itself across a variety of contexts, including types of homicides that would not traditionally be expected to act as grounds for its construction. Second, it serves as an example of the problems with categorizing homicides as instrumental versus expressive. While this act of lethal violence may have begun with the instrumental motive of acquiring the jacket, the grounds under which the killing is ultimately committed appear much more expressive in nature. Third, it reinforces the conclusion of Alder and Polk (1996) concerning the complexity and diversity underlying the achievement of masculinity through violence. For example, this account may represent a drive to achieve status through the taking of others’ possessions as discussed by Anderson (1998). Anderson

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

November 1, 2003

CHANGE AND STABILITY IN THE STRUCTURE OF YOUTH HOMICIDE

187

describes the centrality of nerve to attaining respect, and how nerve can be demonstrated by taking someone else’s belongings. Finally, while not as common, one additional major underlying theme apparent in these narratives that is not unique to homicide offending by youths is jealousy. The man living next door to the victim and suspect said they fought constantly. When it ‘spilled over’ to his apartment, he intervened. He said he’d told them that, if they didn’t love one another, they should separate. When the victim got a restraining order, apparently they were separated for a time. In a letter found at the scene, the suspect blamed the deaths on a man whom he harmed, whom he said was his wife’s lover. The 911 call was placed by the suspect himself. He said “thank you for everything. I kill my wife. I die.” (St. Louis, case #95126) The suspect and the victim’s mother slept in a bedroom with a crib in which the victim and her twin brother slept. The night before this incident, the suspect and the victim’s mother had argued because the suspect had come in late and was ‘high.’ Apparently, the victim’s mother refused to have sex with the suspect because he was too high. The victim’s mother took the victim into their bed at approximately 5:00 a.m. after feeding the victim. At 8:30 a.m., the victim’s mother’s ex-boyfriend, who was also a friend of the victim’s uncle, went to the residence to get his hair cut. Everyone was still asleep, but the ex-boyfriend did visit with the victim’s mother in the hall for approximately 5 minutes. When the victim’s mother returned to her bedroom, the victim was dead. The suspect was very jealous. The suspect had threatened the victim’s mother in past, saying if she saw anyone else he would beat her up (the mother). Apparently, while alone with the baby, the suspect hit the baby in jealous anger. The victim’s mother said the suspect could be violent and had assaulted people. After the baby’s death, the suspect threatened to kill the mother if she told police all she knew. (St. Louis, case #91037)

The significance of jealousy as a motivating factor underlying killings within intimate partner and family relations has received considerable attention in the literature (see Daly and Wilson 1988a; Daly, Wilson, and Weghorst 1982; Polk 1994). The cases described above serve as an important reminder that youth are not immune to sexual jealousy in spite of their young age. Because this process is evident in killings committed by youth and adults alike, it may be a reflection of some of the underlying shared structural contexts across age groups identified in the earlier QCA analyses.

Summary of Findings Our examination of the situational context of youth offending reveals the following key findings.

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

188

November 1, 2003

RETHINKING HOMICIDE

First, homicide structures common to both youth and adult offenders form the dominant trend. This pattern holds over all years combined as well as for each separate decade. However, both adult and juvenile homicides have become more distinct in their structures over time. Second, emerging forms of homicides by youth offenders in the 1990s are characterized by young, Black male offenders, instrumental motives, multiple offenders, guns, and strangers. We find no evidence of a major insurgence of killings by female juveniles or interracial homicides in the 1990s. Third, although killings by younger and older juveniles typically share a common structural context, distinct forms of homicides by younger offenders involve the killing of family members while those by older juvenile offenders are characterized by instrumentally motivated intergroup violence. Dispute-related killings in urban areas form the dominant situational context for both groups of juvenile offenders. Fourth, while exhibiting some change over time, youth homicide does not reflect the image of a superpredator. Homicides by young males in a group setting represent only a small proportion of all U.S. homicides, and their apparently tremendous increase is an artifact of large percentage changes occurring in a low-frequency category (i.e., an increase from 1% to 2% is portrayed as a doubling of these types of offenses). Fifth, a careful reading of narratives of youth homicides in St. Louis reveals some fundamental themes in the processes underlying these killings. The group context of adolescent offending and the prevalence of firearms are evident across a variety of different types of youth homicides. A number of these cases fit the image of honor or character contests, emphasizing the importance of achieving and demonstrating masculinity, respect, and status among individuals at this stage in the life course. Many cases exhibit elements of both instrumental and expressive motivations within the same incident. The use of violence as a means of self-help or social control also emerges as a common theme. Jealousy-motivated killings are identified as a situational context that may be common to both youth and adult homicide offending.

Conclusions and Implications The results of our analyses in this chapter challenge the prevailing image that kids who kill are somehow different from other offenders. To the contrary, we find that in the majority of cases youth and adult offenders share common structural contexts. Structural profiles of youth homicides represent an interplay between male offenders, guns, multiple offenders, and urban environments. The processes underlying these events are dominated by themes of masculine

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

November 1, 2003

CHANGE AND STABILITY IN THE STRUCTURE OF YOUTH HOMICIDE

189

competitiveness, honor, respect, and the desire to remedy perceived wrongs committed against oneself, fellow gang members, family, or friends. These processes are inextricably linked to the structural context of the neighborhoods and communities in which this violence is being carried out. For example, these patterns are consistent with notions of the use of violence as a mechanism for maintaining honor and developing a reputation among disadvantaged populations who face blocked access to more traditional means of demonstrating masculinity and status. The work of Anderson on the “code of the streets” also helps place our findings within the context of inner city communities. Specifically, Anderson (1994, 1998) notes the coexistence of two types of normative systems in the inner city of Philadelphia, one that reflects middle class values (adhered to by the majority of residents of the inner city, the so-called “decent” families), and another that is in opposition to mainstream values (the “street” orientation). Of particular importance is his observation that even adolescents from “decent” families require a working knowledge of the code of the street in order to negotiate their way through public space. In this respect, all inner city adolescents, especially males, become versed in the ways of the street, including the need to defend oneself through displays of toughness or violence. Manners of handling interpersonal disputes are learned early in life and reinforced by the pervasiveness of violence and conflict in these communities. It is within this context that violence becomes a mechanism for acquiring respect. The major challenge facing theoretical explanations of youth violence is to account for how disputes arising in virtually the same situational context and under the same structural conditions in some cases lead to lethal violence while at other times ending nonviolently. Along these lines, the presence of guns and the role of third parties (who may be more common in situations of youth violence given its group-oriented nature) in particular require further exploration. Furthermore, theoretical development is needed to explain how adolescents who are prone to violence (because of strain, lack of self-control, socialization into values conducive to violence, or any other motivational force) commit homicide within particular situational contexts and not others. Explanations of homicide also need to account for the fact that particular processual elements of lethal violence such as confrontational homicides, honor contests, masculine competitiveness, self-help, and jealousy, while typically viewed as elements of explanations of violence by particular subgroups (e.g., males, minorities, youth), are evident in killings across a greater range of subpopulations than is typically assumed or acknowledged. We return to this issue in the next chapter. Further research is also needed with respect to processes of decision making among adolescent violent offenders (see also Fagan and Wilkinson

15:21

P1: JMT 0521832993c07

CB617-Miethe-v1

190

November 1, 2003

RETHINKING HOMICIDE

1998). The void that remains in terms of further developing our understanding of the situational context of lethal violence concerns a need to better comprehend the process of decision making from the point of view of the actors directly involved in these events (as opposed to a researcher’s interpretation of them). For example, there is much to be learned about such diverse situations in youth homicides as killings of cooperative robbery victims, the generation of violent feuds expanding well beyond the original participants, and the willingness of adolescents to enter into situations in which lethal violence is a likely outcome. The development of techniques such as QCA greatly improves our ability to study homicide situations, including features like their qualitative distinctness and diversity. Homicide narratives provide significant insight into the events preceding acts of lethal violence that cannot be gained from homicide data files such as the SHR or Chicago Homicide Project. In combination, these different methods allow us to explore the structure and process of homicides. Important policy implications stem from the findings in this chapter. In particular, the focus of policy initiatives on low-frequency events such as female adolescent offenders, interracial killings by young offenders, and the emergence of a new type of juvenile killer in the form of the superpredator deflects attention from the prime areas that should be of concern: the interplay between masculinity, youth, guns, and urban settings that involve other kids who are similarly situated. Based on the results reported here and the findings from other studies (see Cook and Laub 1998; Fagan and Wilkinson 1998), policy and prevention efforts directed at the proliferation of youth gun violence among minority males in inner cities are paramount in reducing juvenile homicide.

15:21

P1: JMV 0521832993c08

CB617-Miethe-v1

October 20, 2003

C H A P T E R

E I G H T

Racial Differences in Homicide Situations

no issue in criminology has received more attention than the nature and magnitude of racial differences in criminal behavior and criminal processing. Reviewers of the empirical literature often reach conflicting conclusions about the nature and magnitude of racial disparity. Some report major disparities across dispositional decisions, while others conclude that racial differences are overstated (see Walker, Spohn, and Delone 2000; Hawkins 1994; Wilbanks 1987; Georges-Abeyie 1984). Similarly, racial differences in official rates of criminal behavior have been attributed to a variety of sources, including selective police deployment practices in minority communities and a wide range of psychological, social, economic, and historical factors. There is less dispute, however, about the overrepresentation of Blacks as both homicide victims and offenders. In fact, the magnitude of the homicide rate among particular segments of the Black population has led to its declaration as a public health issue. For example, homicide is the leading cause of death among Black males aged 15 to 34 (National Center for Health Statistics 2000). It is the third leading cause of death for White males in this age group. In spite of the concern generated by this problem both within and outside of the Black community, criminologists are not in agreement regarding an explanation for the level differences in homicide across racial groups. While both cultural and structural explanations have been employed to account for differences in violent crime rates between Black and White offenders, the majority of this research has been conducted at the aggregate level. Less is known about how these mechanisms manifest themselves at the situational level. Using various types of homicide data (FBI SHR data from 1976 to 1998; California homicide data from 1987 to 1999; narrative accounts from selected cities), this chapter examines the nature of racial and ethnic differences in offender, victim, and situational elements underlying homicide. 191

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

192

October 20, 2003

RETHINKING HOMICIDE

Rather than using statistical models to assess the relative importance of various predictors of race-specific homicide rates, we use the homicide situation as the unit of analysis to evaluate whether unique combinations of offender, victim, and situational factors have emerged over time that distinguish between Black, White, and Hispanic homicide offenders. Race-specific trends over these three decades are examined in terms of their consistency with various theoretical perspectives, media accounts, and popular conceptions of violence. The observed results are then discussed in terms of their implications for criminological theory and practice.

Trends in Race-Specific Homicide Rates over Time Rates of homicide in the United States have vacillated widely over time. Starting in the 1930s (when systematic data were first collected by the FBI), homicide rates decreased prior to and during World War II, increased in the post-war period, steadily rose throughout the 1960s and 1970s, declined briefly in the early 1980s, generally increased until 1993, decreased successively until the turn of the century, and increased again in 2001 (see FBI 2002; Miethe and McCorkle 2001; Reiss and Roth 1993). Although the temporal pattern for homicide arrest rates is generally similar for both African Americans and Whites, arrest rates are at least 6 times higher for African Americans in each year after World War II and over 10 times higher from the mid-1950s to the mid-1970s (LaFree 1994). In the post-war period, homicide arrests peaked for African Americans in the early 1970s and a decade later for White Americans (LaFree and Drass 1996; LaFree, Drass, and O’Day 1992). African American homicide rates have been characterized by greater variability than White homicide rates through most of the twentieth century (Hawkins 1999). Further subclassification by age and gender shows that homicide rates increased dramatically in the last three decades for male offenders under 25 years old for both racial groups, but this pattern was especially pronounced among young African-American males. The estimated homicide offending rate in 1994 for African-American males between 14 and 17 years old was 140 per 100,000, compared with 16 per 100,000 for young White males, 7 per 100,000 for young African-American females, and 1 per 100,000 for young White females (Fox 1996: Table 2). For White youth, arrest rates for homicide increased from the late 1960s through 1980, whereas among Black youth there was a decline beginning in the early 1970s until the mid-1980s (Messner, Raffalovich, and McMillan 2001). Homicide arrest rates (especially for Black youth) increased dramatically beginning in the mid-1980s and then began a rapid decline in the early 1990s (Cook and Laub 1998). Regardless of their race or gender, homicide rates have decreased since the mid 1970s for offenders over 25 years old (Fox 1996).

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

October 20, 2003

193

Theory and Method in Homicide Research Various theoretical and methodological approaches have been used to examine racial differences in violent crime. Guided by theoretical perspectives that emphasize social structural and cultural differences, aggregate-level data representing national trends and smaller units (like regions, states, SMSAs, counties and cities) have been widely analyzed to assess the impact of the racial composition of the geographical area, and changes in it over time, on rates of homicide and other index crimes (see, for example, Bailey 1984; Blau and Blau 1982; Land, McCall, and Cohen 1990; Messner 1983; Miethe and Meier 1994; Parker 1989). Other aggregate studies have examined the correlates of race-specific models of homicide rates (see, for example, Krivo and Peterson 2000; LaFree and Drass 1996; Messner and Golden 1982; Messner et al. 2001; Ousey 1999; Parker and McCall 1997; Peterson and Krivo 1993; Phillips 1997; Sampson 1987; Smith 1992). Case histories, field research studies, and descriptive historical accounts are other methods used for examining racial differences (see, for reviews, Anderson 1999; Chin 1996; Horowitz 1983; Oliver 1994; McKanna 1997). Of the various theoretical approaches that seek to account for the higher levels of homicide among African Americans, the most popular perspectives focus on either cultural influences or socioeconomic conditions. According to the subculture of violence perspective (Wolfgang and Ferracuti 1967), higher rates of homicide among particular subgroups (like young, urban, African-American males) are attributed to a particular value system that condones violence in certain situations. Harvey (1986) describes the particular experiences of being a young African-American adult as indicative of a subculture of exasperation in which the sense of frustration and powerlessness felt by many African Americans manifests itself through violence and aggression. In spite of evidence indicating a strong positive relationship between percent Black and homicide rates (e.g., Messner 1983), little direct empirical support has been found for this presumed subculture of violence (Loftin and Hill 1974; Parker 1989; Phillips 1997; Sampson 1985a; 1985b; 1987). Social disorganization theory asserts that structural conditions like population mobility, disruptive family relations, and low economic opportunity are associated with higher crime rates because these factors are indicative of weakened social control mechanisms in the community. From this perspective, higher rates of homicide among African Americans living in urban poverty are attributed to lower levels of social control through mainstream institutions, values, and role models (see Sampson and Groves 1989; Shaw and McKay 1942; Short 1997; Wilson 1987). Through the impact of discrimination and inequality on enhancing frustration and subsequent aggression, structural factors such as racial segregation, concentrated disadvantage, and conditions of absolute and relative deprivation have also been identified as

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

194

October 20, 2003

RETHINKING HOMICIDE

criminogenic forces in African-American homicide (Blau and Blau 1982; Corzine and Huff-Corzine 1992; Harvey 1986; Hawkins 1993; Krivo and Peterson 2000; Parker 1989; Phillips 1997). High levels of racial segregation, leading to situations of concentrated disadvantage, have been argued to provide a context in which violence forms a rational response to the plight of individuals confronted with such limited economic opportunities and low future prospects (Parker and McCall 1997; Peterson and Krivo 1993; Phillips 1997). Inequality leads to a weakening of the legitimacy of social norms and impedes the ability of communities to exercise social control (Logan and Messner 1987). Concentrated disadvantage impacts the ability of residents to activate informal social control, in turn leading to increased levels of violence (Sampson, Raudenbush, and Earls 1997). According to Sampson and Wilson (1995), the sources of violent crime do not vary by race, but rather are embedded in structural variations in economic and family organization across communities, cities, and states. Sampson (1987) finds that the impact of Black male joblessness on violent crime is mediated by family disruption. In particular, he suggests that high rates of joblessness among Black males are a major contributing factor to increases in out-of-wedlock births and female-headed households among Black families. Family disruption, in turn, leads to higher crime rates as a result of reduced formal and informal social control, and the impact of marital conflict (which frequently precipitates marital breakup) on delinquent behavior among juveniles. Under these conditions, family disruption is seen as impeding the exercise of social control over juveniles in particular, and thus the effect of family disruption on crime rates is hypothesized to be stronger among this group than adults, but important to both as well. Sampson’s findings support these hypothesized relationships (see also Messner and Sampson 1991; Sampson 1995). More generally, family breakdown has been shown to be an important predictor of violent crime among Blacks and Whites (Parker and McCall 1997).

Racial Differences in Homicide Offending The literature on racial differences in homicide is dominated by aggregatelevel research. The findings of these studies vary considerably. For example, some research finds a differential impact of various sets of predictors on rates of White and Black homicide rates (see Harer and Steffensmeier 1992; Huff-Corzine, Corzine, and Moore 1986; LaFree and Drass 1996; LaFree, Drass, and O’Day 1992; Messner and Golden 1992; Ousey 1999; Parker and McCall 1997; Smith 1992). In other cases, the impact of structural factors is found not to vary by race (e.g., LaFree, Drass, and O’Day 1992; Krivo and Peterson 1996; Sampson 1987; 1995).

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

October 20, 2003

195

Prior studies have demonstrated links between segregation and Black crime rates (Parker and McCall 1997; Peterson and Krivo 1993; Shihadeh and Flynn 1996; Smith 1992), although in one study segregation was found to be related to White homicide rates but not Black homicide rates (Sampson 1985a). The findings regarding relationships between inequality and rates of violent crimes committed by Blacks are also relatively complex. While some studies report weak or no relationships (e.g., Harer and Steffensmeier 1992), others find a positive effect (e.g., Messner and Golden 1992; Shihadeh and Maume 1997; Smith 1992), some report that the effects of inequality are indirect (e.g., Shihadeh and Steffensmeier 1994), and still others report that the relationships vary according to the type of homicide examined (e.g., Corzine and Huff-Corzine 1992). Research has found some support for an effect of intraracial inequality on Black and White youth homicide rates (Messner, Raffalovich, and McMillan 2001). However, results for economic inequality appear to be extremely sensitive to the type of measurement used (LaFree and Drass 1996; Messner, Raffalovich, and McMillan 2001). Individual-level research finds that violence among African Americans varies by place of residence, including southern versus nonsouthern and urban versus rural locations (Hawkins 1999). Rose and McClain (1990) argue that changes in homicide structure among African Americans (i.e., from violence motivated by emotional discord and/or anger to instrumentally oriented violence) occurred more rapidly in urban areas of the northern United States than in the South. By the end of the 1970s, Rose and McClain (1990) contend that robbery homicides by African Americans were increasingly being committed against individuals known to the offender, and frequently involved multiple perpetrators. Similarly, other studies suggest that homicides involving Black offenders are more likely to be committed with a handgun (Wilbanks 1984), to be interracial (Humphrey and Palmer 1986), and to be committed during the commission of another felony (Wilbanks 1986) compared to White homicide offenders. Even though the overwhelming majority of homicides are intraracial, a small body of research examines interracial killings. To the extent that racial segregation limits interracial contact, racial segregation should be negatively related to interracial homicides, a hypothesis that is supported by research on interracial homicide (see Jacobs and Wood 1999; Parker and McCall 1997). Economic competition between Blacks and Whites has also been found to be a motivating factor underlying interracial homicide in some studies (Parker and McCall 1997). With less economic competition occurring between the races in Southern locations (see Jacobs and Wood 1999), these shifting patterns in the nature of homicide may be more evident in the North. Circumstances surrounding interracial killings indicate a greater proportion of homicides involving White offenders and Black victims to be expressively motivated, whereas incidents of Black offenders killing White

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

196

October 20, 2003

RETHINKING HOMICIDE

victims occur more frequently among instrumentally motivated homicides such as those committed during the commission of another felony ( Jacobs and Wood 1999). Although aggregate studies of homicide rates have been widely conducted in previous research, these studies are limited in several respects. First, the high level of colinearity between the racial composition of a geographical area and measures of socioeconomic conditions (like poverty level and single-adult households) makes it difficult to isolate the unique impact of race on homicide rates (see Land et al. 1990; Phillips 1997). Second, most previous studies employ cross-sectional designs, and are thus unable to assess the nature and magnitude of change in racial differences over time. Third, most studies estimate an additive or main effects model that makes the dubious assumption that the impact of race is identical across levels of other variables (see Miethe and Moore 1986). Fourth, aggregate studies, by definition, ignore the situational dynamics that result in the particular social ecology of homicide. Under these varying conditions across studies, it shouldn’t be surprising that substantive conclusions from previous research are mixed about the nature of racial differences in homicide and its correlates.

Changes in the Nature of Homicide over Time There have been much public commentary and concern about changes in the nature of homicide over time. Rather than examining these various changes simultaneously, however, most attention has focused on isolated changes in either the sociodemographic characteristics of offenders, their motivations, their victims, offense attributes, or the situational context of homicide. Both social science research and media accounts suggest that some of these changes in the offender, victim, and situational aspects of homicide may differ for Black and White offenders. Of the various circumstances surrounding homicide, growing public policy and law enforcement efforts have focused on the topic of youth gang violence. Police departments across the country have developed specialized gang enforcement units and several states have passed specific legislation that provides for sentencing enhancements for crimes committed in the furtherance of gangs (Klein 1995; McCorkle and Miethe 1998; Spergel, Curry, Chance, Kane, Ross, Alexander, Simmons, and Oh 1994). Some authors have questioned the magnitude of the gang problem, suggesting that law enforcement officials have overdramatized gang activity and membership to enhance their resource base ( Jackson and Rudman 1993; McCorkle and Miethe 1998; Zatz 1987). While estimates of the prevalence of gang homicides vary widely across jurisdictions and definitions of what constitutes being a “gang” member

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

October 20, 2003

197

(Curry, Ball, and Decker 1996; Decker and Kempf-Leonard 1995; Maxson and Klein 1996; Rosenfeld, Bray, and Egley 1999), there is a widely held stereotypical profile of these offenders and their crimes. Specifically, gang homicides are often assumed to involve young, minority, male offenders who live in large urban areas, use accomplices and firearms to commit expressive acts motivated by anger, frustration, and despair, and victimize persons with similar demographic characteristics (Cummings and Monti 1993; Sanders 1994). The motivations and social ecology of gang violence, however, appear to have changed over time. These changes include a rise in instrumental motivations, female offenders, “mushroom” victims (i.e., innocent bystanders of drive-by shootings), and the emergence of gangs in central cities across the country and in rural areas (Block and Block 1993; Harvey 1986; Sherman, Steele, Laufersweiler, Hoffer, and Julian 1989). Field research studies also indicate that gang violence in public housing projects has changed in the last decade from gang wars to drug wars, with bystanders more widely exposed to danger (Short 1997; Venkatesh 1996). Changes in the economic opportunity structure over the last three decades also provide a basis for speculation about the changing nature of homicides by particular subgroups of African Americans. Faced with relative deprivation and growing economic marginality, instrumentally motivated crimes that result in homicides (like robbery murders and homicides during burglaries) may be especially likely to provide an alternative opportunity structure for young, urban, African-American males (Harvey 1986; Sampson 1987). According to Wilson (1987), the movement of middle- and working-class black families out of urban ghetto neighborhoods in the 1970s and early 1980s has left behind a truly disadvantaged Black underclass that is highly susceptible to violence through its isolation from mainstream institutions, values, and role models. Changes to urban labor markets have further disadvantaged Black inner city residents, with the disappearance of low-skill jobs from these areas and their subsequent replacement with a high-skill professional and service employment sector. Increased racial segregation, joblessness, population turnover, and concentration of poverty in ghettos during this time have also been documented (Sampson and Wilson 1995). Focusing specifically on regional differences in economic opportunities and the motivations for violence, Rose (1984) claims that African-American homicide in large southern urban environments more often represents expressive acts of violence, whereas homicide by these offenders in northern cities is more often designed for instrumental goals. Alternatively, increasing economic marginality may have similar effects on increasing the rate of expressive violence across different geographical areas and racial groups. These assertions about race-specific changes in the nature of homicide, however, have not been empirically evaluated.

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

198

October 20, 2003

RETHINKING HOMICIDE

Homicides among Hispanics Although differences in levels and patterns of violent behavior among Blacks and Whites have received extensive attention in the criminological literature, far less is known about violence committed by other racial and ethnic groups. Part of the reason for the lack of attention given to homicide offending and victimization of groups like Asians and Native Americans is simply the small numbers of cases these groups contribute to the overall homicide rate in any given year. The case of Hispanics, however, is a bit different. In certain large urban areas in the United States, there are sufficient cases of Hispanic homicide to warrant analysis. For example, Latinos make up 40% of homicide victims in Miami (Martinez and Lee 1998). Similarly, as with African Americans, the risk of homicide among Hispanics is relatively high, ranking fourth among the common causes of death among Latino men (Baker 1996). A number of studies report that the risk of homicide victimization among Hispanics is considerably higher than for Anglos (Block 1993; Mercy 1987; Pokorny 1965; Rodriguez 1987; Zahn 1987), although other research reports very similar rates for the two ethnic groups (Martinez 1997). The limited evidence available also suggests that homicides among Hispanics may have a unique character (Martinez and Lee 1999). Bivariate analyses conducted in previous studies indicate that patterns of homicide among Hispanics differ from those of other groups with respect to certain victim, offender, and offense characteristics. For example, Hispanic homicides tend to have a higher involvement of multiple offenders (Block 1993), to be committed by strangers (Block 1993), to be motivated by arguments (Martinez and Lee 1998), to involve males killing other males (Block 1993), to be intraracial (Martinez and Lee 1998), to be street gang–related (Block 1993), to occur on the street (Block 1987; 1993; Loya and Mercy 1985), and to involve firearms (Block 1993; Rodriguez 1987). Hispanics are less likely than non-Hispanic whites to be killed during a robbery (Martinez 1997) or to be killed by a family member (Lebya 1987; Zahn 1987). Hispanic homicides also include fewer female victims (Block 1987; Rodriguez 1987). The gender ratio (homicide victimization risk of males relative to females) has been found to be much higher for Hispanics than other racial/ethnic groups (Block 1993). The findings with respect to age are mixed; some studies report a higher involvement of older offenders in Hispanic homicides (Martinez and Lee 1998) while others report a high risk of homicide offending among Latinos during the ages of 15 to 19 years, dropping sharply thereafter (Block 1993). The risk of becoming a homicide victim among elderly Hispanics is particularly low (Block 1987). The need to examine the intersection of these attributes has been noted elsewhere (e.g., Block 1987; Rodriguez 1987).

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

October 20, 2003

199

The Analysis of Homicide Situations Rather than examining the race-specific predictors of homicide rates, an alternative approach is to focus on the relative similarities and differences in the structure of homicide situations involving different racial groups. In other words, are Black homicides different from White homicides in their combinations of offender, victim, and offense elements? How, and to what extent, have these characteristics changed over time? For many research questions involving crime causation and its distribution, situation-based approaches may be preferred to the traditional aggregate-level analysis because they better capture the social context of crime. Specifically, situation-based approaches require an examination of the convergence of offender, victim, and situational elements in their totality across time and space. Even with extensive disaggregation of the data for subgroup comparisons, the traditional aggregate-level approach can not fully model the complexity of the social ecology that underlies criminal activity and often rests on the tenuous assumption that a variable’s impact is constant across other variables and contexts. In contrast, situation-based approaches begin with the recognition of the fundamental interplay among offender, victim, and situational elements to identify crucial dimensions and combinations of factors that are associated with the occurrence of crime. Aside from their utility for understanding the context of violent crime, many current debates and public commentaries about homicide involve discussions about combinations of factors that may be better addressed through a situational analysis. Expressive acts of violence by minority, urban youth in a group context, for example, have been widely publicized as an emergent and especially prevalent form of homicide in the United States over the last three decades (Bilchik 1999; Cook and Laub 1998; Fox 1996). By examining the combination of offender, victim, and situational factors that define homicide situations over time, we can empirically validate the relative prevalence of these acts and see whether they are a unique form of homicide for particular racial or ethnic groups. Similar assertions about change and stability in other particular elements of Black and White homicides (like an increased use of firearms, group offending, stranger victims, interracial slayings, expressive rage or anger slayings, and rural murders) can also be evaluated more completely through the analysis of homicide situations over time.

Research Questions QCA is used in the current study to examine (1) whether homicide situations involving offenders from different racial and ethnic groups exhibit common or unique patterns in their combination of offender, victim, and

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

200

October 20, 2003

RETHINKING HOMICIDE

situational elements, and (2) whether racial and ethnic differences in the configurations of offender, victim, and situational factors have changed over time. Compared to standard statistical approaches (like contingency table analysis and regression analysis), QCA is designed to uncover patterns of offender, victim, and situational attributes in homicides that uniquely differentiate between African-American, White, and Hispanic offenders. When applied to the study of racial differences in homicide situations, QCA provides answers to several questions about their structural characteristics. First, how many different combinations of offender, victim, and situational characteristics are observed for Black and White homicides in each of the last three decades? Second, how many of these combinations are common to both racial groups? Third, how many of these characteristics are unique to Black homicides and to White homicides? Fourth, have unique forms of Black and White homicides emerged over the last three decades? Fifth, what is the nature and relative prevalence of each unique configuration of race-specific homicide situations? Sixth, what is the nature of Hispanic homicide situations? Are they more similar to Black or White offenders in terms of their underlying structure? By addressing these questions, QCA provides a systematic method for examining the nature and magnitude of the similarities and differences between Black, White, and Hispanic homicides.

Data Description and Methods SHR data are used in this study to examine race-specific changes and stability in the social ecology of homicide situations over time. The primary data involve all homicides involving Black or White offenders for the years 1976 to 1998 (FBI 1976–98). Yearly incidents are combined by decade for comparative purposes. Our analysis is based on a total of 72,563 homicides in the 1970s, 185,756 homicides in the 1980s, and 173,362 homicides in the 1990s. The major variables include offender characteristics, victim characteristics, and offense elements. Offender, victim, and offense elements are used to examine the common and unique configurations underlying the structure of homicide situations for each racial group. The lack of consistent measures of ethnicity over time requires the exclusion of Hispanic offenders from the initial analysis. However, a second set of analyses using homicide data for California for the years 1987 through 1999 permits a comparison of Hispanic with Black and White offenders. Other ethnic and racial groups (such as Asians and Native Americans) are excluded because of their infrequency as offenders. The specific offender characteristics involve their gender and age. Age is dichotomized to compare offenders under 20 with older offenders. Victim

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

October 20, 2003

201

characteristics consist of their race, gender, and age (less or greater than 30 years old). Offense-related attributes include the homicide motive (instrumental versus expressive), the victim–offender relationship (intimate/ family; friend/acquaintance; stranger), interracial offenses, gun use, multiple offenders, region of the country (South versus other regions), and population size (large city over 100,000 vs. other). As shown in Table 8.1, there are no major differences in the individual elements underlying homicide situations between racial groups and over time. For both Black and White offenders, homicides almost always involve male offenders, intraracial attacks, and single offenders. Both groups usually kill nonstrangers during arguments, in large urban areas, with firearms. With the possible exception of urban locations (where Black offending has become more concentrated over time), these main effect trends are similar for each decade and racial groups within them. Contrary to the various sources of rhetoric about fundamental changes in the racial character of violence, these bivariate trends show that most racial differences in offender, victim, and offense characteristics are small in magnitude and persist over time.

Results Black versus White Offenders The trend analysis presented in Table 8.1 indicates that most elements of homicide situations involving Black or White offenders have not changed over time. However, this bivariate analysis may actually mask whether racespecific combinations of factors have emerged over the last three decades. Accordingly, QCA is used to assess the similarity and differences in combinations of offender, victim, and situational elements for each racial group. Across all three decades combined, the majority of homicides (54%) occur in situational contexts that are unique to either Black or White offenders. There is also greater situational clustering among homicides involving Black offenders in the sense that there are a greater proportion of cases falling within a smaller number of structural profiles. In short, the majority of U.S. homicides occur in situations with race-specific signatures, and White offenders appear to kill in a greater diversity of circumstances. The particular offense-related elements that most strongly distinguish between homicides across the two racial groups are multiple offenders, interracial attacks, and urban killings. More specifically, homicide situations unique to White offenders are more than twice as likely to be intraracial (79.4% vs. 29.3%), and are about half as likely to occur in urban locations (29.6% vs. 59.4%). Perhaps most surprisingly, they are also more likely to involve multiple offenders (58.5% vs. 40.6%). Although the prevailing image of

19:38

385,518 116,146 214,615 100,920 120,674 285,667 48,250 283,210 180,660

=Race Family Acquaint Strang Instrum Gun Multoff Urban South

378,165 81,321

Omale O < 20 330,143 219,557

224,257 207,424

Oblack Owhite

Vmale O3098

72,563 185,756 173,362

Total N

Yr7679 Yr8089 Yr9098

Value Labels#

202 30.0 69.3 11.2 74.7 45.3

85.8 23.3 52.7 23.9

79.5 48.1

86.1 20.7

% Black

25.8 62.8 11.1 55.8 38.1

93.1 30.8 46.5 22.8

73.2 53.9

89.2 16.8

% White

26.1 67.2 8.2 70.8 45.5

85.8 26.8 52.0 21.2

78.2 53.5

82.4 16.1

% Black

27.8 63.1 8.9 51.6 38.5

94.5 31.8 44.9 23.3

73.8 54.0

87.9 16.6

% White

Yr 1976–79

32.2 75.0 13.3 76.7 46.0

86.1 20.8 53.6 25.6

80.8 43.6

88.8 25.5

% Black

23.9 65.2 12.4 58.9 35.6

91.5 29.9 47.6 22.5

73.7 53.0

90.0 18.6

% White

Yr 1990–98

CB617-Miethe-v1

Notes: # descriptive labels used in QCA results. ∗ excluded categories. For all variables, 1 = Yes and 0 = No. The proportion of observations in each of the excluded categories can be derived by subtractions. For example, 86.1% of the Black offenders in the 1976–98 period were male so the percentage of Black offenders in this period who are female is 13.9% (100-86.1). Source: FBI SHR Data, 1976–98.

Year of Homicide 1976–79 1980–89 1990–98 Type of Homicide Black Offender White Offender Offender Characteristics Male Offender <20 yrs. old Victim Characteristics Male Victim Age 30 to 98 Victim–Offender Characteristics V–O Same Race V–O Family/Intimate∗ V–O Acquaintances V–O Strangers Offense Characteristics Instrumental Motive Firearm Multiple Offenders Urban South

Variables

Yr 1976–98

Table 8.1. Coding, Value Labels, and Frequencies for All Variables

P1: JMV 0521832993c08 October 20, 2003 19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

October 20, 2003

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

203

Figure 8.1. Proportion of Homicides in Common and Unique Situations

groups of individuals accosting innocent victims is frequently attributed to minority offenders, homicides involving multiple offenders are more often found in particular situational contexts that are relatively unique to White offenders.1 An important question concerns the extent to which patterns such as diversity and situational clustering have changed over time. We pursue this issue by examining homicide situations for Black and White offenders separately for each decade. These results are summarized in Figure 8.1. While the situational contexts of homicides for Black and White offenders have undergone some change over time, stability is the predominant pattern. This can be seen in a variety of ways. First, as shown in Figure 8.1, the proportion of cases falling into unique and common profiles is virtually unchanged across decades. Regardless of the time period, slightly more than half of all homicide cases fall within configurations that are unique to either Black or White offenders. 1

We don’t have a clear explanation for this finding. However, the finding suggests that multiple offenders and the parties involved in these homicides are qualitatively different in urban and less urban areas for Black and White offenders. Distinct intraracial attacks in a group context in less urban areas by White offenders may involve two or three “buddies” in an argument (e.g., bar fights) rather than in a gang context within urban areas. Their uniqueness also lies in the combination of weapons other than firearms, older offenders and victims, and several other particular combinations of attributes that are not found among Black or White offenders who kill in a group context in larger urban environments. Unfortunately, we are unable to explore the nature of this particular finding more closely because of the limitations of the data (e.g., limited detail about the circumstances surrounding the homicide).

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

204

October 20, 2003

RETHINKING HOMICIDE

Second, an examination of the 10 most dominant profiles that are unique to Black and White offenders for each decade reveals a great deal of stability in the nature of the dominant homicide situations for each group over time (see Table 8.2). This is particularly the case among White offenders, where the only notable change is a slight increase in the number of homicide situations occurring in urban contexts. Even though the dominant pattern among Black offenders is also one of stability over time, there are a few clear exceptions. Dominant changes in the profiles for this subgroup involve more instrumental killings, southern locations, and younger victims over time. As suggested by Wilson’s work on Black inner-city communities (see Wilson 1987; 1996), the increase in instrumentally motivated crimes might be expected as a response to the disappearance of legitimate employment opportunities from central cities over the past several decades. The dominant profiles of instrumental homicides committed by Blacks in the 1970s and 1980s occurred in northern locations, but by the 1990s they were equally split between southern and northern regions, which would seem to contradict the prediction by Rose (1984) of an increase in instrumental homicides in northern locations by Blacks. There are also fewer homicides involving female offenders in later decades, a pattern which is consistent with prior research on the changing nature of intimate partner homicides (see Browne, Williams, and Dutton 1999). The argument that stability rather than change is the major theme is evident from the findings of a supplemental QCA analysis in which homicide cases in the 1970s and 1990s are compared within each racial group. For Black offenders, 59.2% of configurations are common to both time periods. These common profiles account for 60% of all homicides committed by Blacks during these decades. A similar pattern emerges with respect to White offenders, where 62.9% of homicide situations are common to both decades. Common profiles across time for White offenders account for 75.4% of all homicides committed by this racial group, reinforcing the pattern noted earlier that the situational context of homicide offending among Whites has remained even more stable over time than that for Black offenders. Finally, a comparison of the victim, offender, and offense attributes that most strongly distinguish between uniquely Black vs. uniquely White offender homicides across time further emphasizes the conclusion that the situational context of lethal violence among both subgroups has remained relatively stable over the past three decades. However, a few patterns of change are also evident. The most dramatic shifts have occurred with respect to gun use, instrumental motives, and intraracial killings (see Figure 8.2). Guns and instrumental motives are more common among homicides that occur in situations unique to Black offenders than White offenders, and this has become increasingly the case in recent times, as is evident

19:38

205

Omale Omale Omale Omale Ofemal Ofemal Omale Omale Ofemal Omale

Omale Omale Omale Omale Ofemal Omale Omale Omale Omale Omale

V < 30 V3098 V3098 V < 30 V3098 V3098 V < 30 V < 30 V3098 V < 30

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race Expr Expr Expr Expr Expr Expr Expr Insrum Expr Expr

Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun

Acqu Acqu Acqu Acqu Family Family Acqu Acqu Family Acqu

Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff

Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban

North North South South South North North North North South

1,315 of 2,099 1,182 of 1,902 1,107 of 1,673 907 of 1,372 446 of 669 412 of 604 371 of 587 295 of 485 284 of 345 184 of 278

Vfemal Vmale Vmale Vfemal Vmale Vfemal Vmale Vmale Vmale Vfemal

V3098 V < 30 V3098 V3098 V3098 V < 30 V3098 V3098 V < 30 V3098

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race Expr Expr Expr Expr Expr Expr Instrum Expr Expr Expr

Gun Gun Gun Gun Gun Gun Nongun Nongun Nongun Nongun

Family Acqu Acqu Family Family Family Strang Acqu Acqu Family

Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff

NonUrb NonUrb NonUrb NonUrb NonUrb NonUrb Urban NonUrb NonUrb NonUrb

South North North North North North North North North North

483 of 794 562 of 790 560 of 776 583 of 666 300 of 420 312 of 393 248 of 361 258 of 351 252 of 342 237 of 273

Percentage of all White offender homicides explained by these dominant profiles: 11% (3,795 of 35,911)

O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O > 20

Profile Ns

(cont.)

61% 71% 72% 88% 71% 79% 69% 74% 74% 87%

% White∗∗

63% 62% 66% 66% 67% 68% 63% 61% 82% 66%

% Black∗

CB617-Miethe-v1

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Vmale Vmale Vmale Vmale Vmale Vmale Vmale Vmale Vmale Vmale

Profile Ns

Percentage of all Black offender homicides explained by these dominant profiles: 18% (6,503 of 36,652)

O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O < 20 O > 20 O > 20 O < 20

White Offenders:

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Black Offenders:

Panel A: 1970s

Table 8.2. Qualitative Comparative Analysis of Homicide Situations Unique to Black and White Offenders, FBI SHR Data, 1976–98

P1: JMV 0521832993c08 October 20, 2003 19:38

206

Omale Omale Omale Omale Omale Omale Omale Ofemal Ofemal Omale

Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale

V < 30 V3098 V < 30 V < 30 V3098 V < 30 V < 30 V < 30 V3098 V < 30

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race Expr Expr Expr Expr Instrum Instrum Instrum Expr Expr Expr

Gun Gun Gun Gun Gun Gun Gun Gun Nongun Gun

Acqu Acqu Acqu Acqu Acqu Strang Acqu Family Family Acqu

Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff

Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban

North South South North North North North South North South

3,541 of 5,627 2,306 of 3,837 2,211 of 3,442 1,145 of 1,660 757 of 1,265 768 of 1,260 858 of 1,257 720 of 1,184 658 of 888 602 of 857

Vmale Vfemal Vmale Vfemal Vfemal Vmale Vmale Vmale Vmale Vfemal

V3098 V3098 V < 30 V3098 V3098 V3098 V3098 V < 30 V3098 V3098

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race Expr Expr Expr Expr Expr Expr Expr Expr Instrum Expr

Gun Gun Gun Gun Gun Gun Nongun Nongun Nongun Nongun

Acqu Family Acqu Family Family Family Acqu Acqu Strang Family

Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff

NonUrb NonUrb NonUrb NonUrb Urban NonUrb NonUrb NonUrb Urban NonUrb

North South North North North South North North North North

1,497 of 1,953 1,348 of 1,934 1,325 of 1,852 1,492 of 1,665 978 of 1,518 799 of 1,228 785 of 1,049 789 of 1,035 576 of 936 767 of 909

Percentage of all White offender homicides explained by these dominant profiles: 11% (10,356 of 93,492)

O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O > 20

Profile Ns

77% 70% 72% 90% 64% 65% 75% 76% 62% 84%

% White∗∗

63% 61% 64% 69% 60% 61% 68% 61% 74% 70%

% Black∗

CB617-Miethe-v1

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Vmale Vmale Vmale Vmale Vmale Vmale Vmale Vmale Vmale Vmale

Profile Ns

Percentage of all Black offender homicides explained by these dominant profiles: 15% (13,566 of 92,264)

O > 20 O > 20 O > 20 O < 20 O > 20 O > 20 O > 20 O > 20 O > 20 O < 20

White Offenders:

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Black Offenders:

Panel B: 1980s

Table 8.2. (cont.)

P1: JMV 0521832993c08 October 20, 2003 19:38

Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale

207

Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale

V < 30 V3098 V < 30 V < 30 V < 30 V < 30 V < 30 V3098 V < 30 V < 30

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race Expr Expr Instrum Expr Instrum Expr Instrum Instrum Expr Expr

Gun Gun Gun Gun Gun Gun Gun Gun Gun Gun

Acqu Acqu Acqu Acqu Acqu Strang Acqu Acqu Family Acqu

Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff

Urban Urban Urban Urban Urban Urban Urban Urban Urban NonUrb

South South North South South South North South South South

3,210 of 4,078 1,890 of 2,888 1,326 of 1,858 1,423 of 1,758 1,090 of 1,196 813 of 1,094 754 of 954 669 of 905 605 of 862 642 of 857

Vmale Vmale Vfemal Vmale Vfemal Vfemal Vmale Vfemal Vfemal Vmale

V3098 V < 30 V3098 V3098 V3098 V3098 V < 30 V3098 V3098 V3098

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race Expr Expr Expr Expr Expr Expr Expr Expr Expr Expr

Gun Gun Gun Gun Gun Gun Nongun Nongun Nongun Nongun

Acqu Acqu Family Acqu Family Family Acqu Family Family Acqu

Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff

NonUrb NonUrb NonUrb NonUrb NonUrb Urban Urban Urban NonUrb NonUrb

South North South North North North North North North North

1,484 of 2,490 1,114 of 1,780 1,138 of 1,589 1,187 of 1,483 1,121 of 1,336 865 of 1,331 710 of 1,270 684 of 1,229 737 of 888 709 of 885

Percentage of all White offender homicides explained by these dominant profiles: 12% (9,749 of 78,021)

O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O > 20 O > 20

Profile Ns

60% 63% 72% 80% 84% 65% 56% 56% 83% 80%

% White∗∗

79% 65% 71% 81% 91% 74% 79% 74% 70% 75%

% Black∗ CB617-Miethe-v1

Notes: ∗ = % Black refers to the percent of Black offenders that are contained within this profile. ∗∗ = % White refers to the percent of white offenders within this profile.

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Vmale Vmale Vmale Vmale Vmale Vmale Vmale Vmale Vmale Vmale

Profile Ns

Percentage of all Black offender homicides explained by these dominant profiles: 13% (12,422 of 95,341)

O > 20 O > 20 O > 20 O < 20 O > 20 O > 20 O < 20 O > 20 O > 20 O < 20

White Offenders:

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Black Offenders:

Panel C: 1990s

P1: JMV 0521832993c08 October 20, 2003 19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

October 20, 2003

208

Gun 1970s

RETHINKING HOMICIDE

Gun 1990s

Instrum 1970s

Instrum 1990s

Interracial Interracial 1970s 1990s

Figure 8.2. Changes over Time in Attributes Distinguishing Unique Profiles

in the widening gap across racial groups in the proportion of homicides unique to each group that are instrumentally motivated and committed with a gun. This pattern fits with media portrayals and criminological research suggesting an increase in firearm homicides and economically motivated crimes among Blacks in recent decades. Interracial homicides are also far more prevalent among profiles unique to Black offenders, but these killings have become less prevalent over time within uniquely Black situational contexts. Nonetheless, the major pattern has been stability in the structure of homicide situations for both Black and White offenders over the last three decades.

Comparative Analysis of Hispanic Homicides It has been suggested that homicides among Hispanics differ from those involving other ethnic groups. However, this conclusion has been drawn largely on the basis of bivariate comparisons. In this section, we examine whether the structures of homicide situations involving Hispanic offenders are unique from those involving either White offenders or Black offenders. We use homicide data from a single state – California – where the proportion of Hispanic offenders is substantial enough to permit these kinds of comparisons.2 In fact, Hispanic offenders make up the greatest percentage of offenders in California between 1987 and 1999. The analyses focus on the following questions: (1) what is the level of diversity and concentration 2

The data for this analysis are derived from the California Homicides Data File, described in Chapter 3. Between 1987 and 1999, 38,467 homicides were reported to the California Criminal Justice Authority. Of these homicides, 38.8% involved Hispanic offenders, 22.4% involved White offenders, 34.2% involved Black offenders, with the remaining 4.7% of cases involving offenders of other racial/ethnic backgrounds.

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

October 20, 2003

209

Figure 8.3. Cases Involving Unique and Common Situations for Hispanic and White Offenders

of homicide situations involving Hispanic offenders; (2) is the structure of homicide situations unique for Hispanic and non-Hispanic White offenders; and (3) is the structure of homicide situations similar across minority groups, specifically Hispanic and Black offenders? We use the findings of prior research on Hispanic homicides to guide our definition of homicide structures for this analysis. In particular, we select victim, offender, and offense attributes that have been found in bivariate analyses to distinguish between Hispanic and non-Hispanic homicides. Accordingly, homicide situations are defined on the basis of the following variables: offender gender, offender age (less than 20; 20–29; 30 and over), victim and offender of the same race or gender, victim–offender relationship (intimate/family; friend/acquaintance; stranger), multiple offender, gun, public outdoor location, and motive (instrumental; expressive; gang).

Hispanic versus White Offenders Using QCA to compare Hispanic and White offenders, we find that homicides by individuals within these subgroups reduce to 989 dominant homicide situations. Over one-third (37.9%) of cases are represented by situations relatively unique to Hispanics and another one-third (32.5%) are accounted for by profiles that are relatively unique to White offenders. The smallest proportion of cases (29.6%) are covered by situations that are common to both groups (Figure 8.3). Thus, these findings strongly support the argument that Hispanic homicides are qualitatively unique, at least as compared to those committed by Whites. The dominant profiles for Hispanic and White offenders are displayed in Panel A of Table 8.3. The situations that are qualitatively unique to Hispanics

19:38

210 acquaint family family acquaint family family family acquaint acquaint family

gun gun nongun nongun gun gun gun gun gun nongun

singoff singoff singoff singoff singoff singoff singoff singoff singoff singoff

otherloc otherloc otherloc otherloc otherloc otherloc otherloc otherloc otherloc otherloc

express express express express express othmotv express express instrum express

gang express express gang gang express gang instrum instrum instrum

348 of 645 410 of 603 238 of 317 165 of 279 118 of 194 158 of 174 136 of 156 68 of 123 57 of 121 60 of 106

(54%) (68%) (75%) (59%) (61%) (91%) (87%) (55%) (47%) (57%)

(98%) (82%) (88%) (98%) (96%) (77%) (98%) (90%) (78%) (74%)

%

CB617-Miethe-v1

samerace samerace samerace samerace samerace samerace samerace diffrace samerace samerace

puboutr puboutr puboutr puboutr puboutr puboutr puboutr puboutr puboutr puboutr

samesex diffsex diffsex samesex samesex diffsex diffsex samesex samesex samesex

multoff singoff singoff singoff multoff singoff multoff multoff multoff singoff

White Offenders: 1. Omale O3098 2. Omale O3098 3. Omale O3098 4. Omale O3098 5. Omale O3098 6. Omale O3098 7. Ofemal O3098 8. Omale O3098 9. Omale O3098 10. Omale O3098

gun gun gun gun gun gun gun gun gun gun

518 of 528 344 of 420 260 of 295 248 of 253 238 of 248 188 of 244 231 of 236 173 of 192 145 of 186 132 of 178

stranger acquaint stranger stranger stranger stranger acquaint stranger stranger acquaint

samesex samesex samesex samesex samesex samesex samesex samesex samesex samesex

Hispanic Offenders: 1. Omale O < 20 2. Omale O2029 3. Omale O2029 4. Omale O < 20 5. Omale O2029 6. Omale O3098 7. Omale O < 20 8. Omale O < 20 9. Omale O2029 10. Omale O2029 samerace samerace samerace samerace samerace samerace samerace samerace samerace samerace

Profile N’s

Panel A: Hispanic vs. White Offenders

Table 8.3. Qualitative Comparative Analysis of Homicide Situations Unique to Hispanic, White, and Black Offenders, California, 1987–99

P1: JMV 0521832993c08 October 20, 2003 19:38

samesex diffsex samesex samesex samesex samesex samesex samesex samesex samesex

samerace samerace samerace diffrace samerace diffrace samerace samerace diffrace diffrace

acquaint family stranger stranger acquaint stranger acquaint acquaint stranger stranger

gun nongun gun gun gun gun gun gun gun gun

gun gun gun gun gun nongun nongun gun gun gun singoff singoff singoff multoff singoff multoff multoff singoff singoff singoff

multoff singoff singoff multoff singoff singoff singoff multoff singoff singoff puboutr otherloc puboutr puboutr puboutr puboutr puboutr otherloc puboutr puboutr

puboutr otherloc puboutr puboutr puboutr puboutr otherloc puboutr otherloc otherloc instrum express instrum instrum instrum instrum instrum instrum instrum instrum

gang express express gang express express express express express express

181 of 312 143 of 223 103 of 177 130 of 175 117 of 170 114 of 159 89 of 156 91 of 150 95 of 149 95 of 136

(58%) (64%) (58%) (74%) (69%) (72%) (57%) (61%) (64%) (70%)

(70%) (65%) (66%) (64%) (66%) (76%) (78%) (71%) (66%) (69%)

%

CB617-Miethe-v1

211

MODEL: Offhisp = Omale + Olt20 + O2029 + =Sex + =Race + Acquaint + Stranger + Gun + Multoff + Puboutr + Instrum + Express + Gang

Black Offenders: 1. Omale O2029 2. Omale O3098 3. Omale O3098 4. Omale O < 20 5. Omale O < 20 6. Omale O2029 7. Omale O2029 8. Omale O2029 9. Omale O2029 10. Omale O < 20

stranger acquaint stranger acquaint stranger acquaint acquaint stranger family stranger

514 of 734 297 of 457 262 of 397 231 of 361 189 of 287 198 of 261 158 of 203 126 of 178 110 of 167 108 of 157

samerace samerace samerace samerace samerace samerace samerace samerace samerace samerace

Hispanic Offenders: 1. Omale O < 20 2. Omale O3098 3. Omale O2029 4. Omale O < 20 5. Omale O3098 6. Omale O2029 7. Omale O2029 8. Omale O2029 9. Omale O2029 10. Omale O2029 samesex samesex samesex samesex samesex samesex samesex samesex diffsex samesex

Profile N’s

Panel B: Hispanic vs. Black Offenders

P1: JMV 0521832993c08 October 20, 2003 19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

212

October 20, 2003

RETHINKING HOMICIDE

Figure 8.4. Attributes Distinguishing Uniquely Hispanic vs. White Offenders

are primarily gun killings in outdoor public locations by young males of other Hispanic males who are strangers. Many of these situations involve multiple offenders and are frequently gang-related. In contrast, unique situations for White offenders typically involve lone, older males who do not commit these violent acts in outdoor public locations. More than half of these dominant profiles involve the killing of a family member, often of the opposite sex, in an expressively motivated act. It is clear when comparing these profiles that the uniqueness of Hispanic homicides lies in the predominance of public outdoor locations, younger offenders, male-on-male killings, guns, and strangers, as well as the lack of homicides perpetrated by family members. This pattern is further emphasized by an examination of the attributes that most strongly distinguish between uniquely Hispanic vs. White offender killings (Figure 8.4). Homicide situations unique to Hispanic offenders are far more likely to involve offenders under the age of 20, to occur in outdoor public locations, and to be gang-related. They are also more likely to involve strangers, multiple offenders, male-on-male killings, and to be committed with a gun. On the whole, then, these QCA results strongly support the contention that Hispanic homicide offending has a qualitatively unique structure. These findings largely support the results of bivariate analyses in the existing literature. It is the combination of these attributes, however, that produces the pattern of predominantly unique homicide situations for Hispanic and White offenders.

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

October 20, 2003

213

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

Unique Hispanic Offender (27.1%)

Common to Both Groups (49.5%)

Unique Black Offender (23.4%)

Figure 8.5. Cases Involving Unique and Common Situations for Hispanic and Black Offenders

Hispanic versus Black Offenders The 28,054 homicide cases involving Hispanic or Black offenders in California between 1987 and 1999 occur in 850 dominant homicide situations. These configurations are comprised of 38% relatively unique to Black offenders, 29% relatively unique to Hispanic offenders, and the remaining one-third common to both groups. Situational profiles that are relatively unique to Hispanic offenders account for just over one-quarter (27.1%) of all homicides. Homicide situations relatively unique to Black offenders represent less than one-quarter (23.4%) of all homicide cases. These findings suggest that Hispanic homicides exhibit greater situational clustering than homicides involving Black offenders, with a greater percentage of cases falling into fewer unique profiles. Roughly half (49.5%) of all homicides occur in situations that are common to both groups (see Figure 8.5). This higher proportion of homicides in common situations suggests that the structure of homicide situations in which Hispanic and Black offenders kill is much less qualitatively distinct than found in our comparison of Hispanics versus non-Hispanic Whites. The situational dynamics underlying homicides by minority offenders is more similar in nature than situations involving White offenders, an issue that is probed further through the examination of homicide narratives in a subsequent section. It is evident from Panel B of Table 8.3 that what makes homicides involving Hispanic offenders qualitatively distinct is quite different depending upon whether the comparison group is White offenders or Black offenders. When comparing Hispanic versus White offenders we found that homicide situations unique to Hispanic offenders are distinguishable by the fact that they are much more likely to involve male offenders under the age of 20

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

214

October 20, 2003

RETHINKING HOMICIDE

Figure 8.6. Attributes Distinguishing Uniquely Hispanic vs. Black Offenders

killing other males, to occur in outdoor public locations, and to be gangrelated, as well as to be committed by strangers, multiple offenders, and with a gun. When compared to Black offenders, however, the uniqueness of homicide situations involving Hispanic offenders comes from the predominance of intraracial killings, expressive and gang motives, and fewer killings in public outdoor locations. The major homicide situations involving Black offenders are predominantly instrumentally motivated male-on-male killings with guns in public outdoor locations. Stranger homicides committed by Hispanic offenders are almost always expressively motivated and involve Hispanic victims, whereas stranger homicides committed by Black offenders are all instrumentally motivated and primarily interracial. As compared to the previous analysis involving Hispanic versus White offenders, there are fewer attributes that distinguish between homicide situations unique to Hispanic versus Black offenders and the differences are not as large. For the most part, however, the same attributes demonstrate importance here in drawing distinctions between unique profiles for Hispanic versus Black offenders. Homicide situations unique to Hispanic offenders are more likely to involve male offenders, perpetrators under the age of 20, same-sex killings, and to be gang-related (see Figure 8.6). Homicide situations unique to Black offenders are nearly three times more likely to be instrumental and are somewhat more likely to involve guns. Multiple offenders and public outdoor locations are not major distinguishing features of homicides committed by members of these minority subgroups, unlike the comparison between Whites and Hispanics.

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

October 20, 2003

215

Thus, the answer to the question of whether homicide situations involving Hispanic offenders are unique in character appears to depend to a large extent on the group to which these offenders are being compared. The qualitatively unique nature of Hispanic homicide is much more evident when the comparison involves White offenders. Hispanic and Black offenders, in contrast, appear to kill in similar situations, perhaps reflecting their shared minority status. Nevertheless some differences between the profiles of these groups are evident, particularly the varying context in which Hispanic and Black offenders kill strangers.

Homicide Narratives The processes underlying lethal violence by Whites, Blacks, and Hispanics are explored further in a qualitative analysis of homicide narratives from Los Angeles, St. Louis, and Miami. The nature of homicides across racial/ethnic groups is examined with respect to two types of homicides: intimate partner killings and confrontational homicides resulting from arguments. Intimate partner homicides are selected because although there has been a great deal of discussion regarding the different situations in which men and women kill an intimate partner, much less attention has focused on racial differences in intimate partner homicides. Confrontational homicides are selected because of their predominance as a homicide motive in general and because displays of masculine competitiveness and identity are likely to be played out within these dispute-type killings. Furthermore, prior research indicates a greater prevalence of killings resulting from arguments in cases involving Black and Hispanic offenders and it has been speculated that these arguments may represent incidents of honor or disrespect in public settings (see Martinez and Lee 1998).

Intimate Partner Homicides According to Polk (1994), two distinct subtypes of homicides underlie the killing of women by men: sexual possession and suicidal masculine depression, both of which involve a desire to control women. The theme of sexual jealousy and ownership is well-noted in the intimate partner homicide literature (see, for example, Daly and Wilson 1988a; Dawson and Gartner 1998; Polk 1994). What is less clear is whether this type of display of masculinity is evident across racial/ethnic groups. A careful reading of homicide narratives for Los Angeles across the twentieth century suggests that masculine sexual possessiveness and jealousy underlie killings of female intimate partners regardless of the race or ethnicity

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

216

October 20, 2003

RETHINKING HOMICIDE

of the perpetrator. Notice the situational factors and developmental sequence underlying the following examples of the killing of female intimate partners: About 10:30 p.m. September 5th call came into the police station that there was a murder at 3rd Ave. Officers answered the call and upon their arrival found the defendant holding victim #1. Officers placed the accused under arrest and went to a residence where they found his wife, victim #2, lying on the floor with multiple wounds on her body and neck. Defendant admitted doing the cutting and killing his wife and stated that it was because she was going with another man. Defendant booked on a charge of Murder. (Los Angeles, September 4, 1930) This killing was the result of a love triangle. The suspect was at the victim’s home when she went to work in the afternoon of 9-13-62. When victim failed to return home at her regular time of 11:30 p.m., the suspect went to his home and obtained his 32-20 revolver and returned, where he observed the victim and another male drive up and park at the curb. The victim got out of the car. Suspect fired two shots at the man in the car, and then shot the victim in the left chest apparently killing the victim instantly. Suspect then fled the scene and returned to his home where he hid the murder weapon. Suspect then returned to the scene where he surrendered to officers. (Los Angeles, September 13, 1962) Victim had been having trouble with her husband for several months, claiming that he was cruel to her. The couple had a 20-months-old baby girl. A week previous to her death, the victim left her husband and went to live with Mr. Smith (alias) and his wife, taking her baby along with her. About 11:30 a.m. the husband called on his wife. Mr. Smith left them alone while he went to pay the landlord the rent due him. When he returned he found the victim lying dead on kitchen floor with her feet beneath the sink and the husband in a dying condition with his head underneath the sink and an Iver Johnson .32 Cal. revolver containing two loaded cartridges and three empty cartridges in his right hand. The victim was killed by her husband who then shot himself. (Los Angeles, July 20, 1936) The victim had been estranged from her husband, the suspect, about a month and a half. At approx. 5:50 p.m. the suspect went to the above address and accosted the victim at a check stand where she worked and, after a few preliminary remarks, he shot her with a .32 Cal. H & R revolver in the chest. Suspect then left the store and went to his residence and shot and killed himself. (Los Angeles, January 14, 1954)

The first and third narratives describe homicides involving Hispanic offenders, the second one involves a Black offender, and in the fourth one the offender is White. However, in all of the above accounts, there is a strong indication that themes of male sexual jealousy and proprietariness are fundamental to the underlying processes leading up to the killings. In several

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

October 20, 2003

217

of these cases, the planning and deliberation behind the acts are clearly evident, such as in the last account. Themes of possessiveness are quite strong in those situations where the victim and offender are estranged, a pattern that is consistent with prior research indicating a particularly high risk of homicide victimization among women who are separated from their partners (see Crawford and Gartner 1992; Crawford, Gartner, and Dawson 1997; Daly and Wilson 1988a; Silverman and Kennedy 1993; Silverman and Mukherjee 1987; Wilson and Daly, 1992; Wilson, Daly, and Wright 1993). The assertion of male dominance and control is also evident in those situations in which the homicide is followed by suicide, thus supporting the finding by Polk (1994) that jealousy and loss of control lead men to kill themselves as well as their partners. Most importantly, the above accounts suggest that the nature of masculinity, which includes a desire to possess and control women, appears to override cultural and structural differences across racial and ethnic groups, subsequently producing very similar underlying processes that lead to the killing of an intimate partner. This finding would seem to imply that both structural and cultural theories that have been developed to explain racial differences in homicide offending have missed the boat when it comes to understanding violence committed against intimate partners. It appears that an emphasis on gender differences in intimate partner homicides is warranted more than race-specific explanations of this type of offending.

Confrontational Homicides Confrontational homicides involve challenges of honor and a willingness to engage in violence in responding to these challenges (Polk 1994). They frequently occur in public (Daly and Wilson 1988a; Polk 1994). These types of homicides are typically thought to be carried out exclusively by males in the sense that they are strongly infused with displays of masculinity and responses to threats against one’s masculine identity. What is less clear is the extent to which confrontational homicides occur among males of different racial and ethnic backgrounds. Wolfgang’s (1958:188–9) discussion of the need for males to defend their name and honor focused largely on men from a particular class (those of lower socioeconomic background), but these honor-based motives were attributed to males of “both races.” Wolfgang’s writing aside, what is implied from the existing criminological literature is that compulsive masculinity is a source of violent confrontations among those who are denied access to traditional avenues of expressing masculinity, especially Black males (e.g., Curtis 1974; Hannerz 1969; Oliver 1994). However, the traditional Hispanic culture of “machismo,” with its stress on masculinity, may imply that confrontational homicides should be relatively common among this subgroup as well.

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

218

October 20, 2003

RETHINKING HOMICIDE

Furthermore, ethnographic studies suggest that violent behavior may be linked to cultural norms concerning the Hispanic male role (Horowitz 1983). There is very little in the way of racial and ethnic comparisons of confrontational homicides, probably because of the limited data available for examining the issue. Most homicide data are available in the form of quantitative variables, such as the SHR, the California Criminal Homicides Data File, and the Chicago Homicide Project. An analysis of themes such as masculine competitiveness and honor contests requires not only having access to the homicide narratives from which the quantitative data files are derived, but that these narratives contain a substantial amount of detail regarding the events leading up to the killing. Many descriptive accounts of homicides do not have sufficient information to assess the extent to which these themes underlie a particular lethal encounter. The St. Louis homicides provide the most detail of those available to us. These narratives permit a comparison of confrontational homicides involving Black and White offenders, but no code exists for Hispanic origin in the file. Thus, for the latter group, we have to rely on narratives for Miami. The homicides in St. Louis between 1990 and 1995 overwhelmingly involve both Black offenders and victims. The narrative accounts of these homicides provide numerous examples of confrontational homicides, many of which appear to stem from identity attacks and threats to honor and masculinity. These themes are illustrated in the following accounts: The victim made a teasing remark and the suspect took offense. The victim was visiting friends when the suspect stopped by to pick up an ID card. He was picking it up from a 19-yr.-old female who was an acquaintance of the victim. As the suspect was leaving, the victim told the girl that she was “picking ‘em young these days.” The suspect asked the victim who he was talking to. The suspect said to the victim, “you better check yourself before you wreck yourself.” Shortly after, when the victim left the apartment, the suspect shot the victim, then fled scene. After being shot, the victim ran into the witness’ apartment and collapsed. (St. Louis, case #93053) The victim and suspect were in drug business together, selling from above location. They had a rival, whom they had known many years, who was ‘big man’ in drug trade in area. As the victim’s and suspect’s business grew, they threatened the rival’s territory. Supposedly there had been drive-bys (no injuries) as a result of rivalry. On the night in question, the victim was driving an auto. The suspect tried to get in the car. (Rival had just been in alley, and the suspect had spoken with him). The victim ‘teased’ the suspect by moving the automobile every time the suspect tried to open the car door. According to a secret witness, the victim was laughing, as if it were just a joke. Apparently, the suspect got fed up. He removed a gun from his waistband and fired into the air. This had no effect on the victim who continued laughing and moving

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

October 20, 2003

219

car forward and backward. The suspect then ran around to the driver’s side, took a combat stance, and fired directly at the victim. The victim was hit. The auto moved forward and crashed. The suspect ran up to car and looked inside. He said ‘oh f–k’ and ran. The rival was still in the area, stopped by a traffic light. The suspect ran up to him and got into his car. The rival drove him from the area. The suspect jumped out, circled around, and ‘discovered’ the victim, then hollered for someone to call police. The suspect was taken to headquarters and initially interviewed as a witness. After the secret witness made a statement, the suspect was arrested. (St. Louis, case #92117) The location of the shooting was the home of a mutual friend, who was asleep in a back bedroom at the time of the shooting. The suspects had been needling the victim about being a punk. The victim lost his temper and threatened suspect #1 with a knife. The victim had suspect #1 pinned down with a knife to his throat. Suspect #2 then pulled a gun from his waistband and told the victim to get off of suspect #1. The victim did and threw down the knife. Suspect #1 then took the gun from suspect #2 and shot the victim. Before being hit, the victim begged suspect #1 not to kill him. The suspects fled scene. (St. Louis, case #93004) Suspect #1 thought the victim ‘dissed’ him in front of a woman. Suspect #1 was talking to a woman acquaintance. She told police they were only friends. The victim was her ex-boyfriend, and they had 3 children together. The victim saw them talking and ordered the woman in the house. He told them both that “the talking was over.” Suspect #1 drove away. He returned to the area with suspect #2 as a passenger in his car. When they saw the victim (who was walking with witness #3) suspect #1 asked the victim why he (the victim) had ‘dissed’ him (suspect #1). Suspect #2 then pointed a rifle out the window of the car, and opened fire. The victim was hit. The suspects fled scene in an auto. (St. Louis, case #94114) The victim was co-owner of a bar and was tending bar. Two Latin males and a Latin female were sitting nearby. The victim’s wife was a former girlfriend of one of the Latin males (the offender). The offender demanded a drink from the bartender (the victim). The victim refused to serve the offender. The offender took out a gun and shot the victim once between the eyes while saying, “Yes, you will.” The two males and the female then fled. (Miami, case #89) The victim had been a passenger in a car driven by witness #4. Witness #2 and witness #3 were also passengers. They pulled into parking lot of an all-night market to buy sodas and snacks. The victim and witnesses were leaving the store when the suspects arrived. Witness #1 was a passenger in this vehicle. They went into store while witness #2 was still inside. Suspect #1 wanted to know what witness #2 was staring at. Witness #2 said he wasn’t staring at anything. According to the store clerk, suspect #1 was the aggressor, and was provoking an argument. Clerk couldn’t see that witness #2 had done anything. All then exited store. The suspects asked witness #4 what his friend’s

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

220

October 20, 2003

RETHINKING HOMICIDE

problem was. Witness #4 said he didn’t answer. One of the suspects punched one of the victim’s party. There was a free-for-all fight, with bottles being thrown. The store clerk called the police. He yelled that the police were coming and everyone got into his respective car. But before witness #4 could pull off lot, suspect #1 kicked out the driver’s side window of his car. Witness #4 drove south and suspect #1 gave chase. Suspect #1 rammed witness #4’s car from behind 3 separate times. Each time, witness #4 drove faster in an attempt to get away from suspect #1. His speed reached between 70–80 mph. But suspect #1 kept pace. The last time suspect #1 rammed him, witness #4 lost control of his car. He skidded sideways, hit a truck, a light pole, and a building, and sent a mailbox into a storefront. The car finally came to rest. All occupants were thrown from car. All were injured. Suspect #1 drove from scene. (St. Louis, case #94029)

There are several common themes underlying the lethal situations described in the above accounts. First, all of the homicides involve some type of confrontation between two or more individuals. However, the events underlying these confrontations appear to stem from two different sources. In some cases the confrontation develops when a person becomes the target of teasing by one or more others, as is evident in the first three accounts. When an individual takes offense to this teasing, the situation becomes ripe for a violent encounter as that person attempts to “save face” after being subject to the humiliation of being made fun of by others (see also Luckenbill 1977). The theme underlying the confrontations in the latter three accounts, however, seems more to reflect issues of masculine competitiveness and identity. The homicide described in the fourth account, for example, illustrates the importance of masculine identity attacks in provoking violence. The perpetrator in this case appears to have viewed the behavior of the victim as a threat to his masculinity in the way he was spoken to by the victim, combined with this having taken place in the presence of a woman. Similar themes appear to underlie the homicide described in the fifth narrative. In the final narrative, the violent encounter begins when a “look” by an individual in one of the two parties is perceived as threatening, an implied challenge that is then met with a counterchallenge. Even though this initial encounter did not involve the victim, it need not be the case that the ultimate victim be a participant in the initial confrontation (Polk 1994). Given that much of the discussion in the criminological literature attributes the use of violence as a means for achieving respect and status and demonstrating masculinity to young Black males who are blocked from traditional avenues of attaining these goals, it is interesting to note that the fifth narrative involves a Hispanic offender and victim and the sixth account involves a White offender and victim. While this does not invalidate arguments regarding the significance of issues surrounding masculinity, respect, and

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

October 20, 2003

221

status for understanding lethal violence committed by Black men, it serves as an important reminder that masculine competitiveness, identity, and the need to acquire status and respect may serve as a provocation of violence across a broader spectrum of males. In fact, confrontations involving character contests and identity attacks can be interracial in nature, as occurs in the following account: Victim (Black male) was driving his car (with others as passengers) when he got involved in a traffic dispute with two unknown Latin males. The argument ended but then the two Latin males followed the victim and the argument broke out again. The victim told his companions that he had been insulted and got out of the car with a gun. However, the Latin males also had guns and they shot first, killing the victim. (Miami, case #245)

The second overarching theme in these accounts is that each case involved more than just two participants. All of the accounts indicate that others were present either as bystanders, witnesses, or participants in the crime. The significance of third parties who are supportive of violence has been noted elsewhere (see Felson, Ribner, and Siegel 1984; Luckenbill 1977; Polk 1994). However, the role of third parties need not be in the form of active encouragement for the likelihood of violence to be increased. If violence can be understood as a mechanism for gaining respect or demonstrating masculine traits such as toughness, an audience – even a passive one – may be integral to such scenarios. In other words, character contests or attempts at “face saving” by definition require the presence of others. Several additional points can be made from an analysis of these narratives that diverge from the prevailing image of confrontational homicides in the existing literature. The first concerns the notion that these types of homicides occur largely between either acquaintances or strangers. While it may be the case that the majority fall within these relationship categories, they are not exclusive to them, as illustrated by the second narrative (case #92117) which involves a victim and offender who are cousins. The issue of confrontational homicides occurring among family members is taken up in the following chapter on victim–offender relationships. The second issue concerns the male-on-male nature of these killings. While men may predominate as victims and offenders in confrontational homicides, women play a variety of influential roles in these types of disputes. For example, it is not uncommon for confrontations to develop in response to insults made against a female companion (Oliver 1994; Polk 1994), as is evident in the following narrative: A friend of the victim had made some remarks about the suspect’s pregnant girlfriend, to the effect that “she is only 16 and already has two kids.” This angered the suspect who had pulled a gun on the victim’s friend earlier in the day. Later, the victim and the victim’s friend were standing near the

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

222

October 20, 2003

RETHINKING HOMICIDE

housing project when the suspect and girlfriend and another female walked by. The victim’s friend then said to the suspect “maybe you should have shot me when you had the chance” and he and the victim put their hands behind their backs, acting like they had weapons. The suspect said, “Man, what’s your problem? I thought this thing was over with you.” The suspect then walked a short distance away, then returned, chasing the victim and friend with a gun, firing shots. The victim and his friend both ran into the projects, but the victim was caught because he was asthmatic and couldn’t outrun the suspect. Many people saw the suspect fire two shots at the victim and friend before they entered the projects, but no one saw the fatal shot fired. The murder weapon was a .22 revolver. (St. Louis, case #90085)

Thus, women may be fundamental to confrontational homicides to the extent that verbal attacks against them can be taken on as a threat against the status and reputation of the male himself, thus setting in motion displays of masculinity that ultimately lead to lethal violence. However, women can take on a more active role than simply being the target of the initial identity attack. In some cases they may issue the threat to masculinity themselves, as is evident in the following account: The suspect had been out all night drinking with friends. When he arrived home, the victim met him at the door and “started arguing.” The suspect said he was tired of the victim’s constant arguments with him and tired of her embarrassing him in front of the neighbors, that the latter had lost respect for him. So, on this date, when the victim “began arguing,” the suspect used his revolver to hit the victim in the head and it discharged. The suspect then put a knife in the victim’s hand to make it look like self-defense. (St. Louis, case #93060)

In this case, it was a female who posed the threat to the male’s status and reputation that subsequently provoked the violence resulting in her death. Thus, while identity attacks and the desire for masculine status are still the underlying themes in the events leading up to this killing, this narrative emphasizes that the source of these threats is not limited to other males. Women are also capable of posing threats to masculinity and provoke confrontations which result in lethal violence. The disrespect shown by a woman to a man can also lead to lethal confrontations between two females, as is evident in the following account: The suspect was at the witness’ home when the victim came by. She was sitting in an auto with a man and called to the house from the car, asking the suspect to join her and the man. The suspect told the victim that she (the victim) disrespected the witness by bringing a man to the house and asking her (the suspect) to join them. She told the victim to leave. The witness also hollered out the window for the victim to leave. The suspect and witness then argued about the victim disrespecting the witness. The suspect and the witness then went to an address which was a crack house. Not clear whether

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

October 20, 2003

223

they were looking for the victim or just ran into her. The suspect and victim argued again over same incident. The victim swung at the suspect’s stomach with an object, not clear if it was a knife. The suspect took a knife from her purse and stabbed the victim. (St. Louis, case #90155)

In summary, the use of violence as a way of gaining status or as a means of redressing displays of disrespect appears to be a major underlying theme in many homicide cases. However, these identity attacks are played out in a variety of different contexts, involving both males and females and involving perpetrators and victims from a variety of racial and ethnic backgrounds. By limiting analyses of confrontational homicides to a particular gender or racial/ethnic group, the nature of masculinity and its relation to lethal violence will be obscured or mispecified.

Summary of Findings The use of QCA to examine SHR data involving Black and White offenders from 1976 to 1998, and California data involving Hispanic, White, and Black offenders from 1987 to 1999, combined with a qualitative analysis of homicide narratives from Los Angeles, St. Louis, and Miami, reveals the following findings. First, the results highlight the importance of disaggregating homicides according to the race and ethnicity of the perpetrator. Although the structures of homicide situations among Whites, Blacks, and Hispanics share some commonalities, the situational contexts of killings by members of these different racial and ethnic subgroups often have a qualitatively distinct nature. Second, the degree to which homicide profiles involving a particular racial or ethnic group can be argued to be qualitatively distinct depends to a great extent on how the comparison is structured. In particular, Hispanic homicides appear to be the most qualitatively distinct, but only when compared to non-Hispanic White offenders. Third, the structure of homicide offending among White and Black offenders has remained relatively stable over time. Changes in the nature of offending for Whites are limited to a mild increase in homicide situations occurring in urban locations. Shifts in the nature of offending among Black offenders involve more instrumental killings, southern locations, and younger victims, as well as fewer female offenders over time. Stability over time, however, is the dominant pattern for homicide situations for each racial group. Fourth, our QCA results indicate that there is a high degree of qualitative uniqueness among Hispanic homicide offending when compared to non-Hispanic Whites, which is largely attributable to the predominance of public outdoor locations, young offenders, male-on-male killings, guns, and

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

224

October 20, 2003

RETHINKING HOMICIDE

strangers among Hispanic offenders, as well as a relative lack of homicides committed against family members. Fifth, the structures of homicide situations involving Hispanic and Black offenders are much more similar than in comparisons of Hispanic and White offenders. The qualitative distinctiveness of homicide situations involving Hispanic offenders compared to Black offenders results from a predominance of intraracial killings, expressive and gang motives, and fewer killings in public outdoor locations. Particularly noteworthy is the differing context of stranger homicides among the dominant profiles for the two groups: stranger homicides involving Hispanic offenders are almost always expressively motivated and intraracial, whereas stranger homicides committed by Black offenders are instrumentally motivated and are largely interracial. Sixth, common themes of male sexual jealousy and proprietariness underlie the killing of female intimate partners, regardless of the race or ethnicity of the perpetrator. Confrontational homicides involving threats to honor and masculine identity occur in a wide variety of contexts, including homicides by White, Hispanic, and Black offenders, interracial homicides, and homicides involving females as victims and offenders.

Conclusions and Implications The analyses in this chapter were stimulated by a desire to fill a significant gap in the literature concerning racial differences in homicide situations. As was noted previously, the overwhelming majority of this research has been in the form of aggregate-level studies that examine the impact of various structural correlates on race-specific homicide rates. What little individual-level research exists has been largely bivariate in nature or is based on speculations that have not been empirically validated. Using QCA and detailed homicide accounts, we examine the issue of whether homicide situations involving Black, White, and Hispanic offenders are unique or common as well as whether there has been a change in the nature of offending across time. We also investigate whether two subtypes of homicide – intimate partner killings and confrontational homicides – share similar processes across racial and ethnic groups. We find that uniqueness and commonality equally characterize homicide situations involving White and Black offenders. Profiles of homicide offending by Hispanics are the most unique, at least when compared to non-Hispanic Whites. In contrast, homicide situations involving Hispanic and Black offenders share a number of similarities. These patterns have significant implications for developing explanations of homicide offending. Along the same lines as findings indicating various sets of predictors have a differential impact on rates of White and Black homicide (see Harer and Steffensmeier 1992; Huff-Corzine, Corzine, and Moore 1986; LaFree and

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

RACIAL DIFFERENCES IN HOMICIDE SITUATIONS

October 20, 2003

225

Drass 1996; LaFree, Drass, and O’Day 1992; Messner and Golden 1992; Parker and McCall 1997; Smith 1992), our finding that half of all homicide cases occur within situational contexts that are relatively unique to either racial group emphasizes the possibility that general theories of crime that assume similar correlates across contexts may prove inadequate as explanations of homicide for different racial subgroups. In other words, the dramatic differences in homicide offending between Blacks and Whites may not simply be a matter of the tremendous divergence these groups experience in terms of the degree to which they are exposed to criminogenic social conditions. At the same time, the finding that roughly half of all homicides committed by White and Black offenders occur in situations common to both racial groups, as well as the conclusion that intimate partner homicides and confrontational homicides exhibit processes that underlie killings across all racial and ethnic groups, implies that race-specific theories may not be necessary for certain subtypes of homicide. In Wilson’s (1991) view, cultural adaptions may occur mostly through processes of social isolation in which residents of impoverished urban neighborhoods are denied access to cultural learning through mainstream social networks. The isolation of young Black males from the White economy and networks for obtaining jobs deprives them of access to the traditional avenues of demonstrating their masculinity (Anderson 1999; Gilmore 1990; Messerschmidt 1986; Polk 1994). As an alternative they prove their manhood by promoting a self-image based on a physical toughness. According to Gibbs and Merighi (1994), the world of crime provides young Black youth with three major sources of approval: a ‘masculine’ sociocultural role; a means of obtaining an income; and an avenue of social mobility within their community. It is within this world that prison comes to be seen as a rite of passage, a reputation for outsmarting the police as an asset, and the carrying of a gun as a symbol of manhood and strength. While we do not deny that the certain subgroups within the population face particularly large obstacles to gaining access to the traditional means of establishing a masculine identity, our findings suggest that masculine identity, maintaining honor, and gaining respect and status are significant themes, which underlie confrontational homicides generally, perhaps suggesting a need to reformulate these arguments to account for violence committed by individuals drawn from a particular social class as opposed to a specific racial or ethnic group. Viewed from the perspective of social class, these central themes of masculine identity, respect, and status may be best considered as a distinct expression of lower class life (see also Polk 1994). Furthermore, it is important to emphasize that the harsh conditions of life in urban ghettos may provide a context in which aggression and violence are fundamental to one’s survival, and the maintenance of self-image cannot be taken to mean that these kinds of behaviors are valued by the residents

19:38

P1: JMV 0521832993c08

CB617-Miethe-v1

226

October 20, 2003

RETHINKING HOMICIDE

of such communities. This point was made by Kornhauser (1978) and bears repeating here. She draws an important distinction between observing subcultural norms and a weakening of societal norms. Rather than supporting the use of violence, individuals living in ghettos plagued by high crime rates may simply grow immune to it because it is such a pervasive aspect of life in these areas. Thus, societal norms proscribing violence become attenuated as opposed to rejected (see also Anderson 1978). Although we find support for the hypothesis that Hispanic homicides have a qualitatively distinct character, this conclusion is tempered by the finding that the uniqueness of killings committed by this ethnic group is to a large extent dependent upon the nature of the comparison. In particular, homicide situations involving Hispanic and non-Hispanic Whites demonstrate a substantial amount of uniqueness, but homicide profiles for Hispanic and Black offenders are much more similar, perhaps suggesting that theoretical explanations of homicide offending need to account for the commonality of homicide situations among minority offenders. The finding that stranger killings committed by Hispanic and Black offenders are very different in their underlying structure serves as an important reminder of what is lost by simple bivariate comparisons across racial/ethnic groups. The dominant situational contexts unique to homicides involving Black offenders defy the common image of killings committed by members of this racial group. Even in the 1990s, the major profiles for Black offenders are all intraracial and committed by a lone offender. More than half involve expressive motives, the majority are committed by offenders over the age of 20, and only one of the top ten situations involves strangers. Thus, the public rhetoric concerning interracial, instrumentally motivated homicides committed by groups of young Black males appears to be preoccupied with a relatively rare form of killing. This same argument applies to the focus on juvenile “superpredators” discussed in the previous chapter. Finally, an overriding conclusion from the analysis of homicide narratives in this chapter concerns the need for more detailed accounts of homicide situations involving the full spectrum of racial and ethnic groups. How threats to masculine identity and the desire to achieve status and respect through the use of violence play out in killings committed by different racial/ethnic groups is a fundamental question for criminologists to pursue. To date, sufficiently detailed narratives for a single location that include an adequate number of homicides by White, Black, and Hispanic offenders are clearly lacking, which severely impedes pursuing a valid answer to this question.

19:38

P1: JMT 0521832993c09

CB617-Miethe-v1

November 3, 2003

C H A P T E R

N I N E

Victim–Offender Relationships

. . . Homicide is a dynamic relationship between two or more persons caught up in a life drama where they operate in a direct, interactional relationship. More so than in any other violation of conduct norms, the relationship the victim bears to the offender plays a role in explaining the reasons for such flagrant violation. (Wolfgang, 1958:203)

classifications of homicide according to the victim–offender relationship have a relatively long history. Victim–offender relationships have been variously defined in terms of the nature of social bonds between the parties, their “relational distance,” and the motives that bring and hold the participants together (see Polk 1994; Silverman and Kennedy 1987; Wolfgang 1958). Social relationships between victims and offenders vary in intensity along a continuum, ranging from intimate familial roles (e.g., spouses, lovers) at one extreme, friends and acquaintances (e.g., neighbors, coworkers) as an intermediate category, and strangers at the other end. Most research has employed a simple dichotomy to characterize the victim–offender relationship (e.g., primary vs. secondary relations, intimates vs. nonintimates, strangers vs. nonstrangers). Whatever the particular classification scheme, the major assumption is that subcategories of victim–offender relationships form the basis for fundamentally distinct homicide situations. This chapter explores the nature of variability in homicide situations within and between different types of victim–offender relationships. QCA is used to examine similarities and differences in the structure of homicide situations for three major homicide subgroups: intimates and family members, friends and acquaintances, and strangers. Subsequent analyses examine the extent to which intimate and family homicides have unique and common profiles. Changes in the structure of intimate homicides over 227

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

228

November 3, 2003

RETHINKING HOMICIDE

time are investigated to shed further light on a trend that has captured the attention of criminologists. Similar comparisons are made for acquaintance versus stranger homicides.

Types of Victim–Offender Relationships There are various ways to categorize the victim–offender relationship in homicide cases. In his seminal work, Wolfgang (1958:204) used eleven categories to measure the victim–offender relationship, including (a) close friends, (b) family relationship, (c) acquaintance, (d) stranger, (e) paramour, mistress, prostitute, (f) sexual rival, (g) enemy, (h) paramour of offender’s mate, (i) felon or police officer, (j) innocent bystander, and (k) homosexual partner. However, this level of detail is relatively rare in previous studies. A more commonly used coding scheme defines primary relationships as those involving relatives, persons romantically linked, and friends, whereas secondary relationships include casual acquaintances and strangers (see Bankston, St. Pierre, and Allen 1985; Decker 1996; Parker and Smith 1979; Smith and Parker 1980). Other researchers have focused on the categories of strangers vs. nonstrangers (see Messner and Tardiff 1985; Riedel 1993; Riedel and Przybylski 1993; Sampson 1987). An alternative classification includes the three major categories of intimates/family, friends/acquaintances, and strangers (see Decker 1993; Polk 1984; Riedel 1987; Rojek and Williams 1993; Williams and Flewelling 1987; 1988). Other studies have included categories such as “rival gangs” and “illegal business relations” to represent the victim–offender relationship (see Block and Block 1992). The category of intimate/family has been further subdivided in other studies to contrast violence committed by intimates compared to other relatives, marital vs. nonmarital intimate partners (Rosenfeld 1997; Wilson and Daly 1992) and blood vs. nonblood relatives (Daly and Wilson 1988a).

The Disaggregation of Homicide by Victim–Offender Relationship The disaggregation of homicides by types of victim–offender relationships has salience for both testing theory and public policy (Maxfield 1989; Williams and Flewelling 1988). At the theoretical level, distinguishing between different victim–offender relationships has the potential to enlighten our understanding of how these relationships either offer protection from, or increase the likelihood of, the occurrence of fatal interactions (Decker 1993). The victim–offender relationship is an important variable in research on personal violence in terms of locating the event within the context of

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

VICTIM–OFFENDER RELATIONSHIPS

November 3, 2003

229

social structures. As Loftin et al. (1987) note, roles like husband, wife, friend, lover, and stranger are intricate social relationships that may characterize homicides sharing a distinctive etiology. The social roles of the victim and offender also largely determine responses to violence on both a moral and legal level (Gottfredson and Gottfredson 1988; Lundesgaarde 1997). Victim–offender relationships are also central to discussions of gender differences in homicide offending. In particular, the large majority of research on female homicide offending focuses exclusively on intimate partner homicides, since that is where women’s involvement in homicide is concentrated. Given this emphasis, however, findings indicating differences in the situations and motivations underlying male- and female-perpetrated homicides are for the most part limited to violence among intimate partners (for an exception see Felson and Messner 1998). Several studies have addressed the possibility that structural covariates of homicide rates may differ depending on the victim–offender relationship (Flewelling and Williams 1999; Parker 1989; Parker and Smith 1979; Parker and Toth 1990; Smith and Parker 1980; Williams and Flewelling 1988). Variation in effects across victim–offender relationship categories has been demonstrated for such factors as economic deprivation (Parker and Smith 1979; Smith and Parker 1980), southernness (Parker and Smith 1979; Parker and Toth 1990; Smith and Parker 1980), racial composition (Parker 1989), infant mortality rates (Parker and Toth 1990), and unemployment (Parker and Toth 1990). The overwhelming majority of research examining the utility of disaggregating homicide rates by victim–offender relationship is macrosocial in nature (for an exception see Messner and Savolainen 2001). Although these types of studies make an important contribution to our understanding of the nature of homicide and the development of policies to combat violent crime, of equal importance are individual-level analyses examining differences in homicides across relationship categories. This latter type of research is, however, comparatively rare. Thus, while we have some idea of the extent to which structural predictors of family, acquaintance, and stranger homicides vary, we know much less about the degree to which the structure of homicide situations differs across these relationship categories. Certain perceptions exist regarding differences in homicides across victim–offender relationships. For example, homicides involving family members are thought to be exclusively expressively motivated, whereas stranger homicides are assumed to be instrumental in their motivation (Loftin 1986; Maxfield 1989; Rojek and Williams 1993). However, several studies have challenged this assumption (Block 1987; Decker 1993; 1996; Riedel 1987; Riedel and Zahn 1985; Williams and Flewelling 1988; Zahn and Sagi 1987). Family homicides are also thought to be less likely to involve a firearm than stranger homicides (Decker 1993; Riedel and Przybylski 1993;

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

230

November 3, 2003

RETHINKING HOMICIDE

Riedel and Zahn 1985; Zahn and Sagi 1987), while stranger homicides have a higher involvement of elderly victims, male victims, multiple offenders, and interracial killings (Decker 1993; Maxfield 1989; Silverman and Kennedy 1987).

Trends in the Victim–Offender Relationship over Time Trends in the nature of the victim–offender relationship over time have been documented in previous research. Available accounts of homicide in Colonial America emphasize the killings of family members and close friends, with relatively few slayings by strangers in the context of single victim and single offender situations mentioned in historical narratives of homicides (see Lane 1997; Monkkonen 1991). The western outlaws and gangsters who robbed banks in the early to mid-1900s gave the impression of high levels of violence between strangers, but stranger homicides appear to have been the exceptional case even in these times. Police and court data reveal several trends over the last half century in the prevalence of homicides within particular types of victim–offender relationships. First, homicides by friends and acquaintances are the modal category throughout this century. Second, homicides involving family members have declined over time. Third, stranger homicides have become more common over time, increasing from about one-eighth of the Philadelphia homicides from 1948 to 1952 (Wolfgang, 1958), to about one-fifth of all U.S. homicides in the 1970s, and to slightly over one-fourth of U.S. homicides in the 1990s. A well-noted trend during the past two decades concerns the decline in levels of intimate partner homicides (Block and Christakos 1995; Dugan, Nagin and Rosenfeld 1999; Greenfield et al. 1998; Lund 2001; Rosenfeld 1997). The decrease in rate of intimate partner killings, however, has not occurred equally among male and female intimate partner offenders, but rather is largely due to a reduction in the rate of women killing their male intimate partners (see Browne and Williams 1989; Greenfield et al. 1998). Trends in female-perpetrated intimate partner homicides also vary across married versus nonmarried couples (Browne and Williams 1993), and have occurred disproportionately among Blacks (Rosenfeld 1997). Several theories and explanations have been offered to account for the relative prevalence of homicides by different types of victim–offender relationships and changes in them over time. For example, the level and rise in stranger homicide in the last three decades are consistent with several theoretical perspectives. Under the theory of displaced aggression, frustrations that emanate from work and family pressures are released in violent forms on any available target. The “legitimate” sources of frustration may have greater immunity from victimization (because they are the “boss,” or the guardians of the offender’s children) so strangers may be the most

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

VICTIM–OFFENDER RELATIONSHIPS

November 3, 2003

231

vulnerable targets. The phenomenon of “road rage” against strangers is often attributed to similar processes. In contrast, the routine activities or lifestyle theory of victimization (see Cohen and Felson 1979; Hindelang, Gottfredson, and Garofalo 1978; Miethe and Meier 1994) attributes the rise in stranger homicide to changes in everyday life that increase individuals’ exposure to motivated offenders, increase their attractiveness as crime victims, and reduce their protection or guardianship. From this perspective, post–World War II increases in public activities and the rise in primary households (e.g., individuals living alone) are the types of macrosocial changes that increase the likelihood of contact with strangers and subsequently the criminal opportunity structure for stranger homicides. Homicides by intimates and family members have been attributed to several factors. These include the greater intensity and volatility of these interpersonal relationships, the greater importance attributed to violations of trust, previous history of conflict and abuses, more physical opportunities for conflict situations to arise among intimate partners and family members because of the higher frequency of social interactions, and the greater lethality of domestic violence as a result of “overkill” (i.e., multiple stab wounds, bullet wounds, or severe beatings) that derives from uncontrolled anger and rage (see Browne, Williams, and Dutton 1999; Daly and Wilson 1982; Gillis 1986; Goode 1969; Silverman and Kennedy 1993). In an effort to understand the declining rate of intimate partner homicide, previous research has examined the association between rates of intimate partner homicide and domestic violence resources (e.g., shelters, hotlines, protective legislation). These studies are based on the assumption that increased awareness and education about domestic violence and changing national, state, and local policy on intimate violence may protect women from abusive males. Declining rates of marriage have also been found to be a contributing factor in the recent reduction in intimate partner homicide (see Rosenfeld 1997), while the rise of separation and divorce has been associated with a decline in domestic homicide historically (see Gillis 1996). The availability of outside interventions for abused women is further associated with a decline in female-perpetrated homicide (Browne and Williams 1989; Dugan, Nagin, and Rosenfeld 1999). However, by providing new criminal opportunities (e.g., greater contact with nonintimate partners), social changes in these protective services may also be linked to other types of homicide situations involving female offenders (e.g., a rise in nonintimate slayings).

Data and Measurement This analysis uses SHR data from 1976 through 1998 to examine unique and common structures of homicide situations for different victim–offender relationships. The initial analysis includes three broadly defined relationship

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

232

November 3, 2003

RETHINKING HOMICIDE

categories: intimate partners and other family members, friends and acquaintances, and strangers.1 This is followed by an investigation of the extent to which homicide situations involving intimate partners differ from killings of other family members. Whether there has been a change in the structure of intimate partner homicide situations over time is also examined. Finally, we compare acquaintance to stranger homicides. For the current analyses, homicide situations are defined on the basis of the following offender, victim, and offense characteristics. Offender attributes include gender, race (White vs. Black vs. other), and age (under 20, 20–29, and 30 or older). Similar variables are included as victim characteristics. Variables measuring whether the homicides are intraracial, intra-age, and intrasex are also included. Offense elements include weapon (gun, knife, other), motive (instrumental, expressive, other), number of offenders (single, multiple), number of victims (single, multiple), and location (large city over 100,000, less urban areas).

Results of Qualitative Comparative Analysis (QCA) Our QCA analysis begins with an examination of whether the major relationship categories used in homicide research (intimate partner/family, friend/acquaintance, and stranger) are unique in terms of combinations of offender, victim, and offense attributes. Following the strategy used in earlier chapters, the term “dominant structures” will be used to describe each distinct combination of individual and offense characteristics satisfying the minimum frequency rule.2

Intimate/Family, Friend/Acquaintance, and Stranger Homicides A comparison of intimate/family to nonintimate/family homicides reveals that of the 524 dominant structures of homicide situations in the SHR between the years 1976 and 1998, 30% are relatively unique to intimate/family homicides, 49% are relatively unique to nonintimate/family homicides, 1

2

The following definitions are used to classify victim–offender relationships: intimate partners (husband, wife, common-law husband, common-law wife, ex-husband, ex-wife, boyfriend, girlfriend, homosexual relationship); other family (mother, father, daughter, son, sister, brother, stepmother, stepfather, stepdaughter, stepson, in-law, other family); friend/ acquaintance (friend, neighbor, employee, employer, acquaintance, other known to victim); and stranger. For this analysis, a minimum frequency of 5 homicides per year (producing an overall minimum of 115 cases) is used for each configuration. Combinations of attributes are defined as “unique” structures in this analysis when the observed cell proportions of the victim–offender relationships exceed the respective unconditional marginal distributions for these categories by 10% or more. “Common” configurations are those in which the marginal distributions are not exceeded by 10%.

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

VICTIM–OFFENDER RELATIONSHIPS

November 3, 2003

233

Figure 9.1. Percentage of Cases in Unique and Common Profiles by Relationship Type

and 21% are common to both. The 156 relatively unique profiles for intimate/family homicides describe roughly one-quarter of all homicides during these years, while the relatively unique profiles for nonintimate/family homicides account for about 60% of homicide cases (see Figure 9.1). The importance of disaggregating homicides by victim–offender relationship is emphasized by the finding that only 16% of cases occur in structural contexts common to both intimate/family and nonintimate/family homicides. An examination of the ten most dominant profiles for intimate/family homicides reveals several interesting patterns (see Table 9.1). Nine out of ten of these profiles are expressive homicides by older offenders against someone of the same age and race but opposite sex committed alone with a gun against a lone victim. They are equally distributed among male and female, and Black and White offenders. When these homicides occur in nonurban locations, they almost always involve a White offender. These ten dominant profiles account for over one-fifth of all intimate/family homicides. When acquaintances are compared to other victim–offender relationships we find that one-fourth of the dominant structures of homicide situations are relatively unique to acquaintances, 40% are relatively unique to nonacquaintance homicides, and the remaining 35% are common to both. Forty-two percent of homicide cases are accounted for by the relatively unique homicide situations for acquaintances.

15:48

234

owhite oblack oblack owhite owhite owhite oblack oblack oblack owhite

gun gun gun gun gun gun knife gun gun gun

omale omale omale omale omale omale omale

off3098 off3098 off3098 off3098 off3098 off3098 off3098 off2029 off3098 off3098

sameage sameage sameage sameage sameage sameage sameage sameage sameage sameage

diffsex diffsex diffsex diffsex diffsex diffsex diffsex diffsex diffsex diffsex

samerace samerace samerace samerace samerace samerace samerace samerace samerace samerace

singoff singoff singoff singoff singoff singoff singoff singoff singoff singoff

singvic singvic singvic singvic singvic singvic singvic singvic singvic singvic

nonurb urban urban urban nonurb nonurb urban urban nonurb urban

3,897 of 4,540 2,580 of 3,838 2,314 of 3,602 2,586 of 3,224 2,286 of 2,697 2,173 of 2,574 1,587 of 2,306 1,305 of 2,234 1,541 of 2,114 1,359 of 1,783

oblack oblack owhite owhite owhite oblack oblack

gun gun gun gun gun gun gun

express express express express express express express

off3098 off2029 off3098 off3098 off2029 offlt20 off3098

sameage sameage sameage sameage sameage sameage sameage

samesex samesex samesex samesex samesex samesex samesex

samerace samerace samerace samerace samerace samerace samerace

singoff singoff singoff singoff singoff singoff singoff

singvic singvic singvic singvic singvic singvic singvic

urban urban urban nonurb urban urban nonurb

10,546 of 14,810 10,167 of 14,016 6,424 of 9,872 6,124 of 8,941 4,996 of 7,701 4,399 of 6,094 4,272 of 5,731

Profile N s

(71%) (73%) (65%) (69%) (65%) (72%) (75%)

%

(86%) (67%) (64%) (80%) (85%) (84%) (69%) (58%) (73%) (76%)

%

CB617-Miethe-v1

1. 2. 3. 4. 5. 6. 7.

express express express express othmotv express express express express express

Profile N s

Percentage of all intimate partner/family homicides explained by these dominant profiles: 22.2% (21,628 of 97,577)

omale omale ofemale omale omale ofemale ofemale ofemale omale ofemale

Panel B: Friend/Acquaintance:

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Panel A: Intimate/Family:

Table 9.1. Qualitative Comparative Analysis of Homicide Situations Unique to Intimate/Other Family, Friend/Acquaintance, and Stranger Homicides, 1976–98

P1: JMT 0521832993c09 November 3, 2003 15:48

omale omale omale omale omale omale omale omale omale omale

express express express

off2029 off2029 off3098

sameage sameage sameage

samesex samesex samesex

samerace samerace samerace

singoff singoff singoff

singvic singvic singvic

nonurb nonurb urban

3,699 of 5,377 3,443 of 4,475 3,113 of 4,173

235

gun gun gun gun gun gun gun gun gun gun

instrum instrum instrum instrum instrum instrum instrum instrum instrum instrum

off2029 off3098 offlt20 off3098 off2029 off3098 off2029 offlt20 off2029 off2029

sameage sameage sameage sameage sameage sameage sameage diffage sameage sameage

samesex samesex samesex samesex samesex samesex samesex samesex samesex samesex

samerace samerace samerace samerace samerace samerace samerace samerace samerace diffrace

singoff singoff singoff singoff singoff singoff singoff singoff multoff singoff

singvic singvic singvic singvic singvic singvic singvic singvic singvic singvic

urban urban urban urban urban urban nonurb urban urban urban

3,258 of 7,752 2,395 of 5,698 1,713 of 4,092 1,997 of 4,005 1,894 of 3,603 972 of 2,209 947 of 2,005 969 of 1,878 985 of 1,861 1,215 of 1,707

Percentage of all stranger homicides explained by these dominant profiles: 19.2% (16,345 of 85,220)

oblack oblack oblack owhite owhite owhite owhite oblack oblack oblack

Profile Ns

(42%) (42%) (42%) (50%) (53%) (44%) (47%) (52%) (53%) (71%)

%

(69%) (77%) (75%)

CB617-Miethe-v1

MODEL: Omale + Oblk + Oothrace + Olt20 + O2029 + =Age + =Race + =Sex + Express + Othmotv + Gun + Knife + Multvic + Multoff + Urban

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

gun gun knife

Percentage of all friend/acquaintance homicides explained by these dominant profiles: 30.1% (57,183 of 189,878)

owhite oblack oblack

Panel C: Stranger:

8. omale 9. omale 10. omale

P1: JMT 0521832993c09 November 3, 2003 15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

236

November 3, 2003

RETHINKING HOMICIDE

Figure 9.2. Cases within the Dominant Unique Profiles for Acquaintance Homicide

An examination of Figure 9.2 indicates that there are essentially two major dominant profiles, which contain over 10,000 homicide cases each, with the remaining eight profiles accounting for considerably fewer cases (i.e., acquaintance homicides demonstrate substantial clustering). These two overwhelmingly dominant profiles involve expressively motivated homicides committed by a lone Black male with a firearm against someone the same age, race, and sex, in an urban location (see also Panel B of Table 9.1). Neither profile involves youth homicide offenders. In fact, there is only one major configuration involving offenders under the age of 20. Thus, although the prevailing image of urban homicides is one of young minority males killing with guns, our findings indicate that the “youth” aspect of this image is not supported. Roughly 30% of acquaintance homicides are captured in these 10 dominant profiles. When comparing stranger and nonstranger homicides, we find that over one-third of the dominant homicide structures are relatively unique to stranger homicides, accounting for over one-fourth of the cases (see Figure 9.1). Another one-third of the profiles are relatively unique to nonstranger homicides, and these profiles contain one-third of the cases. Roughly onethird of homicide situations are common to both, accounting for 40% of the cases. Table 9.1 displays the profiles of the ten most dominant configurations, which account for 19.2% of all stranger homicides. All ten of these profiles are instrumental firearm homicides committed by a male against another male. All but one involve individuals of the same age and race and are committed by one offender in an urban location. The only homicide situation involving a victim and offender who are of different ages is a killing by a youth offender. Thus, of the major relationship categories, we find that stranger homicides are the most unique in terms of having the highest percentage of

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

VICTIM–OFFENDER RELATIONSHIPS

November 3, 2003

237

homicide situations which are unique to that victim–offender relationship category. This suggests that situations involving the killing of strangers exhibit the most diversity in terms of occurring across the greatest range of contexts. This pattern may be a reflection of a wider array of circumstances (i.e., both instrumental and expressive motives) and participants (i.e., random targets and innocent bystanders as well as those engaged in a criminal lifestyle) underlying the situations which lead to the death of a stranger. In contrast, acquaintance homicides show the greatest situational clustering in the sense that there is a higher percentage of cases falling within fewer unique configurations for this subgroup. While only one-quarter of the dominant homicide structures are unique to homicides involving acquaintances, indicating that they are less diverse than either intimate/family or stranger homicides, 42% of homicides nationally that occurred between 1976 and 1998 fall within these 131 homicide situations. Thus, while a broader array of unique situations lead to stranger homicides, these situations are more idiosyncratic in nature in that they account for fewer cases than uniquely acquaintance profiles.

Intimate Partner versus Other Family Homicides The categorization of victim–offender relationships into the three major categories of intimate/family, friend/acquaintance, and stranger is fairly common practice in studies on homicide (see, for example, Decker 1993; Polk 1994; Riedel 1987; Williams and Flewelling 1987, 1988). However, some researchers have raised concerns with respect to the inclusion of intimate partners and other family members within a single category, arguing instead that the two may differ in important ways (Flewelling and Williams 1999). It is possible to examine this question empirically by evaluating the extent to which the structures of homicide situations among these two different groups are similar or different using QCA. The findings of this analysis are presented in Table 9.2.3 The QCA results, which are displayed in Table 9.2, indicate that the majority of configurations are either relatively unique to intimate partner (34%) or other family homicides (50%). Over one-half (56%) of all intimate/family homicides occur in situations unique to intimate partners and over one-third (37%) occur in situations relatively unique to other family members. The small proportion of profiles common to both intimate partner and other family homicides (16% of all profiles) account for only 6.3% of all intimate partner/family homicides. 3

We use a minimum frequency of two per year since the ratio of intimate/family homicides to all other homicides is roughly 2:5.

15:48

238 =Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex Omale Omale Omale Ofem Omale Ofem Ofem Omale Omale Ofem

Expr Expr Expr Expr Othmot Expr Expr Expr Othmot Expr

Gun Gun Gun Gun Gun Gun Knife Gun Gun Gun

O3098 O3098 O3098 O3098 O3098 O3098 O3098 O3098 O3098 O3098

Owht Owht Oblk Oblk Owht Owht Oblk Oblk Owht Owht

Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff

Singvic Singvic Singvic Singvic Singvic Singvic Singvic Singvic Singvic Singvic

NonUrb Urban Urban Urban NonUrb NonUrb Urban NonUrb Urban Urban

3,701 of 3,897 2,432 of 2,586 2,416 of 2,580 2,060 of 2,314 2,130 of 2,286 2,059 of 2,173 1,458 of 1,587 1,457 of 1,541 1,244 of 1,363 1,241 of 1,359

Profile N s

118,628 97,119 54,687 36,305 6,127

Percentage of all intimate partner homicides explained by these dominant profiles: 15.0% (2,993 of 19,969)

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race

(10.6%) 34.0% 50.0% 16.0%

3,541 374 127 187 60

95% 94% 94% 89% 93% 95% 92% 95% 91% 91%

% Intimate∗

(81.9%) 56.3% 37.4% 6.3%

%

%

N

N

Homicide Cases

Configurations

CB617-Miethe-v1

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Intimate Partner Homicides:

Panel B: Unique Profiles within the Most Dominant Homicide Situations

Number of Situations Observed Number of Homicide Situations after N > 46 Cell Frequency Rule Situations Unique to Intimate Partner Homicides Situations Unique to Other Family Homicides Contradictions (Common Situations/Cases) for Both Groups

Panel A: Configurations and Cases (using 10% Difference Rule)

Table 9.2. Qualitative Comparative Analysis of Homicide Situations Unique to Intimate Partner and Other Family Homicides, 1976–98

P1: JMT 0521832993c09 November 3, 2003 15:48

239

=Age =Age =Age =Age =Age =Age =Age =Age =Age =Age

=Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex =Sex Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale

Expr Expr Expr Expr Expr Expr Expr Expr Othmot Expr

Gun Gun Gun Gun Gun Gun Gun Knife Gun Knife

O3098 O3098 O3098 O2029 O3098 O2029 O2029 O3098 O3098 O < 20

Oblk Owht Owht Oblk Oblk Owht Owht Oblk Oblk Oblk

Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff Singoff

Singvic Singvic Singvic Singvic Singvic Singvic Singvic Singvic Singvic Singvic

Urban NonUrb Urban Urban NonUrb Urban NonUrb Urban Urban Urban

1,346 of 2,155 1,516 of 1,794 1,035 of 1,650 971 of 1,444 828 of 963 664 of 941 791 of 895 411 of 583 352 of 570 356 of 524

Percentage of all other family homicides explained by these dominant profiles: 35.8% (978 of 2,735)

=Race =Race =Race =Race =Race =Race =Race =Race =Race =Race

Profile N s

63% 85% 63% 67% 86% 71% 88% 71% 62% 68%

% Family∗∗

MODEL: Intimate= Omale + Oblk + Oothrace + Olt20 + O2029 + =Age + =Race + =Sex + Express + Othmotv + Gun + Knife + Multvic + Multoff + Urban Note: ∗ This column represents the percentage of homicides in the particular profile that involve intimate partners. ∗∗ This column represents the percentage of homicides in the particular profile that involve other family members.

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Other Family Homicides:

P1: JMT 0521832993c09 CB617-Miethe-v1 November 3, 2003 15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

240

November 3, 2003

RETHINKING HOMICIDE

These findings provide strong support for the argument that intimate partner and other family homicides should be examined separately. That the nature of homicides involving these two groups differs is also evident from an examination of the ten most dominant profiles for each type of homicide (see Panel B of Table 9.2). The most dominant situations involving the killing of an intimate partner are homicides where the participants are the same age and race but the opposite sex. They are predominantly expressively motivated and carried out with a gun. They all involve offenders 30 years of age and over. In contrast, the dominant profiles for killings of other family members all involve homicides of males by males. They also consist of offenders across all age ranges.

Trends in Intimate Partner Homicides over Time An issue that has garnered considerable attention from criminologists concerns a decline in intimate partner homicides over the past several decades (Browne and Williams 1993; Dugan, Nagin, and Rosenfeld, 1999; Rosenfeld 1997). The decline has occurred disproportionately among male victims (Browne, Williams, and Dutton 1999; Greenfield et al. 1998; Rosenfeld 1997), and couples who are legally married (Browne, Williams, and Dutton 1999; Rosenfeld 1997). In this section, we use QCA to examine change and stability in the structure of intimate partner homicides over time by comparing intimate partner homicides in the 1970s vs. the 1990s. For the purpose of this analysis, homicide structures are defined on the basis of attributes that have been hypothesized or demonstrated empirically to have changed in recent decades: victim gender, victim race, victim age, offender gender, offender race, offender age, weapon, and whether the individuals involved were spouses, common-law, or boyfriend/girlfriend. On the basis of this definition of homicide situations, we find that over one-third (38%) of these configurations are relatively unique to the 1990s (Figure 9.3). These profiles account for approximately one-quarter (24%) of the homicide cases during this decade. Another 21% of configurations are relatively unique to the 1970s, and these account for another one-quarter of the cases. While 41% of profiles are common to both decades, these comprise more than half (51%) of the intimate partner homicide cases during these decades. Thus, the nature of intimate partner homicides between the 1970s and 1990s is characterized about equally by stability and change. An examination of the ten most dominant profiles for the 1970s and 1990s reveals some rather dramatic differences in the structure of intimate partner homicide situations across decades (see Table 9.3). The configurations in the 1970s are dominated by cases involving older, Black offenders using

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

November 3, 2003

VICTIM–OFFENDER RELATIONSHIPS

241

Figure 9.3. Stability and Change in Intimate Partner Homicide Situations

guns to kill older Black victims to whom they are married. The offenders are almost equally distributed across gender. The unique profiles in the 1990s diverge from those in the 1970s considerably. In this decade they are all killings of girlfriends by boyfriends, with one exception that involves a male homosexual couple. These patterns are consistent with prior research indicating a shift away from female-perpetrated spousal homicides within the category of intimate partner homicides in recent decades (see Dugan, Nagin, and Rosenfeld 1999; Rosenfeld 1997).

Acquaintance versus Stranger Homicides The distinction between acquaintance and stranger homicides is an interesting one. Both involve socially distant relationships, and the fleeting nature of some types of acquaintance relationships may produce a number of commonalities between homicides involving these two groups. Prior research suggests that they share some characteristics, such as being overwhelmingly committed by males (Decker 1993; Silverman and Kennedy 1987; 1993; Zahn and Sagi 1987) and with guns (Decker 1993). This section investigates empirically to what extent these two types of homicide are unique. The structure of homicide situations is defined as follows for this analysis: intrasex, intrarace, intra-age, gun, knife, multiple offender, multiple victim, urban. Given the perception that stranger homicides are predominantly felony-related killings committed by young Black males, these attributes were also used included in the definition of the structure: instrumental motivation, male offender, Black offender, and offender less than 20.

15:48

Ofemale Ofemale Omale Ofemale Omale Omale Ofemale Omale Ofemale Ofemale

Owhite Oblack Owhite Oblack Oblack Oblack Oblack Oblack Oblack Oblack

O3098 O3098 O2029 O3098 O2029 O3098 O2029 O3098 O3098 O3098

Gun Gun Gun Gun Gun Gun Gun Gun Gun Knife

Spouse Spouse Spouse Spouse Spouse Spouse Spouse Commonlaw Commonlaw Commonlaw

242 Omale Omale Omale Omale Omale Omale Omale Omale Omale Omale

O3098 O2029 O3098 O3098 O3098 O2029 O3098 O2029 O3098 O2029

Gun Gun Knife Othwp Othwp Gun Knife Gun Gun Gun

Boy/Girlfriend Boy/Girlfriend Boy/Girlfriend Boy/Girlfriend Boy/Girlfriend Boy/Girlfriend Boy/Girlfriend Boy/Girlfriend Boy/Girlfriend Boy/Girlfriend

MODEL: Yr70v90= Vicmale + Vicwhite + Vicblack + Vic < 20 + Vic2029 + Offmale + Offwhite + Offblack + Off < 20 + Off2029 + Gun + Knife + Spouse + Commonlaw Note: ∗ This column represents the percentage of intimate partner homicides in the particular profile that occurred in the 1970s. ∗∗ This column represents the percentage of intimate partner homicides in the particular profile that occurred in the 1990s.

81% 78% 83% 83% 77% 77% 85% 77% 89% 83%

% 1990s∗∗

51% 63% 48% 51% 53% 46% 67% 56% 69% 49%

% 1970s∗

CB617-Miethe-v1

Percentage of all intimate partner homicides in the 1990s explained by these dominant profiles: 10.1% (2,334 of 23,104)

V3098 V2029 V3098 V3098 V3098 V2029 V3098 V < 20 V2029 V < 20

567 of 704 341 of 436 272 of 328 271 of 326 195 of 252 176 of 228 170 of 201 106 of 138 123 of 138 113 of 137

Vwhite Vblack Vblack Vwhite Vblack Vwhite Vwhite Vblack Vblack Vwhite

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Vfemale Vfemale Vfemale Vfemale Vfemale Vfemale Vfemale Vfemale Vmale Vfemale

Profile N ’s

Intimate Partner Homicides in the 1990s: Owhite Oblack Oblack Owhite Oblack Owhite Owhite Oblack Oblack Owhite

Percentage of all intimate partner homicides in the 1970s explained by these dominant profiles: 24.4% (3,019 of 12,397)

V3098 V3098 V2029 V3098 V2029 V2029 V2029 V3098 V3098 V3098

891 of 1764 618 of 980 273 of 567 208 of 404 209 of 397 161 of 353 218 of 325 146 of 263 174 of 254 121 of 245

Vwhite Vblack Vwhite Vblack Vblack Vblack Vblack Vblack Vblack Vblack

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Vmale Vmale Vfemale Vmale Vfemale Vfemale Vmale Vfemale Vmale Vmale

Profile N s

Intimate Partner Homicides in the 1970s:

Unique Profiles within the Most Dominant Homicide Situations

Table 9.3. Qualitative Comparative Analysis of Stability and Change in Intimate Partner Homicides

P1: JMT 0521832993c09 November 3, 2003 15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

November 3, 2003

243

VICTIM–OFFENDER RELATIONSHIPS

50 40

38.4

40.7 36.4 33.3

28.3

22.9

30 20 10 0 Configurations

Cases

Unique to Acquaintances Unique to Strangers Common to Both Groups

Figure 9.4. Unique and Common Profiles for Acquaintance vs. Stranger Homicides

The QCA results support the notion that the natures of acquaintance and stranger homicides differ. Of the 509 dominant structures of acquaintance and stranger homicides found in the data,4 23% of them are relatively unique to acquaintances, accounting for 38% of cases (Figure 9.4). Another 41% of the profiles are relatively unique to strangers, which comprise 28% of cases. The remaining 36% of profiles are common to both strangers and acquaintances, accounting for one-third of the cases. The strongest attributes distinguishing homicides relatively unique to strangers across all configurations are an instrumental motivation, firearms, interracial, and male offenders, which fit the common conception of stranger killings. Stranger homicides, however, are no more likely than acquaintance homicides to involve a Black offender or a victim and offender who are of the same age or gender. A minimization algorithm was applied to the 50 most dominant profiles for acquaintance and stranger homicides respectively. The algorithm effectively eliminates unnecessary variables for each configuration, producing a final model of the full set of combinations that uniquely produce each outcome. The results from using this minimization algorithm are displayed in Table 9.4. Acquaintance homicides reduce to a set of 24 combinations of between 8 and 12 attributes, while stranger homicides reduce to 18 combinations of between 9 and 11 attributes. This minimized solution demonstrates that the situational context of acquaintance homicides can be reduced to 4

We use a minimum frequency of 4 cases per year for this analysis because the ratio of acquaintance/stranger homicides to all homicides is roughly 4:5.

15:48

SAMEAGE SAMERACE OFFBLACK offlt20 instrumm gun KNIFE multvic multoff+ SAMEAGE SAMERACE OFFBLACK offlt20 instrumm GUN knife multvic multoff+> SAMEAGE SAMERACE OFFMALE OFFBLACK offlt20 instrumm gun multvic multoff+ SAMEAGE SAMERACE OFFMALE OFFBLACK offlt20 instrumm knife multvic multoff+ samesex SAMEAGE SAMERACE offlt20 instrumm gun KNIFE multvic multoff+ samesex SAMEAGE SAMERACE OFFMALE offlt20 instrumm gun multvic multoff+ SAMEAGE SAMERACE OFFMALE offlt20 instrumm gun multvic multoff urban+ SAMEAGE SAMERACE OFFMALE offlt20 instrumm knife multvic multoff urban+ SAMEAGE SAMERACE OFFMALE OFFBLACK offlt20 instrumm GUN knife multoff URBAN+ SAMESEX SAMEAGE SAMERACE OFFMALE OFFBLACK instrumm gun multvic multoff URBAN+ SAMESEX SAMEAGE SAMERACE OFFMALE OFFBLACK instrumm gun KNIFE multvic multoff+ SAMESEX SAMERACE OFFMALE OFFBLACK instrumm GUN knife multvic multoff urban+ SAMESEX SAMEAGE SAMERACE OFFMALE instrumm GUN knife multvic multoff urban+ samesex SAMEAGE SAMERACE OFFBLACK offlt20 instrumm gun multvic multoff URBAN+ samesex SAMEAGE SAMERACE OFFBLACK offlt20 instrumm knife multvic multoff URBAN+ SAMEAGE SAMERACE OFFMALE offblack offlt20 instrumm GUN knife multoff urban+ SAMESEX SAMEAGE SAMERACE OFFMALE offblack instrumm knife multvic multoff urban+ samesex SAMEAGE SAMERACE offlt20 instrumm GUN knife multvic multoff urban+ SAMESEX SAMERACE OFFMALE OFFBLACK OFFLT20 instrumm gun KNIFE multvic multoff URBAN+ SAMESEX SAMEAGE SAMERACE OFFMALE OFFBLACK offlt20 instrumm gun KNIFE multvic URBAN+ SAMESEX SAMEAGE SAMERACE OFFMALE offblack offlt20 instrumm GUN knife MULTVIC urban+ SAMESEX SAMEAGE SAMERACE OFFMALE OFFBLACK offlt20 instrumm GUN knife multvic urban+ SAMESEX SAMEAGE OFFMALE offblack offlt20 instrumm gun KNIFE multvic multoff urban+ samesex SAMEAGE samerace offmale offblack offlt20 instrumm GUN knife multvic multoff URBAN

Acquaintance Homicides

Table 9.4. Minimized Solution of the Most Prevalent Profiles of Uniquely Acquaintance and Uniquely Stranger Homicides, 1976–98

P1: JMT 0521832993c09 CB617-Miethe-v1 November 3, 2003

244

15:48

CB617-Miethe-v1

245

Notes: Variable names in capitals denote the presence of the attribute; lower case denotes the absence of the attribute. “+” in Boolean algebra translates into “or”.

SAMESEX OFFMALE OFFBLACK OFFLT20 INSTRUMM GUN knife multvic URBAN+ SAMESEX SAMERACE SAMEAGE OFFMALE INSTRUMM GUN knife multvic URBAN+ SAMERACE SAMEAGE OFFMALE offblack offlt20 INSTRUMM gun multvic multoff+ SAMERACE SAMEAGE OFFMALE offblack offlt20 INSTRUMM knife multvic multoff+ SAMESEX SAMERACE OFFMALE OFFBLACK INSTRUMM GUN knife multvic multoff URBAN+ SAMESEX SAMERACE SAMEAGE OFFMALE offblack offlt20 INSTRUMM GUN knife URBAN+ SAMESEX SAMEAGE OFFMALE offlt20 INSTRUMM GUN knife multvic multoff URBAN+ SAMESEX SAMERACE SAMEAGE OFFMALE offblack offlt20 INSTRUMM gun multvic URBAN+ SAMESEX SAMERACE OFFMALE offblack OFFLT20 INSTRUMM GUN knife multvic multoff+ SAMESEX SAMERACE SAMEAGE OFFMALE offblack offlt20 INSTRUMM GUN knife multvic+ SAMESEX SAMERACE SAMEAGE OFFMALE offblack offlt20 INSTRUMM GUN knife multoff+ samesex SAMERACE SAMEAGE offlt20 INSTRUMM GUN knife multvic multoff URBAN+ samerace SAMEAGE OFFMALE OFFBLACK offlt20 INSTRUMM knife multvic multoff URBAN+ SAMESEX samerace SAMEAGE OFFMALE OFFBLACK offlt20 GUN knife multvic MULTOFF URBAN+ SAMESEX samerace SAMEAGE OFFMALE OFFBLACK OFFLT20 GUN knife multvic multoff URBAN+ SAMESEX SAMERACE sameage OFFMALE offblack OFFLT20 INSTRUMM gun multvic multoff URBAN+ SAMESEX samerace SAMEAGE OFFMALE OFFBLACK offlt20 INSTRUMM gun multvic multoff URBAN+ SAMESEX samerace SAMEAGE OFFMALE OFFBLACK offlt20 INSTRUMM GUN knife multvic multoff

Stranger Homicides

P1: JMT 0521832993c09 November 3, 2003 15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

246

November 3, 2003

RETHINKING HOMICIDE

a set of profiles that are dominated by lone male Black offenders 20 years of age and older who kill one victim of the same age and race in an expressively motivated act. While many of these unique configurations involve guns, knives are also very common. The utility of looking at combinations of attributes is highlighted by an examination of the six acquaintance profiles that involve a victim and offender who are the opposite sex. While all six profiles involve expressive homicides by offenders 20 and older, they also show considerable diversity in their other attributes. Some involve guns, others knives, some occur in urban locations while others do not, some involve male offenders whereas others are committed by females, some are intraracial while others are interracial. Yet they are different paths (i.e., combinations of attributes) that lead to the same outcome (i.e., an acquaintance homicide). Any useful explanation of homicide offending must be able to account for this wide array of situations in which intersex homicides among acquaintances occur. In contrast to acquaintance homicides, stranger homicides reduce to a set of profiles that involve killings that are predominantly instrumentally motivated between individuals of the same sex, race, and age using guns in urban locations. In the four profiles for stranger homicides that involve a victim and offender of a different race, all are committed by a Black male offender. Only four of the profiles contain homicides committed by individuals under the age of 20. Only one profile entails killings by multiple offenders, and this is not one of the four with a young offender. For the most part, the image of groups of young, Black males randomly attacking White victims is not consistent with the picture of stranger homicides provided by these profiles.

Narrative Accounts of Family and Stranger Homicides While QCA allows us to examine the structure of homicide situations for various types of victim–offender relationships, an inspection of homicide narratives provides some insight into the process underlying these killings. In this section we examine the episodic transactions for family and stranger homicides. Analyses of narrative accounts of intimate partner and acquaintance homicides have been developed in other chapters (Chapters 6 and 8 respectively) and will not be repeated here.

Family Homicides Using QCA we identified major unique profiles of family homicides as being committed by older males against other males in expressively motivated killings with guns. The following narratives largely fit this profile, but as will

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

VICTIM–OFFENDER RELATIONSHIPS

November 3, 2003

247

become evident, they demonstrate that a wide variety of situational processes can underlie this structural context. Suspect claimed the victim had been molesting him since he (suspect) was 13. Suspect had recently told several friends that the victim had been molesting him sexually since he was 13 (until 12, suspect lived in foster homes, reason not given). An adopted daughter also claimed that the victim had molested her and pimped her to other men. She had a 7-yr.-old daughter she claimed had been molested by the victim. However, the victim’s workmates and neighbors claimed the victim frequently had injuries caused by the suspect. It appears the suspect always initiated contact. At time in question, the suspect had asked the victim for money. The victim drove to suspect’s house and the suspect got into the victim’s auto. Shortly after, the suspect got out of car, screaming that the victim was a fag. He then announced that he was going to the victim’s house to “beat V’s ass.” The victim was in bed with his wife when the suspect kicked door open, rushed into bedroom, stabbed the victim, and fled scene. (St. Louis, case #90112) The victims’ mother had gone to a bar with a white man whom she took home afterwards. While they were having sex, the suspect entered armed with a rifle and ran the man out of the house. The man called police. When the police went to above location, the mother left the house. The suspect remained with both victims and a cousin. The suspect sent the cousin away. Numerous police were sent to the location and attempted negotiations (suspect was holding the victims hostage). During the morning, the suspect shot victim #1, victim #2 and then himself. The suspect had been extremely depressed. Approximately two weeks prior, the suspect had left his sister a suicide note. The suspect had been despondent due to illness (sickle-cell anemia), his inability to work because of it, and because he and his wife (the victims’ mother) had been separated. In January of 1990, the suspect had abducted one of the victims and took the child to Illinois. The suspect had threatened the victim’s mother with a rifle on 1–28–90. (St. Louis, case #90014) The suspect’s son apparently had gone out of town and left his car keys with the victim, and told him (the victim) that he could use the car. The suspect, after drinking, directed the victim to give him the keys. The victim refused and the suspect asked him to step outside and fight “if he thought he was tough.” The victim refused to step outside. The suspect left, went upstairs to his room and returned with a rifle, telling the victim to give him the keys as he shot him in the neck. The suspect then walked up to the victim, who was on his knees and shot him twice in the head. The suspect stated that he was very intoxicated at the time of the shooting, but appeared sober to detectives at 4 a.m. the same morning. The suspect admitted killing the victim and stated to detectives, “I did what I had to do because [the victim] thought he was tougher than me.” Witnesses said the victim and suspect had never had problems in the past. (St. Louis, case #90002)

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

248

November 3, 2003

RETHINKING HOMICIDE

Afternoon of day in question, the victim and witness #4 drove the suspect to do some errands. They then returned to the victim’s home and drank beer with another friend. At approximately midnight, the victim offered to drive the suspect and other man home. The victim drove to gas station and asked the men for money for gas. The other man produced $1, but the suspect refused to give anything. He told police he’d given the victim $20 earlier in evening. The victim drove on and he and the suspect argued. The victim then pulled to curb and all 4 men got out (witness #4 had ridden along). The other man was upset by the arguing and walked away. The suspect pulled a gun. Witness #4 pulled a knife. The suspect shot at witness #4 and dropped gun. The victim and the suspect struggled for gun. The suspect shot the victim and fled scene. (St. Louis, case #92044)

These narratives all describe homicides involving family members. They are all killings of males by other males that are expressively motivated and all but one are committed with a gun. However, it is clear from a careful reading of these scenarios that the situational dynamics leading up to these homicides are quite different. The first homicide involves the killing of a parent out of retaliation for what appears to be years of abuse. In such circumstances, the history between the victim and offender may be of paramount importance in understanding and explaining the resort to lethal violence, whereas the actual events immediately preceding the killing may be of only secondary significance. The history of the relationship between the victim and offender also appears to play a central role in the second homicide narrative, which involves the killing of a child by a parent. Numerous factors seem to have coincided in the offender’s life (e.g., the separation from his wife, depression, a debilitating illness), leading the offender to become despondent and take his life, as well as those of his children. The suicide that followed this killing is a key factor in understanding the situational dynamics underlying these types of homicides. In contrast, the latter two homicide narratives appear to be more reflective of confrontational homicides that are believed to frequently occur between friends and acquaintances. However, the case of the dispute over car keys occurred between two brothers and the argument over money occurred between a nephew and uncle. The notion of a character or honor contest is clearly evident in the third narrative, where the brother makes repeated references to the issue of toughness and wanting to prove this toughness in describing his conversation with the victim prior to the homicide and in explaining the killing to the police. Alcohol also appears to be a factor in both of these cases. The idea that character contests may also occur among family members and are not limited to friend/acquaintance homicides emphasizes the importance of examining both the structure and process underlying homicide situations across a wide variety of subgroups

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

VICTIM–OFFENDER RELATIONSHIPS

November 3, 2003

249

and keeping an open mind to the possibility that different victim–offender relationship categories may evince different structures yet share similar underlying processes.

Stranger Homicides QCA identified the major unique profiles of stranger homicides as being instrumentally motivated homicides committed among males with a gun. As the following narratives reveal, the situational events leading up to these killings can be quite varied. The victim and witness had been driving around drinking beer with friends. After dropping off the friends, the victim and witness went to purchase crack. The witness waited while the victim approached two black males to make a purchase. After talking to them, the victim called to the witness and they returned to the auto. As they were getting into their car, they were robbed at gunpoint and the victim was shot. They were robbed by two men, with a third waiting across the street. All three suspects fled. Not sure if the crack purchase was actually made and the robbers were after the crack, or after cash (probably either). The suspects were never identified. (St. Louis, case #90012) The incident occurred in a back yard. The victim and witness #2 wanted to buy cocaine and a friend of the victim’s introduced the victim to witness #1 for that purpose. They met on the street and the victim told witness #1 that he wanted to buy $800 worth. Witness #1 said that he would go get it and return. While obtaining the drugs, witness #1 encountered the suspect and told him he had a deal going down and the amount. Witness #1 returned to the victim and witness #2 and they stepped into a gangway. Witness #2 wanted to taste the drugs and was about to do so when the suspect (who may have been with two others) appeared, armed with a handgun. The suspect shot the victim. Witness #1 was also shot. The suspect went through the victim’s pockets and fled. Witness #1 dropped the drugs trying to get away. Witness #2 also ran. The suspect said he got $400, and gave $150 to a friend who had lent him the gun, then spent the rest on drugs. The suspect had smoked crack just before the robbery. Afterwards, the suspect claimed that witness #1 had told him to rob the victim, but admitted that witness #1 never contacted him for his share. Witness #1 was shot in the buttock. The suspect was arrested in Chicago, Ill. A plastic bag of white powder, glass pipe, and beer bottle were all found at the scene. The murder weapon was a .44 pistol. (St. Louis, case #90063) The suspects and another subject had stolen a car and were driving around looking for something to do. They decided to ‘jack’ another car and followed the victim (the other subject claimed he was not in the car). The victim was with his girlfriend and they stopped and turned their lights off thinking the suspects would pass by. They didn’t, and the victim’s girlfriend scrambled

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

250

November 3, 2003

RETHINKING HOMICIDE

away, not witnessing the shooting, but identifying suspect #1 as the subject who exited the car. It is unknown how much the victim resisted having his car stolen, but by the girlfriend’s statements, he apparently tried to escape the car, but the suspect was intent upon killing him. (St. Louis, case #90081)

All three of these narratives describe instrumentally motivated homicides of males by another male with a gun. However, the sequences of events leading up to the killings vary considerably. While the first two scenarios both involve drug selling, the connection between the victim and offender in the first case appears to have been solely a drug seller–drug purchaser type of relationship. The victim and offender in the second case are also strangers, but there is a network of other individuals participating in the event through whom the two are connected. Thus, while both homicides occurred during the commission of a robbery, they are not “strangers” in the same way. The case described in the third scenario probably fits most closely with people’s perception of a stranger killing, in the sense that the victim was randomly chosen by someone whom they had never had previous contact with before in their lives, essentially making it pure chance that this victim was selected as a car-jacking target. What is not evident from the QCA analysis but is apparent from a careful reading of the complete set of St. Louis homicide narratives is that another common situation involving the killing of strangers pertains to cases in which the “wrong person” was killed. The following narratives serve to illustrate this type of scenario. Suspect #1 had been dating suspect #2 who had a baby by the witness. Suspect #2 still occasionally had sex with the witness, and was beaten by suspect #1 on at least one occasion for doing so. Suspect #2 had also tried to hire a 16-year-old acquaintance to kill suspect #1. The witness deliberately provoked suspect #1 over suspect #2, making phone calls, etc. On the evening in question, the witness and victim were putting speakers on the witness’s car when suspect #1 approached and shot at them. Both men ran. The victim was shot and collapsed. The witness was unhurt. Afterwards, suspect #2 disposed of the weapon for suspect #1, told his family she’d been present. Suspect #2 had pointed out the witness’s car to suspect #1. The victim was a stranger to both suspect #1 and suspect #2. (St. Louis, case #90019) The corner of Lee and Newstead belonged to the Crips. Witness #1, as well as others on the corner, were wearing blue. The victim had asked witness #2 and his friends (with whom the victim had been drinking) to drive her around. She stopped at one other location, at least, before asking to be taken to the corner of Lee and Newstead (where this shooting occurred). Just before the victim and her companions arrived on the corner, suspect #1, suspect #2, and suspect #3 had driven by, waving a red cap, yelling “What’s up Blood,” and making hand signs for Bloods. Someone on the corner waved

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

VICTIM–OFFENDER RELATIONSHIPS

November 3, 2003

251

blue in response. Suspect #2 then drove around the corner. The victim and companions then pulled up to the corner. Suspects had gone around the block and returned, shooting into the crowd on the corner. Not clear whether or not the victim was shot while standing on the sidewalk, or while seated in the auto. Witnesses claimed that the victim was pushed from the car and the car drove over her feet leaving the scene. However, there were no injuries to the victim’s feet. The victim’s companions did flee the scene. They claimed the victim had approached witness #1 on the corner. He denied knowing the victim. The victim prostituted for drugs. She was pregnant with twins at the time she was killed. The victim and suspects had all been drinking prior. Empty beer bottles were found at the scene. (St. Louis, case #90048) Suspect was riding in auto with 3 or 4 others. Witness #7, who was a Crips wannabe called out “What’s up Cuzz?” to people in auto. That greeting is used by Crips and is considered an insult by Bloods, who refer to each other as ‘Blood.’ Suspect and others in auto were Bloods. After calling out greeting, witness #7 took off running. The victim was standing in street visiting with his friends. They had seen witness #7 calling everyone ‘Cuzz’ and heard him shout at occupants of car. As witness #7 ran by the victim’s group, he told them there would be trouble. Auto made U-turn and cruised by the victim’s group. Car stopped and the suspect called out “What’s up Bloods?” then began shooting. The victim’s group started running. The victim was shot. Car sped off. The suspect fired 6 shots, but witnesses claimed at least one other fired from auto also. As the car sped away, a neighbor heard someone in the car say that they had shot the “wrong mother fucker.” (St. Louis, case #90059) Suspect #1 had robbed a dope dealer in above neighborhood and had been having problems in neighborhood ever since. Frequently shot at people in neighborhood from auto, particularly a certain young male. On eve in question, the suspects had been riding with two females and had shot at several fast food restaurants. They drove through above neighborhood and exchanged words with a group of males (What’s up? Who you lookin’ at, etc.). They then parked car and shot at these youths, who dispersed. The victim had been standing in front yard and was shot. Suspect #2 claimed he fired gun in air, was test-firing it before buying it from suspect #1. The victim had been in her front yard with her brother because her aunt and uncle were fighting, she and brother were watching fight through window. (St. Louis, case #91062)

In all of these cases, the killing of the victim is an “accident” in the sense that either they were not the intended target or they were a victim of mistaken identity. Again, however, as noted above with respect to the felonyrelated stranger homicides, these narratives include both “true” stranger homicides in the sense that the victim and offender have absolutely no ties to one another and the victim is simply in the wrong place at the wrong time, and also individuals who, while strangers, are connected through a network

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

252

November 3, 2003

RETHINKING HOMICIDE

of other persons (as in the first scenario), which subsequently places them in harm’s way. A final category of stranger homicides that appear in the set of St. Louis homicide narratives is those in which the offender becomes involved in an ongoing incident between other parties and, for a variety of reasons, engages in lethal violence against one of these individuals. This type of scenario is evident in the following narratives. There was a birthday party. The victim attended with a man who had once been a roommate of the birthday boy. The two men got into an argument, apparently over some property, although two people said it was over the birthday boy’s ex-wife. The argument became a shoving match. The victim, arguer, and one other left. They told friends of the argument, and they all returned to the party. Apparently they increased their numbers so they could fight. There was fighting. The victim was chased by witness #2 and witness #3 who hit the victim pretty badly. They left the victim lying in the street. The suspect then ran up and stabbed the victim as the victim was trying to run away. All the participants had been drinking heavily prior – beer and wine. (St. Louis, case #90038) The suspect did [the killing] for his girlfriend, apparently she or her sister requested it. The suspect’s girlfriend was the victim’s ex-girlfriend. Two groups of young men – the victim and his friends, the suspect and his friends – stopped by to see this woman on eve in question (didn’t stay continuously, lots of coming and going). Approximately 2:00 a.m., the victim and this woman argued, subject unknown. Woman’s sister then said everyone had to leave. As everyone was leaving, woman’s sister called the suspect and one of his friends back to converse privately. The suspect then took up position in outside of gangway and ambushed the victim as the victim walked to car. The victim’s friends then drove the victim to hospital. (St. Louis, case #90151) Victim was in a tavern between 10 p.m. and 10:30 p.m. He was with a friend and bought drinks for other customers. He and friend left. The victim returned at 11 p.m. and ordered a beer, paying with a $5 bill. When beer maid gave him change, he insisted he’d given her a $10 bill. She counted her drawer and determined he couldn’t have. The victim became verbally abusive. The suspect, who was present and dated the barmaid, told her not to worry about the victim, she had done nothing wrong. The victim then became verbally abusive towards the suspect. The two men fought, rolling on the floor. The victim got up and left tavern. He staggered across the street and collapsed, and was seen by a bi-state bus driver, who called the police. (St. Louis, case #92027) Approximately 1 hour before shooting, the victim took a woman to have a check cashed (possibly his girlfriend, unclear). This woman was involved in a running feud with witness #6. The two women encountered one another while cashing checks. They argued and swung at one another. The victim

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

VICTIM–OFFENDER RELATIONSHIPS

November 3, 2003

253

entered on side of his companion and hit witness #6. She and her friend (witness #5) then drove to Murphy Blair where witness #6 told suspect #2 what happened. They then drove to home of the victim. Witness #5 sent one of victim’s nieces inside to tell the victim to step outside. Witnesses #5 and #6 both claimed suspect #2 was supposed to hit the victim in retaliation. Instead, after a few harsh words, the suspect pulled handgun. The victim backed up and/or fell down. Reportedly, suspect #2 stood over the victim and shot him in the head. When witnesses #5 and #6 saw what happened, they jumped in car and started to speed away. Suspect #2 ran to jump in car, but couldn’t catch up. He fell. Suspect #1, a close friend of the victim and who saw earlier shooting then ran up to suspect #2 and shot him in retaliation for suspect #2 shooting the victim. (St. Louis, case #92043)

The homicides described in these narratives are frequently of a retaliatory nature. While the victim and offender in these cases are technically strangers in the sense that before the incident in question, they were unknown to one another, the sequence of events leading up to these killings sometimes involve disputes and/or feuds between individuals who are nonstrangers. It is only through the intervention of this third party that the altercation becomes a stranger homicide. However, these do not fit the typical profile of a stranger homicide. They do, however, emphasize the tremendous significance of third parties in some lethal assaults (see Felson, Ribner, and Siegel 1984; Luckenbill 1977; Polk 1994). The diversity exhibited in these narratives concerning the situational dynamics underlying stranger homicides suggests that further disaggregation may be necessary in order to avoid grouping together distinct subtypes of homicide into broadly defined categories such as stranger killings. Furthermore, any thorough explanation of stranger homicides will need to be able to account for the varied contexts in which individuals become the victims of lethal violence by unknown others.

Summary of Findings The use of QCA to examine SHR data from 1976 to 1998 and a qualitative analysis of homicide narratives from St. Louis reveal the following findings. First, the results support those who argue in favor of disaggregating homicides according to victim–offender relationships. While intimate/family, friend/acquaintance, and stranger killings do share some common homicide structures, we also find a substantial degree of qualitative distinctiveness across these different types. Stranger homicides occur in the largest number of unique situations, suggesting that these exhibit the most diversity. However, the highest proportion of homicides that fall within unique profiles occur among acquaintance killings. Thus, acquaintance homicides exhibit the greatest amount of situational clustering. The findings also indicate that acquaintance and stranger homicides are also qualitatively

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

254

November 3, 2003

RETHINKING HOMICIDE

distinct, further supporting the disaggregation of homicides by victim– offender relationships. Second, the results also emphasize the importance of increasingly refined categorizations of victim–offender relationships. In particular, we find that intimate partner homicides are qualitatively distinct from homicides involving other family members. The reliance on three broad categories of victim–offender relationships (intimate/family; friend/acquaintance; stranger) in much of the homicide literature may need to be reconsidered. Third, we find evidence of both stability and change in the nature of intimate partner homicides over time. While the structure of homicides involving intimate partners shows some consistency over the past few decades, we also find emerging forms of intimate partner homicide in the 1990s that involve the killing of girlfriends by boyfriends. Fourth, narrative accounts of homicides indicate that a wide variety of situational processes underlie family and stranger homicides. There is some evidence that character contests, long thought to be the exclusive property of acquaintances, also occur among family members. Stranger homicides are the outcome of a considerable range of sequences of events, including drug dealing, random attacks against innocent victims, accidental killings of the “wrong” person, and interventions in ongoing disputes between other parties.

Conclusions and Implications The current research was stimulated by the desire to take a different approach to examining the question of the utility of disaggregating homicides by victim–offender relationship. As was noted previously, much of this research has taken the form of aggregate-level studies that examine whether the impact of various structural correlates varies across homicide rates disaggregated by victim–offender relationship. Using QCA and detailed descriptions of homicide cases, we are able to examine this issue from a situational approach. We do so by first examining whether the structures of homicide situations across victim–offender relationship groups are unique or common. This was followed by an inspection of homicide narratives within particular relationship categories. We find a tremendous amount of support regarding the importance of analyzing different relationship subtypes separately. While patterns and trends in the overall homicide rate are important, our findings strongly emphasize the need to consider the possibility that differences across relationship types are being masked by grouping all homicide cases together. This has important implications with respect to the task of theory construction. To the extent that we have found that homicides committed between intimate partners, family members, friends/acquaintances, and strangers are somewhat

15:48

P1: JMT 0521832993c09

CB617-Miethe-v1

VICTIM–OFFENDER RELATIONSHIPS

November 3, 2003

255

qualitatively distinct, the utility of general theories of crime that assume similar correlates across context is brought into question. At the same time, while many of these relationship categories have structures that make them unique from other subtypes, an examination of the situational aspects of these killings through the use of homicide narratives suggests that some processes transcend across relationship categories. For example, evidence of confrontational homicides, long thought to be an exclusive feature of acquaintance killings, is found in narratives for family homicides and stranger homicides. This suggests that perhaps the most fruitful approach to understanding and explaining homicide is going to involve a bringing together of what have largely been separate traditions within the homicide literature: theories that have been developed to explain homicides within particular categories of victim–offender relationships, and theories that have been developed to explain particular subtypes of homicide based on the situational event (e.g., honor contests). What is particularly clear from the current research is the need for more studies that examine both the structure and process of events leading to lethal violence.

15:48

P1: JMV 0521832993c10

CB617-Miethe-v1

November 1, 2003

C H A P T E R

T E N

Conclusions and Implications

homicide situations involve both structure and process. Their structures are defined by the particular combinations of offender, victim, and offense attributes that provide the physical and social context for the homicide. Their process involves the interpersonal and transactional dynamics that ultimately determine whether a homicide occurs within these particular contexts. Through the use of a comparative method (QCA) and actual narrative accounts of homicides, this study was designed to identify the unique and common structures and processes that underlie homicide situations across time and different subgroups. This final chapter summarizes our major findings about homicide situations and their structural and processual features. After reviewing these findings, we discuss their implications for (1) dispelling various myths and misconceptions about homicide; (2)theoretical developments; and (3) public policy issues and violence intervention efforts. We conclude with a discussion of the limitations of the current study and new directions for future research.

Major Findings Our analysis of homicide situations reveals the following major findings about the nature of homicide situations and subgroup variation in them over time: r Homicide situations in the United States from 1976 to 1998 are empirically concentrated among a small subset of possible structures. There are 32,768 possible structures of homicide situations involving the simultaneous classification of 15 dummy variables (215 ), but only 618 of these 256

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

CONCLUSIONS AND IMPLICATIONS

November 1, 2003

257

structural contexts are found to represent at least 5 homicides per year. These 618 distinct homicide situations account for 82% of all homicides from 1976 to 1998. This pattern of situational clustering (i.e., few configurations account for a large proportion of all cases) holds true for each decade. r Most homicides occur in situational contexts that have changed little over the last three decades. Primary attributes characterizing these stable situational contexts involved combinations of intraracial and intra-age group shootings of nonstrangers by a single adult male offender in the course of arguments occurring in large urban areas. Exceptions, representing some emergent trends in the 1990s, include an increased prevalence of situations involving juvenile offenders, multiple offenders, and instrumental motives. r Expressive homicides involving disputes and arguments are more prevalent than instrumental homicides throughout the twentieth century, and both types of homicide motives are unique in their nature and situational context. Instrumental homicides are more often attacks by strangers and multiple offenders, whereas expressive homicides are relatively unique in terms of their predominance as intraracial and intra-age attacks on nonstrangers by single offenders over 30 years old. Narrative accounts further confirm that these lethal disputes often involve what outsiders would call “trivial altercations,” victim provocation, and the presence of alcohol. r Contrary to the popular image of confrontational homicides as public disputes among marginalized males, narrative accounts of homicide suggest that these types of character contests often involve both male and female offenders, occur in both public and private settings, and are not restricted to minority group members. r Even though most of the particular circumstances and motives often thought to represent expressive and instrumental offenses exhibit internal consistency within these general categories, several homicide motives have situational profiles that are common to both categories. These include youth gang violence (often considered an expressively motivated crime) and drug law violations (often viewed as an instrumental offense). Narrative accounts support the QCA results by illustrating how illegal drug transactions are difficult to classify as either instrumental or expressive motives. Thus, although instrumental and expressive homicides typically have distinct signatures, limitations of this dichotomy must be more widely recognized. r Based on Chicago homicides from 1965 to 1994, we find that male and female offenders usually kill within situational contexts that are qualitatively unique for each gender. Males kill predominantly in situations involving acquaintances of the same age, sex, and race who are shot by single

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

258

r

r

r

r

November 1, 2003

RETHINKING HOMICIDE

offenders in disputes in public locations. Females are disproportionately involved in the killing of family members and intimates of the opposite sex within private locations. Both stability and change are found in the dominant homicide situations for each gender over time. For example, emergent contexts for male-perpetrated homicide often involve an increased prevalence of combinations of the following attributes: outdoor locations, guns, multiple offenders, and teenage offenders. Female offenders in the 1990s have shifted away from guns as the lethal weapon and, compared to women in earlier decades, now kill under a more diverse array of idiosyncratic circumstances. Nearly three-fourths of U.S. homicides in the SHR data from 1976 to 1998 occur in situations common to both juvenile and adult offenders. Homicides involving these two age groups, however, have become increasingly unique over time. The relatively unique signatures for youth homicide in the 1990s involve an increased prevalence of particular combinations of Black offenders, interage groups, stranger assaults, multiple offenders, guns, and large-city locations. The vast majority (84%) of all youth homicides occur within situational contexts that have the same structural features for both juvenile offenders under 16 years old and those 16 and older. The unique structural contexts for each age group represent only a small proportion of juvenile homicides. Among these unique profiles, younger juvenile offenders are more often involved in shootings of family members in the course of arguments, whereas instrumental killings of strangers are key elements of the unique signatures for older juvenile offenders. Whether or not racial differences in homicide situations exhibit unique signatures depends on which racial/ethnic groups are being compared. About one-half of the homicides involving either Black or White offenders occur in situations common to both races. The structural context of homicides for these racial groups has exhibited little change over time. Roughly one-half of homicides involving minority group members also share similar structural profiles among Hispanic and Black offenders. In contrast, the majority of homicides take place within unique structural contexts when non-Hispanic Whites are compared with Hispanic offenders. The relative uniqueness of homicides by Hispanic offenders is primarily attributable to the predominance of young male offenders who kill other males with guns in public places. Narrative accounts reveal several themes as major contributory factors in homicides across racial and ethnic groups. For example, male sexual jealousy and proprietariness underlie the killings of female intimate partners regardless of the race or ethnicity of the perpetrator. Confrontational homicides involving threats to honor and masculine identity are central themes in a wide variety of contexts, including homicides

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

CONCLUSIONS AND IMPLICATIONS

November 1, 2003

259

by White, Hispanic, and Black offenders, interracial homicides, and homicides involving females as victims and offenders. r Differences in the nature and distribution of homicide situations are found based on the victim–offender relationship. Stranger homicides are far more diverse in their situational profiles than other types of victim–offender relationships, whereas acquaintance homicides exhibit the greatest amount of situational clustering (i.e., a higher percentage of these homicides are accounted for by a smaller number of unique profiles). r The most dominant situations unique to stranger homicides involve instrumental firearm homicides committed by a male against another male, and these offenses are often intragroup attacks (i.e., same race, same age group) that take place primarily in large cities. The unique profiles for intimate partner and family homicides involve combinations of older adult offenders who kill someone of the same age and race but opposite sex with a gun in the course of an argument. In contrast, acquaintance homicides differ primarily from intimate/family killings as a result of being male-on-male assaults, whereas they differ from stranger killings due to their expressive, rather than instrumental, motivations. Even though they differ in their structural context, narrative accounts reveal that homicides across different victim– offender relationships share some underlying processes. For example, these narratives indicate that evidence of confrontational homicides, long thought to be an exclusive feature of acquaintance killings by men, are found in descriptions of family homicides and stranger killings as well.

Dispelling Myths and Misconceptions Homicide is no different from other social phenomena in that there is a particular conventional wisdom surrounding it. Both public commentators and scholars have made various claims about the nature of homicide, its correlates, and patterns of stability and change over time. How the current study helps dispel many of the myths and misconceptions involving homicide is discussed below.

Homicide as a Simple or Complex Event In many respects, homicide is a relatively simple and straightforward event. It requires merely motive, opportunity, and only a modicum of skill. This position is clearly reflected by Gottfredson and Hirschi (1990:31) who assert that “homicide is perhaps the most mundane and, in our view, most easily explainable crime.”

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

260

November 1, 2003

RETHINKING HOMICIDE

Homicide in public discourse, law, and previous research is treated as a rather monolithic phenomenon, involving at most only a small number of different forms. Typologies never include more than 10 subtypes of homicide scenarios or syndromes (see Block and Christakos 1995; Miethe and McCorkle 2001; Polk 1994), and homicide is often represented as simple dichotomies (e.g., instrumental or expressively motivated; stranger vs. nonstranger victims; adult vs. teen offenders). In fact, Gottfredson and Hirschi (1990:33) contend that homicide comes in “two basic varieties,” one occurring in the commission of other offenses (e.g., felony-homicide situations) and the other “standard” or “typical” form involves the following: people who are known to one another argue over some trivial matter, as they have argued frequently in the past. In fact, in the past their arguments had on occasion led to physical violence, sometimes on the part of the offender, sometimes on the part of the victim. (Gottfredson and Hirschi 1990:33)

Although there are clear advantages to treating homicide as a simple event (e.g., it is easier to describe and summarize in media sound bites to a nontechnical audience), our findings contradict this image in several ways. First, there are approximately 600 nontrivial and distinct situational contexts that underlie U.S. homicides over the last three decades, and an additional 32,000 combinations of offender, victim, and offense elements are never or rarely represented by homicides. If homicide is so simple and easily explained, why does it have 600 dominant structural profiles and why is there variation in prevalence across these particular profiles? Also, why does homicide rarely, if ever, occur in other contexts that provide the basic elements of motive and opportunity? Second, how do you reconcile the image of simplicity of homicide with the findings that most subtypes of homicide situations are characterized by both unique and common profiles across groups and both stability and change over time? In short, temporaland group-specific differences are incompatible with this presumption of a simple structure. Even a cursory review of homicide narratives provides additional evidence against the conventional wisdom that homicide is monolithic in its structure and process. In each narrative account, it seems clear that the homicide would not have transpired at that particular time and place if it were not for the nexus of specific facilitating places, hardware, and others. Furthermore, the fact that none of the particular structural elements has a consistent impact across all contexts offers even more evidence of the limits of a simplified view of the nature of homicide. The alternative view, however, is that homicide is not as complex as we are lead to believe by its large number of distinct structural profiles, or by situational theories of crime events, information-processing and rational choice models of criminal decisionmaking, and subjective and interpretative perspectives of deviance (see Birkbeck and LaFree 1993; Cornish and Clarke

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

CONCLUSIONS AND IMPLICATIONS

November 1, 2003

261

1986; Lofland 1969). From this perspective, homicide’s complexity may be overstated because (1) most distinct structures represent a trivial proportion of homicides (because most homicides are concentrated within a relatively small number of dominant structural contexts), (2) culture and socialization experiences provide the groundwork for structuring and creating shared evaluations and interpretations of objects and events that minimize subjectivity or idiosyncratic reactions to particular proviolence situational cues, and (3) major themes like masculine competitiveness, character or honor contests, self-help and conflict resolution are fundamental aspects of the processual dynamics underlying homicide situations that transcend individuals and context. Our general point is that characterizations of homicide as either simple or complex events are both misrepresentations of its nature. It is a simple and complex event. Some elements of the process and structure of homicide situations are largely invariant across context, while other elements exhibit context-specific effects. This conclusion is not necessarily the most parsimonious result, but it is nonetheless the best representation of the overall nature of homicide situations.

Main Effects or Context-Specific Effects Another instance of the minimization of complexity in criminological research involves the widespread application of the “main effects” model. As we have described previously (see especially Chapter 3), this model specification assumes that the effect of any variable on an outcome is constant across all levels of the other variables. In other words, this specification basically ignores context because it presumes that a variable will have the same impact across all contexts. This preoccupation with main effects is understandable given the elegance of parsimony and the popular demand for simple descriptions, explanations, and solutions. In fact, its widespread use in the face of substantial evidence of its misrepresentation of the social world is a testament to its appeal. Even within the current study, we have resorted to a main effects interpretation of some our findings (e.g., what are the major factors that distinguish unique and common profiles for particular types of homicide situations), largely because this variable-driven interpretative framework is so firmly entrenched in social research that some readers would probably view our study as incomplete without some statements about the relative importance of particular variables across contexts. However, the fact of the matter is that in no instance in the current study have we found a variable that exhibits a constant effect across all contexts, as assumed in the main effects model. In other words, every single offender, victim, and offense attribute that we examined is associated with a subtype of homicide situations in some particular contexts but not when it is

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

262

November 1, 2003

RETHINKING HOMICIDE

combined with other attributes in another context. Under these conditions, interpretations of main effects are misleading and a misrepresentation of the determinants of the situational context of homicide. Although the main effects model is limited both empirically and theoretically (see Becker 1998; Miethe and Moore 1986; Ragin 2000), it is important to note that some structural elements and combinations of them have a strong impact across most contexts. This is especially true of characteristics such as adult male offenders, single offenders, intraracial attacks, and guns as the lethal weapon. However, even among these dominant attributes, their prevalence and discriminatory power are context-specific. For example, a primary contributor to the unique situational contexts for homicides in the 1970s was female offenders who killed their male intimate partners, whereas it was young urban males who most contributed to the emergence of unique homicide situations in the 1990s. Contrary to a main effects specification, the impact of each of these sets of variables is strong in one context and weak in another. Similar to the simple–complex dichotomy, our analysis of homicide situations suggests that their structural elements are both context-specific and diffuse across many contexts. This conclusion would not have been reached using the standard main effects approach because it begins and ends with an assessment of statistical significance of main effects only. In contrast, we can justifiably assert both conclusions in the current study because the comparative approach used here (i.e., QCA) provides an empirical method of determining whether the impact of a variable is best represented as constant across contexts or context-specific. It is the logical progression from maximum complexity in a truth table to the identification of simpler functional forms that contributes to the power of QCA as a methodological tool for accurately describing the nature of the structural contexts of homicide situations.

Stability or Change in Homicide Situations over Time There is some debate in the literature about whether change or stability is the most accurate depiction of homicide and its correlates. Both public commentators and scholars often rely on temporal patterns of homicide rates and individuals’ risks to make inferences about the nature of homicide. However, changes in prevalence are neither a necessary nor sufficient condition for changes in the nature of homicide situations. In other words, homicide situations may change or remain the same over time regardless of whether rates or levels of homicide have gone up or down.1 1

Our results for female offending and youth homicide clearly illustrate the independence of prevalence measures and changes in context. Specifically, there has been no major change

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

CONCLUSIONS AND IMPLICATIONS

November 1, 2003

263

Based on our analysis of homicide situations and subgroup differences within them, our conclusion is that the structures and processes underlying homicide situations are characterized by both change and stability over time. These patterns vary on the basis of the particular type of homicide situation. As described in Chapter 4, the majority of U.S. homicides over the last three decades have occurred in situational contexts that have not changed their structures over time. This pattern of stable structures also holds for three major subtypes of homicide situations: (1) interpersonal disputes, (2) domestic violence, and (3) killings during the commission of other crimes. For both adult and juvenile offenders, the majority of homicides occur in situations that are also stable over time. However, homicides have become increasingly unique over time for each age group. Gender differences offer the most dramatic contrast to these other subgroups; male and female offenders usually kill under qualitatively different structures that have changed over time for each gender. These diverse findings across subgroups provide clear evidence that homicide situations can be variously characterized as both stable and changing. Based on our analysis of homicide narratives in Los Angeles from 1900 to 1975, we find evidence that homicide situations have exhibited patterns of both stability and change over time. Dispute-related homicides have remained primarily intragroup arguments among men over status, money, and sexual jealousy. Felony-type homicide situations are primarily robberyhomicides that are more likely to involve multiple offenders and greater planning, especially when they involve institutions as their victims, a pattern that holds across decades. The major change in structure and process involves a rise in juvenile offending in the last three decades (especially with guns in a group context). The interpersonal and situational dynamics that are described in homicide narratives show a great deal of continuity over time and location (e.g., Las Vegas, Los Angeles, Miami, and St. Louis) even though particular circumstances (e.g., drug-related killings) have become more common as situational contexts for lethal violence in the last two decades in some of these cities. Under these conditions, both stability and change are the dominant empirical patterns.

in the prevalence of female-perpetrated homicide over the last century (see Chapter 4) even though there has been a clear change over time in the situational contexts in which women kill (e.g., women in the 1960s kill their husbands with guns, whereas in the 1990s they more frequently kill their boyfriends and other family members with weapons other than firearms). In contrast, youth-perpetrated homicide increased dramatically over the last three decades and its situational context has also changed (i.e., it has become more urban, with multiple offenders, and with guns). As these two examples illustrate, the situational context for homicide may be changing regardless of its overall prevalence among a particular group.

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

264

November 1, 2003

RETHINKING HOMICIDE

Common or Unique Signatures A final area of debate involves basic claims about whether homicide situations are qualitatively unique or common across various subgroups. Do women and men kill under different contexts? How about teenagers? Are their homicides relatively unique in terms of their motives and other offense attributes? Do racial differences in the prevalence of homicide offending and victimization extend to the combinations of offender, victim, and offense attributes that define their underlying structures? Are instrumental and expressive homicide really two distinct types of homicide situations? The majority of previous research asserts that homicides have unique signatures for different subgroups. For example, it is widely assumed that men and women typically kill under different circumstances, with men killing other men in the course of interpersonal disputes and women killing primarily their family members or intimate partners. Similarly, teen homicide is often considered distinct from adult-perpetrated homicide because of its group context and its assumed impulsive and expressive character. While much literature has addressed subgroup variation in risk of homicide offending, questions about the relative uniqueness and similarity of homicide across subgroups cannot be easily answered from prior research because these studies have not used an analytic procedure that is able to quantify the degree of uniqueness and commonality in homicide situations across subgroups. When we apply QCA to examine unique and common profiles across subgroups, we find evidence of both characterizations of subgroup variation. In other words, subgroups have both unique and common signatures. The best illustration of common signatures across subgroups involves age differences in offending. Contrary to widespread assertions about the unique character of youth homicide (e.g., its irrationality, expressive motives, group context), the vast majority of dominant homicide structures, and cases explained by them, are represented by situations that are common across age groups. The basis of these common profiles are intragroup assaults among males who kill acquaintances during disputes. Narrative accounts further reveal the pervasive theme of masculinity in these confrontational homicides and how they often begin as trivial altercations. However, the structural profiles for each age group have also changed over time, with youth homicide becoming progressively more prevalent within the context of the interplay between large urban locations, firearms, and multiple offenders. Comparative analyses by race and ethnicity also indicate common structural profiles and situational dynamics in homicides among minorities (e.g., comparisons of Hispanics and Blacks), but unique profiles are the predominant trend when comparisons are made between White offenders and minority group members.

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

CONCLUSIONS AND IMPLICATIONS

November 1, 2003

265

These unique and common profiles provide another indication of both the simplicity and complexity of homicide situations. Even within subgroups that are best represented by common situational profiles (e.g., differences between juvenile and adult offenders), there is a sizable minority of homicides that occur within situational contexts that are unique to each group. It is within this context that we emphasize that homicide situations have both unique and common signatures across subgroups. Any adequate theoretical explanation of homicide needs to be able to account for the existence of these two distinct patterns. This dual pattern poses a significant challenge to both general and subgroup-specific theories of violent offending.

Theoretical Advances Theories of homicide are distinguishable by whether they place emphasis on the sources of offender motivation or criminal opportunity. Both types of theories have been used in previous research to explain homicide rates over time and aggregate units (e.g., cities, nations), subgroup variation in the prevalence of homicide, and both change and stability in its prevalence over time. How well do existing theories of offender motivation and criminal opportunity account for the observed findings in the current study?

Predictive Value of Existing Theories Given their widespread application in criminological research, current theories of offender motivation and criminal opportunity should be able to account for our particular findings about the nature of homicide situations. We summarize in this section how existing theories would explain three of the major findings in this study: (1) the stability of the structural context of homicide over time and dominance of profiles involving expressive intragroup homicides by males, (2) the emergence of killings involving young, urban, minority males, and (3) the decline in particular homicide situations over time (e.g., reductions in female-perpetrated intimate partner killings).

Routine Activities and Criminal Opportunity Theory Criminal opportunity theories at the macro-level (e.g., routine activity theory) cannot easily account for the stability of the situational context of homicide over time because they are primarily designed to explain changes in offending and victimization rates. Specifically, societal changes with respect to the nature of daily routines, labor force participation, guardianship patterns, and leisure activities over the last few decades would be expected to produce some change in the types of situations that bring individuals together

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

266

November 1, 2003

RETHINKING HOMICIDE

and ultimately lead to lethal conflict. For example, greater labor force participation on the part of women should expose them to more situations where they will come into conflict with acquaintances and/or strangers. Decreased racial segregation in various aspects of modern life should increase interracial homicides by exposing individuals of all racial backgrounds to more situations of conflict with members of other racial groups. Therefore, we would expect to see fundamental changes in homicide structures over time, but no major shifts consistent with these social trends were found in our analysis. Current formulations of criminal opportunity theories also cannot explain the dominance of male, expressive intragroup homicides. While they might be able to account for the intragroup nature of homicide (since people’s routine activities tend to most frequently bring them into contact with others who are similar in age, race, and gender), they cannot explain why it is only when males converge in time and place that violence occurs (especially when females could be argued to be more “suitable” targets due to their smaller physical stature). These theories are also unable to explain the decline in female-perpetrated intimate partner killings. If the routine activities of women have shifted to outside of the home in recent decades, this may reduce the amount of in-home conflict due to the reduced amount of time spent in this environment. This social change, however, would be expected to produce reduced levels of intimate partner homicide generally, not a much greater reduction in killings by one gender (women) than the other (men).

Strain Theory Merton’s strain theory and its extension by Agnew (1992) can account for both stability and change in homicide situations over time. If violence is a response to strain, then changes in the level of strain experienced by individuals may increase the amount of violence they engage in, but not necessarily alter the types of situations in which they do so. However, strain theory might also predict a shift toward instrumentally motivated killings (especially in Merton’s original version) to the extent that these types of homicides can be perceived as a response to blocked financial or material goals. In fact, strain can explain either expressive or instrumentally motivated homicides, the former being a reaction to the frustration of blocked access to means and the latter being a more rational response to meeting these blocked goals. When applied to gender differences in homicide situations, strain theory can also account for the masculine nature of homicide by pointing to the strain resulting from blocked means to achieving masculine status. Blocked avenues to status could lead to confrontational homicides, especially under Agnew’s version of strain that acknowledges that frustration may result not

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

CONCLUSIONS AND IMPLICATIONS

November 1, 2003

267

only from blocked financial goals, but from goals like status achievement. Furthermore, if homicides committed by young urban minority males involved in the crack market can directly or indirectly be argued to be a response to the frustrations of blocked access to traditional economic opportunities, then strain theory could account for the emergence of this type of homicide as well. Strain theory, however, does a poor job of explaining homicide offending by females in general (since they are more likely to be blocked from access to material goals than men yet are less likely to commit crime). It is also doubtful that strain theory could be reformulated to make any predictions about the unique situational contexts in which women kill or why these might be changing over time.

Social Disorganization Theory The original formulations of social disorganization theory by Shaw and McKay (1942) can be applied to patterns of stability in much the same way as strain theory. Specifically, if violence is a function of a lack of social control in communities characterized by high levels of mobility, population heterogeneity, and concentrated disadvantage, then increases in any of these community characteristics may increase the volume of violence (by leading to further reductions in social control) without necessarily changing the situational context of violence. Theoretical arguments by William Julius Wilson (1987; 1996) and Elijah Anderson (1994; 1998) concerning the changes undergone by inner city communities since the 1960s with respect to the migration of workingand middle-class Black residents out of the inner city and the decline in the number of manufacturing jobs can be perceived as falling within this general theoretical framework. Accordingly, their arguments can also account for increases in the volume of criminal activity but they provide no basis for predicting that the situational context of violence would change as a result of these community structural changes. However, these changing economic opportunities may have facilitated the movement of young, minority urban males into the crack market as sellers (see Blumstein 1995; 2002), thus accounting for the emergence of this type of urban homicide situation. Social disorganization theory has limited utility in explaining other findings about the situational context of homicide. For example, Anderson’s “code of the street” (1999) contends that much male violence within inner city Black communities is linked to public demonstrations of respect and manhood where adolescents and young adults have limited legitimate avenues to explore or express themselves. However, this perspective is not able to explain the diffuseness of masculinity as a major contributory factor to

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

268

November 1, 2003

RETHINKING HOMICIDE

violence across contexts. Similarly, the various forms of social disorganization theory are unable to offer any explanation for killings of intimate partners by either gender or changes in these patterns over time.2

Subculture of Violence Theory The subculture of violence perspective attributes the higher rates of homicide among particular subgroups (like young, urban, African-American males) to a value system that condones violence in particular situations (see Wolfgang and Ferracuti 1967). These subcultures are said to be passed on from generation to generation. Our finding of stability in homicide structures is consistent with these arguments in the sense that the value systems themselves are implied to be invariant. Thus one should expect that the situational context of violence among members of these subcultures will retain a constant character over time. This perspective also offers an explanation for the predominance of profiles involving male, expressive intragroup homicides. The expressive nature of these killings fits with the notion of subcultures in which disputes can frequently escalate to violence. It is also consistent with a variation of this perspective offered by Harvey (1986), who describes the particular experiences of being a young African-American adult as more indicative of a subculture of exasperation where frustration and powerlessness experienced by many Blacks manifest themselves through violence and aggression (see also Oliver 1994). In combination with the work of Anderson (1999) and Wilson (1987; 1996) on the changing structure of inner cities communities, Harvey’s variant of subcultural theory could also be applied as an explanation of the emergence of homicide situations involving young, urban, minority males, who may have become increasingly exasperated in response to these changes. The intragroup aspect of the dominant profiles would be expected given that subcultural members are most likely to interact with one another and therefore confrontations will most often occur between members of this group. This subculture of violence explanation, however, runs into some obstacles when other aspects of our findings are considered. Specifically, why should one expect this subculture to be almost exclusively adhered to among 2

Alternatively, the following argument would be consistent with social disorganization theory. If the rates of joblessness and single parent families in the inner city community have increased over the past three decades, this would be expected to result in subsequent reductions in collective efficacy. The loss of informal control over the behavior of residents living in these neighborhoods could extend to the perpetration of violence against wives if family members, friends, and neighbors are unwilling to intervene in domestic violence situations, this leading to increased rates of intimate partner homicides. Under this interpretation, social disorganization theory could be used to explain patterns of intimate partner homicide.

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

CONCLUSIONS AND IMPLICATIONS

November 1, 2003

269

males compared to females? In this respect, it has little to offer in the way of understanding the existence of unique situational contexts of violence by males and females or changes in them over time.

Rational Choice Theories Rational choice theories could explain the stability of some homicide situations by offering an explanation as to why instrumental intergroup homicides have not become more common. In particular, if criminal behavior is reduced by the threat of certain and severe punishment, the conservative political climate and the various forms of criminal legislation passed in recent times (e.g., “three strikes” laws, bringing back the death penalty, increased penalties for use of firearms, etc.) may have the result of stifling any potential increases in “rationally motivated” violence (such as instrumental killings). By minimizing rational acts, the types of homicides that are thought to be impulsive and unresponsive to the threat of the criminal justice system should remain the dominant (and stable) structures. In contrast, rational choice theories cannot explain the emergence of homicides involving young, minority males in urban areas if these are a reflection of the violence endemic in the crack market. In fact, under the deterrence doctrine, the high penalties associated with crack cocaine and the “wars on drugs” by law enforcement that have been directed primarily at poor, urban minorities should have reduced their involvement in the drug trade and homicides associated with it. Yet, homicides within this particular group of offenders appear to have increased, not decreased, during this period of active drug suppression activities by law enforcement, a finding that is clearly at odds with the principles of deterrence.3 Rational choice theories are better able to explain the decline in femaleperpetrated intimate partner violence to the extent that these homicides are predominantly a response to abusive tactics by males within the context of these relationships. The attention given to domestic violence in recent decades as well as policy initiatives (such as mandatory arrest) directed toward it may have deterred males from engaging in domestic violence, in turn reducing the need for women to defend themselves with lethal violence against abusive husbands.

The Value of Conjunctive Thinking As illustrated in the applications above, existing theories have mixed success in terms of their ability to explain the research findings in this study. There 3

The general topic of how the drug wars discriminate against the minority community is discussed elsewhere (see Tonry 1995).

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

270

November 1, 2003

RETHINKING HOMICIDE

are at least two reasons for this limited explanatory power. First, there is a problem with levels of explanation and the units of analysis in that we are drawing on macro- or micro-level theories to account for mid-level phenomena like the situational context of homicide (see Short 1997 for a discussion of levels of explanation). Second, each of these theories treated separately often ignores what the other considers essential (e.g., motivation theories ignore opportunity and vice versa). Under these conditions, a better explanation for our findings requires some degree of theoretical integration, or what we call an increase in “conjunctive thinking.” Given that our unit of analysis is the situational context, the obvious starting point for explaining the observed results lies in situational-based theories. These theories (e.g., micro-level versions of criminal opportunity theory, Lofland’s (1969) theory of facilitating conditions) focus on the conjunction of offender, victim, and offense attributes in time and space. They place primary importance on context, recognizing that multiple and diverse combinations of attributes may result in similar outcomes. By highlighting multiple causes and the convergence of necessary conditions, most criminologists who adopt a situational approach are concerned with identifying the particular combinations of elements that facilitate the occurrence of crime. In other words, the search is for conjunctive causes that underlie crime situations. The value of existing theories of criminality is that they identify multiple sources of offender motivation. For example, sociological theories emphasize differential opportunity, concentrated disadvantage, economic inequality, social learning, subcultural influences and value conflict, lifestyles, low self-control, masculine competitiveness, and status threats as contributory factors to higher levels of violence in particular communities and within particular groups. Psychological, biological, and evolutionary psychology approaches have also been widely applied to explain the motivational basis for violence (see Daly and Wilson 1988a; Fishbein 2001; Short 1997). The need for conjunctive thinking is clearly illustrated by how our findings about gender differences in homicide situations would be explained by existing theories. For example, Gottredson and Hirschi’s (1990) general theory of crime would assert that the higher concentration of male perpetrators in most homicide situations is due to their lower level of self-control. However, if this is the case, then why do homicide risks for male offending vary so dramatically across contexts (e.g., males account for nearly 100% of the homicides among youth in a group context with guns in urban areas, but in only about one-half of situations of intimate partner homicides). It is only through the conjunction of motive, opportunity, and context that we are able to fully account for these findings of gender differences and the other results in this study. Our call for greater conjunctive thinking in future homicide research is not new. A number of criminologists have made this same point in previous

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

CONCLUSIONS AND IMPLICATIONS

November 1, 2003

271

research (see Meier, Kennedy and Sacco 2001; Messner, Krohn, and Liska 1989; Miethe and Meier 1994). In fact, the value of integrated theory is extremely well illustrated in William Oliver’s (1994:40) explanation of violence by Black males: The integration of the structural-cultural approach, the routine activities approach, and symbolic interactionism allows for the construction of a theoretical framework that describes how macrosociological factors (like social structure, racial discrimination, and poverty) and microsociological factors (like role-orientations, values and norms, routine activities and definitions of the situation) converge and produce disproportionate rates of interpersonal violence among black males. . . . Structural constraints (like access to adequate educational opportunities, chronic unemployment, and low-income status) and compulsive masculinity role-orientations lead many black males to engage in routine activities (like spending an inordinate amount of time frequenting bars and bar settings and participating in activities associated with the use and sale of drugs) that increase the risk of becoming involved in violent confrontations. . . . Structural constraints, compulsive masculinity role-orientations, and the routine activities associated with the enactment of those role-orientations are responsible for the emergence of a criminogenic social context that increases black males’ risk of violence participation.

Consistent with Oliver’s (1994) integrative explanation of violence, we propose that the union of motive, opportunity, and context is the best theoretical framework for understanding the diverse findings found in the current study. For example, this type of conjunctive thinking could provide a theoretical basis to account for the unique situational contexts for homicide by young urban male offenders in the 1990s. More specifically, the convergence of the same kinds of structural constraints (i.e., blocked access to educational opportunities, chronic unemployment, poverty) and cultural factors (i.e., compulsive masculinity) discussed by Oliver may result in young minority males engaging in particular types of routine activities (unstructured socializing with peers [see Osgood et al. 1996], participation in gangs, involvement in the drug market) that are highly conducive to violence. Variations on this approach could potentially form the basis of an understanding of lower class violence more generally. This type of integrative approach could also offer a framework for explaining other context-specific changes. As an example, the decline in femaleperpetrated intimate partner homicides could be examined from the point of view of a conjunction of changes in structural factors (e.g., improved economic status of women, social changes in marriage patterns) and cultural factors (e.g., increased attention to domestic violence and reduced tolerance for this type of behavior, changes in norms surrounding marriage) influencing the changing nature of roles, lifestyles and norms in society as evident in (a) the development of legislation aimed at combating domestic violence and (b) changes in the routine activities of women. These changes

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

272

November 1, 2003

RETHINKING HOMICIDE

may ultimately reduce the exposure of women to violence-prone situations as well as the way they define themselves and their situations of abuse. Similar conjunctive arguments are also necessary to explain the other specific patterns of change and stability of structure and process underlying homicide situations. The ultimate value of this conjunctive approach is that it forces us to look more closely at how motive and opportunity operate to produce homicide events within particular situational contexts.

Public Policy Issues and Violence Intervention A wide variety of short-and long-term intervention efforts have been proposed as public policy for violence prevention. These initiatives have taken various forms, including individual-based treatments for repeat violent offenders, early detection and interventions, family- and school-based programs, increased legal penalties, gun control, gang abatement programs, and wider social welfare programs. The pattern of results observed in this study has public policy implications in each of the following areas.

Issues of Gender, Class, and Race Any reasonable public policy about homicide prevention should be directed at the most prevalent situations in which people are killed. Regardless of decade, the most prevalent situations involve interpersonal disputes among males of the same age and race that involve firearms. These disputes are often played out in public places in the presence of others. The themes of status threats, masculine competitiveness, and character contests are major contributory factors for homicides within public settings, especially among young, urban, minority males. If many lethal disputes derive from frustration and exasperation in innercity neighborhoods characterized by concentrated disadvantage (see Anderson 1999; Wilson 1987), only major social and economic reform may impede this spiraling cycle and culture of violence. Providing adolescent and young adult males with alternative avenues to explore and express themselves (e.g., through job training, educational advancement, the development of nonviolent conflict resolution skills) are some of the basic and immediate intervention efforts that have been proposed in previous research (see Polk 1994). The next most prevalent homicide situation involves interpersonal disputes within the context of family relations and intimate partners. The majority of these situations involve intersex slayings of intimate partners. According to the literature and our findings, homicides within the context of domestic violence have well-defined signatures (i.e., they are disputerelated homicides, often occur in the home, they are the primary context

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

CONCLUSIONS AND IMPLICATIONS

November 1, 2003

273

for female offending, and are associated with long histories of physical and emotional abuse). Various ongoing programs directed at domestic violence policies and practices are clearly justified given the prevalence of homicides within this situational context. However, the shift in the dominant profiles of intimate partner killings toward a greater prevalence of boyfriends killing girlfriends suggests the need for current domestic violence initiatives to be expanded to a wider population of women, particularly nonmarital couples.

Guns and Lethal Violence Guns are a major offense characteristic that is strongly associated with homicide across various situational contexts. The vast majority of U.S. homicides involve firearms as the lethal weapon and this prevalence has not changed appreciably over the last century. Guns are most often the lethal weapon within each major subgroup and there has been a noticeable rise in firearm homicides for some groups, especially young urban males.4 Our review of several thousand homicide narratives provides the basis for two general observations about guns and homicide. First, the dynamics of disputes change dramatically once a firearm becomes involved. The nature of the victim’s reactions to the firearm (e.g., fleeing, hiding, immediate retaliation, standing firm and offering verbal challenges) depends in a large part on the particular elements of the situational context (e.g., the location of the dispute, the victim–offender relationship, bystanders’ reactions). Second, guns are clearly “facilitating hardware” (Lofland 1969) for lethal violence. The narrative accounts are filled with numerous instances of homicides that would have been difficult, if not impossible, to commit if it was not for the availability and use of a firearm. These situations are best represented by (1) drive-by shootings and (2) young males being “picked on” or “dissed” by groups of males, leaving the scene, and then returning with a gun. Within these particular situations, firearms provide a unique opportunity structure for homicide because they are highly lethal weapons, easily used, equalize strength differentials, and may be employed without direct physical contact with the victim. From this perspective, public policy directed at limiting the supply and possession of firearms would be an immediate step in reducing the lethality of many forms of youth violence.5 In fact, there are several current programs 4

5

Several researchers (see Cook and Laub 1998; Fox 1996) have clearly demonstrated that the rise in juvenile homicide over the last three decades is attributed solely to the rise in youth slayings involving firearms. Rates of nongun homicides by juveniles have remained stable over this time period. The alternative view held by opponents of gun control is that the greater possession and supply of guns provides a deterrent effect to potential offenders (i.e, by arming citizens you

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

274

November 1, 2003

RETHINKING HOMICIDE

across the United States (e.g., Operation Ceasefire) that are designed to reduce the number of young males who carry guns (see Kennedy, Braga, and Piehl 2001). Other suggested interventions include (1) closing down the main sources of guns reaching youth and (2) giving youth “face-saving” reasons to not carry guns (see Bilchik 1999). Although previous research has increased our understanding of the emergence of this strong link between guns, youth, and homicide over the last two decades (see Blumstein 1995; 2002; Fagan and Wilkinson 1998), it remains to be seen whether current public policies will be effective in dramatically altering this deadly combination of factors.

Drug-Related Homicides Another offense-related element associated with homicide situations involves illegal drugs. The importance of this situational factor lies primarily in the context of the illegal drug market (e.g., buying and selling) rather than the physiological effects of drugs on an individual’s behavior (e.g., impeding judgment, enhancing risk-taking behaviors). However, illegal drugs affect homicide in both ways (see Goldstein 1985; Goldstein, Brownstein, Ryan, and Bellucci 1989). The dynamic interplay between illegal drug activity (especially the buying and selling of crack cocaine), guns, youths, gangs, and emergent homicide situations is complicated (see Fagan and Wilkinson 1998). However, Alfred Blumstein (2002:16) provides a cogent description of the wider structural dynamics in the late 1980s linking drugs and violence: Because crack markets were run as street markets, especially those operating in inner-city areas, the participants were especially vulnerable to attack by robbers who targeted their sizable assets, either the drugs or the money from the sale of drugs. Calling the police for protection was not an option, so participants in those markets, including recruited juveniles, were likely to carry guns to protect themselves and solve disputes. Once these juveniles started carrying guns, other teenagers who were not involved in the drug markets but went to the same schools or walked the same streets also were more likely to arm themselves. These teenagers felt they needed guns for their own protection, but they also may have believed that weapon possession was a status symbol in the community. This initiated an escalating arms race: As more guns appeared in the streets, there was an increased incentive for juveniles to arm themselves. In light of the much tighter networking of make them less attractive to criminals). We have no way of knowing how many homicides during the last three decades were deterred by the threat of contact with armed citizens or by displays of firearms in dispute-related situations (see Kleck 1991). Our general point is that particular properties of firearms make them an attractive instrument or facilitating hardware for lethal violence.

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

CONCLUSIONS AND IMPLICATIONS

November 1, 2003

275

teenagers than of older people, that diffusion process could proceed quickly. The emergence of teenage gangs – some involved in drug markets – in many cities at about this time contributed to that diffusion.

Given the attention devoted to the drug–homicide connection, it is somewhat surprising that the actual prevalence of these homicides as recorded in the SHR data is relatively low (involving 5% of homicides with a known motive). However, the importance of drugs in understanding homicide within particular contexts may be understated by the methods of counting “drugrelated” homicides. In particular, the undercounting of drug-related homicides in SHR data may derive from two sources: (1) robberies involving drugs may be counted as “robbery” motives not “narcotic law violations” and (2) disputes involving drugs may be counted as “other arguments.” Regardless of its prevalence, our examination of drug-related homicides indicates that they have structural elements that are similar to other lethal disputes (e.g., they disproportionately involve male offenders, guns, large urban areas, victims and offenders of similar age, race, and sex). The processes underlying these homicide situations are also similar to personal robberies and other instrumental felony offenses. It is within this framework that further study of the situational context surrounding drug-related homicides may provide greater insights regarding the utility of the expressive vs. instrumental distinction in classifying homicide motives. From a policy perspective, it is unclear whether significant resource expenditures to reduce illegal drug distributions would actually decrease the prevalence of homicides. There are at least two possible reasons for this uncertainty. First, potential offenders who are thwarted from engaging in illegal drug activities may easily find alternative illegal enterprises for achieving instrumental goals. Second, some disputes over drugs may simply be replaced by disputes over a whole host of other events, especially among persons whose lives are filled with frustration, anger, and despair. Under these types of displacement, even effective drug intervention policies may not necessarily change either the prevalence or nature of homicide situations.

The Focus on Rare Events Criminological research, media attention, and to some extent public policy on violence prevention over the last three decades are often directed at relatively uncommon homicide situations. These include each of the following types of homicide situations: interracial violence, the juvenile “superpredator,” female offenders, gang violence, school shootings, multiple offender attacks, rural crime, and combinations of these attributes. Each of these homicide situations represents only a relatively small minority of homicides, but they receive enormous public attention because of their novelty

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

276

November 1, 2003

RETHINKING HOMICIDE

and shock value. Unfortunately, numerous public policies regarding gang violence, stranger violence, school crime, rural crime, and female murderers have been implemented to deal with these alleged emergent homicide threats. Many of these homicide situations fit the conception of a “moral panic” because the public response to them exceeds the actual gravity of the social problem (see Best 1990; Goode and Ben-Yehuda 1994; McCorkle and Miethe 2001). By feeding upon public fears, various claims about emergent homicide trends become part of the public agenda for crime prevention, misdirecting criminal justice resources away from those areas that are most deserving given the empirical reality of homicide situations. Dominant homicide situations in the United States over the last three decades continue to reflect the intersection of gender, race, class, and urban locations. The most effective public policies on violence intervention are those that enhance our understanding of the situational dynamics in lethal disputes among young, lower class, minority males in inner-city communities and then provide the social and economic resources to eliminate these lingering causes of violence in this particular context.

Limitations and Future Directions Our knowledge of the process and structure underlying homicide situations is affected by the methods and research strategies used to study them. The limitations of the current study and some suggestions for future research are summarized below.

Lethal versus Nonlethal Violence One potential limitation of the current study involves our focus on situations of lethal violence. Clearly, most interpersonal disputes are resolved through nonlethal means and the focus on lethal violence may appear restrictive in scope. Particular offense attributes (e.g., the presence and actions of third parties, the type of weapon used, the particular physical location of the dispute) may also be more important in preventing the escalation of serious assaults to homicides, but our study design limits the ability to isolate factors that differentiate homicidal from nonhomicidal outcomes across situations. Although we recognize the potential problems with extrapolations from lethal violence to nonlethal disputes, there are nonetheless several reasons why our results are still relevant to the general study of interpersonal violence. First, the importance attributed to masculine competitiveness, character or honor contests, and status threats in the current study mirrors findings from research on nonlethal violence (see Deibert and Miethe 2003; Fagan and Wilkinson 1998; Oliver 1994; Short and Strodtbeck 1965). Second,

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

CONCLUSIONS AND IMPLICATIONS

November 1, 2003

277

a considerable number of studies suggest that most homicides are not motivated by a specific intent to kill, thus minimizing the distinctive natures of the etiologies of lethal and nonlethal violence (see Block 1977; Cook 1991; Zimring 1968; 1972). Rather than reflecting a difference in intent, whether a violent act results in the death of the victim is often due to such factors as the lethality of the weapon and access to medical resources (cf. Block 1977; Cook 1991; Doerner 1983; Doerner and Speir 1986; Harries 1997; Kleck and McElrath 1991; Zimring 1968; 1972; for an alternative point of view see Felson and Messner 1996; Kleck 1991). As Cook (1991:4) succinctly states, “. . . many homicides are not the result of sustained, deliberate intent to kill but rather are etiologically indistinguishable from a larger set of assaults and robberies in which the victim does not die.” Zimring and others (e.g., Block 1977; Curtis 1974; Harries 1997) have demonstrated notable similarities when comparing assaults and homicides. Under these conditions, our focus on homicide situations may not be overly restrictive because these deadly situations may be considered a special case of interpersonal violence in general. A more direct test of the comparability of the situational context of lethal and nonlethal violence will be possible with the increased availability and coverage from the FBI’s National Incident Based Recording System (NIBRS). This database contains multiple measures of offender, victim, and offense characteristics for numerous crimes including assaults and homicides. However, the implementation of NIBRS is far from complete, with roughly 20 states currently participating. By applying comparative methods like QCA to these data, future researchers will be better able to empirically evaluate whether lethal and nonlethal violence have qualitatively unique or common structures.

Qualitative Comparative Analysis The current study has relied extensively on the method of Qualitative Comparative Analysis (QCA). As developed and applied by Charles Ragin (1987; 2000), QCA is a case-driven approach that directs attention to causal complexity and conjunctive thinking. It recognizes that the same outcome can be achieved through multiple causal paths, and uses rules of formal logic to identify combinations of attributes that uniquely distinguish between different outcomes. Although QCA is most often used in comparative studies with small sample sizes (i.e., n < 30), we have extended this method to the study of the situational contexts that underlie over 400,000 homicides. As a methodological approach for comparative analysis, QCA is associated with several of the same problems that underlie standard statistical analyses. First, the notion of “garbage in, garbage out” applies equally across methods; the validity of the QCA results depends on the quality of the initial

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

278

November 1, 2003

RETHINKING HOMICIDE

data. Second, model misspecification due to the exclusion of relevant variables is a serious problem in both QCA and traditional forms of statistical modeling. However, QCA is better able to directly evaluate functional forms of relationships among variables by its inclusion of all possible interactions as the initial starting point in the truth table. Third, both methods require applying decision rules for defining outcomes. The use of “absolute” and “relative” rules for determining unique situations in QCA is conceptually similar to the choice of significance levels (e.g., p < .05, p < .01) in standard statistical methods. Given similar problems underlying traditional statistical techniques for modeling crime data, our selection of QCA as the primary analytic approach was based on its specific advantages, particularly its logical compatibility with the model of conjunctive thinking that underlies our conceptualization of homicide situations. Specifically, homicide situations involve particular combinations of offender, victim, and offense elements and any investigation of these situational contexts must treat them holistically. Standard statistical approaches are variable-driven, isolating the unique variation accounted for by each variable, and are often based on the assumption that a variable’s impact is linear and invariant across contexts. Across the entire array of analyses conducted in the current study, we found no situation in which a variable’s impact was the same across all contexts. The value of a conjunctive framework for describing multiple causal pathways and the empirical and theoretical limitations of the main effects model should encourage greater use of QCA for future research on interpersonal violence.

Cross-Cultural Comparisons The current study was based primarily on homicides in the United States and select jurisdictions within it (e.g., California, particular cities). This focus on U.S. homicide, however, raises questions about the generalizability of our findings to other countries. There are several practical and logistical problems with conducting crosscultural research on violence (see Gartner 1990; LaFree 1999). The primary problems involve the lack of availability of national data, limitations on the number and nature of the variables included, restrictions on the historical period covered, and questions about the overall quality of the data collection. Additional research problems include the relative infrequency of homicide in other countries (possibly requiring multi-year data collection activities) and the lack of central data repositories in many countries, which diminish the accessibility of information and its standardization across political units. Even if all of these problems were resolved, there remains the question of what countries to select for comparative analysis. As a result of these constraints, our study is restricted to U.S. homicides over the last three decades.

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

CONCLUSIONS AND IMPLICATIONS

November 1, 2003

279

Although we cannot make any definitive claims regarding the extent to which the current quantitative findings would hold in homicide samples from other countries, the results of our qualitative analysis of homicide narratives suggest that some of the processes underlying situations of lethal violence may transcend geographical boundaries. For example, the themes of masculinity, status, respect, and sexual proprietariness were dominant in the homicide narratives which we examined across several U.S. cities. These same themes have also been identified as critical factors in narrative accounts of homicide in Australia (see Polk 1994) and Canada (see Daly and Wilson 1988a; Dawson and Gartner 1998). The relatively ubiquitous nature of these processual elements further underscores our conclusion that some of these themes crosscut different types of homicide situations, different historical periods, and may even persist across different Western cultures, a pattern that is consistent with an evolutionary perspective on homicide (see Daly and Wilson 1988a). Having demonstrated the utility of QCA as an approach for studying homicide situations and showing how different conclusions can arise when homicide cases are examined holistically, we strongly advocate the application of these comparative methods where possible to homicide data in other countries. The growing availability of national homicide data in other countries (e.g., Great Britain, Australia, Hong Kong) should extend the possibilities of replicating our results in other contexts, which is an important pursuit given some evidence indicating that the nature of homicide in the United States differs from other Western nations (see Mouzos 2000; Polk 1994; Silverman and Kennedy 1993).

Conclusion As we have described in this chapter, homicide takes various shapes and forms. Its situational context involves both structure and process. Homicide is both simple and complex, subgroups kill in both unique and common situational contexts, and the structures of homicide situations have both changed and remained stable over time. Theories of offender motivation or criminal opportunity treated separately are unable to explain this diverse nature of the situational context of homicide. We argue that a fuller understanding of homicide involves conjunctive thinking and attention to the interplay of motive and opportunity within particular situational contexts. Rather than focusing on rare events, effective public policy on violence intervention must focus on the dominant situational contexts in which people kill. These dominant situations are lethal disputes with guns that reflect the interplay of gender, race, class, urban locations, and crime.

16:31

P1: JMV 0521832993c10

CB617-Miethe-v1

November 1, 2003

16:31

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

T1: FCH November 3, 2003

References

Adinkrah, Mensah. 1999. “Spousal Homicides in Fiji.” Homicide Studies 3(3):215– 240. Agnew, Robert. 1992. “Foundations for a General Strain Theory of Crime and Delinquency.” Criminology 30:47–87. Alder, Christine and Ken Polk. 1996. “Masculinity and Child Homicide.” British Journal of Criminology 36:396–411. Alder, Christine and Ken Polk. 2001. Child Victims of Homicide. Cambridge, UK: Cambridge University Press. Allison, Paul D. 1999. Macro Miss (Version 1.05). Web site. [www.ssc.upenn.edu/ ∼allison/miss.sas]. Allison, Paul D. 2000. “Multiple Imputation for Missing Data: A Cautionary Tale.” Sociological Methods and Research 28:301–309. Allison, Paul D. 2002. Missing Data. Thousand Oaks, CA: Sage. Amenta, Edwin, B.G. Carruthers, and Y. Zylan. 1992. “A Hero for the Aged? The Townsend Movement, the Political Mediation Model, and U.S. Old-Age Policy, 1934–1950.” American Journal of Sociology 98:308–39. Amenta, Edwin and Drew Halfmann. 2000. “Wage Wars: Institutional Politics, WPA Wages, and the Struggle for U.S. Social Policy.” American Sociological Review 65(4):506–528. Amenta, Edwin and Jane D. Poulsen. 1994. “Where to Begin: A Survey of Five Approaches to Selecting Independent Variables for Qualitative Comparative Analysis.” Sociological Methods and Research 23(1):22–53. Amir, Menachem. 1988. “Political Terrorism and ‘Common Criminality.’” In R.H. Ward and H.E. Smith (eds.), International Terrorism: Operational Issues (pp. 95–109). Chicago, IL: University of Illinois at Chicago. Anderson, Elijah. 1978. A Place on the Corner. Chicago, IL: University of Chicago Press. Anderson, Elijah. 1994. “The Code of the Streets.” Atlantic Monthly (May):81– 94.

281

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

282

T1: FCH November 3, 2003

REFERENCES

Anderson, Elijah. 1997. “Violence and the Inner City Code of the Street.” In Joan McCord (ed.), Violence and the Inner City. New York: Cambridge University Press. Anderson, Elijah. 1998. “The Social Ecology of Youth Violence.” Crime and Justice: A Review of Research 24:64–103. Anderson, Elijah. 1999. Code of the Street: Decency, Violence, and the Moral Life of the Inner City. New York: W.W. Norton. Archer, Dane and Patricia McDaniel. 1995. “Violence and Gender: Differences and Similarities across Societies.” In R. Barry Ruback and Neil A. Weiner (eds.), Interpersonal Violent Behaviors: Social and Cultural Aspects. New York: Springer Publishing Company. Auerhahn, Kathleen and Robert Nash Parker. 1999. “Drugs, Alcohol, and Homicide.” In M.D. Smith and M.A. Zahn (eds.), Studying and Preventing Homicide: Issues and Challenges (pp. 97–114). Thousand Oaks, CA: Sage. Athens, Lonnie. 1985. “Character Contests and Violent Criminal Conduct: A Critique.” Sociological Quarterly 26:419–431. Bailey, William C. 1984. “Poverty, Inequality, and City Homicide Rates: Some Not So Unexpected Findings.” Criminology 22:531–550. Baker, S.G. 1996. “Demographic Trends in the Chicana/o Population: Policy Implications for the Twenty-First Century.” In D.R. Maciel and I.D. Ortiz (eds.), Chicanas/Chicanos at the Crossroads (pp. 5–24). Tucson, AZ: University of Arizona Press. Bankston, W.B., R.L. St. Pierre, and H.D. Allen. 1985. “Southern Culture and Patterns of Victim-Offender Relationships in Homicide: A Study of Primary and Non-Primary Homicide in Louisiana.” Sociological Spectrum 5:197–211. Barclay, Gordon, Cynthia Tavares, and Arsalaan Siddique. 2001. International Comparisons of Criminal Justice Statistics 1999. Home Office Statistical Bulletin 6/01. London: Home Office. Barker, Roger G. 1965. “Explorations in Ecological Psychology.” American Psychologist 20:1–14. Barker, Roger G. 1968. Ecological Psychology. Stanford, CA: Stanford University Press. Beattie, J.M. 1974. “The Pattern of Crime in England, 1660–1800.” Past and Present (February 1974):47–95. Becker, Howard S. 1998. Tricks of the Trade. Chicago: University of Chicago Press. Bennett, William J., John J. Dilulio, and John P. Walters. 1996. Body Count: Moral Poverty and How to Win America’s War against Crime and Drugs. New York: Simon & Schuster. Best, Joel. 1990. Threatened Children: Rhetoric and Concern about Child Victims. Chicago: University of Chicago Press. Best, Joel. 1999. Random Violence: How We Talk about New Crimes and New Victims. Berkeley, CA: University of California Press. Bilchik, Shay. 1999. Report to Congress on Juvenile Violence Research. Washington, DC: Office of Juvenile Justice and Delinquency Prevention. July 1999. Birkbeck, Christopher and Gary LaFree. 1993. “The Situational Analysis of Crime and Deviance.” Annual Review of Sociology 19:113–137.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

REFERENCES

T1: FCH November 3, 2003

283

Black, Donald J. 1983. “Crime as Social Control.” American Sociological Review 48:34–45. Blau, Judith R. and Peter M. Blau. 1982. “The Cost of Inequality: Metropolitan Structure and Violent Crime.” American Sociological Review 14:114–129. Block, Richard L. Violent Crime. Lexington, MA: Lexington Books. Block, Carolyn Rebecca. 1987. “Lethal Violence in the Chicago Latino Community, 1965 to 1981.” In Jess F. Kraus, Susan B. Sorenson, and Paul D. Juarez (eds.), Proceedings of the Research Conference on Violence and Homicide in Hispanic Communities. University of California, Los Angeles. Block, Carolyn Rebecca. 1993. “Lethal Violence in the Chicago Latino Community.” In Anna Victoria Wilson (ed.), Homicide: The Victim/Offender Connection. Cincinnati, OH: Anderson Publishing. Block, Carolyn R. and Richard L. Block. 1980. Patterns of Change in Chicago Homicide: The Twenties, the Sixties, and the Seventies. Chicago, IL: Statistical Analysis Center, Illinois Law Enforcement Commission. Block, Carolyn R. and Richard L. Block. 1992. “Homicide Syndromes and Vulnerability: Violence in Chicago Community Areas over 25 years.” Studies on Crime and Crime Prevention 1:61–87. Block, Carolyn R. and Richard L. Block. 1993. Street Gang Crime in Chicago. National Institute of Justice Research in Brief, Washington, DC. Block, Carolyn R. and Antigone Christakos. 1995. Major Trends in Chicago Homicide: 1965–1994. Illinois Criminal Justice Information Authority, Chicago. Block, Carolyn R., Christine O. Devitt, Edmund R. Donoghue, Roy J. Dames, and Richard L. Block. 2001. “Are There Types of Intimate Partner Homicide?” In P.H. Blackman, V. L. Leggett, and J.P. Jarvis (eds.), The Diversity of Homicide: Proceedings of the 2000 Annual Meeting of the Homicide Research Working Group. Washington, DC: Federal Bureau of Investigation. Block, Richard L. 1976. “Homicide in Chicago: A Nine-Year Study 1965–1973.” Journal of Criminal Law and Criminology 66(4):496–510. Block, Richard L. and Franklin E. Zimring. 1973. “Homicide in Chicago, 1965– 1970.” Journal of Research in Crime and Delinquency 10:1–7. Blumer, Herbert. 1969. Symbolic Interactionism: Perspective and Method. Los Angeles: University of California Press. Blumstein, Alfred. 1995. “Youth Violence, Guns, and the Illicit-Drug Industry.” The Journal of Criminal Law and Criminology 86(1):10–36. Blumstein, Alfred. 2002. “Why Is Crime Falling – or Is It?” Perspectives on Crime and Justice: 2000–2001. Lecture Series, Vol. 5, March 2002. National Institute of Justice. NCJ187100. Blumstein, Alfred and Daniel Cork. 1996. “Linking Gun Availability to Youth Gun Violence.” Law and Contemporary Problems 59:5–24. Blumstein, Alfred and Joel Wallman. 2000. The Crime Drop in America. New York: Cambridge University Press. Boudouris, James. 1970. Trends in Homicide: Detroit, 1926–1968. Unpublished doctoral dissertation. Wayne State University, Detroit, MI. Brantingham, Paul and Patricia Brantingham. 1984. Patterns in Crime. New York: Macmillan.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

284

T1: FCH November 3, 2003

REFERENCES

Brearley, Harold C. 1932. Homicide in the United States. Chapel Hill: University of North Carolina Press. Brewer, Victoria E., Kelly R. Damphouse, and Cary D. Adkinson. 1998. “The Role of Juveniles in Urban Homicide: The Case of Houston, 1990–1994.” Homicide Studies 2:321–339. Bridenbaugh, Carl. 1955. Cities in Revolt: Urban Life in America, 1743–1776. New York: Knopf. Brown, Richard Maxwell. 1975. Strain of Violence: Historical Studies of American Violence and Vigilantism. New York: Oxford University Press. Brown, Richard Maxwell. 1979. “Historical Patterns of American Violence.” In H.D. Graham and T.R. Gurr (eds.), Violence in America: Historical and Comparative Perspectives. Revised edition. Beverly Hills, CA: Sage. Browne, Angela. 1987. When Battered Women Kill. New York: Macmillan/Free Press. Browne, Angela and Kirk R. Williams. 1989. “Exploring the Effect of Resource Availability and the Likelihood of Female-Perpetrated Homicides.” Law and Society Review 23:75–94. Browne, Angela and Kirk R. Williams. 1993. “Gender, Intimacy, and Lethal Violence: Trends from 1976 through 1987.” Gender and Society 7:78–98. Browne, Angela, Kirk R. Williams, and Donald Dutton. 1999. “Homicide between Intimate Partners.” In M. Dwayne Smith and Margaret A. Zahn (eds.), Studying and Preventing Homicide: Issues and Challenges. Thousand Oaks, CA: Sage Publications. Bugliosi, Vincent. 1974. Helter Skelter: The True Story of the Manson Murders. New York: Norton. Bursik, Robert J., Jr. 1988. “Social Disorganization and Theories of Crime and Delinquency: Problems and Prospects.” Criminology 26:519–551. Bursik, Robert J., Jr. and Jim Webb. 1982. “Community Change and Patterns of Delinquency.” American Journal of Sociology 88:24–42. Campbell, Jacqueline C. 1992. “If I Can’t Have You, No One Can: Issues of Power and Control in Homicide of Female Partners.” In Jill Radford and Diana E.H. Russell (eds.), Femicide: The Politics of Woman Killing. New York: Twayne. Canada, Geoffrey. 1995. Fist, Knife, Stick, Gun. Boston: Beacon. Cantor, David and Lawrence E. Cohen. 1980. “Comparing Measures of Homicide Trends: Methodological and Substantive Differences in the Vital Statistics and Uniform Crime Report Time Series (1933–1975).” Social Science Research 9:121–145. Chambliss, William J. 1967. “Types of Deviance and the Effectiveness of Legal Sanctions.” Wisconsin Law Review, Summer: 703–719. Chen, Jiafang. 1996. “Firearms in the Homicides and Suicides of Youth.” In P.K. Lattimore and C.A. Nahabedian (eds.), The Nature of Homicide: Trends and Changes. Proceedings of the 1996 Meeting of the Homicide Research Working Group (pp. 52–56). Washington, DC: U.S. Department of Justice, Office of Justice Programs, National Institute of Justice. Chilton, Roland. 1987. “Twenty Years of Homicide and Robbery in Chicago: The Impact of the City’s Changing Racial and Age Composition.” Journal of Quantitative Criminology 3:195–214.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

REFERENCES

T1: FCH November 3, 2003

285

Chilton, Roland. 1996. “Urban Arrest Trends for Adult Men and Women: 1960– 1993.” In P.K. Lattimore and C.A. Nahabedian (eds.), The Nature of Homicide: Trends and Changes. Proceedings of the 1996 Meetings of the Homicide Research Group (pp. 84–91). Washington, DC: U.S. Department of Justice, Office of Justice Programs, National Institute of Justice. Chin, L. 1996. “Gang Violence in Chinatown.” In C. Ronald Huff (ed.), Gangs in America, 2nd ed. Thousand Oaks, CA: Sage Publications. Clarke, Ronald V. 1983. “Situational Crime Prevention: Its Theoretical Basis and Practical Scope.” In M. Tonry and N. Morris (eds.), Crime and Justice: An Annual Review of Research (pp. 225–256). Vol. 4. Chicago, IL: University of Chicago Press. Clarke, Ronald V. 1997. Situational Crime Prevention: Successful Case Studies. Harrow and Heston, New York. Cockburn, J.S. 1977. “The Nature and Incidence of Crime in England, 1559– 1625: A Preliminary Survey.” In J.S. Cockburn (ed.), Crime in England, 1550– 1800. Princeton, NJ: Princeton University Press. Cohen, Lawrence E. and David Cantor. 1980. “The Determinants of Larceny: An Empirical and Theoretical Study.” Journal of Research in Crime and Delinquency 17:140–59. Cohen, Lawrence E. and David Cantor. 1981. “Residential Burglary in the United States. Lifestyle and Demographic Factors Associated with the Probability of Victimization.” Journal of Research in Crime and Delinquency 18:113– 27. Cohen, Lawrence E. and Marcus Felson. 1979. “Social Change and Crime Rate Trends: A Routine Activity Approach.” American Sociological Review 44:588– 608. Cohen, Lawrence E., Marcus Felson, and Kenneth C. Land. 1980. “Property Crime Rates in the United States: A Macro-Dynamic Analysis, 1947–77; with ex ante Forecasts for the Mid-1980s.” American Journal of Sociology 86:90– 118. Cohen, Lawrence E., James R. Kluegel, and Kenneth C. Land. 1981. “Social Inequality and Predatory Criminal Victimization: An Exposition and Test of a Formal Theory.” American Sociological Review 46:505–24. Cook, Philip J. 1991. “The Technology of Personal Crime.” Crime and Justice: A Review of Research 14:1–71. Cook, Philip J. and John H. Laub. 1998. “The Unprecedented Epidemic in Youth Violence.” Crime and Justice: A Review of Research 24:27–64. Cook, Philip J. and Mark H. Moore. 1999. “Guns, Gun Control, and Homicide.” In M.D. Smith and M.A. Zahn (eds.), Homicide: A Sourcebook of Social Research (pp. 277–296). Thousand Oaks, CA: Sage. Cooley, Charles Horton. 1964. Human Nature and the Social Order. New York: Schocken Books (1902; revised 1922). Cork, Daniel. 1997. “Examining Space-Time Interactions in City-Level Homicide Data: Crack Markets and the Diffusion of Guns among Youth.” Journal of Quantitative Criminology 5:379–406. Cornish, Derek B., and Ronald V. Clarke. 1986. The Reasoning Criminal: Rational Choice Perspectives on Offending. New York: Springer-Verlag.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

286

T1: FCH November 3, 2003

REFERENCES

Corzine, Jay and Lin Huff-Corzine. 1992. “Racial Inequality and Black Homicide: An Analysis of Felony, Non-Felony, and Total Rates.” Journal of Contemporary Criminal Justice 8:150–165. Coverdill, James E., William Finlay, and Jack K. Martin. 1994. “Labor Management in the Southern Textile Industry: Comparing Qualitative, Quantitative, and Qualitative Comparative Analyses.” Sociological Methods and Research 23:54– 85. Crawford, Maria and Rosemary Gartner. 1992. Women Killing: Intimate Femicide in Ontario, 1974–1990. Toronto: Women We Honor Action Committee. Crawford, Maria, Rosemary Gartner, and Myrna Dawson. 1997. Intimate Femicide in Ontario, 1991–1994. Toronto: Women We Honor Action Committee. Cummings, S. and D. J. Monti. 1993. Gangs: The Origins and Impact of Contemporary Youth Gangs in the United States. Albany, NY: SUNY Press. Curry, G. David., Richard A. Ball, and Scott H. Decker. 1996. “Estimating the National Scope of Gang Crime from Law Enforcement Data.” In C. Ronald Huff (ed.), Gangs in America (2nd ed.). Thousand Oaks, CA: Sage. Curtis, Lynn. 1974. Criminal Violence: National Patterns and Behavior. Lexington, MA: Heath. Daly, Martin and Margo Wilson. 1982. “Homicide and Kinship.” American Anthropologist 84:372–378. Daly, Martin and Margo Wilson. 1988a. Homicide. New York: Aldine de Gruyter. Daly, Martin and Margo Wilson. 1988b. “Evolutionary Social Psychology and Family Homicide.” Science 242:519–524. Daly, Martin and Margo Wilson. 1999. “An Evolutionary Psychological Perspective on Homicide.” In M. Dwayne Smith and Margaret A. Zahn (eds.), Studying and Preventing Homicide: Issues and Challenges. Thousand Oaks, CA: Sage Publications. Daly, Martin, Margo Wilson, and Suzanne J. Weghorst. 1982. “Male Sexual Jealousy.” Ethology and Sociobiology 3:11–27. Davidson, R.N. 1989. “Micro-Environments of Violence.” In D. Evans and D. Herbert (eds.), The Geography of Crime (pp. 59–85). London: Routledge. Dawson, Myrna and Rosemary Gartner. 1998. “Male Proprietariness or Despair? Examining the Gendered Nature of Homicides Followed by Suicides.” Paper Presented at the Annual Meeting of the American Society of Criminology. Washington, DC. November. Decker, Scott H. 1993. “Exploring Victim-Offender Relationships in Homicide: The Role of Individual and Event Characteristics.” Justice Quarterly 10(4):585– 612. Decker, Scott H. 1996. “Deviant Homicide: A New Look at the Role of Motives and Victim-Offender Relationships.” Journal of Research in Crime and Delinquency 33:427–449. Decker, Scott H. and K. Kempf-Leonard. 1995. “Constructing Gangs: The Social Definition of Youth Activities.” In M. Klein, C. Maxson, and J. Miller (eds.), The Modern Gang Reader (pp. 112–137). Los Angeles: Roxbury. Deibert, Gini R. and Terance D. Miethe. 2003. “Character Contests and Dispute Related Offenses.” Deviant Behavior 24:245–267.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

REFERENCES

T1: FCH November 3, 2003

287

Dilulio, John. 1995. “The Coming of the Super-Predators.” Weekly Standard (November 27, 1995):23. Dilulio, John. 1996. How to Stop the Coming Crime Wave. New York: Manhattan Institute. Dobrin, Adam, Brian Wiersema, Colin Loftin, and David McDowall. 1996. Statistical Handbook on Violence in America. Phoenix, AZ: The Oryx Press. Doerner, William G. 1983. “Why does Johnny Reb Die When Shot? The Impact of Medical Resources upon Lethality.” Sociological Inquiry 53:1–15. Doerner, William G. and John Speir. 1986. “Stitch and Sew: The Impact of Medical Resources upon Criminally Induced Lethality.” Criminology 24:319–330. Drago, Harry Sinclair. 1970. The Great Range Wars: Violence on the Grasslands. Lincoln, NE: University of Nebraska Press. Drass, Kriss A. and Terance D. Miethe. 2001. “Qualitative Comparative Analysis and the Study of Crime Events.” In R.F. Meier, L.W. Kennedy, and V.F. Sacco (eds.), The Process and Structure of Crime. Advances in Criminological Theory (pp. 125–140), Vol. 9. New Brunswick, NJ: Transaction Publishers. Drass, Kriss A. and Charles C. Ragin. 1989. QCA: Qualitative Comparative Analysis. Evanston, IL: Center for Urban Affairs and Policy Research, Northwestern University. Drass, Kriss A. and J.W. Spencer. 1987. “Accounting for Presentencing Recommendations: Typologies and Probation Officers’ Theory of Office.” Social Problems 34:277–293. Dugan, Laura, Daniel S. Nagin, and Richard Rosenfeld. 1999. “Explaining the Decline in Intimate Partner Homicide: The Effects of Changing Domesticity, Women’s Status, and Domestic Violence Resources.” Homicide Studies 3(3):187–214. Dykstra, Robert R. 1968. The Cattle Towns. New York: Alfred A. Knopf. Eckberg, Douglas. 1995. “Estimates of Early Twentieth-Century U.S. Homicide Rates: An Econometric Forecasting Approach.” Demography 32:1–16. Elikann, Peter T. 1999. Superpredators: The Demonization of Our Children by the Law. Reading, MA: Perseus Books. Erickson, Kai T. 1966. Wayward Puritans: A Study in the Sociology of Deviance. New York: Wiley. Fagan, Jeffrey and Deanna L. Wilkinson. 1998. “Guns, Youth Violence, and Social Identity in Inner Cities.” In Mark Moore and Michael Tonry (eds.), Crime and Justice: An Annual Review of Research (pp. 105–188). Vol. 24. Chicago: University of Chicago Press. Farrington, David P. 1998. “Predictors, Causes, and Correlates of Male Youth Violence.” In Mark Moore and Michael Tonry (eds.), Crime and Justice: An Annual Review of Research (pp. 421–475). Vol. 24. Chicago: University of Chicago Press. Fattah, Ezzat A. 1991. Understanding Criminal Victimization: An Introduction to Theoretical Victimology. Scarborough, Ontario: Prentice-Hall Canada. Federal Bureau of Investigation (FBI). 1950–2002. Crime in the United States: Uniform Crime Reports. Selected years. Washington, DC: U.S. Government Printing Office.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

288

T1: FCH November 3, 2003

REFERENCES

Federal Bureau of Investigation’s Supplementary Homicide Reports (SHR). 1976–1998. Supplementary Homicide Reports. Available through Ann Arbor, MI: Inter-University Consortium for Social and Political Rearch (see Fox 2000). Federal Bureau of Investigation. 1984. Uniform Crime Reporting Handbook. Washington, DC: Government Printing Office. Feeney, Floyd. 1986. “Robbers as Decision-Makers.” In Derek B. Cornish and Ronald V. Clarke (eds.), The Reasoning Criminal: Rational Choice Perspectives on Offending (pp. 53–71). New York: Springer-Verlag. Felson, Richard B. 1981. “An Interactionist Approach to Aggression.” In J. Tedeschi (ed.), Impression Management Theory and Social Psychological Research. New York: Academic Press. Felson, Richard B. 1993. “Predatory and Dispute-Related Violence: A Social Interactionist Approach.” In Ronald V. Clarke and Marcus Felson (eds.), Routine Activity and Rational Choice (pp. 103–125). New Brunswick, NJ: Transaction Publishers. Felson, Richard B. and Steven F. Messner. “To Kill or not to Kill: Lethal Outcomes in Injurious Attacks.” Criminology 34:519–545. Felson, Richard B. and Steven F. Messner. 1998. “Disentangling the Effects of Gender and Intimacy on Victim Precipitation in Homicide.” Criminology 36(2):405–424. Felson, Richard B., Stephen A. Ribner, and Merryl S. Siegel. 1984. “Age and the Effect of the Third Parties during Criminal Violence.” Sociology and Social Research 68(4):452–62. Felson, Richard B. and Henry J. Steadman. 1983. “Situational Factors in Disputes Leading to Criminal Violence.” Criminology 21:59–74. Felson, Richard B. and James T. Tedeschi. 1993. Aggression and Violence: Social Interactionist Perspectives. Washington, DC: American Psychological Association. Ferdinand, Theodore N. 1967. “The Criminal Patterns of Boston Since 1849.” American Journal of Sociology 73:84–99. Fishbein, Diana. 2001. Biobehavioral Perspectives in Criminology. Belmont, CA: Wadsworth. Flewelling, Robert L. 1996. “Exploring the Recent Surge in Youth Homicide Rates: Geographic Variations.” In P.K. Lattimore and C.A. Nahabedian (eds.), The Nature of Homicide: Trends and Changes. Proceedings of the 1996 Meeting of the Homicide Research Working Group (pp. 58–62). Washington, DC: U.S. Department of Justice, Office of Justice Programs, National Institute of Justice. Flewelling, Robert L. and Kirk R. Williams. 1999. “Categorizing Homicides: The Use of Disaggregated Data in Homicide Research.” In M. Dwayne Smith and Margaret A. Zahn (eds.), Studying and Preventing Homicide: Issues and Challenges. Thousand Oaks, CA: Sage Publications. Fox, James Allan. 1996. Trends in Juvenile Violence: A Report to the United States Attorney General on Current and Future Rates of Juvenile Offending. Washington, DC: Bureau of Justice Statistics.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

REFERENCES

T1: FCH November 3, 2003

289

Fox, James Alan. 2000. Uniform Crime Reports [United States]: Supplementary Homicide Reports, 1976–1998 [Computer file]. ICPSR version. Boston, MA: Northeastern University, College of Criminal Justice [producer], 2000. Ann Arbor, MI: Inter-University Consortium for Political and Social Research [distributor], 2000. Fox, William. 1998. Social Statistics. 3rd ed. Bellevue, WA: MicroCase Corporation. Frank, G. 1967. The Boston Strangler. New York: The New American Library. Friedman, Lawrence M. and Robert V. Percival. 1981. The Roots of Justice: Crime and Punishment in Alameda County, California, 1870–1910. Chapel Hill, NC: University of North Carolina Press. Furnham, Adrian and Michael Argyle. 1981. The Psychology of Social Situations. Selected Readings. Oxford: Pergamon Press. Gabor, Thomas, M. Baril, M. Cusson, D. Elie, M. Leblanc, and A. Normandeau. 1987. Armed Robbery: Cops, Robbers, and Victims. Springfield, IL: Charles C. Thomas Publishing. Gartner, Rosemary. 1990. “The Victims of Homicide: A Temporal and CrossNational Comparison.” American Sociological Review 55:92–106. Gartner, Rosemary, K. Baker, and Fred C. Pampel. 1990. “Gender Stratification and the Gender Gap in Homicide Victimization.” Social Problems 37:593– 612. Gelles, Richard J. 1987. Violent Home. Newbury Park, CA: Sage Publications. Georges-Abeyie, Daniel. 1984. The Criminal Justice System and Blacks. New York: Clark Boardman. Gest, Ted and D. Friedman. 1994. “The New Crime Wave.” U.S. News and World Report 117:26 (August 29). Gibbs, Jewelle Taylor and Joseph R. Merighi. 1994. “Young Black Males: Marginality, Masculinity and Criminality.” In Tim Newburn and Elizabeth A. Stanko (eds.), Just Boys Doing Business? Men, Masculinities and Crime. London: Routledge. Gillis, A.R. 1986. “Domesticity, Divorce and Deadly Quarrels: An Exploratory Study of Integration-Regulation and Homicide.” In Timothy Hartnagel and Robert Silverman (eds.), Critique and Explanation: Essays in Honor of Gwynne Nettler. New Brunswick, NJ: Transaction. Gillis, A.R. 1996. “So Long as They Both Shall Live: Marital Dissolution and the Decline of Domestic Homicide in France, 1852–1909.” American Journal of Sociology 101:1273–1305. Gilmore, David D. 1990. Manhood in the Making. New Haven: Yale University Press. Given, James Buchanan. 1977. Society and Homicide in Thirteenth-Century England. Stanford, CA: Stanford University Press. Godwin, Grover Maurice. 2001. Criminal Psychology and Forensic Technology: A Collaborative Approach for Effective Profiling. Boca Raton, FL: CRC Press. Goetting, Ann. 1987. “Homicidal Wives: A Profile.” Journal of Family Issues 8(3):332–341.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

290

T1: FCH November 3, 2003

REFERENCES

Goetting, Ann. 1995. Homicide in Families and Other Special Populations. New York: Springer. Goffman, Erving. 1959. The Presentation of Self in Everyday Life. Garden City, NY: Doubleday. Goffman, Erving. 1967. Interaction Ritual: Essays in Face-to-Face Behavior. Chicago, IL: Aldine Publishing Company. Goldstein, Paul J. 1985. “The Drugs-Violence Nexus: A Tripartite Conceptual Framework.” Journal of Drug Issues 14:493–506. Goldstein, Paul J., H.H. Brownstein, P.J. Ryan, and P.A. Bellucci. 1989. “Crack and Homicide in New York City, 1988: A Conceptually Based Event Analysis.” Contemporary Drug Issues 16:651–687. Gottfredson, Michael R. and Don. M. Gottfredson. 1988. Decision Making in Criminal Justice: Toward the Rational Exercise of Discretion. New York: Plenum Press. Gottfredson, Michael R. and Travis Hirschi. 1990. A General Theory of Crime. Stanford, CA: Stanford University Press. Goode, Eric and Nachman Ben-Yehuda. 1994. Moral Panics: The Social Construction of Deviance. Cambridge, MA: Blackwell. Goode, W. 1969. “Violence among Intimates. U.S. National Commission on the Causes and Prevention of Violence. Task Force on Individual Acts of Violence.” Crimes and Violence 13:941–977. Gove, Walter R., Michael Hughes, and Michael Geerken. 1985. “Are Uniform Crime Reports a Valid Indicator of the Index Crimes? An Affirmative Answer with Minor Qualifications.” Criminology 23:451–501. Greenfield, L.A., Michael R. Rand, D. Craven, P.A. Klaus, C.A. Perkins, C. Ringel, G. Warchol, C. Maston, and James A. Fox. 1998. Violence by Intimates: Analysis of Data on Crimes by Current or Former Spouses, Boyfriends, and Girlfriends (NCJ167237). Washington, DC: U.S. Department of Justice. Griffin, Larry and Charles C. Ragin. 1994. “Some Observations on Formal Methods of Qualitative Analysis.” Sociological Methods and Research 23:4–21. Gump, P.V. 1971. “The Behavior Setting: A Promising Unit for Environmental Designers.” Landscape Architecture 61:130–134. Gurr, Ted Robert. 1979. “On the History of Violent Crime in Europe and America.” In H.D. Graham and T.R. Gurr (eds.), Violence in America: Historical and Comparative Perspectives. Beverly Hills, CA: Sage. Gurr, Ted Robert. 1981. “Historical Trends in Violent Crime: A Critical Review of the Evidence.” In M. Tonry and N. Morris (eds.), Crime and Justice: An Annual Review of Research (pp. 295–353). Chicago, IL: University of Chicago Press. Gurr, Ted Robert. 1989. “Historical Trends in Violent Crime: Europe and the United States.” In T.R. Gurr (ed.), Violence in America: The History of Crime (pp. 21–54), Vol. 1. Newbury Park, CA: Sage Publications. Hanawalt, Barbara A. 1979. Crime and Conflict in English Communities, 1300–1348. Cambridge, MA: Harvard University Press. Hannerz. U. 1969. Soulside: Inquiries into Ghetto Culture. New York: Columbia University Press.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

REFERENCES

T1: FCH November 3, 2003

291

Harer, Miles D. and Darrell Steffensmeier. 1992. “The Differing Effects of Economic Inequality on Black and White Rates of Violence.” Social Forces 70:1035– 1054. Harries, Keith D. 1980. Crime and the Environment. Springfield, IL: Thomas. Harries, Keith D. 1997. Serious Violence. 2nd ed. Springfield, IL: Charles C. Thomas. Hartnagel, Timothy. 1982. “Modernization, Female Social Roles, and Female Crime: A Cross-National Investigation.” The Sociological Quarterly 23(4):477– 490. Harvey, William B. 1986. “Homicide among Young Black Adults: Life in the Subculture of Exasperation.” In D.F. Hawkins (ed.), Homicide among Black Americans (pp. 153–171). Lanham, MD: University Press of America. Hawkins, Darnell F. 1993. “Crime and Ethnicity.” In B. Forst (ed.), The SocioEconomics of Crime and Justice. Armonk, NY: M.E. Sharpe. Hawkins, Darnell F. 1994. “The Analysis of Racial Disparities in Crime and Justice: A Double-Edged Sword.” In Enhancing Capacities and Confronting Controversies in Criminal Justice (NCJ-145318; pp. 48–49). Washington, DC: U.S. Department of Justice. Hawkins, Darnell F. 1999. “What Can We Learn from Data Disaggregation? The Case of Homicide and African Americans.” In M. Dwayne Smith and Margaret A. Zahn (eds.), Studying and Preventing Homicide: Issues and Challenges. Thousand Oaks, CA: Sage Publications. Hawley, Amos. 1950. Human Ecology: A Theory of Community Structure. New York: Ronald Press. Heide, Kathleen M. 1999. “Youth Homicide: An Integration of Psychological, Sociological, and Biological Approaches.” In M. Dwayne Smith and Margaret A. Zahn (eds.), Studying and Preventing Homicide: Issues and Challenges. Thousand Oaks, CA: Sage Publications. Hepburn, John R. 1973. “Violent Behavior in Interpersonal Relationships.” Sociological Quarterly 14:419–429. Hewitt, John D. 1988. “The Victim-Offender Relationship in Convicted Homicide Cases: 1960–1984.” Journal of Criminal Justice 16:25–33. Hicks, Alexander. 1994. “Qualitative Comparative Analysis and Analytic Induction: The Case of the Emergence of the Social Security State.” Sociological Methods and Research 23:86–113. Hindelang, Michael S., Michael Gottfredson, and James Garofalo. 1978. Victims of Personal Crime. Cambridge, MA: Ballinger. Hindus, Michael Stephen. 1980. Prison and Plantation: Crime, Justice, and Authority in Massachusetts and South Carolina, 1767–1878. Chapel Hill, NC: University of North Carolina Press. Holinger, Paul C. and Elaine H. Klemen. 1982. “Violent Deaths in the United States, 1900–1975.” Social Science and Medicine 16:1929–1938. Horowitz, Ruth. 1983. Honor and the American Dream – Culture and Identity in a Chicago Community. New Brunswick, NJ: Rutgers University Press. Horowitz, Ruth and Gary Schwartz. 1974. “Honor, Normative Ambiguity, and Gang Violence.” American Sociological Review 39:238–251.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

292

T1: FCH November 3, 2003

REFERENCES

Hough, Michael. 1987. “Offenders’ Choice of Targets: Findings from Victim Surveys.” Journal of Quantitative Criminology 3:355–369. Howell, James C. and J. David Hawkins. 1998. “Prevention of Youth Violence.” Crime and Justice 24:263–315. Huff-Corzine, Lin, Jay Corzine, and David C. Moore. 1986. “Southern Exposure: Deciphering the South’s Influence on Homicide Rates.” Social Forces 64:906– 924. Humphrey, John A. and Stuart Palmer. 1986. “Race, Sex, and Criminal Homicide Offender-Victim Relationships.” In D.F. Hawkins (ed.), Homicide among Black Americans (pp. 57–67). Lanham, MD: University Press of America. Inciardi, James A., Alan A. Block, and Lyle A. Hallowell. 1977. Historical Approaches to Crime: Research Strategies and Issues. Beverly Hills, CA: Sage Publications. Jacobs, Wilbur R. 1972. Dispossessing the American Indian: Indians and Whites on the Colonial Frontier. New York: Scribner. Jackson, Pamela and C. Rudman. 1993. “Moral Panic and the Response to Gangs in California.” In S. Cummings and D. Monti (eds.), Gangs: The Origin and Impact of Contemporary Youth Gangs in the United States (pp. 257–275). New York: State University of New York Press. Jacobs, David and Katherine Wood. 1999. “Interracial Conflict and Interracial Homicide: Do Political and Economic Rivalries Explain White Killings of Blacks or Black Killings of Whites?” American Journal of Sociology 105(1):157– 90. Jenkins, Philip. 1994. Using Murder: The Social Construction of Serial Homicide. New York: Aldine de Gruyter. Jensen, Vickie. 2001. Why Women Kill: Homicide and Gender Equality. Boulder, CO: Lynne Rienner Publishers Inc. Jones, Ann. 1980. Women Who Kill. New York: Holt, Rinehart, and Winston. Jurik, Nancy and Peter Gregware. 1992. “A Method for Murder: The Study of Homicide by Women.” Perspectives on Social Problems 4:179–201. Jurik, Nancy and Russ Winn. 1990. “Gender and Homicide: A Comparison of Men and Women Who Kill.” Violence and Victims 5:227–242. Kaplun, D. and R. Reich. 1976. “The Murdered Child and His Killers.” American Journal of Psychiatry 133(7):809–813. Katz, Jack. 1988. Seductions of Crime. New York: Basic Books. Kennedy, David M., Anthony A. Braga, and Anne M. Piehl. 2001. “Developing and Implementing Operation Ceasefire.” In Reducing Gun Violence: The Boston Gun Project’s Operation Ceasefire. NIJ Research Report (NCJ 188741: pp. 5–53). Washington, DC: U.S. Department of Justice. Kennedy, Leslie W. and David R. Forde. 1999. When Push Comes to Shove: A Routine Conflict Approach to Violence. Albany, NY: State University of New York Press. Kennedy, Leslie W., and Vincent F. Sacco. 1996. Crime Counts: A Criminal Event Analysis. Scarborough, Ontario: Nelson Canada. Kleck, Gary. 1991. Point Blank: Guns and Violence in America. Hawthorne, NY: Aldine.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

REFERENCES

T1: FCH November 3, 2003

293

Kleck, Gary and Karen McElrath. 1991. “The Effects of Weaponry on Human Violence.” Social Forces 69(3):669–692. Klein, Malcom W. 1995. The American Street Gang: Its Nature, Prevalence, and Control. New York: Oxford University Press. Kornhauser, Ruth. 1978. Social Sources of Delinquency. Chicago, IL: University of Chicago Press. Kovandzic, T.V., L.M. Vieraitis, and M.R. Yeisley. 1998. “The Structural Covariates of Urban Homicide: Reassessing the Impact of Income Inequality and Poverty in the Post-Reagan Era.” Criminology 36:569–600. Krivo, Lauren J. and Ruth D. Peterson. 2000. “The Structural Context of Homicide: Accounting for Racial Differences in Process.” American Sociological Review 65:547–559. LaFree, Gary D. 1998. Losing Legitimacy: Street Crime and the Decline of Social Institutions in America. Boulder, CO: Westview. LaFree, Gary D. 1999. “A Summary and Review of Cross-National Comparative Studies of Homicide.” In M.D. Smith and M.A. Zahn (eds.), Homicide: A Sourcebook of Social Research (pp. 125–145). Thousand Oaks, CA: Sage. LaFree, Gary D. and Christopher Birkbeck. 1991. “The Neglected Situation: A Cross-National Study of the Situational Characteristics of Crime.” Criminology 29:73–98. LaFree, Gary D., Kriss A. Drass, and Patrick O’Day. 1992. “Race and Crime in Postwar America: Determinants of African-American and White Rates, 1957– 1988.” Criminology 30:157–188. LaFree, Gary D. and Kriss A. Drass. 1996. “The Effect of Changes in Intraracial Income Inequality and Educational Attainment on Changes in Arrest Rates for African Americans and Whites, 1957 to 1990.” American Sociological Review 61:614–634. LaFree, Gary D. and Kriss A. Drass. 2002. “Counting Crime Booms among Nations: Evidence for Homicide Victimization Rates, 1956 to 1998.” Criminology 40:769–800. Land, Kenneth C., Patricia L. McCall, and Lawrence E. Cohen. 1990. “Structural Covariates of Homicide Rates: Are There Invariances across Time and Social Space?” American Journal of Sociology 95:922–963. Lane, Roger. 1997. Murder in America: A History. Columbus, OH: Ohio State University. Lane, Roger. 1979. Violent Death in the City: Suicide, Accident, and Homicide in Nineteenth Century Philadelphia. Cambridge, MA: Harvard University Press. Langan, Patrick A. and John M. Dawson. 1995. “Spouse Murder Defendants in Large Urban Counties.” BJS Special Report, NCJ-153256. Washington, DC: Bureau of Justice Statistics. Leyba, Cynthia. 1987. “Homicides in Bernalillo County: 1978–1982.” In Jess F. Kraus, Susan B. Sorenson, and Paul D. Juarez (eds.), Proceedings of the Research Conference on Violence and Homicide in Hispanic Communities. University of California, Los Angeles. Lofland, John. 1969. Deviance and Identity. Englewood Cliffs, NJ: Prentice Hall.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

294

T1: FCH November 3, 2003

REFERENCES

Loftin, Colin. 1986. “The Validity of Robbery–Murder Classifications in Baltimore.” Violence and Victims 1:191–204. Loftin, Colin and Richard H. Hill. 1974. “Regional Culture and Homicide: An Examination of the Gastil-Hackney Thesis.” American Sociological Review 39:714–724. Loftin, Colin, K. Kindley, S. Norris, and B. Wiersema. 1987. “An Attribute Approach to Relationships between Offenders and Victims in Homicide.” Journal of Criminal Law and Criminology 78:259–271. Logan, John R. and Steven F. Messner. 1987. “Racial Residential Segregation and Suburban Violent Crime.” Social Science Quarterly 68:510–525. Loya, Fred. and James A. Mercy. 1985. The Epidemiology of Homicide in the City of Los Angeles, 1970–79. UCLA, Centers for Disease Control. Department of Health and Human Services, Public Health Service. Luckenbill, David F. 1977. “Criminal Homicide as a Situated Transaction.” Social Problems 25:176–186. Luckenbill, David F. and Daniel P. Doyle. 1989. “Structural Position and Violence: Developing a Cultural Explanation.” Criminology 27:419–436. Lund, Laura E. 2001. “Intimate Partner Homicide Trends in California, 1990 through 1998.” In P.H. Blackman, V.L. Leggett, and J.P. Jarvis (eds.), The Diversity of Homicide: Proceedings of the 2000 Annual Meeting of the Homicide Research Working Group. Washington, DC: Federal Bureau of Investigation. Lundsgaarde, Henry P. 1977. Murder in Space City: A Cultural Analysis of Houston Homicide Patterns. New York: Oxford University Press. Lynch, James P. 1987. “Routine Activity and Victimization at Work.” Journal of Quantitative Criminology 3:283–300. MacDonald, J.M. 1975. Armed Robbery: Offenders and Their Victims. Springfield, IL: Charles C. Thomas Publishing. Maltz, Michael D. 1999. Bridging the Gaps in Police Crime Data. Washington, DC: U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. Mann, Cormae Richey. 1996. When Women Kill. Albany, NY: State University of New York Press. Martinez, Ramiro, Jr. 1997. “Homicide among Miami’s Ethnic Groups: Anglos, Blacks, and Latinos in the 1990s.” Homicide Studies 1:17–34. Martinez, Ramiro, Jr. 2002. Latino Homicide: Immigration, Violence, and Community. New York: Routledge. Martinez, Ramiro, Jr. and Matthew T. Lee. 1998. “Immigration and the Ethnic Distribution of Homicide in Miami, 1985–1995.” Homicide Studies 2(3):291– 304. Martinez, Ramiro, Jr. and Matthew T. Lee. 1999. “Extending Ethnicity in Homicide Research: The Case of Latinos.” In M. Dwayne Smith and Margaret A. Zahn (eds.), Studying and Preventing Homicide: Issues and Challenges. Thousand Oaks, CA: Sage Publications. Maxfield, Michael G. 1989. “Circumstances in Supplementary Homicide Reports: Variety and Validity.” Criminology 27:671–695.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

REFERENCES

T1: FCH November 3, 2003

295

Maxson, Cheryl L. 1999. “Gang Homicide.” In M.D. Smith and M.A. Zahn (eds.), Homicide: A Sourcebook of Social Research (pp. 239–254). Thousand Oaks, CA: Sage. Maxson, Cheryl L., M.A. Gordon, and Malcolm W. Klein. 1985. “Differences between Gang and Non-Gang Homicides.” Criminology 23:209–222. Maxson, Cheryl L. and M.W. Klein. 1996. “Defining Gang Homicide: An Updated Look at Member and Motive Approaches.” In C. Ronald Huff (ed.), Gangs in America (2nd ed.). Thousand Oaks, CA: Sage. McCorkle, Richard C. and Terance D. Miethe. 1998. “The Political and Organizational Response to Gangs: An Examination of a ‘Moral Panic.’” Justice Quarterly 15(1):41–64. McCorkle, Richard C. and Terance D. Miethe. 2001. Panic: The Social Construction of the Street Gang Problem. Upper Saddle River, NJ: Prentice Hall. McGrath, Roger. 1984. Gunfighters, Highwaymen, and Vigilantes: Violence on the Frontier. Berkeley, CA: University of California Press. McKanna, Clare V., Jr. 1997. Homicide, Race, and Justice in the American West, 1880– 1920. Tucson, AZ: University of Arizona Press. Mead, George Herbert. 1934. Mind, Self, and Society. Chicago: University of Chicago Press. Meier, Robert F., Leslie W. Kennedy, and Vincent F. Sacco. 2001. The Process and Structure of Crime: Crime Events and Crime Analysis. Advances in Criminological Theory, Vol. 9. New Brunswick, NJ: Transaction. Mercy, James A. 1987. “Assaultive Injury among Hispanics: A Public Health Problem.” In Jess F. Kraus, Susan B. Sorenson, and Paul D. Juarez (eds.), Proceedings of the Research Conference on Violence and Homicide in Hispanic Communities. University of California, Los Angeles. Messerschmidt, James W. 1986. Capitalism, Patriarchy and Crime: Toward a Socialist Feminist Criminology. Totowa, NH: Rowman and Littlefield. Messner, Steven F. 1983. “Regional and Racial Effects on the Urban Homicide Rate. The Subculture of Violence Revisited.” American Journal of Sociology 88: 997–1007. Messner, Steven F., Glenn Deane, and Mark Beaulieu. 2002. “A LogMultiplicative Association Model for Allocating Homicides with Unknown Victim-Offender Relationships.” Criminology 40:457–480. Messner, Steven F. and Reid M. Golden. 1992. “Racial Inequality and Racially Disaggregated Homicide Rates: An Assessment of Alternative Theoretical Explanations.” Criminology 30:421–447. Messner, Steven F., Marvin D. Krohn, and Allen E. Liska. 1989. Theoretical Integration in the Study of Deviance and Crime: Problems and Prospects. Albany, NY: State University of New York Press. Messner, Steven F., Lawrence E. Raffalovich, and Richard McMillan. 2001. “Economic Deprivation and Changes in Homicide Arrest Rates for White and Black Youths, 1967–1998: A National Time-Series Analysis.” Criminology 39(3):591– 614. Messner, Steven F. and Richard Rosenfeld. 2001. Crime and the American Dream. 3rd ed. Belmont, CA: Wadsworth.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

296

T1: FCH November 3, 2003

REFERENCES

Messner, Steven F. and Robert J. Sampson. 1991. “The Sex Ratio, Family Disruption, and Rates of Violent Crime: The Paradox of Demographic Structure.” Social Forces 69:693–713. Messner, Steven F. and Jukka Savolainen. 2001. “Gender and the VictimOffender Relationship in Homicide: A Comparison of Finland and the United States.” International Criminal Justice Review 11:34–57. Messner, Steven F. and Kenneth Tardiff. 1985. “The Social Ecology of Urban Homicide: An Application of the ‘Routine Activities’ Approach.” Criminology 23:241–267. Messner, Steven F. and Judith R. Blau. 1987. “Routine Leisure Activities and Rates of Crime: A Macro-Level Analysis.” Social Forces 65:1035–52. Miethe, Terance D. and Kriss A. Drass. 1999. “Exploring the Social Context of Instrumental and Expressive Homicides: An Application of Qualitative Comparative Analysis.” Journal of Quantitative Criminology 15(1):1–21. Miethe, Terance D., Michael Hughes, and David McDowall. 1991. “Social Change and Crime Rates: An Evaluation of Alternative Theoretical Approaches.” Social Forces 70:165–185. Miethe, Terance D. and Richard C. McCorkle. 2001. Crime Profiles: The Anatomy of Dangerous Persons, Places, and Situations. 2nd ed. Los Angeles: Roxbury. Miethe, Terance D. and Robert F. Meier. 1990. “Criminal Opportunity and Victimization Rates: A Structural-Choice Theory of Criminal Victimization.” Journal of Research in Crime and Delinquency 27:243–66. Miethe, Terance D. and Robert F. Meier. 1994. Crime and Its Social Context: Toward an Integrated Theory of Offenders, Victims, and Situations. Albany, NY: State University of New York Press. Miethe, Terance D. and Charles A. Moore. 1986. “Racial Differences in Criminal Processing: The Consequence of Model Selection on Conclusions about Differential Treatment.” Sociological Quarterly 27(2):17–237. Miethe, Terance D., Mark Stafford, and J. Scott Long. 1987. “Social Differentiation in Criminal Victimization: A Test of Routine Activity/Lifestyle Theories.” American Sociological Review 52:184–94. Mischel, Walter. 1968. Personality and Assessment. New York: John Wiley and Sons. Monkkonen, Eric H. 1975. The Dangerous Class: Crime and Poverty in Columbus, Ohio, 1860–1885. Cambridge, MA: Harvard University Press. Monkkonen, Eric H. 1991. Crime and Justice in American History. The Colonies and Early Republic. Vols. 1 and 2. Ed. by E.H. Monkonnen. Westport, CT: Meckler Publishing. Monkkonen, Eric H. 1994. Police Departments, Arrests and Crime in the United States, 1860–1920 [U.S. Cities, 1880–1890] [Computer file]. Compiled by Eric Monkkonen, University of California, Los Angeles. ICPSR ed. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [producer and distributor], 1994. Monkkonen, Eric H. 2001. Murder in New York City. Berkeley, CA: University of California Press. Moore, Mark H. and Michael Tonry. 1998. “Youth Violence in America.” In M. Moore and M. Tonry (eds.), Crime and Justice: An Annual Review of Research (pp. 1–24), Vol. 24. Chicago, IL: University of Chicago Press.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

REFERENCES

T1: FCH November 3, 2003

297

Mosher, Clayton J., Terance D. Miethe, and Dretha M. Phillips. 2002. The Mismeasure of Crime. Thousand Oaks, CA: Sage. Mouzos, Jenny. 2000. Homicidal Encounters: A Study of Homicide in Australia, 1989– 1999. Canberra, Australia: Australian Institute of Criminology. Mucheno, Michael C., Peter R. Gregware, and Kriss A. Drass. 1991. “Court Management of AIDS Disputes: A Sociolegal Analysis.” Law and Social Inquiry 16:737–74. Nash, Jay Robert. 1973. Bloodletters and Badmen: A Narrative Encyclopedia of American Criminals from the Pilgrims to the Present. New York: M. Evans and Company. National Center for Health Statistics (NCHS). 1993. Vital Statistics for the United States. Washington, DC: Government Printing Office. National Center for Health Statistics (NCHS). 2000. “Leading Causes of Death Reports, 1999–2000.” Data Source: National Center for Health Statistics Vital Statistics System. Tables produced and made available through the Office of Statistics and Programming, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention. O’Brien, Robert. 1987. “The Interracial Nature of Violent Crime: A Reexamination.” American Journal of Sociology 92:817–835. Oliver, William. 1994. The Violent Social World of Black Men. New York: Lexington. Osgood, D. Wayne, J.K. Wilson, P.M. O’Malley, J.G. Bachman, and L.D. Johnson. 1996. “Routine Activities and Individual Deviant Behavior.” American Sociological Review 61:635–655. Ousey, Graham C. 1999. “Homicide, Structural Factors, and the Racial Invariance Assumption.” Criminology 37(2):405–426. Pampel, Fred C. and Kirk R. Williams. 2000. “Intimacy and Homicide: Compensating for Missing Data in the SHR.” Criminology 38:661–680. Parker, Karen F. and Patricia L. McCall. 1997. “Adding Another Piece to the Inequality-Homicide Puzzle: The Impact of Structural Equality on Racially Disaggregated Homicide Rates.” Homicide Studies 1:35–60. Parker, Robert Nash. 1989. “Poverty, Subculture of Violence, and Type of Homicide.” Social Forces 67:983–1007. Parker, Robert Nash. 1995. “Bringing ‘Booze’ Back in: The Relationship Between Alcohol and Homicide.” Journal of Research in Crime and Delinquency 32:3–38. Parker, Robert Nash and Linda Rebhun. 1995. Alcohol and Homicide: A Deadly Combination of Two American Traditions. Albany, NY: State University of New York Press. Parker, Robert Nash and M. Dwayne Smith. 1979. “Deterrence, Poverty, and Type of Homicide.” American Journal of Sociology 85:614–624. Parker, Robert Nash and A.T. Toth. 1990. “Family, Intimacy, and Homicide: A Macro-Social Approach.” Violence and Victims 5:195–210. Perkins, Craig and Patsy Klaus. 1996. Criminal Victimization, 1994. Washington, DC: Bureau of Justice Statistics. Pervin, Lawrence A. 1978. “Definitions, Measurements, and Classifications of Stimuli, Situations, and Environments.” Human Ecology 6:71–105. Petee, Thomas A., Greg S. Weaver, Jay Corzine, Lin Huff-Corzine, and Janice Wittekind. 2001. “Victim–Offender Relationships and the Situational Context of Homicide.” In P.H. Blackman, V.L. Leggett, and J.P. Jarvis (eds.),

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

298

T1: FCH November 3, 2003

REFERENCES

The Diversity of Homicide: Proceedings of the 2000 Annual Meeting of the Homicide Research Working Group. Washington, DC: Federal Bureau of Investigation. Petersilia, Joan. 1980. “Criminal Career Research: A Review of Recent Evidence.” In N. Morris and M. Tonry (eds.), Crime and Justice: An Annual Review of Research (pp. 321–379), Vol. 2. Chicago, IL: University of Chicago Press. Peterson, Ruth D. and Lauren J. Krivo. 1993. “Racial Segregation and Urban Black Homicide.” Social Forces 71:1001–1026. Phillips, Julie A. 1997. “Variation in African-American Homicide Rates: An Assessment of Potential Explanations.” Criminology 35:527–560. Plint, T. 1851. Crime in England. London: Charles Gilpin. Polk, Kenneth. 1994. When Men Kill: Scenarios of Masculine Violence. New York: Cambridge University Press. Pokorny, A. 1965. “Human Violence: A Comparison of Homicide, Aggravated Assault, Suicide, and Attempted Suicide.” Journal of Criminal Law, Criminology, and Police Science 56:488–497. Prassel, Frank Richard. 1972. The Western Peace Officer: A Legacy of Law and Order. Norman, OK: University of Oklahoma Press. Prentky, R., Cohen, M., and Seghorn, T. 1985. “Development of a Rational Taxonomy for the Classification of Rapists: The Massachusetts Treatment Center System.” Bulletin of American Academy of Psychiatry and Law, 13(1):39– 70. Quetelot, L. Adolph. 1842. Treatise on Man and the Development of His Faculties. Edinburgh, Scotland: S.W. and R. Chambers. Ragin, Charles C. 1987. The Comparative Method: Moving beyond Qualitative and Quantitative Strategies. Berkeley, CA: University of California Press. Ragin, Charles C. 1992. Introduction to Qualitative Comparative Analysis. Institute for Policy Research. Northwestern University. Ragin, Charles C. 2000. Fuzzy-Set Social Science. Chicago, IL: The University of Chicago Press. Ragin, Charles C., Susan E. Mayer, and Kriss A. Drass. 1984. “Assessing Discrimination: A Boolean Approach.” American Sociological Review 49:221–34. Rasche, Christine E. 1993. “‘Given Reasons for Violence in Intimate Relationships.” In A.V. Wilson (ed.), Homicide: The Victim/Offender Connection (pp. 75– 100). Cincinnati, OH: Anderson Publishing Co. Redfield, H.V. 1880. Homicide, North and South. Philadelphia: J.B. Lippincott. Regoeczi, Wendy C., Terance D. Miethe and Kriss A. Drass. 2001. “Missing Data in Supplementary Homicide Reports: An Application of Qualitative Comparative Analysis to Explore Patterns of Change and Stability over Time.” Paper presented at the annual meetings of the American Society of Criminology, Atlanta, GA. Regoeczi, Wendy C. and Marc Riedel. 2000. “Estimating Stranger Homicides: Los Angeles and Chicago.” Paper Presented at the Annual Meeting of the American Society of Criminology. San Francisco. November. Regoeczi, Wendy C. and Marc Riedel. 2003. “The Application of Missing Data Estimation Models to the Problem of Unknown Victim/Offender Relationships in Homicide Cases.” Journal of Quantitative Criminology 19(3):155–183.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

REFERENCES

T1: FCH November 3, 2003

299

Reiman, Jeffrey H. 2001. The Rich Get Richer and the Poor Get Prison: Ideology, Class, and Criminal Justice. 6th ed. Boston: Allyn and Bacon. Reiss, Albert J., Jr. and Jeffrey A. Roth. 1993. Understanding and Preventing Violence: Vol. 3. Social Influences. Washington, DC: National Academy Press. Riedel, Marc. 1987. “Symposium on Stranger Violence: Perspectives, Issues, and Problems [Special issue].” Journal of Criminal Law and Criminology 78(2). Riedel, Marc. 1993. Stranger Violence: A Theoretical Inquiry. New York: Garland Publishing. Riedel, Marc. 1999. “Sources of Homicide Data: A Review and Comparison.” In M.D. Smith and M.A. Zahn (eds.), Homicide: A Sourcebook of Social Research (pp. 75–95). Thousand Oaks, CA: Sage. Riedel, Marc and Joel Best. 1998. “Patterns in Intimate Partner Homicide: California, 1987–1996.” Homicide Studies 2:305–320. Riedel, Marc and Roger J. Przybylski. 1993. “Stranger Murders and Assault: A Study of a Neglected Form of Stranger Violence.” In A.V. Wilson (ed.), Homicide: The Victim/Offender Connection. Cincinnati, OH: Anderson. Riedel, Marc and Wendy C. Regoeczi. 2001. “Trying to Relate: Victims, Offenders, and Their Classification in Homicide Research.” Paper Presented at the Annual Meeting of the American Society of Criminology. Atlanta. Riedel, Marc and Margaret A. Zahn. 1985. The Nature and Patterns of American Homicide. National Institute of Justice Report. Washington, DC: Government Printing Office. Robinson, John P., Vladimir G. Andreyenkov, and Vasily D. Patrushev. 1988. The Rhythm of Everyday Life: How Soviet and American Citizens Use Time. Boulder, CO: Westview. Rodriguez, Orlando. 1987. “Hispanics and Homicide in New York City.” In Jess F. Kraus, Susan B. Sorenson, and Paul D. Juarez (eds.), Proceedings of the Research Conference on Violence and Homicide in Hispanic Communities. Los Angeles: University of California. Rojek, Dean. G. and James L. Williams. 1993. “Interracial vs. Intraracial Offenses in terms of the Victim/Offender Relationship.” In A.V. Wilson (ed.), Homicide: The Victim/Offender Connection. Cincinnati, OH: Anderson. Roncek, Dennis W. 1981. “Dangerous Places: Crime and Residential Environment.” Social Forces 60(1):74–96. Roncek, Dennis W. and Pamela A. Maier. 1991. “Bars, Blocks, and Crimes Revisited: Linking the Theory of Routine Activities to the Empiricism of ‘Hot Spots.’” Criminology 29:725–754. Rosa, Joseph G. 1969. The Gunfighter: Man or Myth? Norman, OK: University of Oklahoma Press. Rose, Harold M. 1984. “Black-On-Black Homicides: Overview and Recommendations.” In D. Georges-Abeyie (ed.), The Criminal Justice System and Blacks (pp. 61–74). New York: Clark Boardman Company. Rose, Harold M. and Paula D. McClain. 1990. Race, Place, and Risk: Black Homicide in Urban America. Albany, NY: State University of New York Press. Rose, Harold, M. 1993. “Elevated Homicide Risk Environments in a Single African-American Community: The Case of Milwaukee.” In C.R. Block and

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

300

T1: FCH November 3, 2003

REFERENCES

R.L. Block (eds.), Questions and Answers in Lethal and Non-Lethal Violence. Proceedings of the Second Annual Workshop of the the Homicide Research Working Group (pp. 119–130). Washington, DC: U.S. Department of Justice, Office of Justice Programs, National Institute of Justice. Rosenberg, Ruth and Raymond A. Knight. 1988. “Determining Male Sexual Offender Subtypes Using Cluster Analysis.” Journal of Quantitative Criminology 4:383-410. Rosenberg, Ruth, Raymond A. Knight, R.A. Prentky, and A. Lee. 1988. “Validating the Components of a Taxonomic System for Rapists: A Path Analytic Approach.” Bulletin of the American Academy of Psychiatry and Law, 16(2):169– 185. Rosenfeld, Richard. 1997. “Changing Relationships between Men and Women: A Note on the Decline of Intimate Partner Homicide.” Homicide Studies 1:72– 83. Rosenfeld, Richard, Timothy M. Bray, and Arlen Egley. 1999. “Facilitating Violence: A Comparison of Gang-Motivated, Gang-Affiliated, and NonGang Youth Homicides.” Journal of Quantitative Criminology 15(4):495– 516. Rule, Ann. 1980. The Stranger Beside Me. New York: New American Library. Sacco, Vincent F. and Leslie W. Kennedy. 1994. The Criminal Event: An Introduction to Criminology. Scarborough, Ontario: Nelson Canada. Sacco, Vincent F. and Leslie W. Kennedy. 2002. The Criminal Event: Perspectives in Space and Time. 2nd ed. Belmont, CA: Wadsworth/Thompson Learning. Sampson, Robert J. 1985a. “Race and Criminal Violence: A Demographically Disaggregated Analysis of Urban Homicide.” Crime and Delinquency 31:47– 82. Sampson, Robert J. 1985b. “Neighborhood and Crime: The Structural Determinants of Personal Victimization.” Journal of Research in Crime and Delinquency 22(1):7–40. Sampson, Robert J. 1987. “Urban Black Violence: The Effect of Male Joblessness and Family Disruption.” American Journal of Sociology 93:348–382. Sampson, Robert J. 1995. “Unemployment and Imbalanced Sex Ratios: RaceSpecific Consequences for Family Structure and Crime.” In M. Belinda Tucker and Claudia Mitchell-Kernan (eds.), The Decline in Marriage among African Americans: Causes, Consequences, and Policy Implications. New York: Russell Sage Foundation. Sampson, Robert J. and W. Bryon Groves. 1989. “Community Structure and Crime: Testing Social-Disorganization Theory.” American Journal of Sociology 94:774–802. Sampson, Robert J. and Janet L. Lauritsen. 1990. “Deviant Lifestyles, Proximity to Crime, and the Offender–Victim Link in Personal Violence.” Journal of Research in Crime and Delinquency 27:110–139. Sampson, Robert J. and Janet L. Lauritsen. 1993. “Violent Victimization and Offending: Individual-, Situational-, and Community-Level Risk Factors.” In A.J. Reiss and J. Roth (eds.), Understanding and Preventing Violence. Volume 3: Social Influences. Washington, DC: National Academy Press.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

REFERENCES

T1: FCH November 3, 2003

301

Sampson, Robert J., Stephen W. Raudenbush, and Felton Earls. 1997. “Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy.” Science 277:918–920. Sampson, Robert J. and William Julius Wilson. 1995. “Toward a Theory of Race, Crime, and Urban Inequality.” In John Hagan and Ruth D. Peterson (eds.), Crime and Inequality. Stanford, CA: Stanford University Press. Sampson, Robert J. and John D. Wooldredge. 1987. “Linking the Micro- and Macro-Level Dimensions of Lifestyle – Routine Activity and Opportunity Models of Predatory Victimization.” Journal of Quantitative Criminology 3:371– 93. Sanders, William B. 1994. Gangbangs and Drive-Bys: Grounded Culture and Juvenile Gang Violence. New York: Aldine de Gruyter. Savitz, Leonard, Korni Swaroop Kumar, and Margaret A. Zahn. 1991. “Quantifying Luckenbill.” Deviant Behavior 12:19–29. Schafer, Joseph L. 1997. Analysis of Incomplete Multivariate Data. London: Chapman and Hall. Schuerman, Leo and Solomon Kobrin. 1986. “Community Careers in Crime.” In A.J. Reiss, Jr. and M. Tonry (eds.), Crime and Justice: A Review of Research (pp. 67–100). Vol. 8. Chicago, IL: University of Chicago Press. Schwartz, T. 1981. The Hillside Strangler: A Murderer’s Mind. New York: Doubleday. Shaw, Clifford and Henry McKay. 1942. Juvenile Delinquency and Urban Areas. Chicago, IL: University of Chicago Press. Sherman, Lawrence W., Patrick R. Gartin, and Michael E. Buerger. 1989. “Hot Spots of Predatory Crime: Routine Activities and the Criminology of Place.” Criminology 27:27–55. Sherman, Lawrence W., Leslie Steele, Deborah Laufersweiler, Nancy Hoffer, and Sherry Julian. 1989. “Stray Bullets and Mushrooms: Random Shootings of Bystanders in Four Cities, 1977–1988.” Journal of Quantitative Criminology 5:297–316. Shibutani, Tamotsu. 1961. Society and Personality. Englewood Cliffs, NJ: PrenticeHall. Shihadeh, Edward S. and Nicole Flynn. 1996. “Segregation and Crime: The Effect of Black Social Isolation on the Rates of Black Urban Violence.” Social Forces 74:1325–1352. Shihadeh, Edward S. and Michael O. Maume. 1997. “Segregation and Crime: The Relationship between Black Centralization and Urban Black Homicide.” Homicide Studies 1(3):254–280. Shihadeh, Edward S. and Darrell J. Steffensmeier. 1994. “Economic Inequality, Family Disruption, and Urban Black Violence: Cities as Units of Stratification and Social Control.” Social Forces 73(2):729–751. Short, James F., Jr. 1995. “The Level of Explanation Problem in Criminology.” In R.F. Meier (ed.), Theoretical Methods in Criminology. Beverely Hills, CA: Sage Publications. Short, James F., Jr. 1997. Poverty, Ethnicity, and Violent Crime. Boulder, CO: Westview.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

302

T1: FCH November 3, 2003

REFERENCES

Short, James F., Jr. 1998. “The Level of Explanation Problem Revisited.” The American Society of Criminology 1997 Presidential Address. Criminology 36:3– 36. Short, James F., Jr. and Fred L. Strodtbeck. 1965. Group Process and Gang Delinquency. Chicago, IL: University of Chicago Press. Sickmund, Melissa, Howard N. Snyder, and Eileen Poe-Yamagata. 1997. Juvenile Offenders and Victims: 1997 Update on Violence. Washington, DC: National Center for Juvenile Justice. Siegel, Larry. 1998. Criminology: Theories, Patterns, and Typologies. 6th ed. Belmont, CA: West/Wadsworth Publishing Company. Silverman, Ira J., Manuel Vega, and Terry A. Danner. 1993. “The Female Murderer.” In Anna V. Wilson (ed.), Homicide: The Victim/Offender Connection. Cincinnati, OH: Anderson. Silverman, Robert A. and Leslie W. Kennedy. 1987. “Relational Distance and Homicide: The Role of the Stranger.” Journal of Criminal Law and Criminology 78:272–308. Silverman, Robert A. and Leslie W. Kennedy. 1993. Deadly Deeds: Murder in Canada. Scarborough: Nelson Canada. Silverman, Robert A. and S.K. Mukherjee. 1987. “Intimate Homicide: An Analysis of Violent Social Relationships.” Behavioural Sciences and the Law 5:37– 47. Simon, Rita James. 1975. Women and Crime. Lexington, MA: Lexington Books. Singer, Simon I. 1981. “Homogeneous Victim–Offender Populations: A Review and Some Research Implications.” The Journal of Criminal Law and Criminology 72(2):779–788. Skocpol, Theda. 1984. “Emerging Agendas and Recurrent Strategies in Historical Sociology.” In T. Skocpol (ed.), Vision and Method in Historical Sociology. New York: Cambridge University Press. Skogan, Wesley G. and Michael M. Maxfield. 1981. Coping with Crime: Individual and Neighborhood Reactions. Beverly Hills, CA: Sage. Smith, M. Dwayne. 1992. “Variations in Correlates of Race-Specific Urban Homicide Rates.” Journal of Contemporary Criminal Justice 8:137–149. Smith, M. Dwayne and Robert Nash Parker. 1980. “Type of Homicide and Variation in Regional Rates.” Social Forces 59:136–147. Smith, M. Dwayne and Margaret A. Zahn (eds.). 1999. Homicide: A Sourcebook of Social Research. Thousand Oaks, CA: Sage. Spergel, Irving, G.D. Curry, R. Chance, C. Kane, R. Ross, A. Alexander, E. Simmons, and S. Oh. 1994. Gang Suppression and Intervention: Problem and Response. Washington, DC: U.S. Department of Justice, Office of Juvenile Justice and Delinquency Prevention. Stahura, John M. and John J. Sloan, III. 1988. “Urban Stratification of Places, Routine Activities, and Suburban Crime Rates.” Social Forces 66:1102– 18. Steel, Edward M. 1991. “Criminality in Jeffersonian America – A Sample.” In E.H. Monkkonen (ed.), Crime and Justice in American History (pp. 721–726), Vol. 2. Westport, CT: Meckler Publishing.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

REFERENCES

T1: FCH November 3, 2003

303

Steffensmeier, Darrell and Emilie Allan. 1992. “The Nature of Female Offending: Patterns and Explanations.” In Ruth T. Zaplin (ed.), Female Offenders: Critical Perspectives and Effective Interventions. Gaithersburg, MD: Aspen Publishers. Steffensmeier, Darrell and Dana L. Haynie. 2000. “The Structural Sources of Urban Female Violence in the United States: A Macrosocial GenderDisaggregated Analysis of Adult and Juvenile Homicide Offending Rates.” Homicide Studies 4(2):107–134. Straus, Murray A., Richard J. Gelles, and Suzanne. K. Steinmetz. 1980. Behind Closed Doors: Violence in the American Family. New York: Doubleday/Anchor. Sudnow, David. 1965. “Normal Crimes: Sociological Features of the Penal Code in a Public Defender Office.” Social Problems 12:255–276. Sullivan, T. and P. Maiken. 1983. The Killer Clown. New York: Grosset & Dunlap. Sutherland, Edwin H. 1947. Principles of Criminology (4th ed.). Philadelphia: Lippincott. Swanton, B. 1989. Violence and the Public Contact Workers. Australian Institute of Criminology, Australia. Swigert, Victoria L. and Ronald A. Farrell. 1976. Murder, Inequality, and the Law. Lexington, MA: D.C. Heath. Taylor, Ralph B. 1997. Breaking away from Broken Windows: Baltimore Neighborhoods and the Nationwide Fight against Crime, Grime, Fear, and Decline. Boulder, CO: Westview. Tedeschi, James T. and Richard B. Felson. 1994. Violence, Aggression, and Coercive Actions. Washington, DC: American Psychological Association. Tennenbaum, Abraham N. 1993. “The Supplemental Homicide Report: A Neglected but Valuable Source for Homicide Research.” In C.R. Block and R.L Block (eds.), Questions and Answers in Lethal and Non-Lethal Violence: Proceedings from the Second Annual Workshop of the Homicide Research Working Group (pp. 77– 86). Washington, DC: U.S. Department of Justice, Office of Justice Programs, National Institute of Justice. Thomas, Charles W., and Williams, J.S. 1977. “Actors, Actions, and Deterrence: A Reformulation of Chambliss’s Typology of Deterrence.” In M. Riedel and P.A. Vales (eds.), Treating the Offender: Problems and Issues. New York: Praeger Publishers. Tita, George E., Jacqueline Cohen, and Alfred Blumstein. 1996. “Exploring the Gang-Drug-Gun Nexus: Evidence from Homicides in Pittsburgh.” Paper presented at the Annual Meeting of the American Society of Criminology, Chicago. Tomsen, Stephen. 1997. “A Top Night: Social Protest, Masculinity and the Culture of Drinking Violence.” British Journal of Criminology 37(1):90–102. Tonry, Michael. 1995. Maligned Neglect – Race, Crime, and Punishment in America. New York: Oxford University Press. Venkatesh, Sudhir A. 1996. “The Gang and the Community.” In C.R. Huff (ed.), Gangs in America, 2nd ed. Newbury Park, CA: Sage. Vold, George B., Thomas J. Bernard, and Jeffrey B. Snipes. 1998. Theoretical Criminology. 4th ed. New York: Oxford University Press.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

304

T1: FCH November 3, 2003

REFERENCES

von Hentig, Hans. 1948. The Criminal and His Victim. New Haven, CT: Yale University Press. Voss, Harwin L. and John R. Hepburn. 1968. “Patterns in Criminal Homicide in Chicago.” Journal of Criminal Law, Criminology, and Police Science 59:449– 508. Walker, Samuel, Cassia Spohn, and Miriam DeLone. 2000. The Color of Justice: Race, Ethnicity, and Crime in America. 2nd ed. Belmont, CA: Wadsworth Thomson Learning. Wallace, A. 1986. Homicide: The Social Reality. Sydney, Australia: New South Wales Bureau of Crime and Statistics. Ward, D.E., M. Jackson, and R.E. Ward. 1979. “Crimes of Violence by Women.” In Freda Adler and R.J. Adler (eds.), The Criminology of Deviant Women. Boston: Houghton Mifflin. Warr, Mark. 2002. Companions in Crime: The Social Aspects of Criminal Conduct. Cambridge, UK: Cambridge University Press. Weisheit, Ralph. 1993. “Structural Correlates of Female Homicide Patterns.” In A.V. Wilson (ed.), Homicide: The Victim/Offender Connection (pp. 191–206). Cincinnati, OH: Anderson. White, Norman, Richard Rosenfeld, and Carolyn Phillips. 2002. “Social Networks in Criminal Victimization.” Paper Presented at the Annual Meeting of the American Society of Criminology, Chicago. Whittingham, M.D. 1981. “Vandalism: The Urge to Damage and Destroy.” Canadian Journal of Criminology, 23(1):69–73. Wickham-Crowley, Timothy P. 1991. “Qualitative Comparative Approach to Latin American Revolutions.” International Journal of Comparative Sociology 32:82–109. Wilbanks, William. 1983. “The Female Homicide Offender in Dade County, Florida.” Criminal Justice Review 8:9–14. Wilbanks, William. 1984. Murder in Miami: An Analysis of Homicide Patterns and Trends in Dade County (Miami), Florida, 1917–1983. Lanham, MD: University Press of America. Wilbanks, William. 1987. The Myth of a Racist Criminal Justice System. Monterey, CA: Brooks/Cole. Wilkinson, Deanna L. and Jeffrey Fagan. 1996. “The Role of Firearms in Violence ‘Scripts’: The Dynamics of Gun Events among Adolescent Males.” Law and Contemporary Problems 59:55–90. Wilkinson, Deanna L. and Jeffrey Fagan. 2001. “A Theory of Violent Events.” In Robert F. Meier, Leslie W. Kennedy, and Vincent F. Sacco (eds.), The Process and Structure of Crime: Criminal Events and Crime Analysis. Advances in Criminological Theory. Vol. 9. New Brunswick, NJ: Transaction. Williams, Kirk R. and Robert F. Flewelling. 1987. “Family, Acquaintance, and Stranger Violence: Alternative Procedures for Rate Calculations.” Criminology 25(3):543–560. Williams, Kirk R. and Robert F. Flewelling. 1988. “The Social Production of Criminal Homicide: A Comparative Study of Disaggregated Rates in America Cities.” American Sociological Review 54:421–431.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

REFERENCES

T1: FCH November 3, 2003

305

Wilson, James Q. 1975. Thinking about Crime. New York: Basic Books. Wilson, Margo and Martin Daly. 1985. “Competitiveness, Risk Taking, and Violence: The Young Male Syndrome.” Ethology and Sociobiology 6:59–73. Wilson, Margo and Martin Daly. 1992. “Til Death Do Us Part.” In Jill Radford and Diane E.H. Russell (eds.), Femicide: The Politics of Woman Killing. New York: Twayne. Wilson, Margo I. and Martin Daly. 1993. “Spousal Homicide Risk and Estrangement.” Violence and Victims 8:3–16. Wilson, Margo and Martin Daly. 1995. “An Evolutionary Psychological Perspective on Male Sexual Proprietariness and Violence against Wives.” In Neil Weiner and R.B. Ruback (eds.), Interpersonal Violent Behaviors: Social and Cultural Aspects. New York: Springer. Wilson, Margo, Martin Daly, and Christine Wright. 1993. “Uxoricide in Canada: Demographic Risk Patterns.” Canadian Journal of Criminology (July):263– 291. Wilson, William Julius. 1987. The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. Chicago, IL: The University of Chicago Press. Wilson, William Julius. 1991. “Studying Inner-City Social Dislocations: The Challenge of Public Agenda Research.” American Sociological Review 56:1– 14. Wilson, William Julius. 1996. When Work Disappears: The World of the New Urban Poor. New York: Knopf. Wolfgang, Marvin. 1958. Patterns of Criminal Homicide. Philadelphia: University of Pennsylvania Press. Wolfgang, Marvin and Franco Ferracuti. 1967. The Subculture of Violence: Toward an Integrated Theory in Criminology. London: Tavistock. Zahn, Margaret A. 1987. “Homicide in Nine American Cities: The Hispanic Case.” In Jess F. Kraus, Susan B. Sorenson, and Paul D. Juarez (eds.), Proceedings of the Research Conference on Violence and Homicide in Hispanic Communities. University of California, Los Angeles. Zahn, Margaret A. and Katherine M. Jamieson. 1996. “Changing Patterns of Homicide and Social Policy.” In P.K. Lattimore and C.A. Nahabedian (eds.), The Nature of Homicide: Trends and Changes. Proceedings of the 1996 Meeting of the Homicide Research Working Group (pp. 176–179). Washington, DC: U.S. Department of Justice, Office of Justice Programs, National Institute of Justice. Zahn, Margaret A. and Patricia L. McCall. 1999. “Trends and Patterns of Homicide in the 20th-Century United States.” In M. Dwayne Smith and Margaret A. Zahn (eds.), Studying and Preventing Homicide: Issues and Challenges. Thousand Oaks, CA: Sage Publications. Zahn, Margaret A. and P.C. Sagi. 1987. “Stranger Homicides in Nine American Cities.” Journal of Criminal Law and Criminology 78:377–397. Zatz, Marjorie S. 1987. “Chicano Youth Gangs and Crime: The Creation of a Moral Panic.” Contemporary Crises 11:129–158. Zimring, Franklin E. 1968. “Is Gun Control Likely to Reduce Violent Killings?” University of Chicago Law Review 35:21–37.

16:14

P1: FCH/FFX 0521832993REF

P2: FCH/FFX QC: FCH/FFX CB617-Miethe-v1

306

T1: FCH November 3, 2003

REFERENCES

Zimring, Franklin E. 1998. American Youth Violence. New York: Oxford University Press. Zimring, Franklin E. 1972. “The Medium is the Message: Firearm Caliber as a Determinant of Death from Assault.” Journal of Legal Studies 1:97–124. Zimring, Franklin E., Satyanshu K. Mukherjee, and Barrick Van Winkle. 1983. “Intimate Violence: A Study of Intersexual Homicide in Chicago.” The University of Chicago Law Review 50(2):910–930.

16:14

P1: IYP CB617-IND

CB617-Miethe-v1

November 11, 2003

Author Index

Adinkrah, M., 130, 281 Adkinson, C.D., 41, 284 Agnew, R., 266, 281 Allan, E., 131, 303 Alder, C., 176, 184, 186, 281 Alexander, A, 196, 302 Allen, H.D., 228, 282 Allison, P.D., 43, 281 Amenta, E., 10, 51, 53, 142, 161, 281 Amir, M., 102, 281 Anderson, E., 12, 94, 121, 168, 182, 184, 186, 189, 193, 225, 226, 267, 268, 272, 281, 282 Andreyenkov, V.G., 33, 299 Archer, D., 133, 282 Argyle, M., 16, 289 Auerhahn, K., 8, 10, 82, 282 Athens, L., 22, 116, 282 Bachman, J., 271, 297 Baker, K., 130, 289 Baker, S.G., 198, 282 Bailey, W.C., 193, 282 Ball, R., 197, 286 Bankston, W.B., 228, 281 Barclay, G., 5, 63, 282 Baril, M., 103, 289 Barker, R.G., 16, 282 Beattie, J.M., 38, 62, 282 Beaulieu, M., 81, 295

Becker, H.S., 11, 51, 53, 262, 282 Bellucci, P., 274, 290 Ben-Yehuda, N., 6, 276, 290 Bennett, W.J., 34, 167, 177, 282 Best, J., 6, 99, 131, 276, 282, 299 Bilchik, S., 199, 274, 282 Birkbeck, C., 1, 2, 8, 16, 17, 22, 23, 59, 97, 105, 142, 260, 282, 293 Black, D.J., 184, 283 Blau, J.R., 34, 46, 193, 194, 283, 296 Blau, P.M., 34, 193, 194, 283 Block, A., 37, 292 Block, C.R., 6, 8, 21, 22, 41, 42, 48, 72, 77, 85, 102, 103, 115, 131, 137, 161–162, 197, 198, 228, 229, 230, 260, 283 Block, R.L., 6, 8, 21, 41, 48, 77, 85, 102, 103, 137, 143, 197, 228, 277, 283 Blumer, H., 16, 283 Blumstein, A., 5, 41, 93, 166, 175, 176, 181, 267, 274, 283 Boudouris, J., 6, 283 Braga, A.A., 274, 292 Brantingham, P., 6, 8, 37, 41, 63, 67, 73, 78, 81, 283 Bray, T.M., 41, 180, 197, 300 Brearley, H.C., 7, 12, 37, 38, 39, 73, 75, 76, 78, 79, 130, 284 Brewer, V.E., 41, 284 Bridenbaugh, C., 67, 284

307

21:17

P1: IYP CB617-IND

CB617-Miethe-v1

308 Brown, R.M., 5, 65, 67, 68, 69, 70, 77, 284 Browne, A., 93, 121, 131, 132, 160–162, 204, 230, 231, 240–241, 284 Brownstein, H., 274, 290 Bugliosi, V., 4, 284 Buerger, M.E., 17, 18, 301 Bursik, R.J., 18, 45, 284 Campbell, J.C., 130, 284 Canada, G., 182, 284 Cantor, D., 25, 41, 284, 285 Carruthers, B.G., 10, 142, 161, 281 Chambliss, W.J., 102, 284 Chance, R., 196, 302 Chen, J., 6, 39, 284 Chilton, R., 6, 41, 284, 285 Chin, L., 193, 285 Christakos, A., 6, 8, 21, 22, 41, 72, 77, 102, 103, 115, 131, 161–162, 230, 260, 283 Clarke, R.V., 27, 102, 260, 285 Cockburn, J.S., 4, 62, 63, 285 Cohen, J., 41, 303 Cohen, M., 102, 298 Cohen, L.E., 6, 25, 26, 33, 41, 46, 48, 97, 133, 193, 196, 231, 284, 285, 293 Cork, D., 181, 283, 285 Cook, P.J., 34, 94, 165, 166, 167, 168, 169, 176, 178, 179, 190, 192, 199, 273, 277, 285 Cooley, C.H., 16, 285 Cornish, D.B., 27, 102, 260, 285 Corzine, J., 138, 194, 195, 224, 286, 292, 297 Coverdill, J.E., 10, 49, 53, 286 Craven, D., 230, 240, 290 Crawford, M., 217, 286 Cummings, S., 197, 286 Curry, G.D., 196, 197, 286, 302 Curtis, L., 217, 277, 286 Cusson, M., 103, 289 Daly, M., 90, 99, 116, 130, 131, 132, 133, 156, 157, 160–162, 182, 184, 187, 215, 217, 228, 231, 270, 279, 286, 305

November 11, 2003

AUTHOR INDEX

Dames, R., 137, 283 Damphouse, K.R., 41, 284 Davidson, R.N., 8, 16, 17, 18, 286 Dawson, J., 155, 157, 293 Dawson, M., 155, 156, 215, 217, 279, 286 Deane, G., 80, 295 Decker, S.H., 21, 40, 46, 78, 79, 102, 103, 104, 138, 197, 228, 229, 230, 237, 240–241, 286 Devitt, C., 137, 283 Deibert, G.R., 276, 286 DeLone, M., 39, 191, 304 DiIulio, J., 34, 167, 177, 282, 287 Dobrin, A., 6, 40, 287 Doerner, W., 277, 287 Donoghue, E., 137, 283 Doyle, D., 2, 294 Drago, H.S., 5, 71, 287 Drass, K.A., 6, 10, 21, 42, 47, 50, 53, 54, 55, 63, 101, 141, 142, 161, 298 Dugan, L., 230, 231, 240, 241, 287 Dutton, D., 93, 131, 204, 231, 284 Dykstra, R.R., 12, 37, 287 Earls, F., 194, 301 Eckberg, D., 39, 287 Egley, A., 41, 180, 197, 300 Elie, D., 103, 289 Elikann, P.T., 176, 287 Erickson, K.T., 38, 65, 287 Fagan, J., 1, 94, 168, 180, 181, 182, 184, 189, 190, 274, 276, 287, 304 Farrell, R., 38, 303 Farrington, D.P., 168, 287 Fattah, E., 11, 287 Federal Bureau of Investigation (FBI), 5, 7, 8, 12, 19, 34, 40, 62, 72, 73, 74, 121, 130, 165, 166, 168, 170, 171, 175, 191, 192, 200, 202, 205, 223, 231, 232, 241, 287, 288 Feeney, F., 103, 288 Felson, M., 6, 25, 26, 33, 46, 97, 133, 231, 285 Felson, R.B., 21, 22, 25, 101, 103, 104, 127, 182, 221, 229, 241, 277, 288, 303

21:17

P1: IYP CB617-IND

CB617-Miethe-v1

November 11, 2003

309

AUTHOR INDEX

Ferdinand, T.N., 40, 288 Ferracuti, F., 133, 193, 268, 305 Finlay, W., 10, 49, 53, 286 Fishbein, D., 270, 288 Flewelling, R.L., 40, 42, 46, 79, 81, 138, 228, 229, 237, 288, 304 Flynn, N., 195, 301 Forde, D.R., 1, 292 Fox, J.A., 74, 166, 167, 169, 177, 192, 199, 230, 240, 273, 288, 289, 290 Fox, W., 54, 289 Frank, G., 4, 289 Friedman, D., 177, 289 Friedman, L.M., 5, 71, 289 Furnham, A., 16, 289 Gabor, T., 103, 289 Garofalo, J., 18, 33, 133, 231, 291 Gartin, P., 17, 18, 301 Gartner, R., 130, 155, 156, 215, 217, 279, 286, 289 Geerken, M., 41, 290 Gelles, R.J., 4, 102, 289, 303 Georges-Abeyie, D., 191, 289 Gest, T., 177, 289 Gibbs, J.T., 181, 225, 289 Gillis, A.R., 231, 289 Gilmore, D.D., 225, 289 Given, J.B., 5, 62, 289 Godwin, G.M., 129, 289 Goetting, A., 4, 121, 130, 132, 289, 290 Goffman, E., 16, 184, 290 Golden, R., 193, 194, 195, 225, 295 Goldstein, P.J., 274, 290 Gottfredson, D., 229, 290 Gottfredson, M.R., 18, 33, 127, 133, 184, 229, 231, 259, 260, 270, 290, 291 Goode, E., 6, 276, 290 Goode, W., 231, 290 Gordon, M., 41, 295 Gove, W.R., 41, 290 Greenfield, L.A., 230, 240, 290 Gregware, P., 49, 53, 121, 132, 154, 159–161, 162, 292, 297

Griffin, L., 49, 290 Groves, W.B., 18, 45, 193, 300 Gump, P.V., 16, 290 Gurr, T.R., 4, 5, 38, 39, 62, 63, 71, 290 Halfmann, D., 10, 53, 281 Hallowell, L., 37, 292 Hanawalt, B.A., 62, 290 Hannerz. U., 217, 290 Harer, M.D., 194, 195, 224, 291 Harries, K.D., 7, 40, 73, 277, 291 Hartnagel, T., 130, 291 Harvey, W.B., 193, 194, 197, 268, 291 Hawkins, D.F., 168, 191, 192, 194, 195, 291, 292 Hawley, A., 16, 291 Haynie, D., 131, 303 Heide, K.M., 94, 168, 291 Hepburn, J., 22, 41, 291, 304 Hewitt, J.D., 138, 291 Hicks, A., 53, 54, 291 Hill, R.H., 193, 294 Hindelang, M.S., 18, 25, 26, 33, 133, 231, 291 Hindus, M.S., 38, 67, 291 Hirschi, T., 127, 133, 184, 259, 260, 270, 290 Hoffer, N., 197, 301 Holinger, P.C., 6, 291 Horowitz, R., 22, 193, 218, 291 Hough, M., 26, 27, 292 Howell, J.C., 168, 292 Huff-Corzine, L., 138, 194, 195, 224, 286, 292, 297 Hughes, M., 18, 41, 45, 290, 296 Humphrey, A., 195, 292 Inciardi, J., 37, 292 Jacobs, W.R., 195, 196, 292 Jackson, M., 132, 304 Jackson, P., 6, 196, 292 Jacobs, D., 68, 292 Jamieson, K.M., 6, 41, 305 Jenkins, P., 99, 292 Jensen, V., 131, 292

21:17

P1: IYP CB617-IND

CB617-Miethe-v1

310 Johnson, L., 271, 297 Jones, A., 130, 292 Julian, S., 197, 301 Jurik, N., 49, 121, 130, 131, 132, 143, 154, 159, 161–162, 292 Kane, C., 196, 302 Kaplun, D., 131, 292 Katz, J., 22, 72, 121, 292 Kempf-Leonard, C., 197, 286 Kennedy, D.M., 274, 292 Kennedy, L.W., 1, 3, 6, 8, 27, 40, 41, 73, 81, 105, 130, 131, 217, 227, 230, 231, 240–241, 271, 279, 292, 295, 300, 302 Kindley, K., 229, 294 Klaus, P., 8, 83, 230, 240, 290, 297 Kleck, G., 34, 272, 274, 277, 292, 293 Klein, M.W., 41, 196, 197, 293, 295 Klemen, E., 6, 291 Kluegel, J., 25, 285 Knight, R., 48, 102 Kobrin, S., 18, 301 Kornhauser, R., 45, 226, 293 Kovandzic, T.V., 46, 293 Krivo, L.J., 45, 47, 193, 194, 195, 293, 298 Krohn, M., 271, 295 Kumar, K., 116, 301 LaFree, G.D., 1, 2, 6, 12, 16, 17, 22, 23, 32, 33, 37, 41, 47, 59, 63, 97, 105, 142, 192, 193, 194, 195, 224, 225, 260, 278, 282, 293, 300 Land, K.C., 25, 37, 46, 48, 193, 196, 285, 293 Lane, R., 4, 6, 12, 37, 38, 39, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 80, 81, 83, 125, 130, 131, 166, 230, 293 Langan, P.A., 155, 157, 293 Laub, J.H., 94, 165, 166, 167, 168, 169, 176, 178, 179, 190, 192, 199, 273, 285 Lauferweiler, D., 197, 301 Lauritsen, J., 15, 16, 18, 34, 76, 167, 300 LeBlanc, M., 103, 289

November 11, 2003

AUTHOR INDEX

Lee, A., 102, 300 Lee, M., 7, 73, 198, 215, 294 Leyba, C., 198, 293 Liska, A., 271, 295 Lofland, J., 16, 23, 24, 25, 26, 97, 134, 135, 138, 159–161, 261, 270, 273, 293 Loftin, C., 6, 40, 193, 229, 287, 294 Logan, J.R., 194, 294 Long, J.S., 25, 296 Loya, F., 198, 294 Luckenbill, D.F., 2, 8, 22, 40, 49, 90, 116, 121, 132, 146, 159–162, 182, 220, 221, 241, 294 Lund, L.E., 294 Lundsgaarde, H.P., 41, 157, 159, 229, 294 Lynch, J.P., 17, 18, 294 MacDonald, J.M., 103, 294 Maier, P., 18, 299 Maiken, P., 4, 303 Maltz, M.D., 42, 294 Mann, C.R., 143, 294 Martin, J., 10, 49, 53, 286 Martinez, R., 7, 15, 73, 198, 215, 294 Maston, C., 230, 240, 290 Maume, M.O., 195, 301 Maxfield, M.G., 18, 40, 42, 103, 104, 121, 138, 154, 228, 229, 230, 294, 302 Maxson, C.L., 41, 197, 295 Mayer, S., 53, 142, 298 McCall, P.L., 6, 39, 40, 48, 72, 166, 167, 193, 194, 195, 196, 225, 293, 297, 305 McClain, P., 12, 195, 299 McCorkle, R.C., 6, 8, 48, 69, 83, 85, 102, 141, 161, 296 McDaniel, P., 133, 282 McDowall, D., 6, 18, 40, 45, 287, 296 McElrath, K., 273, 293 McGrath, R., 5, 71, 295 McKanna, C.V., 5, 12, 37, 38, 39, 71, 75, 78, 82, 130, 193, 295 McKay, H., 17, 267, 301 McMillan R., 192, 193, 195, 295 Mead, G.H., 16, 295

21:17

P1: IYP CB617-IND

CB617-Miethe-v1

November 11, 2003

311

AUTHOR INDEX

Meier, R.F., 1, 3, 8, 16, 18, 19, 25, 26, 27, 45, 46, 105, 127, 193, 231, 271, 295, 296 Mercy, J.A., 198, 294, 295 Merighi, J., 181, 225, 289 Messerschmidt, J.W., 99, 225, 295 Messner, S.F., 15, 31, 46, 47, 80, 192, 193, 194, 195, 225, 228, 229, 271, 277, 288, 294, 295, 296 Miethe, T.D., 1, 6, 8, 10, 16, 18, 19, 21, 25, 26, 27, 39, 41, 42, 45, 46, 48, 50, 53, 54, 55, 69, 71, 83, 85, 101–141, 142, 161, 298 Mischel, W., 16, 296 Monkkonen, E.H., 6, 65, 75, 76, 77, 230, 296 Monti, D., 197, 286 Moore, C.A., 39, 196, 262, 296 Moore, D., 194, 224, 292 Moore, M.H., 34, 45, 165, 168, 285, 296 Mosher, C.J., 41, 71, 297 Mouzos, J., 73, 279, 297 Mucheno, M.C., 53, 297 Mukherjee, S.K., 137, 217, 302, 306 Nagin, D., 230, 231, 240, 241, 287 Nash, J.R., 4, 297 National Center for Health Statistics (NCHS), 8, 83, 191, 297 Normandeau, A., 103, 289 Norris, S., 229, 294 O’Brien, R., 34, 297 O’Day, P., 192, 194, 225, 293 Oh, S., 196, 302 Oliver, W., 121, 182, 193, 217, 221, 268, 271, 276, 297 O’Malley, P., 271, 297 Ousey, G.C., 193, 194, 297 Osgood, D.W., 271, 297 Palmer, S., 195, 292 Pampel, F.C., 81, 130, 289, 297 Parker, K.F., 193, 194, 195, 225, 297 Parker, R.N., 8, 82, 102, 193, 228, 229, 282, 297, 302

Patrushev, V.D., 33, 299 Percival, R., 5, 71, 289 Perkins, C., 8, 83, 230, 240, 290, 297 Pervin, L.A., 16, 297 Petee, T.A., 138, 297 Petersilia, J., 102, 298 Peterson, R.D., 45, 47, 193, 194, 195, 293, 298 Phillips, C., 183, 304 Phillips D., 41, 71, 297 Phillips, J.A., 193, 194, 196, 298 Piehl, A.M., 274, 292 Plint, T., 38, 298 Poe-Yamagata, E., 5, 166, 167, 169, 302 Polk, K., 21, 23, 39, 41, 49, 73, 81, 90, 99, 101, 103, 116, 121, 127, 131, 132, 133, 138, 146, 160–162, 176, 182, 184, 186, 187, 215, 217, 220, 221, 225, 227, 228, 237, 241, 260, 272, 279, 281, 298 Pokorny, A., 198, 298 Poulsen, J.D., 10, 51, 53, 281 Prassel, F.R., 5, 71, 298 Prentky, R., 48, 102, 298, 300 Przybylski, R., 228, 229, 230, 299 Quetelot, L.A., 38, 298 Raffalovich, L., 192, 193, 195, 295 Ragin, C.C., 10, 11, 47, 48, 49, 50, 51, 53, 55, 90, 142, 146, 262, 277, 287, 290, 298 Rand, M., 230, 240, 290 Rasche, C.E., 156, 298 Raudenbush, S., 194, 301 Redfield, H.V., 12, 37, 298 Rebhun, L., 8, 82, 297 Regoeczi, W.C., 43, 48, 54, 55, 81, 298, 299 Reich, R., 131, 292 Reiman, J.H., 39, 299 Reiss, A.J., 2, 8, 192, 299 Ribner, S., 182, 221, 241, 288 Riedel, M., 5, 6, 37, 40, 41, 43, 46, 48, 71, 79, 81, 131, 228, 229, 230, 237, 298, 299

21:17

P1: IYP CB617-IND

CB617-Miethe-v1

312 Ringel, C., 230, 240, 290 Robinson, J.P., 33, 299 Rodriguez, O., 198, 299 Rojek, D.G., 228, 229, 299 Roncek, D.W., 18, 299 Rosa, J.G., 5, 71, 299 Rose, H.M., 12, 102, 195, 197, 204, 299 Rosenberg, R., 48, 102, 300 Rosenfeld, R., 15, 31, 41, 132, 135, 180, 183, 197, 228, 230, 231, 240 Roth, J., 2, 8, 192, 299 Rudman, C., 6, 196, 292 Rule, A., 4, 300 Ryan, P.J., 274, 290 Sacco, V.F., 1, 3, 8, 27, 105, 292, 295, 300 Sagi, P.C., 229, 230, 240–241, 305 Sampson, R.J., 7, 15, 16, 18, 25, 34, 45, 47, 75, 167, 193, 194, 195, 197, 228, 296, 300, 301 Sanders, W.B., 197, 301 Savitz, L., 116, 301 Savolainen, J., 229, 296 Schafer, J.L., 43, 301 Schwartz, G., 22, 291 Schwartz, T., 4, 301 Schuerman, L., 18, 301 Seghorn, T., 102, 298 Shaw, C., 17, 193, 301 Sherman, L.W., 17, 18, 197, 301 Shibutani, T., 16, 301 Shihadeh, E.S., 195, 301 Short, J.F., Jr., 2, 4, 7, 15, 16, 30, 73, 179, 180, 193, 197, 270, 276, 301, 302 Sickmund, M., 5, 166, 167, 169, 302 Siddique, A., 5, 63, 282 Siegel, L., 102, 221, 302 Siegel, M., 182, 241, 288 Silverman, I., 132, 302 Silverman, R., 6, 40, 41, 73, 81, 130, 131, 132, 217, 227, 230, 231, 241, 279, 302 Simmons, E., 196, 302 Simon, R.J., 130, 302 Singer, S.I., 7, 8, 34, 76, 167, 302 Skocpol, T., 49, 302 Skogan, W.G., 18, 302

November 11, 2003

AUTHOR INDEX

Sloan, J.J., III, 46, 302 Smith, M.D., 2, 12, 15, 30, 39, 72, 102, 193, 195, 225, 228, 229, 297, 302 Snipes, J.B., 11, 303 Snyder, H., 5, 166, 167, 169, 302 Speir, J., 277, 287 Spergel, I., 196, 302 Spencer, J.W., 53, 287 Spohn, C., 39, 191, 304 Stafford, M., 25, 296 Stahura, J.M., 46, 302 Steadman, H., 22, 288 Steel, E.M., 68, 302 Steele, L., 197, 301 Steffensmeier, D., 131, 194, 195, 291, 301, 303 Steinmetz, S., 4, 303 St. Pierre, R., 228, 282 Straus, M.A., 4, 303 Strodtbeck, F.L., 179, 180, 276, 302 Sudnow, D., 38, 303 Sullivan, T., 4, 303 Sutherland, E.H., 16, 303 Swanton, B., 102, 303 Swigert, V.L., 38, 303 Tardiff, K., 228, 296 Tavares, C., 5, 63, 282 Taylor, R.B., 18, 303 Tedeschi, J.T., 25, 182, 288, 303 Tennenbaum, A.N., 40, 303 Thomas, C.W., 102, 303 Tita, G.E., 41, 303 Tomsen, S., 186, 303 Tonry, M., 165, 168, 269, 296, 303 Toth, A., 229, 297 Vega, M., 132, 302 Venkatesh, S.A., 197, 303 Vieraitis, L., 46, 293 Vold, G.B., 5, 303 von Hentig, H., 11, 304 Voss, H., 41, 304 Wallace, A., 156, 304 Walker, S., 39, 191, 304 Walters, J.P., 34, 167, 177, 282

21:27

P1: IYP CB617-IND

CB617-Miethe-v1

November 11, 2003

313

AUTHOR INDEX

Warchol, G., 230, 240, 290 Ward, D.E., 132, 304 Ward, R.E., 132, 304 Warr, M., 179, 304 Weaver, G., 138, 297 Weisheit, R., 131, 304 Weghorst, S., 187, 286 White, N., 183, 304 Whittingham, M.D., 102, 304 Wilbanks, W., 6, 9, 28, 37, 38, 39, 41, 44, 83, 124, 132, 154, 191, 195, 304 Wickham-Crowley, T.P., 53, 304 Wiersema, B., 6, 40, 229, 287, 294 Wilkinson, D.L., 1, 94, 168, 181, 182, 184, 189, 190, 274, 276, 287, 304 Williams, J., 102, 228, 229, 299, 303 Williams, K.R., 40, 42, 79, 81, 93, 131, 138, 204, 228, 229, 230, 237, 240–241, 284, 288, 297, 304 Wilson, J.K., 271, 297 Wilson, M., 90, 91, 99, 116, 130, 131, 132, 133, 156, 157, 160–162, 182, 184, 187, 215, 217, 228, 270, 279, 286, 305

Wilson, W.J., 12, 31, 197, 204, 225, 267, 268, 272, 301, 305 Wilson, J.Q., 32, 305 Winkle, B.V., 137, 306 Winn, R., 130, 131, 143, 292 Wittekind, J., 138, 297 Wolfgang, M., 7, 8, 20, 38, 40, 73, 80, 81, 83, 90, 118, 119, 130, 131, 132, 133, 143, 193, 217, 227, 228, 230, 268, 305 Wood, K., 195, 196, 292 Wooldredge, J., 25, 301 Wright, C., 131, 217, 305 Yeisley, M., 46, 293 Zahn, M.A., 2, 6, 12, 15, 30, 39, 40, 41, 72, 116, 166, 167, 198, 229, 230, 240–241, 299, 301, 302, 305 Zatz, M.S., 6, 196, 305 Zimring, F.E., 41, 78, 102, 137, 143, 168, 277, 283, 306 Zylan, Y., 10, 142, 281

21:27

P1: IYP CB617-IND

CB617-Miethe-v1

November 11, 2003

21:27

P1: IYP CB617-INDS

CB617-Miethe-v1

November 11, 2003

Subject Index

Age patterns in homicide, 165–190 correlates of youth violence, 166–168 dominant homicide situations, 168, 169–173, 174, 176, 188, 258 drug market violence, 175, 180–181 emergent forms of youth homicide, 175, 178, 188 group context of youth violence, 178, 179–180, 181–182, 184–186, 188 guns and violence, 167, 169, 170, 189, 190 juvenile “superpredators,” 167, 176–179, 188 historical patterns, 166–167, 170 masculinity and youth homicide, 181, 182, 186, 188, 189 multiple offenders in youth homicide, 167, 169, 170 narrative descriptions of, 179–187 offender trends, 7, 166 policy and prevention efforts, 165, 190 research questions, 168 situational context differences for young and older juveniles, 175–176, 258 stability and change over time, 174–175, 258 summary of current findings, 187–188

unique and common situational profiles for, 169–173, 174, 258 victim trends, 166–167 youth homicide and motive, 166–167, 169 Alcohol and drug use. See Offense characteristics in homicide “Bachelor” subculture, 66, 69, 75, 166 Canada, 8, 40, 63, 73, 168, 279 Case studies, 3, 4, 10, 49 “Character contests,” 22, 71, 90, 116–119, 125, 128, 132, 146, 161, 162, 188, 221, 248, 254, 257, 272 see also Masculinity and violence; Confrontational homicides Clearance rates 41 “Code of the street,” 184, 189, 267 Coding of variables, 55, 83–84, 105, 138–141, 168, 202, 228 Combinatorial and conjunctive approaches for studying homicide situations, 23–28, 51, 53, 59, 91, 132, 134–136, 137, 161, 162, 269–272 Concentrated disadvantage, 193–194, 267, 270, 272

315

21:30

P1: IYP CB617-INDS

CB617-Miethe-v1

316 Confrontational homicides, 21, 22, 30, 90, 101, 116–119, 121, 128, 182, 184, 189, 215, 217–223, 224, 225, 248, 255, 257, 259, 264, 266, 268, 271 see also Character contests; Instrumental and expressive homicides Crime events, 25, 27–28, 105, 127, 135, 260 Dangerous places, 17–19, 26, 135 times, 17–19, 64, 135 Data sources, 5, 12–13, 36–41, 44 current study, 12–14, 41–44, 137–138, 200–201, 231–232 California Homicides Data File, 13, 42, 191, 200, 208, 210, 213, 218, 223 Chicago Homicide Project, 13, 22, 42, 55, 56, 137, 138, 139, 190, 218 court records, 4, 12, 36, 38–39, 62, 65, 66, 68, 79 homicide narratives, 3, 4, 5, 13, 41, 42, 43–44, 49, 73, 79, 116–125, 127, 246–253, 257, 258, 259, 260, 263, 273, 279 in Las Vegas, 3, 13, 42, 44, 120 in Los Angeles, 3, 13–14, 42, 43, 101, 116, 118, 120, 121–123, 157–160, 216, 263 in Miami, 3, 9, 13, 28–29, 42, 44, 124, 128, 154, 221 in St. Louis, 3, 13, 22, 42, 44, 119–120, 179–187, 218–220, 221–223, 247–253 mortality records, 12, 36, 39–40 newspaper accounts, 12, 36, 37–38 police reports, 12, 28, 36, 40–41 see also Supplementary Homicide Reports Supplementary Homicide Reports (SHR), 8, 13, 40, 41, 42–43, 83–86, 105, 165, 168, 190, 218, 223, 231, 253, 275 Defense of others, 119–121

November 11, 2003

SUBJECT INDEX

Domestic violence, 88, 154–160, 231, 269, 271 gender differences in, 91, 147, 153, 154–160 histories of past abuse, 155, 156, 157 narrative descriptions of, 154–160 prevalence of, 4, 147, 153, 231, 272–273 structural profiles of, 91, 147, 153, 272–273 stable and unique situational contexts for, 91, 153, 157–160, 263, 272–273 see also Victim–offender relationships; Race and ethnic patterns, intimate partner homicide Drugs and homicide drug market violence, 175, 180–181, 267, 269, 271, 274–275 narrative descriptions of, 124, 154, 180–181, 249 see also Offense characteristics, alcohol and drug use England, 4–5, 8, 61, 62–64, 73, 78 Expressive homicides. See Instrumental and expressive homicides; Confrontational homicides Facilitating hardware, 23, 24, 25, 97, 134, 135, 136, 162, 260, 273, 274 places, 23, 24, 25, 97, 134, 135, 136, 162 others, 23, 24–25, 97, 134, 135, 136, 162, 260 Feuding in American history, 68, 70 Gangs, 20, 21, 22, 75, 81, 103, 115, 118–119, 180, 183, 189, 196–197, 274 youth violence and, 21, 84, 99, 103, 115, 125, 126, 179, 180, 196 instrumental and expressive motives, 21, 84, 103, 115, 126, 127, 257 Gender patterns in homicide, 130–164 arrest trends, 7, 73, 130

21:30

P1: IYP CB617-INDS

CB617-Miethe-v1

SUBJECT INDEX

correlates of female offending, 132, 146–147 dominant situations, 143–153, 257–258 emergent forms of homicide situations, 147–153 historical patterns, 130, 131 instrumental homicides and, 153–154 intimate partners and domestic violence, 131, 154–160 male possessiveness and proprietariness, 21, 156, 159, 215–217 narrative descriptions of, 153–160 offender trends, 7, 73, 130 policy implications, 163–164 research questions about, 136–137 situational dynamics, 153–160 social changes and, 147–153 stability and change in structure over time, 147–153 theoretical perspectives, 132–136 unique and common structures, 143–153 see also Masculinity and violence Guns and lethal violence, 273–274 historical trends, 34–35, 81 see also Offense characteristics; Age patterns in homicide; Guns and lethal violence Historical trends and homicide, 61–72 American history, 64–72 seventeenth and eighteenth century, 65–68 nineteenth century, 68–71 twentieth century, 71–72 England, 62–64 unique situations in the 1970s and 1980s, 92–93 unique situations in the 1990s, 93–94 Honor contests. See Character contests; Confrontational homicides Infanticide, 73 Instrumental and expressive homicides, 101–129

November 11, 2003

317 changes over time in unique profiles of, 112–114 confrontational homicides and character contests, 116–119 crime prevention implications of, 102 definitions of, 101–102, 103–104 dominant structures, 109–112 historical trends, 77–79 internal consistency of classifying, 114–116, 257 narrative descriptions of, 116–125 policy implications, 128–129 prevalence of, 105 research questions, 104 summary of current findings, 126–127 unique and common structures of, 107–114, 257 usefulness of classification of, 102 Interaction effects in statistical models, 6, 36–45, 47–48 as a specification of functional form Interpersonal disputes and arguments. See Confrontational homicides; Character contests; Instrumental and expressive homicides Masculinity and violence, 23, 90, 91, 98–99, 163, 181, 182, 186, 188, 189, 215, 217, 218–223, 225, 226, 258, 264, 266, 267, 271, 279 masculine competitiveness, 21–22, 98, 131, 134, 189, 215, 218–223, 261, 270, 272, 276 see also Gender patterns in homicide Main effect model, 6, 10–14, 36–45, 47, 59, 91, 92, 108, 109–112, 132, 196, 261–262 functional form, 45, 97, 132, 167 limitations of, 47, 97, 109, 114, 132 Missing data, 43, 83, 137 imputation methods for, 43, 83, 137 variables used in imputation, 43 trends without, 105, 137

21:30

P1: IYP CB617-INDS

CB617-Miethe-v1

318 Motive and circumstances, see also Instrumental and expressive homicides; Confrontational homicides; Character contests; Robbery homicide Multiple offenders, 81–82 see also Offense characteristics; Age patterns in homicide, group context of youth violence Myths and misconceptions, 259–265 common or unique signatures, 264–265 homicide as simple or complex event, 259–261 main effects or context-specific effects, 261–262 stability or change in homicide, 262–263 Native American violence, 66–68, 70, 71 Offender characteristics, 8, 61–75, 83, 105, 138, 168, 200 cross-cultural patterns, 62–63, 64, 72–75, 278–279 historical patterns, 61–75 social class of, 74–75, 76, 189, 197, 217, 225, 267, 272–273 see also Age patterns in homicide; Gender patterns in homicide Offense characteristics in homicide, 9, 12, 26, 61–72, 77–83, 84, 138, 141, 166–167, 168, 232 alcohol and drug use, 8, 12, 19, 20, 28, 29, 42, 69, 72, 77, 82, 115, 118, 125, 154, 159, 248, 257 historical patterns, 4–5, 61–72, 77–83, 166–167, 168 intraracial and interracial attacks, 81 number of offenders, 81–82 see also Age patterns in homicide, group context for youth violence number of victims, 81–82 spatial and temporal location, 4–5, 82–83 weapon used, 81, 273–274 see also Guns and lethal violence

November 11, 2003

SUBJECT INDEX

Public policy, 99, 102, 128–129, 163–164, 272–276 drug-related homicides, 274–275 focus on exceptional cases, 163, 275–276 issues of gender, class, and race, 163–164, 272–273 guns and lethal violence, 273–274 Qualitative approaches for studying homicide, 49–50, 60 Qualitative Comparative Analysis (QCA), 10–11, 36, 50–60, 87, 105–106, 277–278 advantages of, 53–54, 60, 278 definitions of unique and common profiles, 54–55, 58, 106 deterministic and relative rules, 54–55, 58–59, 142, 146 illustration of, 52–53, 55–60 limitations of, 54, 277–278 linguistic qualifiers in the interpretation of, 90 logic of, 50–53 methodological issues in, 53–55 minimum cell frequency rules in, 55, 59–60, 84, 85, 87, 88, 106, 142, 146, 147, 152, 171, 232 previous applications of, 10–11, 53 software for, 52 truth tables in, 52, 55–59 Quantitative studies homicide trends and correlates in, 6–7 Race and ethnic patterns in homicide, 191–226 aggregate studies of, 193, 194–195 arrest trends, 7, 74 changes in homicide situations over time, 196–197 confrontational homicides and, 217–223 correlates of, 193–195 dominant profiles, 258 emergent structures, 113–114

21:30

P1: IYP CB617-INDS

CB617-Miethe-v1

SUBJECT INDEX

Hispanic homicides characteristics of, 198, 208–209 historical patterns, 74, 75 instrumental and expressive motives, 113–114 intimate partner homicides and, 215–217 intraracial and interracial patterns, 195–196 narrative descriptions of, 215–223 offender trends, 74, 75, 192, 201 regional differences in, 195–196 research questions, 199–200 situational approach, 199 stability and change in structures over time, 201–208, 223–224 summary of current findings, 223–224 theoretical explanations, 193–194 unique and common situational contexts, 201–215, 223–224, 226 Black vs. White offenders, 201–208 Hispanic vs. White offenders, 209–212 Hispanic vs. Black offenders, 213–215 victim trends Regional differences, 70–71, 102, 197, 204 Robbery homicide frequency of, 78–79 historical trends in, 63, 67, 78–79, 263 narrative accounts of, 121–125, 185–186, 249–250 see also Instrumental and expressive homicides Situational clustering, 59, 60, 98, 143–146, 152, 161, 201, 213, 236, 237, 253, 256–257, 259 Situations analysis of, 1–3, 8–10 definitions of, 8–9, 16–30 diversity of homicide, 9, 20–28, 85–86, 107 dominant profiles of homicide, 88, 89, 110–111, 135, 144–145,

November 11, 2003

319 148–149, 150–151, 169–173, 174, 205–206, 210–211, 234–235, 236, 242 structure of homicide, 8–10, 11–12, 29–30, 104–105, 240–241 stability and change in structures over time, 87, 88–94, 112–114, 148–149, 150–151, 174, 203, 205–206, 241, 242 trends over time in homicide, 26, 87–94, 104–105, 240–241 unique and common profiles in homicide, 26, 86, 87–89, 94, 107–111, 112, 115, 135, 144–145, 148–149, 150–151, 172–173, 174, 203, 205–206, 209, 210–211, 213, 233, 234–235, 236, 238–239, 242, 243, 244–245, 264–265 see also Age patterns in homicide; Gender patterns in homicide; Instrumental and expressive homicides; Race and ethnic patterns in homicide; Victim–offender relationship Signatures, 20, 28, 30, 38, 52, 264–265 See also Situations, unique and common profiles Supplementary Homicide Reports. See Data sources Theoretical explanations, 2, 15–33, 94–99, 127–128, 132–136, 161–162, 193–194, 230–231, 265–272 anomie-strain, 31–32, 98, 266–267 criminal opportunity, 25–27, 33–35, 97–98, 135–136, 161–162, 193–194, 265–266 combinatorial or conjunctive theories, 23–28, 161–162, 269–272 declines in legitimacy of social institutions, 32–33 macro and micro approaches, 30–35, 193–194 rational choice theories, 269

21:30

P1: IYP CB617-INDS

CB617-Miethe-v1

320 Theoretical (cont.) routine activities and lifestyles, 25–27, 33–35, 97–98, 135–136, 161–162, 193–194, 265–266 self-control and general theory of crime, 127–128, 133, 270 social disorganization, 17–18, 135–136, 193–194, 265–266, 267–268 subculture of violence, 134–135, 193, 268–269 symbolic interactionism, 16–17, 23–25, 97, 134–135, 161–162, 193 Third parties, 21, 24–25, 27, 159, 189, 221, 253, 276 see also Facilitating, others Typologies and syndromes of homicide, 11, 19–22, 23, 48–49, 50, 53, 85–86, 102, 129, 161, 260 Urban violence, 63, 68, 83, 84, 94 historical patterns in, 63, 69, 70, 83 Victim characteristics, 9, 12, 75–77, 83, 138, 168, 201 see also Age patterns in homicide; Gender patterns in homicide; Race and ethnic patterns in homicide; Victim–offender relationship Victim–offender relationship, 79–80, 227–255 acquaintances, 79–80, 241–246 classification of, 228–230, 232 correlates of homicide differences by, 229 dominant structural profiles, 232–233, 259 acquaintance homicides, 233–236, 237, 241–246, 259

November 11, 2003

SUBJECT INDEX

intimate/family homicides, 104–105, 233, 237–241, 259 stranger homicides, 236–237, 241–246, 259 family and intimate partners, 59, 79–80, 104–105, 132, 231, 237–241, 246–249 explanations for, 230–231 gender differences in, 229 historical patterns, 79–80, 230 narrative description of, 12 family homicides, 246–249 stranger homicides, 249–253 stability and change in structure of homicide situations intimate partners, 104–105, 240–241, 254 stranger homicides, 79–80, 241–246, 249–253 explanations for, 230–231 summary of current findings, 253–254 trends over time, 79–80, 104–105, 230–231, 240–241 unique and common structures, 232–233, 253–254, 255 acquaintance homicides, 233–236, 237, 241–246 intimate/family homicides, 104–105, 233, 237–241 stranger homicides, 236–237, 241–246 Victim precipitation, 12, 20, 23, 157 Vigilante groups, 4, 67, 68, 69, 70, 71, 81 Western frontier violence, 4, 5, 39, 68, 69, 70–71, 82 Youth homicide, 165–190 see also Age patterns in homicide

21:30