Building Methodological Bridges

BUILDING METHODOLOGICAL BRIDGES RESEARCH METHODOLOGY IN STRATEGY AND MANAGEMENT Series Editors: David J. Ketchen, Jr...

Author: Donald D. Bergh and David J. Ketchen (editors)

149 downloads 953 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

BUILDING METHODOLOGICAL BRIDGES

RESEARCH METHODOLOGY IN STRATEGY AND MANAGEMENT Series Editors: David J. Ketchen, Jr. and Donald D. Bergh Previous Volumes: Volumes 1–5: Research Methodology in Strategy and Management – Edited by David J. Ketchen, Jr. and Donald D. Bergh

RESEARCH METHODOLOGY IN STRATEGY AND MANAGEMENT VOLUME 6

BUILDING METHODOLOGICAL BRIDGES EDITED BY

DONALD D. BERGH University of Denver, USA

DAVID J. KETCHEN, Jr. Auburn University, USA

United Kingdom – North America – Japan India – Malaysia – China

Emerald Group Publishing Limited Howard House, Wagon Lane, Bingley BD16 1WA, UK First edition 2011 Copyright r 2011 Emerald Group Publishing Limited Reprints and permission service Contact: [email protected] No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. No responsibility is accepted for the accuracy of information contained in the text, illustrations or advertisements. The opinions expressed in these chapters are not necessarily those of the Editor or the publisher. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-1-78052-026-1 ISSN: 1479-8387 (Series)

Emerald Group Publishing Limited, Howard House, Environmental Management System has been certified by ISOQAR to ISO 14001:2004 standards Awarded in recognition of Emerald’s production department’s adherence to quality systems and processes when preparing scholarly journals for print

CONTENTS LIST OF CONTRIBUTORS

vii

INTRODUCTION: BUILDING METHODOLOGICAL BRIDGES

ix

PART I: BRIDGES BETWEEN MACRO AND MICRO RESEARCH REDISCOVERING THE INDIVIDUAL IN STRATEGY: METHODOLOGICAL CHALLENGES, STRATEGIES, AND PROSPECTS Steven W. Floyd and Rebekka Sputtek INTO THE GREAT WIDE OPEN: BRIDGING THE MICRO–MACRO DIVIDE IN THE ORGANIZATIONAL SCIENCES M. Ronald Buckley, Maria Riaz Hamdani, Anthony C. Klotz and Sorin Valcea

3

31

STRATEGIC HUMAN RESOURCE STAFFING AND ORGANIZATION RESEARCH: ARE THEY ONE-SIZEFITS-ALL ENDEAVORS? Philip Bobko and Denise Potosky

69

LEADER–MEMBER EXCHANGE (LMX) MEASUREMENT: EVIDENCE FOR CONSENSUS, CONSTRUCT BREADTH, AND DISCRIMINANT VALIDITY Dana L. Joseph, Daniel A. Newman and Hock-Peng Sin

89

ENTREPRENEURIAL MOBILITY Mike Wright v

137

vi

CONTENTS

PART II: BRIDGES TO STRONGER DESIGNS AND ANALYSIS MULTI-STUDY PACKAGES IN ORGANIZATIONAL SCIENCE RESEARCH Wayne A. Hochwarter, Gerald R. Ferris and T. Johnston Hanes

163

TEMPLATES AND TURNS IN QUALITATIVE STUDIES OF STRATEGY AND MANAGEMENT Ann Langley and Chahrazad Abdallah

201

THE REVOLUTION WITH A SOLUTION: ALL IS NOT QUIET ON THE STATISTICAL AND METHODOLOGICAL MYTHS AND URBAN LEGENDS FRONT Robert J. Vandenberg

237

QUALITATIVE COMPARATIVE ANALYSIS AND STRATEGIC MANAGEMENT RESEARCH: CURRENT STATE AND FUTURE PROSPECTS Thomas Greckhamer and Kevin W. Mossholder

259

WORTH A THOUSAND WORDS: PHOTOGRAPHS AS A NOVEL METHODOLOGICAL TOOL IN STRATEGIC MANAGEMENT Joshua L. Ray and Anne D. Smith

289

LIST OF CONTRIBUTORS Chahrazad Abdallah

Birkbeck College, University of London, London, UK

Philip Bobko

Departments of Management and Psychology, Gettysburg College, Gettysburg, PA, USA

M. Ronald Buckley

Division of Management and Entrepreneurship, Price College of Business, University of Oklahoma, Norman, OK, USA

Gerald R. Ferris

Department of Management, Florida State University, Tallahassee, FL, USA

Steven W. Floyd

McIntire School of Commerce, University of Virginia, Charlottesville, VA, USA

Thomas Greckhamer

Ourso College of Business, Louisiana State University, Baton Rouge, LA, USA

Maria Riaz Hamdani

Department of Management, College of Business Administration, University of Akron, Akron, OH, USA

T. Johnston Hanes

School of Labor and Employment Relations, University of Illinois at Urbana-Champaign, Champaign, IL, USA

Wayne A. Hochwarter

Department of Management, Florida State University, Tallahassee, FL, USA

Dana L. Joseph

Department of Psychology, University of Illinois at Urbana-Champaign, Champaign, IL, USA

vii

viii

LIST OF CONTRIBUTORS

Anthony C. Klotz

Division of Management and Entrepreneurship, Price College of Business, University of Oklahoma, Norman, OK, USA

Ann Langley

HEC Montre´al, Montre´al, Canada

Kevin W. Mossholder

College of Business, Auburn University, Auburn, AL, USA

Daniel A. Newman

Department of Psychology, University of Illinois at Urbana-Champaign, Champaign, IL, USA

Denise Potosky

Great Valley School of Graduate Professional Studies, Management Division, Pennsylvania State University Malvern, PA, USA

Joshua L. Ray

Department of Management, College of Business, University of Tennessee, Knoxville, TN, USA

Hock-Peng Sin

Department of Management, The Eli Broad Graduate School of Management, Michigan State University, East Lansing, MI, USA

Anne D. Smith

Department of Management, College of Business, University of Tennessee, Knoxville, TN, USA

Rebekka Sputtek

Institute of Management, University of St. Gallen, St. Gallen, Switzerland

Sorin Valcea

Department of Management, School of Business, Washburn University, Topeka, KS, USA

Robert J. Vandenberg

Department of Management, Terry College of Business, University of Georgia, Athens, GA, USA

Mike Wright

Center for Management Buyout Research, University of Nottingham, Nottingham, UK

INTRODUCTION: BUILDING METHODOLOGICAL BRIDGES Welcome to volume 6 of Research Methodology in Strategy and Management! In creating this series about eight years ago, our belief was that the organizational sciences needed a forum wherein leading scholars could openly express their views about important and emerging issues within research methods. In particular, we wanted the book series to serve as a metaphorical bridge between areas of inquiry that could beneﬁt from increased interaction with each other. This sixth volume of the series recalls these roots by being built around the theme of ‘‘Building methodological bridges.’’ Within this overarching theme, this volume includes two subthemes. The ﬁrst is ‘‘Bridges between Macro and Micro Research.’’ We were fortunate to be able to attract ﬁve excellent chapters that contribute to this theme. Steven W. Floyd and Rebekka Sputtek outline key challenges in capturing the role of individuals within strategy making, as well as providing useful responses to these challenges. M. Ronald Buckley, Maria Riaz Hamdani, Anthony C. Klotz, and Sorin Valcea strike a complementary chord in describing ways that scholars can bridge the micro–macro divide that characterizes much of the organizational sciences. Philip Bobko and Denise Potosky tackle key issues within one of the areas where the micro–macro divide has been bridged most frequently within past inquiry: strategic human resource management. Leadership is another domain within which micro and macro ideas often come together. Dana L. Joseph, Daniel A. Newman, and HockPeng Sin contribute to understanding how to capture one of the leading perspectives on leadership: Leader–Member Exchange. In a ﬁnal piece, Mike Wright examines entrepreneurial mobility, a topic that inherently requires integration of macro and micro ideas. A study’s conclusions can only be as robust as the design and analysis that led to its conclusions. With this in mind, the second subtheme of this volume is ‘‘Bridges to Stronger Designs and Analysis.’’ Five outstanding contributions contribute to this theme. First, Wayne A. Hochwarter, Gerald R. Ferris and T. Johnston Hanes draw on their own vast successes to explain ix

x

INTRODUCTION

how using multi-study designs can enhance the research enterprise. Ann Langley and Chahrazad Abdallah provide insights on conducting qualitative research. Robert J. Vandenberg builds on his own experience as a renowned methodologist and former editor of Organizational Research Methods to debunk a series of myths about methods. Thomas Greckhamer and Kevin W. Mossholder outline how qualitative comparative analysis has contributed, and can continue to contribute, to strategy research. In the ﬁnal and perhaps most path-breaking chapter of this volume, Joshua L. Ray and Anne D. Smith build upon the folk wisdom that a picture is worth a thousand words. In particular, they demonstrate how photographic images can be an incredibly valuable means of building knowledge. We think this chapter will inspire scholars to add a new data source to their array of research tools. Not only do images offer an untapped dimension for understanding organizations, taking pictures is a heck of lot more fun than downloading data or coding text! As you can see, this volume builds important methodological bridges. This is evident not only through the topics covered, but also in the backgrounds of the scholars who have contributed chapters. The roster includes thought leaders not only from strategic management but also from organizational behavior, human resource management, and entrepreneurship. Looking to future volumes, we have decided to expand upon our general theme of bridging by reaching out to and enlisting volume editors who can bring new insights to Research Methodology in Strategy and Management. We are looking to attract volume editors who vary along important dimensions, including scholarly background, geographic location, years of experience, and personal demographics. Please email us if you have an idea for a theme that could provide the basis for a volume. In the meantime, we are conﬁdent that the chapters offered in the current volume will build bridges for the organizational sciences toward greater methodological rigor and creativity. Donald D. Bergh David J. Ketchen Jr. Editors

PART I BRIDGES BETWEEN MACRO AND MICRO RESEARCH

REDISCOVERING THE INDIVIDUAL IN STRATEGY: METHODOLOGICAL CHALLENGES, STRATEGIES, AND PROSPECTS Steven W. Floyd and Rebekka Sputtek ABSTRACT Purpose – The purpose of this chapter is to advance research that relies on information from or about individuals and their role in strategic outcomes. Methodology/approach – We start by identifying three streams of strategy research that call for individual-level data (upper echelons, micro-foundations, and strategy-as-practice) and proceed by examining the methods employed across 43 recent empirical studies. Findings – Our analysis addresses three key challenges faced by researchers in these domains and catalogues the strategies used to surmount them. Social implications – By helping to improve methods for research on individuals and strategy, this chapter advances understanding of how people throughout the organization may contribute to strategic outcomes. Originality/value of paper – Our chapter is one of the ﬁrst to analyze methods across three research domains that have heretofore been Building Methodological Bridges Research Methodology in Strategy and Management, Volume 6, 3–30 Copyright r 2011 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-8387/doi:10.1108/S1479-8387(2011)0000006004

3

4

STEVEN W. FLOYD AND REBEKKA SPUTTEK

considered separate. In addition to describing what has been done, we suggest opportunities for improvement, frequently by ﬁnding ways to cross-fertilize methods from one stream into another. Keywords: individual; strategy; upper echelons theory; microfoundations; strategy-as-practice The individual has always been at the center of strategic management thought. Although early writers did not necessarily think of individuals as a level of analysis in a scientiﬁc sense, they wrote mostly about the problems and concerns of chief executives and general managers (e.g., Andrews, 1971; Barnard, 1938; Penrose, 1959). As the domain matured in the direction of scientiﬁc enterprise, however, many researchers shifted attention away from individuals and management, per se. This shift was a natural outgrowth of the role of industrial organization economics in the development of theory. With the emergence of resource-based theory (Barney, 1991), however, particularly that branch concerned with dynamic capabilities (Teece, Pisano, & Shuen, 1997), economics refocused the ﬁeld’s agenda on organizations and those who manage them. Behavioral constructs are now widely considered a basis for understanding economic outcomes, including competitive advantage and proﬁtability. Indeed deﬁnitions of key constructs identify the sources of success in managerial terminology. For example, Zahra, Sapienza, and Davidsson (2006) deﬁne dynamic capabilities ‘‘as the abilities to reconﬁgure a ﬁrm’s resources and routines in the manner envisioned and deemed appropriate by its principal decision-maker(s)’’ (p. 918). More broadly but with a similar managerial focus, Eisenhardt and Martin (2000) describe dynamic capabilities as the ‘‘y speciﬁc and identiﬁable processes such as product development, strategic decision making, and alliancing’’ (abstract). Given this orientation, it is not surprising that signiﬁcant research interest has turned to the ‘‘micro-foundations’’ of capability and competitive advantage (Felin & Foss, 2009; Teece, 2007). Independently and with considerable impact in its own right, strategy researchers with a more sociological or psychological bent have sought to explain economic performance in terms of managerial phenomena, including strategic decision-making (Frederickson, 1984), corporate renewal (Burgelman, 1983), top management team (TMT) consensus (Bourgeois III, 1980), among other variables. Indeed, Hambrick and Mason’s (1984) inﬂuential paper launched a substantial body of work that takes managers’ psychological percepts as central variables.

Rediscovering the Individual in Strategy

5

Parallel to the micro-foundations (MF) and upper echelons (UE) streams of research, which emanate largely from North America, scholars in Europe have opened a conversation on the micro-process and practices associated with managing strategy (Jarzabkowski, 2003; Johnson, Melin, & Whittington, 2003; Whittington, 2006). As researchers in this stream of work are prone to say, the focus is on ‘‘what people actually do.’’ In particular, strategyas-practice (SAP) research spotlights the activities and practices of strategists (Whittington, 2006). Doubtless there are other strands of strategy research that take managers and individuals as central, but we have identiﬁed the UE, MF, and practice streams as especially relevant and impactful to the broader ﬁeld. Our purpose in this chapter is to advance research in these three domains, ﬁrst by describing what has been done to meet the methodological challenges presented by each line of research, and second by identifying ways that one stream can learn from another. We use opportunities for cross-pollination as the basis of suggestions for future research.

CONCEPTUAL BACKGROUND AND APPROACH We begin with a brief discussion of the assumptions and theoretical propositions that undergird research in the three domains. Apologizing in advance for oversimplifying, we seek here to trace the broad outlines. The purpose is to lay a backdrop for deﬁning the central challenges of research design and methodology for research seeking to connect individuals to strategy. We begin with UE. UE theory assumes that executives behave in strategic decision-making situations according to their personalized interpretation of these situations. This personalized interpretation springs from executives’ experiences, values, and personality (Hambrick, 2007; Hambrick & Mason, 1984). UE theory deﬁnes a given situation as a set of cues perceived by an executive; individuals’ perceptions, in turn, are inﬂuenced by their unique cognitive make-up, including especially their values that focus attention and color interpretation. As Hambrick and Mason put it: ‘‘y The decision maker brings a cognitive base and set of values to a decision, which create a screen between the situation and his/her eventual perception’’ (Hambrick & Mason, 1984, p. 195). Thus, researchers should look for the antecedents of strategic choice in cognitive representations, including perceptions that are heavily

6

STEVEN W. FLOYD AND REBEKKA SPUTTEK

inﬂuenced by an individual’s values. Objective demographic variables may correlate highly with executive perceptions, but the principal message of UE theory is that top managers’ perceptions, values, and experiences should play a central role in explanations of outcomes like strategy and performance. While the UE perspective is well established, research that springs explicitly from the idea that organizational (capabilities) have their source in MF, including individuals, is relatively recent. Many studies have been precipitated by Teece’s (2007) focus on: ‘‘The micro-foundations of dynamic capabilities – the distinct skills, processes, procedures, organizational structures, decision rules, and disciplines – which undergird enterprise-level sensing, seizing, and reconﬁguring of capacities [that] are difﬁcult to develop and deploy’’ (abstract). According to this view, the sources of competitive advantage are located at the level of individual action and interaction (Felin & Foss, 2009). Still in its early stages, MF research has the potential to inﬂuence a wide range of topics in strategic management. To date, initial empirical interest seems to have been generated mostly around relationships between individuals and organizational capability (e.g., Holcomb, Holmes, & Connelly, 2009). Finally, published SAP research is often traced to a journal’s special issue in 2003 wherein the coeditors observed that the strategic management ﬁeld should focus more on the activities and interactions of people who ‘‘do strategy’’ (Johnson et al., 2003). A ﬂurry of papers followed this special issue and led to the development of a substantial community focused on the role of the ‘‘everyday activities of organizational life y relating to strategic outcomes’’ (SAP, 2010). Whittington (2006) articulated a conceptual foundation for this body of work in his depiction of the dynamic inﬂuences across institutionalized strategy practices, practitioners of strategy, and speciﬁc instances of strategy praxis. Thus, while this perspective incorporates the macrostructural context, the strategists’ activities and their interactions ﬁgure prominently in theory development and empirical studies (e.g., Mantere, 2005). We organize our effort to examine the methods employed in these three streams of research around a set of three descriptive and three evaluative questions. (The latter are shown below in parentheses.) The origin of these questions is the recognition that at a fundamental level research on UE, MF, and SAP seeks to address a similar set of issues. Central among these are Who are the individuals that affect strategy-related outcomes? How do these individuals do it? These questions, in turn, point to three methodological challenges faced across the three domains.

Rediscovering the Individual in Strategy

7

(1) Who does the method identify as relevant individuals? (Are they the right people?); (2) How does the method connect these individuals to organizational outcomes? (How does it account for multilevel inﬂuences?); and (3) What does the method actually measure? (Is it the right phenomenon?) We see these questions as representing key challenges for research on the role of individuals in strategy outcomes and therefore as a useful basis for describing it. For each set of questions, we attempt to describe how the challenges have been met. This provides a basis for our synthesis, critique, and suggestions for future research. Since we seek to describe and critique, rather than test hypotheses or build new theory, our method is to examine in some depth a signiﬁcant body of recently published research in the three domains. The decision to examine studies across these three particular research streams has the advantage of tapping into a diverse set of methodologies. This work emanates from different disciplinary orientations – psychology, economics, and sociology. As a result, one ﬁnds a rich set of methods, ranging from large scale, survey designs to single case studies, and participant observation. Our choice of which speciﬁc articles to include was constrained to papers published in leading journals and driven by the desire to represent recent work across the three streams. Initially, we set our sights on a 5-year time horizon, but in order to incorporate studies that illustrated a unique approach to one of the challenges, we extended our scope selectively back to the mid-1990s for UE research and to 2003 for SAP. The MF work is new enough that a 5-year time frame is adequate. The result is a judgmental sample of 43 papers that are representative of the state-of-the-art. In that sense, they provide a reasonable basis for our study of methods in these domains.

STRATEGIES FOR MEETING THE CHALLENGES OF METHOD Table 1 summarizes our observations of the strategies used in the studies we reviewed to overcome the challenges of ‘‘who, how, and what’’ for each of the three research streams. The ﬁrst column identiﬁes the challenge, and the second column describes the methodological strategies employed by the studies we reviewed to meet the challenge. We proceed by describing the challenges and strategies within each of the three domains.

8

STEVEN W. FLOYD AND REBEKKA SPUTTEK

Table 1.

Strategies Observed for Meeting the Challenges of Method Across Three Domains.

Challenges Who does the method study?

Methodsa Upper echelons (UE) Sampling CEOs from SME populations companies (Boone, de Brabander, & van Witteloostuijn, 1996; Delgado-Garcia & De La Fuente-Sabate´, 2010; Khan & Manopichetwattana, 1989; Marginson & McAulay, 2008; Nadkarni & Herrmann, 2010; Peterson, Martorana, Smith, & Owens, 2003; Roth, 1995; Simsek, Heavey, & Veiga, 2010; Simsek, Veiga, Lubatkin, & Dino, 2005; Wally & Baum, 1994) Sampling CEOs from a speciﬁc industry (i.e., Spanish banking) (Delgado-Garcia & De La Fuente-Sabate´, 2010) Ask CEOs to identify TMT members (cannot take this for granted) (Ling, Simsek, Lubatkin, & Veiga, 2008; Smith et al., 1994) Micro-foundations (MF) Executive directors and management teams of branches of a Dutch ﬁnancial services ﬁrm in a study of team attributes, transformation leadership, and organizational ambidexterity (Jansen, George, Van den Bosch, & Volberda, 2008) Head coaches of NFL teams as surrogates for managers in ﬁrms in a study of managerial ability and ﬁrm performance (Holcomb et al., 2009) Automotive engineers to assess individual knowledge transfer (teaching strategies) and individual absorptive capacity effects on individual knowledge compared to collective (managerial) coordination (Zheng & Anand, 2009) Strategy-as-practice (SAP) TMT as key actors who employ practices in the work of strategy (Jarzabkowski, 2003) Strategic managers at the periphery as well as the center (Regne´r, 2003) Interview TMT to identify population of middle managers (Currie & Procter, 2005; Pappas & Wooldridge, 2007)

How does the method bridge multiple levels of analysis?

UE ‘‘Inside-out approaches’’ CEO W Organizational outcomes/performance (Billett & Qian, 2008; Delgado-Garcia & De La Fuente-Sabate´, 2010)

9

Rediscovering the Individual in Strategy

Table 1. (Continued ) Challenges

Methodsa CEO core self-evaluation, narcissism W leadership behaviors W organizational outcomes including performance (Li & Tang, 2010; Resick, Weingarden, Whitman, & Hiller, 2009; Simsek et al., 2010) CEO tenure W top management team behavior W ﬁrm performance (Simsek, 2007) CEO þ team level þ ﬁrm level W TMT behavioral integration (Simsek et al., 2005) CEO tenure W TMT risk-taking þ entrepreneurial initiative W ﬁrm performance (Lubatkin, Simsek, Yan, & Veiga, 2006) UE ‘‘Outside-in approaches’’ Organizational environment W individual short termism W individual outcome (Marginson & McAulay, 2008) Organizational environment W CEO identiﬁcation with the corporate elite W strategic help provided (McDonald & Westphal, 2010) UE ‘‘Inside-In approaches’’ CEO behavior W CEO self-perceptionW CEO behavior (Billett & Qian, 2008) Entrepreneur characteristic W entrepreneur expectations of performance (Cassar, 2010) MF Transformational leadership behavior of executive director as moderator of relationship between team attributes (shared vision, social integration, contingency rewards) and organizational ambidexterity (Jansen et al., 2008) Individual-level ability W team performance statistics/payroll (resource productivity) W organizational performance (% wins) (Holcomb et al., 2009) Individual and collective levels of teaching and absorptive capacity as inﬂuences on capability transfer of receiving organization (Zheng & Anand, 2009) SAP Inferring antecedents of cohort behaviors and organizational outcomes based on interviews of individuals (Currie & Procter, 2005)

10

STEVEN W. FLOYD AND REBEKKA SPUTTEK

Table 1. (Continued ) Challenges

Methodsa Practices of TMTs, meetings, activities, and practices (direction setting, resource allocation, monitoring/ control) within organizations (Jarzabkowski, 2003) Middle managers relative centrality in intraorganizational social network as a basis of new ideas and inﬂuence that foster divergent strategic activity (Pappas & Wooldridge, 2007) Deductive managerial activities at the center respond to issue and create strategy; inductive managerial activities at the periphery respond to issue and create strategy (Regne´r, 2003) Speciﬁc strategy talk (practices) of managers within ﬁrms creates strategy (Vaara, Kleymann, & Seristo¨, 2004)

(What) does the method measure? UE Self-report questionnaire scales of CEO psychological data (Delgado-Garcia & De La Fuente-Sabate´, 2010; Marginson & McAulay, 2008; Nadkarni & Herrmann, 2010; Simsek et al., 2010, 2005), and historiometrics (Resick et al., 2009) Self-report demographic data (Forbes, 2005; Marginson & McAulay, 2008; McDonald & Westphal, 2010; Simsek et al., 2005) Archival demographic data (Billett & Qian, 2008; Chatterjee & Hambrick, 2007; Hayward & Hambrick, 1997) CEO report, survey measures of TMT risk-taking, initiative, and ﬁrm performance (Lubatkin et al., 2006) Participant observation as method to access ‘‘intimate details’’ of TM- and board-level behaviors (Parker, 2007). MF Observe transformational leadership in executive director by surveying other members of the management team (Jansen et al., 2008) Organizational performance measured by % of wins, ratio of team statistics (e.g., rushing yards gained) to payroll as measures of resource productivity, coach’s career win %, total years as head coach, # of awards, etc., as measure of managerial ability (Holcomb et al., 2009) Individual-level knowledge and absorptive capacity of engineers measured by surveying engineering managers (Zheng & Anand, 2009)

11

Rediscovering the Individual in Strategy

Table 1. (Continued ) Challenges

Methodsa SAP Analysis of 3 cases to assess organizational contingencies inﬂuences middle managers strategic behavior (Currie & Procter, 2005) Differentiating social structures (e.g., role expectations) from agency and enabling conditions of actor-driven behavior (Mantere & Sillince, 2007) Measuring different types of centrality and peerreported divergent strategic activity (Pappas & Wooldridge, 2007) Case-based, triangulated data to observe strategic activities at the center (e.g., HQ) and periphery (e.g., developmental unit and subsidiary) that create strategy in response to speciﬁc issues (Regne´r, 2003) Ethnographic examination of micro-practices (e.g., episodic routines and conversations that ‘‘discipline the client’’) versus ‘‘discrete and conscious activity of managers’’ (Rouleau, 2005) Observation of press, PR, online forums plus interviews of top-, middle- and low-level managers in examination of discursive elements of strategy related to strategic alliance; strategy deﬁned as discourse activity (Vaara et al., 2004)

a

Studies are categorized as upper echelons (UE), micro-foundations (MF), and strategy-aspractice (SAP).

Who Does the Method Identify as Relevant Individuals? (Are These the Right People?) Identifying the right individuals to study in UE research is less difﬁcult from a theoretical perspective than from the view of a methodologist. Theoretically, the focal persons are a rather small number of people – chief executives, board members, and top managers – at the very top of the organization. The problem for the researcher is that many of these individuals, particularly those in large public companies, are reluctant subjects, for the most part declining to be interviewed, complete surveys, or otherwise participate in research. Hambrick and Mason (1984) anticipated this problem and advocated the use

12

STEVEN W. FLOYD AND REBEKKA SPUTTEK

of demographic characteristics as reasonable surrogates for certain psychological variables. The validity of this approach has been questioned, however, and scholars have suggested a focus on more direct measures of UE personalities (Hambrick, 2007; Marko´czy, 1997). To overcome the difﬁculties of accessing top managers of large companies, many of the studies in our review use subjective self-report data obtained from managers below senior levels (Marginson & McAulay, 2008; Peterson et al., 2003) or from senior managers in smaller companies (Boone et al., 1996; Delgado-Garcia & De La Fuente-Sabate´, 2010; Khan & Manopichetwattana, 1989; Marginson & McAulay, 2008; Nadkarni & Herrmann, 2010; Peterson et al., 2003; Roth, 1995; Simsek et al., 2010, 2005; Wally & Baum, 1994). This latter strategy relies on the assumption that CEOs of large companies differ little from CEOs of small companies. While they may be similar in some respects, CEOs of small companies are more often also founders and owners. Governance research conducted within small and medium enterprises (SMEs) (e.g., Zahra, Neubaum, & Naldi, 2007) suggests this may create profound differences in both perspectives, and possibly, also in priorities. Depending on the theoretical and empirical scope within which one is trying to generalize, using a sample of SMEs represents an approximation of the population that includes large companies. Sampling managers lower in the hierarchy as substitutes for top managers overcomes this problem but raises others. In particular, lower level managers have nowhere near the inﬂuence of top managers, and this difference alone is likely to change their perceptions relevant to strategy. Based on the studies we examined, it appears that direct access to psychological measures for members of the UE of large companies is lacking in current research. While this conﬁrms Hambrick and Mason’s (1984) original intuition, it suggests a limitation on the empirical boundaries of accumulated knowledge in the UE domain. In short, if the goal is to generalize to TMTs within large corporations, this strand of UE research often appears to target the wrong people. The architects of the UE perspective anticipated the problems associated with researcher access. Another approach is the use of data about top managers in the form of demographics as surrogates for top managers’ perceptions and mindset (Forbes, 2005), either as a basis for assessing explanatory variables or as controls (Marginson & McAulay, 2008; McDonald & Westphal, 2010; Simsek et al., 2005). A related set of studies draws on archival access for demographic data (Billett & Qian, 2008; Chatterjee & Hambrick, 2007; Hayward & Hambrick, 1997). One study collects subjective (by the researcher) psychological data using historiometric analyses (Resick et al., 2009); another

Rediscovering the Individual in Strategy

13

relies on coding of a qualitative data set by researchers to measure executives’ psychological proﬁle (Simon & Houghton, 2003). In one respect, research seeking to establish the MF of organizational capability or dynamic capability (Felin & Foss, 2009) has the opposite problem compared to UE. With the last, it is easy to identify the right individuals; the challenge is to get access. With MF, the real challenge is identifying the ‘‘right’’ individuals; access is a secondary concern. Indeed, there seems to be little consensus with respect to where to look within organizations to ﬁnd a theoretically relevant set of individuals, and few studies have reported problems of access. Research in this domain has used data from bank branch managers (Jansen et al., 2008), automotive engineers (Zheng & Anand, 2009), and head coaches of NFL football teams (Holcomb et al., 2009) as a source or center of the organization’s capability and predictor of performance. Research in the SAP domain shares the identiﬁcation problem. That is, the need for a theoretical justiﬁcation for sampling from a particular population of individuals. Put differently: Who practices strategy? Early studies focused on top managers as strategy practitioners (Jarzabkowski, 2003). More recent research has expanded the deﬁnition of strategy practitioners to include those at the ‘‘periphery’’ of the organizations, arguing that strategic activities at the periphery differ sharply from activities at the center (Regne´r, 2003). Several studies in this domain have focused on the roles, activities, and practices of middle-level managers (Mantere, 2005). Deﬁnitions of who counts as a middle manager differ signiﬁcantly from one study to another. In the absence of a clear theoretical answer on the issue of whom to observe, one approach has been to interview members of top management, asking them to identify strategic actors in the middle ranks (Currie & Procter, 2005; Pappas & Wooldridge, 2007).

How Does the Method Connect Individuals to Organizational Outcomes? (How Does It Account for Multilevel Inﬂuences?) From the outset, Hambrick and Mason (1984) recognized a sequence of causal links required to justify the impact of top manager psychological characteristics on organization outcomes. Perceptions and decision-making processes are key bridging mechanisms that explain how executive characteristics such as personality and experience relate to the strategic choices of the organization’s most powerful individuals, and hence, the performance outcomes that follow from these choices.

14

STEVEN W. FLOYD AND REBEKKA SPUTTEK

UE researchers seem to follow three directions in tracing such relationships across multiple levels of analysis: inside-out, outside-in, and inside-in. Inside-out approaches make inferences about variables at the individual level in order to explain organizational outcomes. In some cases, more than one level is identiﬁed to bridge the distance between individual and organizational variables. Outside-in approaches use organizational or environmental factors as explanatory variables at the individual level. Inside-in approaches avoid the problem altogether by focusing on independent and dependent variables that are all on an individual level of analysis. Inside-out approach appears to be the most common logic. Typical are studies investigating CEO locus of control and small ﬁrm performance (Boone et al., 1996), CEO affective traits, strategic conformity and performance conformity (Delgado-Garcia & De La Fuente-Sabate´, 2010), CEO hubris and premiums paid in acquisitions (Hayward & Hambrick, 1997), TMT experience and decision-making speed (Forbes, 2005), managerial locus of control and organizational innovation (Khan & Manopichetwattana, 1989), and CEO overconﬁdence and the extent of pioneering/risky product introductions (Simon & Houghton, 2003). A second set of studies in this category takes a more complex approach. They consider the inﬂuence of a mediator or moderator in relationships between CEO characteristics and organizational outcomes. Studies focus, for example, on strategic ﬂexibility as a full mediator of CEO personality– ﬁrm performance relationships (Nadkarni & Herrmann, 2010). Other authors investigate the mediating role of transformational and transactional leadership (Resick et al., 2009) on relationships between narcissism and core self-evaluation on CEO inﬂuence at the organizational level. Seeing the problem more in terms of contextualizing inferences than transcending levels of analysis, Simsek et al. (2010) set up environmental dynamism as a moderator in the relationships between CEO core selfevaluation and (organizational level) entrepreneurial orientation. Similarly, Li and Tang (2010) seek to contextualize the relationship between CEO hubris and ﬁrm risk-taking by considering the mediating role of managerial discretion. In one of the few studies where the design explicitly reasons up the micro-, meso-, macro- ladder, Simsek et al. (2005) analyze links between TMT behavior integration, CEO collectivistic orientation, team level goal preference diversity, and ﬁrm level performance. Outside-in approaches examine how factors in the organization’s external environment (e.g., stock markets) inﬂuence managers’ attitudes (e.g., shorttermism) (Marginson & McAulay, 2008) or how the extent of a ﬁrm’s international interdependence relates to CEO characteristics that are closer

Rediscovering the Individual in Strategy

15

to the theoretical ideal (Roth, 1995). Another outside-in study incorporates CEO identiﬁcation with the corporate elite as a mediator in the relationship between board control and the degree of strategic help provided by the CEO (McDonald & Westphal, 2010). Finally, two studies we reviewed took what one might call an inside-in approach – studying all variables on one level of analysis. One of these investigated CEO acquisition behavior and its inﬂuence on CEO overconﬁdence and subsequent CEO acquisition behavior (Billett & Qian, 2008). Another examined overoptimism as an antecedent of an executive’s positive expectations for ﬁrm performance (Cassar, 2010). While the theory and measures in these studies consistently focused on a single level of analysis, it may be worth noting that the individual-level outcome examined in both studies has an obvious organizational-level referent, that is, acquisition strategy and ﬁrm performance, respectively. Empirical work in the MF area is still relatively scarce, but recent studies seem to have recognized the challenge of bridging multiple levels of analysis from the outset. Jansen et al. (2008), for example, examine transformational leadership at the individual level of analysis, not as an independent variable, but as a moderator. The argument is that transformational leadership on the part of the managing director strengthens relationships between certain team attributes (e.g., shared vision) and organizational-level ambidexterity. The method in this case draws questionnaire data from multiple respondents and develops measures of independent and dependent variables from different individuals. In a totally different context, but with a similar sensitivity toward the multilevel challenge, Holcomb and Holmes (2009) use empirical measures that trace relationships between an individual coach’s ability, team performance, and then to organizational performance. SAP researchers have been creative in meeting the multilevel challenge. One qualitative study addresses the problem by using interview data to infer relationships between middle manager cohort behavior and organizational outcomes (Currie & Procter, 2005). Another study introduces strategy meetings as a unit of analysis and suggests such episodes are likely contexts for investigating strategy practices, such as direction-setting or monitoring and control (Jarzabkowski, 2003). In a similar shift away from traditional micro–macro distinctions, Regne´r (2003) analyzes interview data from managers in the center (e.g., in the headquarters unit) and periphery (e.g., in a project group), asserting that one deﬁnes the context for the other. Vaara et al. (2004) shift the unit of analysis again; this time focusing on particular strategic issues or themes as a basis for describing the discourse or ‘‘talk’’ associated with strategy creation. What can be said about all these studies is

16

STEVEN W. FLOYD AND REBEKKA SPUTTEK

that the methods are largely qualitative and therefore tend to substitute interpretation for methods in meeting the challenge created by multiple levels of analysis.

What Does the Method Actually Measure? (Is It the Right Phenomenon?) In our earlier discussion of how UE researchers address the problem of accessing top managers of large companies, we noted the solutions included sampling from archival or self-report demographic data. The inferential leaps that must be made to accept demographic indicators as measures of top manager psychological or cognitive states may be criticized. One alternative is drawing samples of individuals from related but different populations (e.g., managers lower in the management hierarchy, CEOs from American SMEs, or executives from the Spanish banking sector). This approach has the advantage of yielding more direct measures of individual psychological variables like core self-evaluation (Simsek et al., 2010) and CEO affect (Delgado-Garcia & De La Fuente-Sabate´, 2010). Moreover, since variables like personality appear to be set at a relatively early stage of life and endure, student samples represent another potential way for UE research to meet the challenge of access and employ direct measures of the variables of interest (Hambrick, 2007). Many of the researchers associated with MF research are aiming to relate an individual-level measure to organizational capabilities. As a result, scholars in this emerging domain tend to assess either managerial ability directly based on indicators like the number of performance awards (Holcomb et al., 2009) or variables that are easily connected to capabilities or dynamic capabilities (change), such as transformational leadership behavior (Jansen et al., 2008) and engineering knowledge (Zheng & Anand, 2009). The need for considerable theory development remains in this research stream, and until such theory is in place, it is possible to imagine a host of individual-level variables as MF for capability at the unit or organizational level of analysis. SAP scholars are less interested in attributes of individuals per se and more interested in what the individuals do. In the studies we examined, ‘‘what they do’’ is interpreted variously as strategic behavior, activities, practices, and talk. Three studies assessed some sort of middle manager strategic behavior, conceptualized in the form of strategic roles (Currie & Procter, 2005; Pappas & Wooldridge, 2007) or individual agency beyond the strategic roles (Mantere & Sillince, 2007). Studies of practices focus on a

Rediscovering the Individual in Strategy

17

range of activities, including those micro-processes (cf. interactions) associated with episodic routines, conversations, and discourse (Rouleau, 2005; Vaara et al., 2004). What one observes in this work is a wide range of conceptual and empirical distance between focal individuals and the speciﬁc measure. One study employs the detailed, up-close observation of ethnography (Rouleau, 2005); another infers strategy ‘‘talk’’ from news accounts and internal, archival sources (Vaara et al., 2004). In between these extremes, interviews and surveys of managers are often deployed to measure individual behavior. A strength in this work is its disciplined approach to incorporating context and boundary conditions in the examination of practices. Again, however, context means different things in each study – social networks, industry, meetings, episodes, and even roles themselves. While a continuation of such method diversity has beneﬁts, in the long run, one might hope for a more cumulative approach.

SUGGESTIONS FOR FUTURE RESEARCH Table 2 summarizes suggestions for future research. Many of these are inspired by the studies we observed; some are of our own making; and several suggest transfers of method from one domain to another. Unlike Table 1, the suggested strategies in Table 2 are not identiﬁed with particular domains. This is because our goal in making each of these suggestions is to trigger useful reﬂections on method across all three domains. In that sense, we seek to cross-pollinate method strategies from one domain to another. Who – Studying the Right Individuals An overarching question facing anyone concerned with individual-level methods in strategy research is whether the theoretical lens they are using provides an unambiguous basis for identifying relevant individuals. It would seem to go without saying that unless a study taps the appropriate individual or individuals in the analytical process, the potential for valid conclusions arising from the research is substantially diminished. Yet, studies differ widely in terms of who they study, and perhaps more importantly, studies vary widely in the degree of speciﬁcity and rigor they use to resolve the question of who. As the more mature research stream, UE research is rather deﬁnitive in identifying top managers and board members as the primary focus of study,

18

STEVEN W. FLOYD AND REBEKKA SPUTTEK

Table 2. Challenge

Suggestions for Future Research. Suggestions for Future Research

Who does the method study?

Need for theory to better guide the ‘‘who’’ question, for example, is top management the place to look for microfoundations of capability – is the NFL coach metaphor more apt for CEOs or middle managers? (Holcomb et al., 2009) Accessibility of SMEs (Ling et al., 2008; Lubatkin et al., 2006), nonproﬁt organizations (Maitlis & Lawrence, 2003; Pappas & Wooldridge, 2007), and public sector (Currie & Procter, 2005) may offer opportunities but may impose bounds on external validity. Identifying the right managers requires methods to account for differential levels of inﬂuence and power of individual managers on strategy (Maitlis & Lawrence, 2003; Pappas & Wooldridge, 2007, cf. strategic champions). For large cohorts, a random sample may be an alternative (Mantere, 2005). Look for managers in what they do or produce as well as who they are, for example, coordination systems for collective teaching and learning (Zheng & Anand, 2009), transformational leadership (Ling et al., 2008), politics and discourse (Maitlis & Lawrence, 2003), manager exploration/exploitation activities (Mom, Van Den Bosch, & Volberda, 2007) divergent strategic activity (Pappas & Wooldridge, 2007), strategy creation (Regne´r, 2003), sensemaking and sensegiving practices (Rouleau, 2005), and strategy discourse (Vaara et al., 2004)

How does the method bridge multiple levels of analysis?

UE studies lose credibility when they ‘‘stretch the causal thread’’ from individual cognition (even the CEO’s) to organizational outcomes; more mediators are needed at the group (operating team, department, initiative) and organizational level, for example, use of resources and discretion as mediators of managerial ability (Holcomb et al., 2009) Case studies may rely too much on respondent accounts in the attribution of multilevel effects (e.g., Currie & Procter, 2005) but multiple data sources and rigorous analysis can tease out new multilevel effects and identify causal mechanisms, for example, TMT actors W strategy practices W organizational outcomes (Jarzabkowski, 2003) There is too little research that looks at social interactions across managerial levels – vertical interactions are especially crucial for strategy making (Mom et al., 2007)

19

Rediscovering the Individual in Strategy

Table 2. (Continued ) Challenge

Suggestions for Future Research Need for research that focuses on dynamics and consequences of strategic action; ‘‘decision story’’ development and analysis as a tool (Maitlis & Lawrence, 2003). Strategy-making episodes (e.g., meetings) and entrepreneurial strategic initiatives (Lubatkin et al., 2006) offer units of analysis for contextualizing individual strategic actions and outcomes over time (Hendry & Seidl, 2003)

What does the method measure?

Need validity studies to quantify the validity of background and demographic variables as surrogates for psychological states (Marko´czy, 1997) Strategic signiﬁcance of everyday activities may only be seen retrospectively, for example, R&D efforts of engineers; researchers should focus on activities related to ‘‘crucial events’’ in realizing a strategy. Managers may not be aware of the relevance of their activity to strategy, capability, etc. (Rouleau, 2005); strategizing is bound by various types of discourse that the people involved are not always aware of (Vaara et al., 2004)

for example. The scope of research on MF and SAP, on the other hand, would seem to include managers and professionals at many levels of the hierarchy and in many different subunits. Although some lack of consistency can be expected early in the development of a research stream, scholars working on MF and SAP may also want to consider the beneﬁts of a unifying conceptual model or ‘‘umbrella construct’’ (Floyd, Cornelissen, Wright, & Delios, 2011). Hambrick and Mason (1984) launched UE research within a speciﬁc conceptual framework, and the adoption of and reference to this framework in subsequent research has been a signiﬁcant factor in producing a uniquely coherent research stream. Thus, if MF is a search for the source of capability in organizations, then research may beneﬁt from a shared framework that limits differences in who (what individuals) are studied across researchers and potentially increases the cumulative quality of future work. Lacking such a framework, researchers are likely to rely on sampling designs based on different logics, study different individuals, and produce

20

STEVEN W. FLOYD AND REBEKKA SPUTTEK

ﬁndings that are at least difﬁcult to compare and at worst wholly inconsistent. In an interesting study of NFL football organizations, Holcomb and Holmes (2009) deﬁne managerial capability as the knowledge, skill, and experience of managers, and identify coaches as the subject. While the study is useful in demonstrating effects of managerial capability, the analysis leaves open questions about where the coach-equivalent might reside in a large complex ﬁrm. Is the coach a top manager, middle manager, or simply a team leader? Managers in each of these and other locations may be associated with managerial capabilities, as well as other operating capabilities. What would be helpful to future researchers is an organizing framework, perhaps one that began with a deﬁnition of capabilities, and proceeded to articulate a set of categories or types of individuals and their relationship to organizationallevel capabilities. Existing work may provide clues about the dimensions of such a framework. In particular, part of what it means to be associated with managerial capabilities has to do with power and inﬂuence. In addition to formal authority, this may mean tracking individuals’ level of informal inﬂuence as a way to differentiate the relevant subject cohort, for example, study managers who are central actors (Pappas & Wooldridge, 2007). In large cohorts, random selection may play a role (Mantere, 2005), but it seems likely that a more theory-driven approach would lead to a more coherent set of ﬁndings, for example, concerning middle-level managers relationship to capability. SAP research may also beneﬁt from a common framework. In this case, the need is for a framework that deﬁnes practices. If strategists are individuals who practice strategies, then the burden for researches shifts from identifying subjects (‘‘the right individuals’’) as a starting point to identifying what individuals do that comprises strategy practice. This shift from who to what can be extrapolated into the other domains. UE research can be accused of being preoccupied with who top managers are, that is, their demographic characteristics. By assuming that elites are the locus of strategic activity, this research stream has neglected the activity itself as an antecedent of ﬁrm performance. Studies we reviewed both within and outside UE suggest a number of possibilities for studying activities in the UE, including transformational leadership (Ling et al., 2008) politics and discourse (Maitlis & Lawrence, 2003), manager exploration/exploitation activities (Mom et al., 2007) divergent strategic activity (Pappas & Wooldridge, 2007), strategy creation (Regne´r, 2003), sensemaking and sensegiving (Rouleau, 2005), and strategy ‘‘talk’’ (Vaara et al., 2004). Again, since the CEOs and TMTs of large ﬁrms may not be open to studies that probe into their activities, researchers may

Rediscovering the Individual in Strategy

21

have to ﬁnd data in SMEs (Ling et al., 2008; Lubatkin et al., 2006), nonproﬁts (Maitlis & Lawrence, 2003; Pappas & Wooldridge, 2007), and organizations in the public sector (Currie & Procter, 2005). Doing so will increase opportunities for access, but as noted above, may also impose bounds on external validity. In sum, even though this is a chapter about method, our ﬁrst methodological suggestion is for conceptual frameworks that would encourage greater harmony of methods in MF and SAP research. A more consistent set of methods for identifying and accessing relevant individuals will, in turn, increase the cumulative nature of ﬁndings in these domains. Second, research in all three domains could beneﬁt from methods, such as social network analysis, that differentiate individuals in terms of their relative connection to strategy. Third, another way to identify the right individuals, suggested by SAP but relevant to all three, is ﬁrst to identify a set of activities that impact strategy and then focus research on the individuals that perform these activities.

How – Accounting for Multilevel Inﬂuences Despite the clarity with which Hambrick and Mason’s (1984) original framework identiﬁed the causal mechanisms that explain the links between individual or TMT characteristics and organizational outcomes, the methods in many UE studies rely on variables at only two levels of analysis. While recent examples suggest an increasing awareness among UE researchers of the need to take account of both internal and external variables as mediators and moderators of relationships between individuals and organizational-level outcomes (e.g., Simsek, Veiga, & Lubatkin, 2007), more mediators at the group level (operating team, department, and initiative) and moderators at the organizational/network level should be identiﬁed, for example, use of resources and managerial discretion (Holcomb et al., 2009). Perhaps because SAP lacks the conceptual foundations for deducing factors that explain multilevel relationships, case data and induction have become a rich source for better understanding the mechanisms between variables at the individual and organizational levels of analysis (cf. Parker, 2007). Qualitative researchers should be cautious about relying too much on respondent accounts in the attribution of multilevel effects (Currie & Procter, 2005), but multiple data sources and rigorous analysis can tease out new multilevel effects and identify causal mechanisms (Jarzabkowski, 2003). One rather obvious multilevel relationship relevant to individualorganizational bridging are the interactions across levels of the management

22

STEVEN W. FLOYD AND REBEKKA SPUTTEK

hierarchy. Theory suggests such vertical interactions are crucial in the formation of strategy (Floyd & Lane, 2000), and their relevance is apparent in all three research streams: UE – top managers implement decisions through middle and lower levels; MF – middle managers propose capability-building initiatives to top managers; SAP – strategy practices engage top and middle managers in interaction episodes. Yet, only one study in our review examined them (Mom et al., 2007). We would encourage the development of methods that facilitate observation and analysis across multiple levels of the management hierarchy. The strategy episode is a unit of analysis suggested in SAP (Hendry & Seidl, 2003) that has relevance to MF and UE. Such episodes (e.g., strategic planning meetings) are the locus of the managerial activity and interactions that are intended by the participants to have signiﬁcant consequences for the organization. In that sense, such episodes have a degree of face validity as a bridge in the multilevel chain. Entrepreneurial strategic initiatives (Lechner, Frankenberger, & Floyd, 2010; Lubatkin et al., 2006) and other developmental groups (McGrath, 2001) are another unit of analysis with high face validity for connecting individual strategic actions to organizational outcomes (Hendry & Seidl, 2003) in all three research domains. Both episodes and initiatives have the added advantage of incorporating time into the analysis. Episodes are deﬁned by temporal and geographic space (Hendry & Seidl, 2003), and initiatives evolve through stages of variation, selection, and retention (Burgelman, 1983). This suggests the use of longitudinal designs that connect individual activity to strategy outcomes over time. Maitlis and Lawrence (2003) employ a ‘‘decision story’’ methodology, for example, that focuses on the dynamics and consequences of strategic action. Much more could be done to develop methods for observing strategic activities episodically and/or within strategic initiatives. In sum, research in all three domains has employed methods that bridge multiple levels of analysis, and the availability of mediators and moderators is likely to make reviewers less tolerant of methods that make gross inferential leaps. Despite recent developments, all three domains would beneﬁt from qualitative research that suggests new accounts of multiplelevel effects. Strategy episodes and strategic initiatives offer ways to span the temporal bridge between individual activity and strategy outcomes.

What – Measuring Right Phenomenon The criticism of UE research for using background or demographic variables as surrogates for psychological variables is well known (Hambrick, 2007;

Rediscovering the Individual in Strategy

23

Marko´czy, 1997), and even its strongest advocates have acknowledged that more direct measures of executives’ values, personality, perceptions, and other mental states would be preferable (Hambrick, 2007). From the start, Hambrick and Mason (1984) acknowledged that, for all their objectivity and accessibility, demographic variables ‘‘contain more noise than purer psychological measures’’ (p. 196). If such purer data is not accessible, however, indirect measures may be considered an acceptable, even necessary, alternative. In this context, it is striking that our review identiﬁed several studies that were successful in overcoming the access problem, surveying CEOs, and other executives to collect data for their studies (Delgado-Garcia & De La Fuente-Sabate´, 2010; Forbes, 2005; Simsek et al., 2005). Although none of these drew samples from large US companies, each was able to get top managers to respond to relatively detailed items, some of which appear to present threats to personal privacy and corporate conﬁdentiality equal to or greater than questions about personality or other mental states. DelgadoGarcia and De La Fuente-Sabate´ (2010) were successful in getting bankers to report on their affective traits. Forbes (2005) successfully asked managers to report when they ﬁrst began considering an important strategic decision and when a commitment to a speciﬁc choice was made – items that respondents could easily interpret as a measure of their decision-making performance. Using an equally complex, nine-item scale, Simsek et al. (2005) successfully measured behavioral integration; these items asked about collaboration and information exchange within the TMT – factors that reﬂect sensitive social issues within the team. Given the success of these studies in gaining access to executives’ inner thoughts and self-perceptions, one may wonder why in every case they maintained the use of demographic surrogates. Delgado-Garcia and De La Fuente-Sabate´ (2010) measured education as a surrogate for risk-taking. Forbes (2005) used age as a surrogate for an individual’s time perceptions and willingness to take risks. Simsek et al. (2005) employed tenure, functional and educational diversity of the TMT as a means of measuring a team’s homogeneity of perspectives. While it is doubtful that any of these researchers saw their use of demographics as surrogates, the arguments in the papers suggest as much. One can ask: Since access difﬁculties had been overcome, wouldn’t more direct measures of the phenomena have been preferable? Although we uncovered no explanation in the articles themselves, it is very likely that the research teams have answers to this question. It may also be the case, however, that the use of demographic surrogates has become self-perpetuating in UE research. Thus, in order to be consistent with earlier

24

STEVEN W. FLOYD AND REBEKKA SPUTTEK

studies, newer studies – even when they have access to more direct measures – continue to use surrogates. Rather than recommending that researchers abandon using demographic variables altogether, we would suggest that scholars follow Marko´czy’s (1997) effort to quantify the validity of these surrogates. In her study, Marko´czy (1997) found that background characteristics explained only 17.5 percent of variance in managers’ beliefs. She rightly interpreted this as an unacceptably low level of validity, leading to her call that research ‘‘accepts no substitutes.’’ Future research should follow in Marko´czy’s (1997) footsteps. Indeed a whole program of research could be organized around validity studies of background variables as indicators of executives’ mental states. In many cases, established scales exist for phenomena that have historically been measured with surrogates (e.g., personality, risk-taking propensity, and belief structures), and these suggest themselves as a starting point for such research. Work of this kind is important not only to compute measurement validities. Use of an invalid measure also threatens statistical conclusion validity. Put simply, if the measure is not measuring what it is intended to measure, then covariance with other variables provides no evidence for hypothesized relationships. What is perhaps most noteworthy about the MF and SAP studies we reviewed is the wide range of variables employed to track individuals’ and their role in strategy outcomes, including transformational leadership behavior, teaching activity, exploration/exploitation activities, and network centrality among many others. The range of data collection methodologies is equally diverse, ranging from archival data, self-report surveys, social network analysis, interviews, observation, and ethnography. Since an equally wide range of theories underlie these studies, it is not surprising that they draw on different variables and data collection strategies. The question one can ask, however, is: How well does the measurement strategy ﬁt the theoretical framework? On the whole, it appears that both MF and SAP have employed observation and measurement approaches that are very proximate to the phenomena of interest. In some cases, this has involved the development or adoption of survey-based measurement scales; in others, qualitative analysis of observations, interviews, and archival data has been the measurement strategy. For the most part, these two research streams cannot be accused of using surrogates or creating a gap between measurement and theory. In some respects, however, the closeness of the measures to phenomena may be a function of the theoretical diversity found in these domains. Without a consensus, researchers are free to develop theory that ﬁts the

Rediscovering the Individual in Strategy

25

observations they have available. While this generates a rich set of empirical results, it limits the potential for cumulative work and undermines the impact of individual studies. After all, subsequent research will build on – and cite – prior research to the extent it investigates similar questions using similar measures. Thus, our recommendation for MF and SAP is to be inspired by UE, not in the use of surrogates, but in the development of measures and methods of observation that are more coherent. Understandably, progress on this front is connected to the development of conceptual frameworks for research in these domains, but a shared understanding of what the central variables are and how they can be measured will facilitate such progress. The work we reviewed suggests several key variables where more robust measures could be developed, including managerial capability and strategy practice. These are key variables, but their measurement to date has been idiosyncratic. In addition to the need for more agreement about what is important and how to measure it, MF and SAP face two other more speciﬁc issues. The ﬁrst of these revolves around the question of measuring strategy practices and/or activities. The problem is knowing what is really ‘‘strategic’’ in the sense that said practices or activities impact strategy. An activity may look strategic, for example, a strategic planning episode, but have little or no relevance to realized strategy; alternatively, a practice may look nonstrategic, for example, playing a round of golf with the chief executive, but have signiﬁcant inﬂuence on strategic decisions (Mintzberg, Ahistrand, & Lampel, 1998). In many cases, one can only discern what is strategic retrospectively, for example, once it has become clear that the conversation on the golf course was a key turning point in a decision process. Similar observations may be made about research that seeks to observe the sources of capability. While it may seem obvious that a football coach’s ability is relevant to the managerial capability within the organization, what about the decision to ﬁre the previous coach? That, too, is related to management of the organization but its relevance to organizational-level managerial capability may only become apparent retrospectively. A related problem is that individuals may not be aware of the relationships between their activity and strategy outcomes. This is perhaps especially true of individuals outside the UE, but even senior executives may not recognize their behavior as strategic. Contemporaneously or retrospectively, for example, the chief executive may never come to appreciate the strategic signiﬁcance of the aforementioned round of golf. And if the individuals who are involved cannot identify their behavior as strategic or as related to critical organizational capability, where does that leave the researcher? More deﬁnitive theory would

26

STEVEN W. FLOYD AND REBEKKA SPUTTEK

help, but that begs the question: What theory? We take as the focus of our ﬁnal recommendation the problem of identifying individual activities, practices, or behaviors as ‘‘strategic.’’ If the goal is to connect activity to outcome, then a longitudinal design is inevitable. In particular, one needs a biographical account of individuals’ histories that provides data to objectively derive linkages from activity to outcome and do so with a sufﬁcient number of individuals to detect a pattern. For individuals of high social status (chief executives of large public companies), biographical information is often collected in the business media and in internal documents (e.g., minutes of meetings and ofﬁcial calendars). For others, diaries may be a method for developing autobiographical data (Mintzberg, 1973). Such data could be subjected to qualitative analysis in order to develop theory around potential patterns. Moreover, given sufﬁcient detail and multiple sources, such data can be transformed into objective measures using historiometric analysis (Resick et al., 2009). This method uses trained raters and measurement scales with established psychometric properties. Thus, although the challenges of collecting such data are nontrivial, systematic analysis of managerial biographies may provide the means for both theory-building and theory-testing research. In sum, in this ﬁnal section of the chapter, we argued that UE research would beneﬁt from a systematic program of validity studies wherein demographic surrogates were examined as measures of psychological variables. In addition, we suggested that MF and SAP research should work toward the development of a small set of key variables and measurement approaches that could be the basis of more programmatic and cumulative work. Finally, we suggested biographical data and analyses such as historiometrics as a ﬁrst step toward the problem of recognizing truly ‘‘strategic’’ managerial practices and/ or activities.

CONCLUSION In this chapter, we have tried to catalogue the methods employed by researchers seeking to examine the individual’s relationship to strategy. In the process, we hope to have illuminated the state-of-the-art and crosspollinate ideas from three research streams that, we believe, have much in common. Each of the domains has already developed a great number of interesting and important ﬁndings. We hope this chapter will help to consolidate this effort and support the development of even better methods in future research.

Rediscovering the Individual in Strategy

27

REFERENCES Andrews, K. R. (1971). New horizons in corporate strategy. McKinsey Quarterly, 7(3), 34–43. Barnard, C. (1938). Functions of the executive. Cambridge, MA: Harvard University Press. Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17(1), 99. Billett, M. T., & Qian, Y. (2008). Are overconﬁdent CEOs born or made? Evidence of selfattribution bias from frequent acquirers. Management Science, 54(6), 1037–1051. Boone, C., de Brabander, B., & van Witteloostuijn, A. (1996). CEO locus of control and small ﬁrm performance: An integrative framework and empirical test. Journal of Management Studies, 33(5), 667–699. Bourgeois, L. J., III. (1980). Performance and consensus. Strategic Management Journal, 1(3), 227–248. Burgelman, R. A. (1983). A process model of internal corporate venturing in the diversiﬁed major ﬁrm. Administrative Science Quarterly, 28(2), 223–244. Cassar, G. (2010). Are individuals entering self-employment overly optimistic? An empirical test of plans and projections on nascent entrepreneur expectations. Strategic Management Journal, 31(8), 822–840. Chatterjee, A., & Hambrick, D. C. (2007). It’s all about me: Narcissistic chief executive ofﬁcers and their effects on company strategy and performance. Administrative Science Quarterly, 52(3), 351–386. Currie, G., & Procter, S. J. (2005). The antecedents of middle managers’ strategic contribution: The case of a professional bureaucracy. Journal of Management Studies, 42(7), 1325–1356. Delgado-Garcia, J. B., & De La Fuente-Sabate´, J. M. (2010). How do CEO emotions matter? Impact of CEO affective traits on strategic and performance conformity in the Spanish banking industry. Strategic Management Journal, 31(5), 562–574. Eisenhardt, K. M., & Martin, J. A. (2000). Dynamic capabilities: What are they? Strategic Management Journal, 21(10/11), 1105. Felin, T., & Foss, N. J. (2009). Organizational routines and capabilities: Historical drift and a course-correction toward microfoundations. Scandinavian Journal of Management, 25(2), 157–167. Floyd, S. W., Cornelissen, J., Wright, M., & Delios, A. (2011). Process and practices of strategising and organising: Review, development, and the role of bridging and umbrella constructs. Journal of Management Studies (Forthcoming). Floyd, S. W., & Lane, P. J. (2000). Strategizing throughout the organization: Managing role conﬂict in strategic renewal. Academy of Management Review, 25(1), 154–177. Forbes, D. P. (2005). Managerial determinants of decision speed in new ventures. Strategic Management Journal, 26(4), 355–366. Frederickson, J. W. (1984). The comprehensiveness of strategic decision processes: Extension, observations, future directions. Academy of Management Journal, 27(3), 445–466. Hambrick, D. C. (2007). Upper echelons theory: An update. Academy of Management Review, 32(2), 334–343. Hambrick, D. C., & Mason, P. A. (1984). Upper echelons: The organization as a reﬂection of its top managers. Academy of Management Review, 9(2), 193–206. Hayward, M. L. A., & Hambrick, D. C. (1997). Explaining the premiums paid for large acquisitions: Evidence of CEO hubris. Administrative Science Quarterly, 42(1), 103–127.

28

STEVEN W. FLOYD AND REBEKKA SPUTTEK

Hendry, J., & Seidl, D. (2003). The structure and signiﬁcance of strategic episodes: Social systems theory and the routine practices of strategic change. Journal of Management Studies, 40(1), 175–196. Holcomb, T. R., Holmes, R. M., Jr., & Connelly, B. L. (2009). Making the most of what you have: Managerial ability as a source of resource value creation. Strategic Management Journal, 30(5), 457–485. Jansen, J. J. P., George, G., Van den Bosch, F. A. J., & Volberda, H. W. (2008). Senior team attributes and organizational ambidexterity: The moderating role of transformational leadership. Journal of Management Studies, 45(5), 982–1007. Jarzabkowski, P. (2003). Strategic practices: An activity theory perspective on continuity and change. Journal of Management Studies, 40(1), 23–55. Johnson, G., Melin, L., & Whittington, R. (2003). Micro strategy and strategizing: Towards an activity-based view. Journal of Management Studies, 40(1), 3–22. Khan, A. M., & Manopichetwattana, V. (1989). Innovative and noninnovative small ﬁrms: Types and characteristics. Management Science, 35(5), 597–606. Lechner, C., Frankenberger, K., & Floyd, S. W. (2010). Task contingencies in the curvilinear relationships between intergroup networks and initiative performance. Academy of Management Journal, 53(4), 865–889. Li, J., & Tang, Y. I. (2010). CEO hubris and ﬁrm risk taking in China: The moderating role of managerial discretion. Academy of Management Journal, 53(1), 45–68. Ling, Y. A. N., Simsek, Z., Lubatkin, M. H., & Veiga, J. F. (2008). Transformational leadership’s role in promoting entrepreneurship: Examining the CEO-TMT Interface. Academy of Management Journal, 51(3), 557–576. Lubatkin, M. H., Simsek, Z., Yan, L., & Veiga, J. F. (2006). Ambidexterity and performance in small- to medium-sized ﬁrms: The pivotal role of top management team behavioral integration. Journal of Management, 32(5), 646–672. Maitlis, S., & Lawrence, T. B. (2003). Orchestral maneuvers in the dark: Understanding failure in organizational strategizing. Journal of Management Studies, 40(1), 109–139. Mantere, S. (2005). Strategic practices as enablers and disablers of championing activity. Strategic Organization, 3(2), 157–184. Mantere, S., & Sillince, J. A. A. (2007). Strategic intent as a rhetorical device. Scandinavian Journal of Management, 23(4), 406–423. Marginson, D., & McAulay, L. (2008). Exploring the debate on short-termism: A theoretical and empirical analysis. Strategic Management Journal, 29(3), 273–292. Marko´czy, L. (1997). Measuring beliefs: Accept no substitutes. Academy of Management Journal, 40(5), 1228–1242. McDonald, M. L., & Westphal, J. D. (2010). A little help here? Board control, CEO identiﬁcation with the corporate elite, and strategic help provided to CEOs and other ﬁrms. Academy of Management Journal, 53(2), 343–370. McGrath, R. G. (2001). Exploratory learning, innovative capacity, and managerial oversight. Academy of Management Journal, 44(1), 118–131. Mintzberg, H. (1973). A new look at the chief executive’s job. Organizational Dynamics, 1(3), 20–30. Mintzberg, H., Ahistrand, B., & Lampel, J. (1998). Strategy safari: A guided tour through the wilds of strategic management. New York: The Free Press. Mom, T. J. M., Van Den Bosch, F. A. J., & Volberda, H. W. (2007). Investigating managers’ exploration and exploitation activities: The inﬂuence of top-down,

Rediscovering the Individual in Strategy

29

bottom-up, and horizontal knowledge inﬂows. Journal of Management Studies, 44(6), 910–931. Nadkarni, S., & Herrmann, P. O. L. (2010). CEO personality, strategic ﬂexibility, and ﬁrm performance: The case of the Indian business process outsourcing industry. Academy of Management Journal, 53(5), 1050–1073. Pappas, J. M., & Wooldridge, B. (2007). Middle managers’ divergent strategic activity: An investigation of multiple measures of network centrality. Journal of Management Studies, 44(3), 323–341. Parker, L. D. (2007). Boardroom strategizing in professional associations: Processual and institutional perspectives. Journal of Management Studies, 44(8), 1454–1480. Penrose, E. (1959). The theory of the growth of the ﬁrm. New York: Wiley. Peterson, R. S., Martorana, P. V., Smith, D. B., & Owens, P. D. (2003). The impact of chief executive ofﬁcer personality on top management team dynamics: One mechanism by which leadership affects organizational performance. Journal of Applied Psychology, 88(5), 795. Regne´r, P. (2003). Strategy creation in the periphery: Inductive versus deductive strategy making. Journal of Management Studies, 40(1), 57–82. Resick, C. J., Weingarden, S. M., Whitman, D. S., & Hiller, N. J. (2009). The bright-side and the dark-side of CEO personality: Examining core self-evaluations, narcissism, transformational leadership, and strategic inﬂuence. Journal of Applied Psychology, 94(6), 1365–1381. Roth, K. (1995). Managing international interdependence: CEO characteristics in resourcebased framework. Academy of Management Journal, 38(1), 200–231. Rouleau, L. (2005). Micro-practices of strategic sensemaking and sensegiving: How middle managers interpret and sell change every day. Journal of Management Studies, 42(7), 1413–1441. SAP. (2010). Available at http://www.s-as-p.org/about Simon, M., & Houghton, S. M. (2003). The relationship between overconﬁdence and the introduction of risky products: Evidence from a ﬁeld study. Academy of Management Journal, 46(2), 139–149. Simsek, Z. (2007). CEO tenure and organizational performance: An intervening model. Strategic Management Journal, 28(6), 653–662. Simsek, Z., Heavey, C., & Veiga, J. F. (2010). The impact of CEO core self-evaluation on the ﬁrm’s entrepreneurial orientation. Strategic Management Journal, 31(1), 110–119. Simsek, Z., Veiga, J. F., & Lubatkin, M. H. (2007). The impact of managerial environmental perceptions on corporate entrepreneurship: Towards understanding discretionary slack’s pivotal role. Journal of Management Studies, 44(8), 1398–1424. Simsek, Z., Veiga, J. F., Lubatkin, M. H., & Dino, R. N. (2005). Modeling the multilevel determinants of top management team behavioral integration. Academy of Management Journal, 48(1), 69–84. Smith, K. G., Smith, K. A., Sims, H. P., Jr., O’Bannon, D. P., Scully, J. A., & Olian, J. D. (1994). Top management team demography and process: The role of social integration and communication. Administrative Science Quarterly, 39(3), 412–438. Teece, D. J. (2007). Explicating dynamic capabilities: The nature and microfoundations of (sustainable) enterprise performance. Strategic Management Journal, 28(13), 1319–1350. Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic Management Journal, 18(7), 509–533.

30

STEVEN W. FLOYD AND REBEKKA SPUTTEK

Vaara, E., Kleymann, B., & Seristo¨, H. (2004). Strategies as discursive constructions: The case of airline alliances. Journal of Management Studies, 41(1), 1–35. Wally, S., & Baum, J. R. (1994). Personal and structural determinants of the pace of strategic decision making. Academy of Management Journal, 37(4), 932. Whittington, R. (2006). Completing the practice turn in strategy research. Organization Studies, 27(5), 613–634. Zahra, S. A., Neubaum, D. O., & Naldi, L. (2007). The effects of ownership and governance on SMEs’ international knowledge-based resources. Small Business Economics, 29(3), 309–327. Zahra, S. A., Sapienza, H. J., & Davidsson, P. (2006). Entrepreneurship and dynamic capabilities: A review, model and research agenda. Journal of Management Studies, 43(4), 917–955. Zheng, J. Z., & Anand, J. (2009). A multilevel perspective on knowledge transfer: Evidence from the Chinese automotive industry. Strategic Management Journal, 30(9), 959–983.

INTO THE GREAT WIDE OPEN: BRIDGING THE MICRO–MACRO DIVIDE IN THE ORGANIZATIONAL SCIENCES M. Ronald Buckley, Maria Riaz Hamdani, Anthony C. Klotz and Sorin Valcea ABSTRACT Purpose – The purpose of this chapter is to establish some of the reasons why there exists a chasm between micro and macro disciplines of organizational sciences. We aim to suggest some fecund areas for bridging the gap between the micro and macro side of our science. Methodology/Approach – In this chapter, we have polled our colleagues to ascertain the areas that they believe have the most potential to bridge the micro–macro divide. In addition, we have reviewed extant literature to identify some of the areas where bridging work has already started. Findings – Through our survey and literature review, we have identiﬁed a number of areas which can help in narrowing the micro–macro divide. Social Implications – By suggesting some ways to bridge the micro– macro divide, this chapter helps in setting future research agenda that will help in viewing organizational problems from multiple lenses. Our work Building Methodological Bridges Research Methodology in Strategy and Management, Volume 6, 31–68 Copyright r 2011 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-8387/doi:10.1108/S1479-8387(2011)0000006006

31

32

M. RONALD BUCKLEY ET AL.

also encourages the scholars from various disciplines to explore ways that can integrate the broad disciplines of organizational sciences. Originality/Value of Paper – We have attempted to take the pulse of researchers in management disciplines concerning the chasm between micro and macro disciplines, and we have tried to integrate this information with the bridging research that has already been reported. Moreover, we have suggested a number of reasons why this gap is so difﬁcult to remediate. We discuss how bridging the gap is connected to the way in which we train, develop, and reward nascent scholars in our ﬁeld. Keywords: cross-disciplinary research; organizational studies; micro– macro research Many of us talk frequently about the ‘‘organizational sciences.’’ So much so that it is intuitive to visualize the organizational sciences as a monolithic research area. However, this is certainly not the case. Organizational sciences encompass many different disciplines making it a polytheistic ﬁeld. As Mahoney (1985, p. 15) stated, ‘‘The so-called organization sciences encompass scholarship in the more traditional social sciences of economics, sociology, psychology, political science, and related disciplines.’’ As a multidisciplinary ﬁeld (cf. Zammuto & Connolly, 1984), typiﬁed by different methodologies (e.g., qualitative and quantitative), different levels of analysis (e.g., individual, group, and organization), and different levels of paradigm development (e.g., sociology vs. economics), we must inevitably recognize the fragmentation among these disciplines and resign ourselves to the difﬁculties of achieving consensus. As the volume of information generated from these disparate areas grows, the integration of these disciplines be comes more and more difﬁcult to manage effectively. Owing to the plethora of information that is available in the organizational sciences, it has become impossible (and would be terribly inefﬁcient) to treat each piece of information individually. Thus, most have chosen to focus their work in an area based on their interests, abilities, market demands, or necessity. We in the organizational sciences have, for the sake of parsimony, developed a number of general groupings that we utilize to facilitate the communication, interpretation, dissemination, and understanding of that information generated in the research process. Two such groupings that are extensively used in the organizational sciences are the ‘‘micro approach to management’’ and the ‘‘macro approach to management.’’ In our persistent attempts to advance knowledge in the organizational sciences, these two

Bridging the Micro–Macro Divide in the Organizational Sciences

33

constructs have ostensibly developed sacrosanct status. When this is conﬂated with the fact that most research scholars generally conduct research efforts within a single level of analysis, a false opposition between researchers who study similar organizational phenomena from different levels of analysis (micro vs. macro) becomes evident. With respect to the macro and micro approaches to management, we have evolved into what is commonly known as the bifurcation fallacy. That is, many researchers now assume that the micro and macro approaches to management are two mutually exclusive possibilities when, in fact, spanning them is, in fact, feasible (see Hackman, 2003), and such integration would likely result in the creation of new knowledge. Our circumstances may also be emblematic of what is commonly known as a false dichotomy, creating an either micro or macro approach to issues in the organizational sciences. Of course, it is comfortable for us to think in terms of either-or’s. Unfortunately, with respect to the micro–macro divide in the organizational sciences, the world is much more complicated than the approach that many of us have taken in our attempts to develop knowledge in management. And therein lays the crux of one of our most intractable problems in the organizational sciences. This quandary was succinctly illustrated in this anecdote told by Steve Kerr in his keynote address for the Society of Industrial and Organizational Psychology in 2009: I ﬁrst ran into these dichotomies while teaching at Ohio State. We had an open faculty slot, and the department chair wanted to know who was micro and who was macro. It was apparently easy for everybody else to classify themselves, but I was studying reward systems then and had joined the ﬂourishing Ohio State leadership program, and I didn’t see how you could get very far in either rewards or leadership if you only read the micro or only read the macro literature. It just seemed kind of odd. If I had been better read I would have used Russ Ackoff’s response. When confronted with a similar question, Russ observed that ‘‘nature is not organized in the same manner as universities.’’

WITH OR WITHOUT YOU: A BIT OF HISTORY ON THE MICRO–MACRO DIVIDE A short reprise of the historical roots of organizational science reveals that it wasn’t always like this. Early researchers in management, unconstrained by the methodological requirements of organization science research today, appear to have rather effortlessly and seamlessly merged the micro and macro aspects of investigating organizational phenomena (see Hatch, 1997 for a review). Frederick Taylor’s introduction of Scientiﬁc Management set the stage for management control systems by focusing on the efﬁciency of each

34

M. RONALD BUCKLEY ET AL.

worker. Henri Fayol’s contributions spanned from the speciﬁc responsibilities of managers to the importance of organizational culture, especially among the rank and ﬁle workers in organizations. Max Weber’s concept of the bureaucracy is still used today to refer to certain organizational structures, but he also delved into human resources topics such as managerial training and ﬁling systems. Chester Barnard wrote primarily about the executive level of management, yet his contribution is pivotal for the development of the ﬁeld of organizational behavior in that he emphasized communication and motivation from the top-down. The study of organizations has progressed extensively since the seminal work of these early pioneers. Now, the Academy of Management boasts nearly 20,000 members across 25 divisions and interest groups. The four largest divisions are organizational behavior (OB) and human resources (HR) (i.e., the micro side), and business policy and strategy (BPS) and organization and management theory (OT) (i.e., the macro side). It is not surprising then that, when the micro–macro dichotomy in management is discussed, the ‘‘micro’’ often refers to OB and HR while the ‘‘macro’’ refers to BPS and OT (e.g., Hitt, Beamish, Jackson, & Mathieu, 2007; Short, Palmer, & Ketchen, 2003). Somewhere in the recent history of management thought, scholars began to feel that they can live with or without other critical disciplines of the organization sciences. A brief look at the history of these four major disciplines sheds some understanding on the development of this perspective. Organizational behavior was born from the mother disciplines of primarily psychology, but also sociology, and anthropology (Mahoney, 1985). Owing to this heritage, methodological rigor is a sacrosanct founding principle of the area, albeit sometimes at the expense of dependent variables that are both meaningful and relevant to practitioners. The study of HR emerged from industrial/organizational psychology, which itself was developed to study the most practical of business challenges, most notably that of mental testing for personnel selection purposes for both the United States Military Forces and business organizations (Landy & Shankster, 1994). Much of OT’s history has been written by sociologists, philosophers, political scientists, and economists, as well as practitioner/theorists such as the aforementioned Fayol (Hatch, 1997). Finally, the formal study of business strategy emerged from three very practical forces – the need by organizations for more sophisticated tools to cope with the instable business environment, the birth of strategic consulting groups, and the evolution and prevalence of the diversiﬁed ﬁrm (Rumelt, Schendel, & Teece, 1991). Given the diverse origins of these four disciplines, it is not surprising that many silos exist in their methodologies, constructs, and principles. In our

Bridging the Micro–Macro Divide in the Organizational Sciences

35

opinion, that should not give pause; given the different formative exercises experienced by each discipline, there should be silos. Clearly, there are areas in micro, such as cognition and training, and areas in macro, such as population ecology and strategic groups, that will primarily inform research and practice on their ‘‘side of the house.’’ However, given the shared ancestry of certain elements of these disciplines, such as HR and strategy’s focus on practical application and OB and OT’s sociological concerns, there are many areas where ‘‘bridge building’’ might be a worthwhile endeavor. Exemplars of this type of scholarship abound, from top management team research to the strategic human resources literature. The purpose of this chapter is to look at the process through which some of these bridges have been successfully built, and then extend this successful bridge building to constructs that are currently happily residing in silos in one discipline. To do this, we ﬁrst share the barriers and opportunities to bridge-building research identiﬁed by a sample survey of the membership of the Academy of Management. Next, we develop a model of crossing the divide based on some constructs that have successfully climbed up and down the ladder between micro and macro domains of research. We then identify constructs from OB/HR and from strategy/OT that are ripe for making the multilevel move and apply our model to illustrate how both sides can continue to push the frontiers of knowledge in the other. It is our hope that in venturing into the great wide open between the micro and macro disciplines of management research, we will provide a blueprint for future bridges to be built, thereby increasing the relevance and insightfulness of our work to one another and to our partners in industry.

MORE THAN A FEELING: A SURVEY OF OUR COLLEAGUES The micro–macro divide continues to be a salient topic of interest for management scholars. Indeed, Roberts, Hulin, and Rousseau (1978) recognized that boundaries such as these ‘‘hamper integration of information from different disciplines; thus, a signal considered of paramount importance in one discipline is often ignored as noise in the other’’ (pp. 27–28). This quote cuts to the core of this issue; we have a tendency to treat as noise that which falls outside of our areas of interest. In the three decades since, much insightful work has been done to understand and address this schism created by paradigm adoption and paradigm exclusion (e.g., Hitt et al., 2007;

36

M. RONALD BUCKLEY ET AL.

Rousseau, 1985; Pfeffer, 1993). Despite this progress, there is debate concerning the existence of the divide, its causes, and the amount of attention it should receive. To assess the pulse of our colleagues in the management disciplines concerning the micro–macro divide and ﬁnd evidence that this problem is actually more than a feeling, we collected data from the three largest divisions of the Academy of Management – (OB) organizational behavior, (HR) human resources management, and (BPS) strategy. Their responses to a number of questions about both their side of the divide and the other side of the divide provided a framework for our analysis of the current state of the schism within the disciplines of organizational science. The survey link was sent to all members of the OB, HR, and BPS divisions through list server email. These divisions have 2,601, 1,201, and 2,255 members who subscribe to their listservs, respectively. We received 292 responses for a very conservative overall response rate of 4.8% (given there is likely signiﬁcant overlap between these groups, especially OB and HR). Of these, 192 respondents clearly identiﬁed themselves as either OB or HRM, 95 identiﬁed themselves as BPS, and 5 saw themselves as either both OB and BPS or neither. Our initial question concerned respondent perception of the importance of cross-disciplinary research between the micro and macro disciplines for the future of management research. An overwhelming majority (91%) of respondents felt that this type of research was either important or very important for the future of our ﬁeld. So we can conﬁdently conclude that the micro–macro divide is more than a feeling; it is an important matter to our colleagues and should be looked at in more detail.

I WANNA HOLD YOUR HAND: OPPORTUNITIES FOR CROSS-DISCIPLINE BRIDGE BUILDING We also asked researchers to highlight the most interesting theories in management as well as the most important constructs and theories on their side of the divide and on the other side of the divide. In doing this, our goal was to identify the theories and constructs that have been most successful in bringing micro and macro researchers together and the areas that may be most conducive to making the step up or leap down the organizational science ladder. Table 1 summarizes the responses to this series of questions. It is clear that most respondents from both sides have a broad set of ideas from their own discipline that they feel can contribute to the other side of

Resource-bBased view Decision making Social exchange Theory Identity theories Cultural issues Motivation theories Goal setting theory Teams/groups Institutional theory Behavioral theory of ﬁrm Agency theory Leadership HR management

Most interesting theory or idea in management

All Respondents

Leadership Teams/groups Decision making Human resources Identity Motivation Emotion Learning Culture Ethics Justice Trust Climate Turnover/retention

OB/HR ideas that can contribute to strategy Resource-based view Agency theory

Most inﬂuential strategy theory on OB/HR Resource-based view Entrepreneurship Opportunity Agency theory Governance Network theory Governance Competitive advantage Learning Real options

Leadership Decision making TMT/upper echelons

Most inﬂuential OB/HR theory on strategy

Strategy Respondents Only Strategy ideas that can contribute to OB/HR

Micro–Macro Survey Results.

OB/HR Respondents Only

Table 1.

Bridging the Micro–Macro Divide in the Organizational Sciences 37

38

M. RONALD BUCKLEY ET AL.

the divide. However, respondents cited far fewer concepts from the other side of the divide that have contributed to their own discipline. This is not surprising, as our ﬁeld encourages and rewards specialization, but it is worth noting that 41% of the BPS respondents and 52% of the OB/HR respondents did not even answer the questions requesting an acknowledgement of the contributions of the other side of the divide to their own area of expertise. However, there is some agreement on certain theories and ideas that are likely to bridge the cross-discipline divide, including the resource-based view (RBV) of the ﬁrm (Barney, 1986), decision making, agency theory, leadership studies, and teams research. In these areas, both OB/HR and BPS researchers seem to agree on the beneﬁt of work in these areas to both sides of the schism. Taken together, however, these areas account for only 34% of the responses, and many other bridging opportunities are identiﬁed. In sum, these responses point out some notable bridge-crossing successes but also suggest that many possible niches of cross-disciplinary research have yet to be explored and exploited.

ANOTHER BRICK IN THE WALL: BARRIERS TO BRIDGE BUILDING We also asked survey participants to describe the most pressing factors that impede cross-disciplinary research in OB/HR and BPS. The most common responses, ranked in order, were publishing risks, methodological challenges, theory base differences, and doctoral education. Each of these challenges has been discussed in previous literature. Here, we brieﬂy incorporate the speciﬁc concerns of our respondents with this previous work. The risks associated with trying to successfully publish cross-discipline research in premier outlets were the most commonly cited bricks in the wall between micro and macro scholarship. This type of work is seen as being considerably more time-intensive and energy-intensive, yet there exists considerable uncertainty as to how it will be received by journal reviewers. One respondent put it simply by saying that it ‘‘takes time to develop an understanding of how issues are framed and regarded in related subdisciplines, so the payoff is not there.’’ For example, a strategy researcher integrating motivational theories may be unsure how the reviewers at a strategy-oriented journal will receive and evaluate this type of work. Will

Bridging the Micro–Macro Divide in the Organizational Sciences

39

they agree that this type of work is a signiﬁcant contribution? Or will it be evaluated as something that is misplaced? Furthermore, does it clearly ﬁt within the guidelines of the journal’s mission? Indeed, almost a third of the rejections at our major scholarly journals are desk rejections and perhaps those who have attempted this type of research may have inﬂated these numbers. Furthermore, an author may also be uneasy about the type of reviewer that will evaluate such an article. If the article is assigned to a motivational researcher, that person may balk at the extension of this work to such a high level. If the article is assigned to a strategy reviewer, he or she may question the contribution to the macro-level literature. In sum, ‘‘This work looks to be difﬁcult – therefore not a very promising investment for a researcher’s time. This is likely especially the case for younger researchers who are still trying to build a reputation and earn tenure.’’ As Rynes (2006) describes, journal boards must take deliberate action to encourage openness to new ideas, otherwise the review process gravitates toward ‘‘conservatism and incrementalism.’’ On the basis of the responses we received, it would appear that many authors feel that journals are already wallowing in this rut as it pertains to maintaining a welcoming attitude with respect to crossdiscipline research. Second, methodological challenges discourage some researchers from engaging in macro-micro spanning research. OB/HR researchers are challenged enough with ever-increasing requirements such as multisource data, longitudinal designs, and large sample sizes at the individual level and agreement concerns at the team level. In the same way, strategy researchers struggle with sample sizes, levels of analysis, and statistical anomalies with large scale data. For many, expanding the scope and rigor of the study to include variables more than one level away has become a signiﬁcant challenge. One respondent commented, ‘‘May require a mixed methods methodology which is far more time consuming.’’ Overall, a consensus reports that this type of research is more time consuming and is accompanied with a smaller probability of a publication payoff. Third, as shown in Table 1, OB/HR researchers only identiﬁed two inﬂuential strategy theories, while respondents from strategy merely highlighted a handful of inﬂuential theories from the micro side, two of which – leadership and top management teams – are clearly very similar. These ﬁndings could demonstrate that there simply is consensus among researchers as to what the most inﬂuential theories from the other discipline have most impacted their side of the divide. While this may partially be the case, many of the qualitative responses alluded to the fact that perhaps this seemingly narrow perception of cross-discipline inﬂuence indicates a lack of

40

M. RONALD BUCKLEY ET AL.

understanding of the other side’s work and contribution. This drives a situation where, according to one respondent, ‘‘It is difﬁcult to ﬁnd micro people who know enough macro ‘language’ and research gaps and to ﬁnd macro people who know enough micro ‘language’ and research gaps to converse well enough to ﬁnd areas of fruitful collaboration and overlap.’’ As this respondent has so ably pointed out, the solution to the dearth of knowledge of the other side of the divide is not going to come from researchers simply spending more time learning the theories and advances in the other domain. After all, it is challenging enough to stay on top of the most recent developments in our own research streams. Part of the solution likely lies in collaborations between macro and micro researchers. There are signs that mechanisms are being built to facilitate this type of collaborative work; a new interest group, Strategizing Activities and Practice, formed before the 2010 Academy of Management meeting in Montreal. Part of the mission of this group is to study ‘‘How do such things as knowledge, identity, and emotions affect the strategic work of individuals?’’ Clearly, this is an encouraging signal for the future of micro–macro spanning research in the organizational sciences. Related to this, the structure of doctoral training often does not foster a sufﬁcient understanding of the corresponding side of the micro–macro divide. This was apparent in some of the open-ended responses that alluded to the fact that many programs do not require micro or macro students to even take a doctoral seminar in a BPS or OB/HR seminar (in a ‘‘competing’’ discipline), respectively. Indeed, this under-identiﬁed model of doctoral training will, necessarily, in most cases result in a lifetime of specialized work. As one respondent put it, ‘‘The entire career process for researchers is geared to slicing knowledge into ever thinner slices for publication, ﬁrstly as a doctoral thesis and then as articles in journals that are themselves highly specialized.’’ While this statement is true, there is no simple way to remediate this issue. First, Ph.D. students are under growing pressure to specify their interests and develop research streams earlier and earlier in their doctoral education. Whereas, for many of us who received their degrees some time ago, the ﬁrst few years of doctoral training was a time for exploration and experimentation in the entire ﬁeld of management, students now feel obligated to choose an area of specialization as soon as possible so they can begin developing a pipeline of research as soon as possible. Applicants for assistant professor jobs invariably have a slide denoting their ‘‘research stream’’ when all they possess is a tightly deﬁned research trickle. As a well-known senior management scholar (Professor Pamela Perrewe of Florida State University) recently commented in a

Bridging the Micro–Macro Divide in the Organizational Sciences

41

presentation to doctoral students and faculty ‘‘Research streams used to be something you worry about when you go up for tenure, now it is something that you have to worry about as a doctoral student.’’ It has taken a number of years for our ﬁeld to trend in this direction, and we believe it is unlikely that this trend will reverse anytime soon. The incentives that used to exist for a doctoral student to spend time developing a broad understanding of all the areas of management seem to no longer be potent. Fifth, and ﬁnally, a frequently cited reason for the gap in scholarship between OB/HR and strategy can best be described as ‘‘political reasons.’’ Indeed, responses such as ‘‘many are committed to protect the ‘unique’ knowledge of their discipline from ‘intruders’’’ and ‘‘ﬁelds develop vested interests, cultures and boundaries (institutional characteristics) that often impede cross-disciplinary research’’ capture the value of staying on your own turf, especially for nascent scholars. It is politically unwise and exceedingly difﬁcult to attempt to make contributions to the other side when many buy into the notion of the purity of their side of the divide and will protect their side from outsiders who try to make contributions. This barrier is not especially surprising, as both nascent and longtime contributors to organizational sciences are trained to be critical of new ideas, lest we succumb to construct proliferation and the weak theoretical grounding that accompanies such expansion. Furthermore, one of the strengths of the organizational sciences is the large number of outstanding scholars that are extremely passionate about their area of research. When this passion and critical mindset are combined, the barriers to entry encountered when one attempts to contribute to an established research area become daunting. The beneﬁts this brings to academic rigor have a large downside – it serves to signiﬁcantly reduce the odds of success for authors attempting to bridge the micro–macro divide.

IT’S ALL BEEN DONE BEFORE: SUCCESS STORIES IN BRIDGING THE DIVIDE There have been a number of insightful, well-designed, and thought-out studies that have bridged the micro and macro chasm. In this section, our aim is to highlight some of the topics that have helped in bridging the micro–macro research gap. To review the research we examined articles published in a subset of the leading scholarly journals such as Academy of Management

42

M. RONALD BUCKLEY ET AL.

Journal, Academy of Management Review, Strategic Management Journal, Journal of Applied Psychology, Personnel Psychology, Organizational Behavior and Human Decision Processes, Strategic Entrepreneurship Journal, Administrative Science Quarterly, and Management Science. We did not limit our search to any particular timeframe. Our search indicates that articles guiding on various theoretical and methodological issues in conducting multilevel and cross-level research are available since the 1980s (e.g., Mossholder & Bedeian, 1983; Rousseau, 1985). However, articles actually incorporating a cross-disciplinary framework have primarily started appearing in the research literature since late 1990s and onward.

Micro–Macro Articles Deﬁned We want to clarify at the onset the deﬁnition of micro–macro articles that we have used for article selection in this review. We deﬁne micro–macro articles as those articles that combine theoretical concepts from various management disciplines to examine an organizational phenomenon. We consider micro–macro articles to be cross-disciplinary in nature. Thus, for our discussion, an article with multiple levels of analysis is not necessarily a micro–macro article. As explained by Molloy, Ployhart, and Wright (2011), two major types of divides exist in the management literature – a system-level divide and a disciplinary divide. A system-level divide concerns the hierarchical nature of entities involved in a theory. That is, it relates to whether the primary focus of the theory is individuals, groups, organizations, industries, or the larger economic or social sphere. On the contrary, a disciplinary divide concerns the fundamental organized principles of knowledge that are used as a primary source for theory building. In the management literature, these disciplines can be seen as consisting of various specializations such as strategy, organizational theory, organizational behavior, human resources management, entrepreneurship, and international business. All these disciplines trace their roots back to one of the three foundational areas of economics, sociology, and psychology. A multilevel article that spans various system-level divides does not necessarily bridge various disciplines. Thus, for our discussion a singlediscipline, multilevel article will not be classiﬁed as a micro–macro article. Some example of single-discipline, multilevel articles include multilevel articles in group work (e.g., Barry & Stewart, 1997; Tasa, Seijts, & Taggar, 2007; Zyphur, Narayanan, Koh, & Koh, 2009), multilevel models in

Bridging the Micro–Macro Divide in the Organizational Sciences

43

organization climate (e.g., Fedor, Caldwell, & Herold, 2006; Mayer, Nishii, Schneider, & Goldstein, 2007; Taras, Steel, & Kirkman, 2010), and various organizational theory and strategy articles (e.g., Holm, 1995; Hough, 2006; Mani, Antia, & Rindﬂeisch, 2007; Misangyi, Elms, Greckhamer, & Lepine, 2006). These articles explore the effects of a phenomenon at multiple levels by primarily employing the theoretical framework of their own disciplines. Similarly, a study design to investigate the effects of micro-level variable on some macro-level indicator such as proﬁts or sales characterizes a multilevel approach and seldom employs cross-disciplinary frameworks. For example, McKay, Avery, and Morris (2009) demonstrated that subordinate and supervisor perceptions of a pro-diversity environment impact retail store sales. Using a multilevel model of organizational climate (e.g., Zohar, 2000), they reported that that greatest sales growth was found when both the subordinate and the supervisor perceived a pro-diversity environment (McKay et al., 2009). In a similar fashion, employees’ customer service attitudes and behaviors have been linked to unit-level performance when there is a customer service climate (Grizzle, Zablah, Brown, Mowen, & Lee, 2009). These articles utilize variables at multiple levels but essentially remain within one primary conceptual domain to explain an organizational phenomenon; so while they are multilevel, they are not cross-disciplinary. On the basis of this conceptualization of micro–macro articles, we searched for areas where cross-disciplinary work has already started (Table 2). As we have suggested, various disciplines in management have a distinct system-level focus (e.g., organizational behavior researchers are interested in individuals and groups and strategy scholars are interested in organizations and industries). Therefore, in general, micro–macro articles (i.e., cross-disciplinary articles) are multilevel in nature as well. This accounts for the complexity of theoretical and methodological challenges involved in crafting micro–macro articles. Despite these challenges, our review indicates that many promising areas of cross-disciplinary work in the organizational sciences have emerged over the past many years. Top management team research, strategic human resources management, social networks, international management, and organizational learning are some of the areas where theoretical lenses from various management disciplines are being used to cross the micro–macro schism. In the section below, we brieﬂy discuss some of these areas to draw a picture of the bridge that has already started to be built by a number of insightful management scholars.

Macro Literature

Agency theory and structuralist view of board power and control

Corporate diversiﬁcation and theory of macro or corporate control

Authors

Westphal (1998)

Rowe & Wright (1997)

Theory of micro controls exerted by human resources management function

Interpersonal inﬂuence behaviors (e.g., ingratiation and persuasion)

Future Areas of Research

Can board members learn to avoid CEO ingratiation and persuasion behavior toward board members biases in their monitoring behavior are associated with changes in resulting from managerial inﬂuence board structure that increase the tactics? How upward managerial inﬂuence board’s independence from management. Moreover, CEO’s tactics may vary across cultures? What type of behavioral process inﬂuence behaviors moderate the effect of increased structural board and structural context can enhance independence on corporate strategy board effectiveness? and CEO compensation policy How related and unrelated ﬁrms The type of corporate (macro) may beneﬁt differently from the use controls used by related or of various practices (e.g., unrelated diversiﬁed ﬁrms results in innovative HRM practices) over a relative emphasis on either the long run? ﬂexibility or ﬁt among HRM practices. The ﬁrms diversiﬁed in related areas emphasize ﬂexibility and those diversiﬁed in unrelated areas emphasize ﬁt. This emphasis on ﬂexibility or ﬁt, in turn, has implications for the use of HRM (micro) controls such as clan, behavior, and outcome controls. The related ﬁrms exhibit the use of all three types of micro controls, while unrelated ﬁrms exhibit a relative emphasis on the use of outcome controls

Key Cross-Disciplinary Assertions

Summary of Exemplary Cross-Disciplinary Studies.

Micro Literature

Table 2.

44 M. RONALD BUCKLEY ET AL.

Corporate governance and varieties of capitalism

Population ecology

Aguilera, Rupp, Williams, and Ganapathi (2007)

Eckhardt and Ciuchta (2008)

Organizations are pressured to engage Is there a linkage between in CSR by many different actors employee perceptions of CSR and (e.g., employees, government, employee behaviors, commitment NGOs, consumers, and and job satisfaction? How actors’ motives at different shareholders) each driven by instrumental, relational, and moral levels may interact to predict motives increased CSR and, consequently, positive social change? What factors limit or enhance an The population-level outcomes of individual selection models in entrepreneur’s ability to modify entrepreneurship foster a multistage selection criteria, as well as how selection process that evolves modiﬁcations to selection criteria through time. More speciﬁcally, might differentially inﬂuence the ﬁtness of the population of entrepreneurship can be viewed as a opportunities being pursued at a funneling process that starts with a speciﬁc point in time? pool of opportunities at a speciﬁc What is the relationship between point in time. This pool of the speciﬁc characteristics of opportunities is winnowed through opportunities or selection settings staged selection events, yielding a subset of successfully exploited and speciﬁc organizing strategies opportunities. Those opportunities such as business models? that are successfully exploited become inputs into established evolutionary theories that start with the existence of an organization. Those opportunities that are not successfully exploited at a given point in time may be available for exploitation in a following period

Organizational justice

Theories of differences in individual attributes, prior knowledge, and cognition Individual opportunity nexus literature in entrepreneurship

Bridging the Micro–Macro Divide in the Organizational Sciences 45

Macro Literature

Competitive strategy and ﬁrm performance

Authors

Acquaah (2007)

Theories of an individuals’ social capital and networking.

Micro Literature The social capital developed from managerial networking and social relationships with various stakeholders (i.e., top executives in other ﬁrms, government ofﬁcials, and community leaders) enhance organizational performance. However, the impact of social capital on organizational performance differs between ﬁrms that pursue the different competitive strategies (low-cost, differentiation, and combination of low-cost and differentiation) and those who do not pursue those strategies

Key Cross-Disciplinary Assertions

Table 2. (Continued )

Do various sources of social capital differ in inﬂuencing long-term performance of organizations and other non-economic outcomes (e.g., social change)?

Future Areas of Research

46 M. RONALD BUCKLEY ET AL.

Bridging the Micro–Macro Divide in the Organizational Sciences

47

Top Management Teams One promising area for a cross-disciplinary management research is top management team research in strategy. Many strategy scholars have utilized theories of individual behavior to explain variance in ﬁrm-level outputs due to top executives’ decisions. The seminal work of Hambrick and Mason (1984) encouraged many strategy researchers to explore the impact of individual differences in top management teams. This focus has primarily resulted in the use of demographic characteristics as proxies for deep-level individual differences that shape team processes (e.g., Carpenter & Fredrickson, 2001; Sanders & Carpenter, 1998; Wiersema & Bantel, 1992). Some researchers have also focused on team processes such as communication (O’Reilly, Snyder, & Boothe, 1993) and social integration (Smith et al., 1994). However, the focus is shifting toward developing a more in-depth understanding of individual behavior in top management teams. For example, Hambrick (1994) proposed the concept of behavioral integration to capture how team collaboration, information exchange, and focus toward joint decision-making impact various strategic-level outcomes. Extending this work, Simsek, Veiga, Lubatkin, and Dino (2005) proposed various individual-level (e.g., CEO’s collectivist orientation), team-level (e.g., team diversity, size, and tenure), and ﬁrm-level (e.g., ﬁrm size, performance, and age) antecedents of behavioral integration in executive teams. Using predictors at all three levels explained maximum variance in the behavioral integration variable. Clearly, top management team research is an area that has been developing stronger and stronger ties between the macro and micro domains for nearly three decades.

Executive Compensation Executive compensation is another example where, when it comes to crossdiscipline research, it’s all been done before. Researchers have used concepts from economics, psychology, and sociology to study the antecedents and consequences of compensation in various management disciplines. Within the area of top management teams, micro–macro theories have been combined to investigate executive compensation (e.g., Carpenter & Wade, 2002; Fredrickson, Davis-Blake, & Sanders, 2010; Westphal, 1998; Zajac & Westphal, 1995). Westphal (1998) integrated micro–macro work by proposing that the inﬂuence tactics (e.g., ingratiation and persuasion) used by chief executive ofﬁcers shape the relationship between board structure and

48

M. RONALD BUCKLEY ET AL.

organizational outcomes (e.g., corporate diversiﬁcation and CEO’s compensation). He argued that considering the role of interpersonal inﬂuence tactics may help explain previous inconsistent ﬁndings regarding the relationship between board independence and diversiﬁcation. He proposed that increases in the board’s independence from management will be associated with higher levels of CEO ingratiation and persuasion behavior toward board members and that such inﬂuence behaviors, in turn, offset the effect of increased structural board independence on corporate strategy and CEO compensation policy. Similarly, Carpenter and Wade (2002) used complementary perspectives from theories of power, resources dependence, and upper echelons to propose how pay of non-CEO executives is inﬂuenced by a number of multilevel factors such as the ﬁrm’s strategy and related resource allocation decisions, the functional background of the ﬁrm’s CEO, and an incumbent’s managerial experience, functional expertise, and tenure.

Strategic Human Resources Management Another well-established micro–macro area of investigation extends out of the function of human resources management. The concept of strategic human resources management has provided a rich framework to explain how organizational practices generate higher-level inﬂuences. As explained by Wright, Dunford, and Snell (2001), the development of strategic human resources management has not only helped both strategy and human resources scholars in their own disciplines but also aided in bridging the micro–macro gulf. The application of the RBV of the ﬁrm (Barney, 1991) from the strategy literature has provided theoretical guidance to the assertion that people can be a source of competitive advantage. Likewise, human resources management research has in turn advanced strategy research by providing solid empirical support for the RBV. The merger of the RBV and human resources management has been tremendously popular and has resulted in numerous articles explaining how human resource practices at the individual level impact ﬁrm-level outcomes (e.g., Becker & Gerhart, 1996; Birdi et al., 2008; Gong, Law, & Chang, 2009; Huselid, Jackson, & Schuler, 1997), social networks of top executives (Collins & Clark, 2003), micro control systems (Snell, 1992), international human resources management practices (e.g., Milliman, Von Glinow, & Nathan, 1991; Taylor, Beechler, & Napier, 1996), organizational context (Soo Min, Morgeson, & Campion, 2008), and multiple stakeholder climate (Chuang & Liao, 2010).

Bridging the Micro–Macro Divide in the Organizational Sciences

49

However, RBV is not the only theory that has spanned the human resources management and strategy literatures. For example, Rowe and Wright (1997) proposed that business diversiﬁcation is related to micro controls (e.g., clan, behavior, and outcome) exerted by human resources management function at a divisional level. They argued that related and unrelated diversiﬁcation requires either ﬂexibility or ﬁt. As related ﬁrms require ﬂexibility, these ﬁrms use all three types of human resources controls. However, unrelated ﬁrms only require ﬁt and thus, solely focus on outcome controls.

Corporate Social Responsibility The recent developments in the business sector have heightened interest in the area of corporate social responsibility (CSR). Not surprisingly, scholars have applied multiple paradigms to understand the concept of CSR. Aguilera, Rupp, Williams, and Ganapathi (2007) provided a multilevel theoretical model to understand why business organizations are increasingly engaging in CSR. The authors examined CSR at individual, organizational, country, and transnational levels. In doing so, authors combined theories from ethics, international business, organizational behavior, sociology, and legal studies. More speciﬁcally, drawing from the organizational justice, corporate governance, and varieties of capitalism literatures, the authors proposed that various stakeholders (e.g., government agencies, NGOs, shareholders, and employees) have multiple motives (e.g., instrumental, moral, and relational) to facilitate organizational engagement in the acts of social responsibility. Similarly, a multidisciplinary work by De Luque, Washburn, Waldman, and House (2008) examined relationships between CEO values (economic values or broader social values), followers’ perceptions of CEOs, followers’ in-role and extra-role behaviors (individual-level outcomes), and ﬁrm performance. They found that CEO’s possessing a broader deﬁnition of ﬁrm’s value (i.e., values emphasizing stakeholders and social performance) were perceived to be visionary by the followers, and consequently these followers engaged more in extra-role behaviors. These extra-role behaviors were found to be positively related to ﬁrm performance.

Social Networks Social capital and social networks have also sparked the interest of researchers from various disciplines. Traditionally, social capital has been

50

M. RONALD BUCKLEY ET AL.

an area of investigation in organizational behavior (e.g., Bolino, Turnley, & Bloodgood, 2002; Dess & Shaw, 2001; Leana & Van Buren, 1999; Mehra, Kilduff, & Brass, 2001; Oh, Chung, & Labiance, 2004; Seibert, Kraimer, & Liden, 2001) and organizational theory (e.g., Brass, 1984; Brass, Galaskiewicz, Greve, & Tsai, 2004; Brass, 1985; Burt, 1997; Nahapiet & Ghoshal, 1998; Tsai & Ghoshal, 1998; Zeitz, 1980). Recently however, researchers have begun to use these concepts to integrate various disciplines. For example, Acquaah (2007) applied social capital theory in the strategy literature to examine how a manager’s social capital (e.g., relationships with politicians and higher-level executives at other ﬁrms) increases organizational performance. These effects were moderated by an organization’s strategy. Inkpen and Tsang (2005) combined social network and knowledge transfer literature to show how various types of social networks (e.g., intracorporate networks, strategic alliances, and industrial districts) impact the nature of social capital embedded in a network and how social capital in turn shapes various conditions inﬂuencing knowledge transfer and learning. In addition to knowledge transfer, network and social capital theories have also been used to explain diverse topics such as ﬁrms’ competitive behavior (Gnyawali & Madhavan, 2001), strategic change (Haynes & Hillman, 2010), managerial performance (Moran, 2005), inter-unit linkages in newly formed business units in a company (Wenspin, 2000), business unit performance (Carmeli & Azeroual, 2009), and executive compensation (Geletkanycz, Boyd, & Finkelstein, 2001). Some researchers have also found negative effects of social capital at the organizational level of analysis (e.g., Blyler & Coff, 2003; McDonald & Westphal, 2003; Molina-Morales & Martı´ nez-Ferna´ndez, 2009). International business is another broad scholarly domain where social capital research has been successfully applied. Kostova (1999) linked theories of social capital with theories of multinational corporations (MNCs) in a model spanning individual, organizational, and country levels. She used a cross-disciplinary perspective to explicate the factors determining the success of transfer of strategic organizational practices. In her model, success of transfer primarily depends on an individual employee and is embedded in social, organizational, and relational contexts. She used institutional and resource dependence perspectives to explain how knowledge transfer can inﬂuence an employee’s commitment, job satisfaction, and psychological attachment through implementation of the practice in the recipient unit. In a subsequent article, Kostova and Roth (2003) developed a theory of the creation of social capital in foreign subunits of MNCs. They argue that the required levels and forms of social capital are determined by

Bridging the Micro–Macro Divide in the Organizational Sciences

51

the nature of interdependence between headquarters and subunits. In their micro–macro model of social capital formation, the motivations and personal characteristics of boundary spanners help translate their individual social capital to a sub-unit’s social capital.

Organizational Learning Considering the increased focus on learning and emergence of the knowledge economy, it is not surprising that much needed bridging of theories has started in the area of organizational learning in strategy and individual and team-based learning in organizational behavior. For instance, Zhao and Anand (2009) used a multilevel, cross-disciplinary framework to explain knowledge transfer in multinational ﬁrms. They provided support for their hypothesis that in organizations, the processes of teaching and absorptive capacity exist both at an individual level and at a collective level, and that collective mechanisms are more effective in transferring individual knowledge. Huy (1999) built a model to explain how individuals’ behaviors in organizations impacts macro-level factors in the organization. He used theories of emotional intelligence to explain how employees’ capabilities to deal with emotions facilitate their personal adaptation to change. Similarly, at the macro level, an organization’s capability to respond to human emotions is related to its capability to successfully go through a radical change. Both organizations and individuals transform though the isomorphic processes of emotional dynamics and change dynamics.

Decision Making Decision making has traditionally combined theories from economics and psychology. In a recent article, Guler (2007) added theoretical concepts from organizational theory to examine the investment decisions made in organizations. Guler (2007) asserts that political and institutional inﬂuences also play a critical part in leading executives to avoid terminating unsuccessful investments. She used qualitative and quantities methodologies to examine multilevel inﬂuences on sequential investment decisions in the US venture capital industry. The results indicate that pressures to continue investments despite evidence that expected returns are declining stem from multiple sources. She found that with increasing rounds of investments,

52

M. RONALD BUCKLEY ET AL.

individual reluctance to terminate investments stems from intraorganizational politics and coercive and normative pressures from co-investors and limited partners, in addition to various emotional, cognitive, and group biases.

Entrepreneurship In entrepreneurship, many scholars have combined theories from various disciplines to explain organizational phenomenon. Eckhardt and Ciuchta (2008) examined how micro-level theories inﬂuence macro-level outcomes. More speciﬁcally, it is proposed that entrepreneurs’ personal traits and decisions regarding which opportunities to pursue interact with the nature of available opportunities to give rise to a multistage selection process that has implications for the for populations of available opportunities and technologies. Similarly, Simons and Roberts (2008) examined how an entrepreneur’s previous work experiences inﬂuences an entrepreneur’s likelihood to pursue a new market and the organizational form of the entrepreneur would use to exploit the market. In fact, the study of entrepreneurship is, almost by necessity, a cross-disciplinary endeavor. As an example, the recent explosion of research on entrepreneurial opportunities has pulled from theories at every level of analysis (see Busenitz et al., 2003).

SHOULD I STAY OR SHOULD I GO?: ACADEMIC BACKGROUND AND THE DECISION TO BRIDGE THE DIVIDE At this point, we would also like to say a few words on the background of researchers engaged in cross-disciplinary research. Our purpose is to investigate whether academic background hinders or aids in bridging the gap. That is, when researchers have a research idea that spans multiple disciplines, how does their primary domain inﬂuence their decision whether to stay or go? When we reviewed the disciplinary background of those involved in micro–macro research papers, we noticed an encouraging trend toward research collaboration. For instance, the study of top management teams has inspired researchers from various disciplines to collaborate. Consider the article by Simsek et al. (2005); it is a result of collaboration between group of researchers belonging to various disciplines of strategy,

Bridging the Micro–Macro Divide in the Organizational Sciences

53

organizational behavior, and entrepreneurship. Strategic human resources management is another area where scholars from various disciplines have collaborated. For example, Rowe and Right’s (1997) work is a result of the joint efforts of strategy and human resources management scholars, yet again reinforcing the trend of cross-disciplinary partnership. However, collaboration is not a necessary condition for cross-disciplinary work. Indeed, many micro–macro papers have been authored by scholars who have a single disciplinary focus. In the top management team research, scholars from both micro (e.g., O’Reilly et al., 1993; Barrick, Bradley, KristofBrown, & Colbert, 2007) and macro domains (e.g., Carpenter & Fredrickson, 2001; Sanders & Carpenter, 1998) have made contributions to this area by bringing forward aspects from their own speciﬁc domains. Similarly, in the area of international business, there is a trend toward extending beyond one’s domain and exploring theories from other management disciplines (e.g., Huy, 1999; Kostova, 1999). Some of these researchers have published with colleagues in other disciplines that probably has exposed these individuals to cross-management disciplines. In other cases, industry and consulting experience may have forced researchers to face the pervasiveness of crossdisciplinary phenomena in the practitioner world. It is a paradox worth noting that many articles calling for bridging of macro–micro gap and providing guidance on multilevel theory building and analysis are primarily published by organizational behavior scholars (e.g., Chan, 1998; House, Rousseau, & Thomas-Hunt, 1995; Klein, Tosi, & Cannella, 1999; Morgeson & Hofmann, 1999). However, our review indicates that except for the area of strategic human resources management, the majority of articles employing a cross-disciplinary approach is related to macro topics (strategy or organizational theory) and has primarily been authored by macro researchers. These micro–macro articles are either featured in broad-management journals (e.g., Academy of Management Review) or specialized strategy or entrepreneurship journals (e.g., Strategic Management Journal, Journal of Business Venturing). Perhaps one reason for the dearth of cross-disciplinary articles in the organizational behavior domain is the lack of focus toward the context in which micro phenomenon are investigated. The need to emphasize context in the organizational behavior research has been mentioned in the past by other researchers (e.g., Cappelli & Sherer, 1991; Mowday & Sutton, 1993; O’Reilly, 1991; Staw, 1991). Mowday and Sutton (1993) reviewed past research in organizational behavior and found that over the years the contributions were primarily incremental and methodological in nature. Cappelli and Sherer (1991) shared similar views by concluding that over the course of a number of

54

M. RONALD BUCKLEY ET AL.

years, organizational behavior research has been systematically abandoning contextual arguments – the very focus that distinguishes it from the ﬁeld of psychology – and has been focusing more on cognitive processes. Indeed, it appears that strategy researchers may be less likely to be comfortable assuming away people than OB researchers are assuming away organizations. As argued by many scholars (e.g., Mowday & Sutton, 1993; Pfeffer, 1991), borrowing theories from other disciplines (such as sociology, political science, and strategy) would help bring the focus back to the context in organizational behavior research.

BREAK ON THROUGH (TO THE OTHER SIDE): POTENTIAL BRIDGING OPPORTUNITIES On the basis of our reading of the literatures and our survey, we would like to suggest a number of areas that may be amenable to bridging the micro–macro chasm. The ﬁrst suggestion concerns agent–principal relationships. One of the particularities of existing micro–macro work that becomes immediately apparent is that it tends to happen at the intersection of the two worlds – the CEO or the top management team. This is not surprising, especially when considering that, as revealed by our review, the majority of bridging work is conducted by strategy scholars, who usually operate with datasets that consist of a sole observation for each organization. This makes the study of CEOs intuitively appealing, since there is only one CEO for each organization and no multilevel analysis needs to be performed. Considering this preference, perhaps the best place to start stimulating more multidisciplinary research is in areas that involve the CEO and the Top Management Team (TMT). One such area of study is the issues of agent–principal relationships. Agent–Principal Relationships Agency Theory (Jensen & Meckling, 1976) suggests that top managers’ interests may diverge from those of the shareholders who appointed them and that control or governance structures need to be imposed on the former (referred to as the agent) such that the latter (referred to as principals) not incur losses. This view has been criticized by proponents of Stewardship Theory (e.g., Davis, Schoorman, & Donaldson, 1997), who suggest that there may be individual differences in top managers that play an important role in determining whether they pursue their own interests at the expense of

Bridging the Micro–Macro Divide in the Organizational Sciences

55

the shareholders’ interests. Nevertheless, this point of view has received little attention. In fact, while there are some studies that take a stewardship perspective on the agent–principal relationship (e.g., Lee & O’Neill, 2003), we are not aware of any article that has actually tested the propositions advanced by Davis et al. (1997). Davis and colleagues (1997) proposed that the tendency of CEOs to be stewards – that is, to represent the interests of shareholders as well as they can – may depend to some extent on aspects other than governance structures, including individual differences such as higher order needs, sources of motivation, organizational commitment, or cultural values. There are obvious difﬁculties in exploring these relationships. On one hand, previous work on governance has relied to a great extent on sources of secondary data available in large databases. A study involving individual differences would necessitate an actual survey with the consequential risk of low response rates or range restriction within the dataset. On the other hand, because strong situations tend to wash out the effects of individual-level variables, the data would need to contain sufﬁcient variance in both the strength of governance mechanisms and the individual characteristics that inﬂuence stewardship behavior, if any interactions are to be observed between these two predictors. Nevertheless, such studies would greatly improve our understanding of agent–principal relationships. For example, Lee and O’Neill (2003) found that stewardship behaviors were more likely in Japanese ﬁrms than in US ﬁrms, but could not account for these differences. They suggested differences in factors such as culture, risk of employment, and markets for corporate control as possible explanations. However, there may also be individual differences in the type of CEOs that Japanese ﬁrms attract and groom. A study comparing the values and needs of CEOs in various cultures may reveal differences in the prototypical top manager that might account for a preference for agentic versus stewardship behavior. In other words, is the fact that agency theory better depicts US ﬁrms, while stewardship theory better depicts Japanese ﬁrms simply a consequence of the different type of people that are promoted to the CEO position in the two different cultures? Looking at this problem from both micro and macro perspectives will likely provide us with some insight that is currently lacking.

Citizenship The second area that holds promise as a fecund area for researchers to explore from both the micro and macro perspective is the study of

56

M. RONALD BUCKLEY ET AL.

organizational citizens and corporate citizens. This presents an interesting opportunity for bridging the micro–macro crevasse. Interestingly, constructs that have similar meaning have evolved separately in the micro and macro literature. Much research has been done at the individual level looking at behaviors that go above and beyond the prescribed role requirements of employees (e.g., Podsakoff, MacKenzie, Paine, & Bachrach, 2000). These behaviors are referred to as organizational citizenship behaviors (Organ, 1988) and have been generally associated either with the altruistic nature of employees or, more often, with the reciprocating of good treatment of employees by their employing organizations. Alternatively, Van Dyne, Graham, and Dienesch (1994) proposed a conceptualization rooted in the political philosophy notion of civic citizenship. When looking at the macro conceptualization of citizenship, we ﬁnd both similarities and differences with the micro conceptualizations of citizenship. While both the micro and the macro constructs imply that the respective entity (i.e., employee in the micro literature and organization in the macro literature) engages in some form of discretionary behavior not prescribed in its role. The employee or the ‘‘good soldier’’ engages in extrarole behaviors so as to beneﬁt the organization and its members (Williams & Anderson, 1991), while the corporate citizen does so to protect the citizenship right of individuals in the society (Matten & Crane, 2005). Thus, the conceptualization of the employee citizen in micro literature more closely resembles President John F. Kennedy’s view on citizens who should be concerned with what they can do for their country and less with the rights that they themselves have, while the conceptualization of the corporate citizen in macro literature ascribes to corporations a role similar to national governments in providing, enabling, or channeling the rights of individual citizens within or outside the organization (Matten & Crane, 2005). While there are obvious differences in these two conceptualizations, there are also striking similarities in research ﬁndings that suggest phenomena at the micro level may have parallels at the macro level, and vice versa. For example, some micro researchers have argued that employees may engage in citizenship behaviors for instrumental or for impression management purposes (e.g., Bolino, 1999). At the macro level, researchers have found that some organizations engage in CSR practices – an aspect of corporate citizenship – for window dressing purposes, while others integrate these practices with their core strategy (Weaver, Trevino, & Cochran, 1999). It is perhaps worthwhile to explore whether the consequences of having such different motivations will also exhibit parallel ﬁndings at the micro and

Bridging the Micro–Macro Divide in the Organizational Sciences

57

macro level. For example, at the micro level, supervisors have been observed to reward citizenship behaviors less if they attribute a selﬁsh motivation to the behavior (Grant, Parker, & Collins, 2009). It is possible that stakeholders also respond less positively to CSR activities that are perceived to be superﬁcial. Parallels are also apparent when it comes to the pressures to behave as better citizens. Micro researchers have noted that in some organizations behaviors that are generally understood as discretionary citizenship behaviors are actually perceived to be in-role behaviors by employees (Morrison, 1994). Similarly, at the macro level some organizations have proactively begun to manage CSR expectations by use of a triple bottom line that includes environmental sustainability and social performance as performance indicators alongside the more traditional indicator of economic proﬁtability (Hart & Milstein, 2003). Moreover, pressures to be a better citizen exists for both employees within an organization and for the organization itself. Van Dyne and Ellis (2004) identiﬁed the phenomenon of job creep, where extra-role behaviors that once were discretionary are subsequently taken for granted and eventually expected by employers. Similarly, at the macro level, institutional pressures push organizations to engage in CSR practices (Aguilera et al., 2007; Campbell, 2007). Van Dyne and Ellis (2004) suggested that employees who feel such pressures may respond by pressuring other employees to pick up their fair share of responsibilities. Similarly, we may expect corporations to lobby governments to share in the load of providing and enabling citizenship rights. For example, corporations may request government aid in order to protect employee’s property rights over their pension fund, which may otherwise be infringed should the company go bankrupt. A central question in the citizenship research in both the micro and the macro literature is whether citizenship behavior interferes in a positive or negative way with the actual performance of the citizen, be it job performance for the individual or economic performance of the organization. Research ﬁndings at the micro level are equivocal; in some studies OCB is positively related to objectively measured performance (e.g., Podsakoff, Ahearne, & MacKenzie, 1997), while in others the relationship is negative (e.g., Podsakoff & MacKenzie, 1994). Moreover, some researchers argue that citizenship behaviors may come at the expense of in-role behavior, with the potential to lower in-role performance and thus the career success of the ‘‘good soldier’’ (Bergeron, 2007). This line of thought has its correspondence in the macro literature by scholars who argue for the focus of ﬁrms on economic outcomes and on shareholders’

58

M. RONALD BUCKLEY ET AL.

interests above all else. There is also research suggesting that CSR may actually increase the market value of the ﬁrm (McWilliams & Siegel, 2001). In a recent article, Mackey, Mackey, and Barney (2007) proposed that corporate citizenship behaviors can increase or decrease the market value of ﬁrms depending on the demand and supply for socially responsible corporate behaviors. This interesting take on citizenship may inform studies of citizenship behaviors at the micro level. Some evidence for this already exists; Bommer, Dierdorff, and Rubin (2007) found that when citizenship behaviors were common (i.e., in high supply) they were no longer rewarded by supervisors. It is probable that the supply and demand of citizenship behaviors may inﬂuence whether such behaviors have a positive or a negative effect on both employee outcomes and organizational outcomes. For example, helping other employees may be in more demand when new hires are brought on board, but may hinder performance when there is no demand for helping and when time could be better spent on in-role task performance. Similarly, working extra hours when needed may provide a company with an edge in tough times, but when employees begin to see extra work primarily as a means to differentiate one another in the eyes of the boss, potentially dysfunctional behaviors (e.g., conﬂict among employees) may soon follow (Bolino, Valcea, & Harvey, 2010; Van Dyne & Ellis, 2004). An interpretation of our survey results indicates that there exists great potential for cross-disciplinary research. What we have suggested here are only a few areas in which bridging studies might be probative. It is our intention that this may encourage micro and macro researchers to develop cross-disciplinary research teams and explore other avenues of research that might be efﬁcacious in bridging the aforementioned chasm.

DON’T STOP BELIEVIN’: THE FUTURE OF MICRO–MACRO RESEARCH We do not pretend to be the ﬁrst to have grappled with this issue. In fact, Molloy et al. (2011) have detailed the speciﬁc intellectual steps that they believe are required to successfully craft theories that might be effective in bridging disciplinary and system-level divides. We believe our work is a supplement to these efforts in that we have attempted to take the pulse of researchers in management disciplines concerning these issues and to integrate this information with the bridging research that has already been reported.

Bridging the Micro–Macro Divide in the Organizational Sciences

59

Overall, initiatives taken through academic platforms such as Academy of Management (AOM) meetings, design of Ph.D. programs, and incentives would encourage scholars from various disciplines to interact more often. It is only through mutual exploration of common questions and common problems that we can foresee bridging of theoretical gaps. As Molloy et al. (2011) make explicit through their analysis, ‘‘organization’’ is the common construct binding together all management disciplines. The reality is complex; for some researchers, organization is a micro entity (e.g., strategy or entrepreneurship), while for others it is a macro entity (e.g., organizational behavior or human resources management), and for still others it can either be micro or macro (e.g., organizational theory, international business). Actually, an organization is BOTH a micro and macro entity. Looking from only one side under identiﬁes myriad possible explanations and possibilities. Continuing to ignore what lays on the other side of the chasm fails to recognize the possibilities held by those constructs for enriching knowledge on our own side of the chasm and disregards the potential of corresponding areas to add to the explanatory power of our organizational endeavors. Recognizing the gulf between the diverse management disciplines is not enough – it has preceded us and it still exists. There are a number of broader initiatives, some quite simple, many less so, that our ﬁeld would need to take. Fundamentally, we need to reappraise the ways in which we socialize and train researchers to interact with other researchers in our academic community. We need to ensure that our doctoral trainees are exposed to the dominant theoretical approaches in both the micro and the macro areas. To those colleagues who toil in doctoral programs not requiring your students to digest a doctoral seminar in both the micro and the macro areas we have a simple message: You are not preparing your doctoral students as well as you intend nor are you capitalizing on their inquisitiveness and potential. We understand that students only have a ﬁnite amount of time available for doctoral seminars. That point made, we ask our fellow Ph.D. level educators this question: Is it of more educational value to have a student take a third or fourth micro (macro) seminar than to have them set aside at least one of those seminars for an education-broadening seminar in macro (micro)? Ensuring the availability of these types of seminars will help our future prote´ge´es develop a broader based repertoire of theoretical approaches and enable them to possess the tool kit for developing research that bridges the micro–macro gap. However, the solution to this issue lies deeper than merely providing our students with a more diverse course background but this would be a step in the right direction.

60

M. RONALD BUCKLEY ET AL.

There are a number of both intellectual and economic incentives that are facilitating the development of narrow specialized skills among members of our ﬁeld. We seem to have forgotten Kerr’s (1995) admonitions in his ‘‘Folly of rewarding A while hoping for B.’’ We have naively rewarded a generation of researchers with both intellectual and economic rewards for developing narrow specialized skills while bemoaning the lack of new grand theory. We are too optimistic in hoping that a continuation of similar inputs will result in different outputs. A recent article by Certo, Sirmon, and Brymer (2010) reports that since 1988–2008, the average time to publish ﬁve top-tier articles has increased from 5 years to 10 years. These are the realities associated with the publish or perish paradigm that motivate our research endeavors. If it has become signiﬁcantly more difﬁcult to publish in one’s primary area of focus, would scholars be willing to explore crossdisciplinary work that requires increased time and resources investment and is accompanied by signiﬁcantly higher publication risks? In the face of such realities, can we develop some institutional incentives that can be associated with cross-disciplinary work? For example, would programs be willing to weigh more favorably a scholarly work that is cross-disciplinary in nature and might, more realistically, model organizational truth? As many of us teach our students about organizational change, hoping for a result other than the status quo in the absence of a mechanism for pushing employees out of their comfort zones is simply futile. Importantly, though, we in no way intend to discourage developing specialized knowledge, as one can only be master of one trade and this is how we generate knowledge in the organizational sciences. We do, however, suggest that we broaden what we mean by specialized knowledge to encompass contributions that span various levels of organizational analysis. Furthermore, there is a need to develop mechanisms that encourage masters of different trades to interact and share knowledge. We must dispel the notion that you are a radical in your ﬁeld if you utilize constructs and concepts from a different management area to both develop and increase explanatory power. On this point we might make a few suggestions for the academic journals as well: (1) there has been a trend in journals toward the development of special issues. We recommend that journals might encourage more special issues that are centered on ‘‘bridging’’ themes; and (2) journal mission statements might need to be further developed in a direction to where they might be encouraging of ‘‘bridging’’ work or, at minimum, accepting of such work. There are a number of more practical actions that young scholars in the organizational sciences can take to engage them in cross-disciplinary

Bridging the Micro–Macro Divide in the Organizational Sciences

61

research in their careers. First, take a portfolio approach to your research, instead of the typical pipeline approach. That is, diversify your research projects in low-, medium-, and high-risk clusters as you would with your ﬁnancial portfolio. This way, your focused research stream can be developed alongside higher risk cross-disciplinary projects. Second, take every opportunity to interact with practitioners in the ﬁeld. Take advantage of your college’s advisory board, the local chamber of commerce or rotary club, or even your spouse’s ofﬁce party! The ‘‘real world’’ is suffused with micro–macro spanning challenges, and your research will be enriched by immersing yourself in it. Third, spend some time diving deeply into knowledge in areas completely unrelated to your discipline. Weick’s (1993) colorful and captivating theory of what happens when sensemaking fails would not have been as impactful had it not been for his deep understanding of the Mann Gulch disaster, which on the surface has nothing to do with organizations. Furthermore, management giant Peter Drucker was dedicated to immersing himself in a completely new subject unrelated to management every year or so throughout his entire life (Drucker & Wartzman, 2010). Finally, look for the opportunities for synergistic research collaborations at your front door. We have seen many new faculty members become so focused on their tenure clock that they don’t take the time to socialize with the other scholars in their own department. This is inexcusable. We implore junior faculty to explicitly schedule time into every week or month to build relationships with the other members of their department, college, and university. At worst, you’ll end up strengthening your community of scholars, and at best, you’ll help build the bridge across in the great wide open of the organizational sciences. Much work has been done over the past 30 years to provide researchers with the methodological and theoretical tools to bridge the micro–macro divide in the organizational sciences. As we have highlighted, these efforts have paid off, as a number of constructs now stretch across the chasm between the two poles. However, as these problems have been surmounted, other barriers such as risks in the publishing process, the current academic reward system, and ever-narrowing knowledge silos have arisen. These challenges are certainly not insoluble, and it is our hope that by calling attention to them and providing some suggestions for overcoming them, the future of the organizational sciences will look a bit brighter. In other words, after peering into the great wide open, we see many opportunities for future researchers to use the micro and macro perspectives together in concert to make large advances on the frontiers of knowledge in the organizational sciences.

62

M. RONALD BUCKLEY ET AL.

REFERENCES Acquaah, M. (2007). Managerial social capital, strategic orientation, and organizational performance in an emerging economy. Strategic Management Journal, 28, 1235–1255. Aguilera, R. V., Rupp, D. E., Williams, C. A., & Ganapathi, J. (2007). Putting the s back in corporate social responsibility: A multilevel theory of social change in organizations. Academy of Management Review, 32, 836–863. Barney, J. B. (1986). Organizational culture: Can it be a source of sustained competitive advantage? Academy of Management Review, 11, 656–665. Barney, J. B. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17, 99–120. Barrick, M. R., Bradley, B. H., Kristof-Brown, A. L., & Colbert, A. E. (2007). The moderating role of top management team interdependence: Implications for real teams and working groups. Academy of Management Journal, 50(3), 544–557. Barry, B., & Stewart, G. L. (1997). Composition, process, and performance in self-managed groups: The role of personality. Journal of Applied Psychology, 82, 62–78. Becker, B., & Gerhart, B. (1996). The impact of human resource management on organizational performance: Progress and prospects. Academy of Management Journal, 39, 779–801. Bergeron, D. M. (2007). The potential paradox of organizational citizenship behavior: Good citizens at what cost? Academy of Management Review, 32, 1078–1095. Birdi, K., Clegg, C., Patterson, M., Robinson, A., Stride, C. B., & Wall, T. D. (2008). The impact of human resource and operational management practices on company productivity: A longitudinal study. Personnel Psychology, 61, 467–501. Blyler, M., & Coff, R. W. (2003). Dynamic capabilities, social capital, and rent appropriation: Ties that split pies. Strategic Management Journal, 24, 677–686. Bolino, M. C. (1999). Citizenship and impression management: Good soldiers or good actors? Academy of Management Review, 24, 82–98. Bolino, M. C., Turnley, W. H., & Bloodgood, J. M. (2002). Citizenship behavior and the creation of social capital in organizations. Academy of Management Review, 27, 505–522. Bolino, M., Valcea, S., & Harvey, J. (2010). Employee, manage thyself: The potentially negative implications of expecting employees to behave proactively. Journal of Occupational and Organizational Psychology, 83(2), 325–345. Bommer, W. H., Dierdorff, E. C., & Rubin, R. S. (2007). Does prevalence mitigate relevance? The moderating effect of group-level OCB on employee performance. Academy of Management Journal, 50, 1481–1494. Brass, D. (1984). Being in the right place: A structural analysis of individual inﬂuence in an organization. Administrative Science Quarterly, 29, 518–539. Brass, D. J. (1985). Mens and womens networks: A study of interaction patterns and inﬂuence in an organization. Academy of Management Journal, 28, 327–343. Brass, D. J., Galaskiewicz, J., Greve, H. R., & Tsai, W. (2004). Taking stock of networks and organizations: A multilevel perspective. Academy of Management Journal, 47, 795–817. Burt, R. S. (1997). The contingent value of social capital. Administrative Science Quarterly, 42, 339–365. Busenitz, L., West, G., Shepherd, D., Nelson, T., Chandler, G., & Zacharakis, A. (2003). Entrepreneurship research in emergence: Past trends and future directions. Journal of Management, 29, 285–308.

Bridging the Micro–Macro Divide in the Organizational Sciences

63

Campbell, J. L. (2007). Why would corporations behave in socially responsible ways? An institutional theory of corporate social responsibility. Academy of Management Review, 32, 946–967. Cappelli, P., & Sherer, P. D. (1991). The missing role of context in OB: The need for a mesolevel approach. In: L. L. Cummings & B. M. Staw (Eds), Research in organizational behavior (Vol. 13, pp. 55–110). Greenwich, CT: JAI Press. Carmeli, A., & Azeroual, B. (2009). How relational capital and knowledge combination capability enhance the performance of work units in a high technology industry. Strategic Entrepreneurship Journal, 3, 85–103. Carpenter, M., & Fredrickson, J. (2001). Top management teams, global strategic posture, and the moderating role of uncertainty. Academy of Management Journal, 44, 533–545. Carpenter, M., & Wade, J. (2002). Microlevel opportunity structures as determinants of nonCEO executive pay. Academy of Management Journal, 45, 1085–1103. Certo, S. T., Sirmon, D. G., & Brymer, R. (2010). Competition and scholarly productivity in management: Investigating changes in scholarship from 1988 to 2008. Academy of Management Learning and Education, 9(4), 591–606. Chan, D. (1998). Functional relations among constructs in the same content domain at different levels of analysis: A typology of composition models. Journal of Applied Psychology, 83, 234–246. Chuang, C. H., & Liao, H. (2010). Strategic human resource management in service context: Taking care of business by taking care of employees and customers. Personnel Psychology, 63, 153–196. Collins, C. J., & Clark, K. D. (2003). Strategic human resource practices, top management team social networks, and ﬁrm performance: The role of human resource practices in creating organizational competitive advantage. Academy of Management Journal, 46, 740–751. Davis, J., Schoorman, F., & Donaldson, L. (1997). Toward a stewardship theory of management. Academy of Management Review, 22, 20–47. de Luque, M., Washburn, N., Waldman, D., & House, R. (2008). Unrequited proﬁt: how stakeholder and economic values relate to subordinates’ perceptions of leadership and ﬁrm performance. Administrative Science Quarterly, 53, 626–654. Dess, G. G., & Shaw, J. D. (2001). Voluntary turnover, social capital, and organizational performance. Academy of Management Review, 26, 446–456. Drucker, P. F., & Wartzman, R. (2010). The Drucker lectures: Essential lessons on management, society, and economy. New York: McGraw-Hill. Eckhardt, J., & Ciuchta, M. (2008). Selected variation: the population-level implications of multistage selection in entrepreneurship. Strategic Entrepreneurship Journal, 2, 209–224. Fedor, D., Caldwell, S., & Herold, D. (2006). The effects of organizational changes on employee commitment: A multilevel investigation. Personnel Psychology, 59, 1–29. Fredrickson, J. W., Davis-Blake, A., & Sanders, W. M. G. (2010). Sharing the wealth: Social comparisons and pay dispersion in the CEO’s top team. Strategic Management Journal, 31, 1031–1053. Geletkanycz, M. A., Boyd, B. K., & Finkelstein, S. (2001). The strategic value of CEO external directorate networks: Implications for CEO compensation. Strategic Management Journal, 22, 889–898. Gnyawali, D., & Madhavan, R. (2001). Cooperative networks and competitive dynamics: A structural embeddedness perspective. Academy of Management Review, 26, 431–445.

64

M. RONALD BUCKLEY ET AL.

Gong, Y., Law, K. S., & Chang, S. (2009). Human resources management and ﬁrm performance: The differential role of managerial affective and continuance commitment. Journal of Applied Psychology, 94, 263–275. Grant, A. M., Parker, S., & Collins, C. (2009). Getting credit for proactive behavior: Supervisor reactions depend on what you value and how you feel. Personnel Psychology, 62, 31–55. Grizzle, J., Zablah, A., Brown, T., Mowen, J., & Lee, J. (2009). Employee customer orientation in context: How the environment moderates the inﬂuence of customer orientation on performance outcomes. Journal of Applied Psychology, 94, 1227–1242. Guler, I. (2007). Throwing good money after bad? Political and institutional inﬂuences on sequential decision making in the venture capital industry. Administrative Science Quarterly, 52, 248–285. Hackman, J. (2003). Learning more by crossing levels: Evidence from airplanes, hospitals, and orchestras. Journal of Organizational Behavior, 24(8), 905–922. Hambrick, D. C. (1994). Top management groups: A conceptual integration and reconsideration of the ‘‘team’’ label. In: L. L. Cummings & B. M. Staw (Eds), Research in organizational behavior (Vol. 16, pp. 171–214). Greenwich, CT: JAI Press. Hambrick, D. C., & Mason, P. A. (1984). Upper echelons: The organization as a reﬂection of its top managers. Academy of Management Review, 9, 193–206. Hart, S., & Milstein, M. (2003). Creating sustainable value. Academy of Management Executive, 17, 56–69. Hatch, M. J. (1997). Organizational theory: Modern, symbolic, and postmodern perspectives. New York: Oxford University Press. Haynes, K. T., & Hillman, A. (2010). The effect of board capital and CEO power on strategic change. Strategic Management Journal, 31, 1145–1163. Hitt, M., Beamish, P., Jackson, S., & Mathieu, J. (2007). Building theoretical and empirical bridges across levels: Multilevel research in management. Academy of Management Journal, 50, 1385–1399. Holm, P. (1995). The dynamics of institutionalization: Transformation processes in Norwegian ﬁsheries. Administrative Science Quarterly, 40, 398–422. Hough, J. R. (2006). Business segment performance redux: A multilevel approach. Strategic Management Journal, 27, 45–61. House, R., Rousseau, D. M., & Thomas-Hunt, M. (1995). The meso paradigm: A framework for the integration of micro and macro organizational behavior. In: L. L. Cummings & B. M. Staw (Eds), Research in organizational behavior (Vol. 17, pp. 71–114). Greenwich, CT: JAI Press. Huselid, M. A., Jackson, S. E., & Schuler, R. S. (1997). Technical and strategic human resource management effectiveness as determinants of ﬁrm performance. Academy of Management Journal, 40, 171–188. Huy, Q. N. (1999). Emotional capability, emotional intelligence, and radical change. Academy of Management Review, 24, 325–345. Inkpen, A. C., & Tsang, E. W. K. (2005). Social capital, networks, and knowledge transfer. Academy of Management Review, 30, 146–165. Jensen, M., & Meckling, W. (1976). Theory of the ﬁrm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3, 305–360. Kerr, S. (1995). On the folly of rewarding A, while hoping for B. Academy of Management Executive, 9, 7–16.

Bridging the Micro–Macro Divide in the Organizational Sciences

65

Klein, K. J., Tosi, H., & Cannella, A. A. (1999). Multilevel theory building: Beneﬁts, barriers, and new developments. Academy of Management Review, 24, 243–248. Kostova, T. (1999). Transnational transfer of strategic organizational practices: A contextual perspective. Academy of Management Review, 24, 308–324. Kostova, T., & Roth, K. (2003). Social capital in multinational corporations and a micro-macro model of its formation. Academy of Management Review, 28, 297–317. Landy, F. J., & Shankster, L. J. (1994). Personnel selection and placement. Annual Review of Psychology, 45, 261–296. Leana, C. R., & Van Buren, H. J. (1999). Organizational social capital and employment practices. Academy of Management Review, 24, 538–555. Lee, P. M., & O’Neill, H. M. (2003). Ownership structures and R&D investments of US and Japanese ﬁrms: Agency and stewardship perspectives. Academy of Management Journal, 46, 212–225. Mackey, A., Mackey, T. B., & Barney, J. B. (2007). Corporate social responsibility and ﬁrm performance: Investor preferences and corporate strategies. Academy of Management Review, 32, 817–835. Mahoney, T. A. (1985). Journal publishing and the organizations sciences: An analysis of exchanges. In: L. L. Cummings & P. J. Frost (Eds), Publishing in the organizational sciences (pp. 14–34). Homewood, IL: Richard D. Irwin, Inc. Mani, S., Antia, K. D., & Rindﬂeisch, A. (2007). Entry mode and equity level: A multilevel examination of foreign direct investment ownership structure. Strategic Management Journal, 28, 857–866. Matten, D., & Crane, A. (2005). Corporate citizenship: Toward an extended theoretical conceptualization. Academy of Management Review, 30, 166–179. Mayer, D., Nishii, L., Schneider, B., & Goldstein, H. (2007). The precursors and products of justice climates: Group leader antecedents and employee attitudinal consequences. Personnel Psychology, 60, 929–963. McDonald, M., & Westphal, J. (2003). Getting by with the advice of their friends: CEOs’ advice networks and ﬁrms’ strategic responses to poor performance. Administrative Science Quarterly, 48, 1–35. McKay, P., Avery, D., & Morris, M. (2009). A tale of two climates: Diversity climate from subordinates’and managers’perspectives and their role in store unit sales performance. Personnel Psychology, 62, 767–791. McWilliams, A., & Siegel, D. (2001). Corporate social responsibility: A theory of the ﬁrm perspective. Academy of Management Review, 26, 117–127. Mehra, A., Kilduff, M., & Brass, D. J. (2001). The social networks of high and low self-monitors: Implications for workplace performance. Administrative Science Quarterly, 46, 121–146. Milliman, J., Von Glinow, M. A., & Nathan, M. (1991). Organizational life cycles and strategic international human resource management in multinational companies: Implications for congruence theory. Academy of Management Review, 16, 318–339. Misangyi, V. F., Elms, H., Greckhamer, T., & Lepine, J. A. (2006). A new perspective on a fundamental debate: A multilevel approach to industry, corporate, and business unit effects. Strategic Management Journal, 27, 571–590. Molina-Morales, F. X., & Martı´ nez-Ferna´ndez, M. T. (2009). Too much love in the neighborhood can hurt: How an excess of intensity and trust in relationships may produce negative effects on ﬁrms. Strategic Management Journal, 30, 1013–1023.

66

M. RONALD BUCKLEY ET AL.

Molloy, J. C., Ployhart, R. E., & Wright, P. M. (2011). The myth of ‘‘the’’ micro–macro divide: Bridging system-level and disciplinary divides. Journal of Management, 37(2), 581–609. Moran, P. (2005). Structural vs. relational embeddedness: Social capital and managerial performance. Strategic Management Journal, 26, 1129–1151. Morgeson, F. P., & Hofmann, D. A. (1999). The structure and function of collective constructs: Implications for multilevel research and theory development. Academy of Management Review, 24, 249–265. Morrison, E. W. (1994). Role deﬁnitions and organizational citizenship behavior: The importance of the employee’s perspective. Academy of Management Journal, 37, 1543–1567. Mossholder, K. W., & Bedeian, A. G. (1983). Cross-level inference and organizational research: Perspectives on interpretation and application. Academy of Management Review, 8(4), 547–558. Mowday, R., & Sutton, R. (1993). Organizational behavior: Linking individuals and groups to organizational contexts. Annual Review of Psychology, 44, 195–229. Nahapiet, J., & Ghoshal, S. (1998). Social capital, intellectual capital, and the organizational advantage. Academy of Management Review, 23, 242–266. O’Reilly, C. (1991). Organizational behavior: Where we’ve been, where we’re going. Annual Review of Psychology, 42, 427–458. O’Reilly, C., Snyder, R., & Boothe, J. (1993). Effects of executive team demography on organizational change. In: G. P. Huber & W. H. Glick (Eds), Organizational change and redesign: Ideas and insights for improving performance (pp. 147–175). New York: Oxford University Press. Oh, H., Chung, M., & Labiance, G. (2004). Group social capital and group effectiveness: The role of informal socializing ties. Academy of Management Journal, 47, 860–875. Organ, D. W. (1988). Organizational citizenship behavior: The good soldier syndrome. Lexington, MA: Lexington Books. Pfeffer, J. (1991). Organization theory and structural perspectives on management. Journal of Management, 17, 789–803. Pfeffer, J. (1993). Barriers to the advance of organizational science: Paradigm development as a dependent variable. Academy of Management Review, 18, 599–620. Podsakoff, P. M., Ahearne, M., & MacKenzie, S. B. (1997). Organizational citizenship behavior and the quantity and quality of work group performance. Journal of Applied Psychology, 82, 262–270. Podsakoff, E. M., & MacKenzie, S. B. (1994). Organizational citizenship behaviors and sales unit effectiveness. Journal of Marketing Research, 31, 351–363. Podsakoff, P. M., MacKenzie, S. B., Paine, J. B., & Bachrach, D. G. (2000). Organizational citizenship behaviors: A critical review of the theoretical and empirical literature and suggestions for future research. Journal of Management, 26, 513–563. Roberts, K., Hulin, C., & Rousseau, D. M. (1978). Developing an interdisciplinary science of organizations. San Francisco: Jossey-Bass. Rousseau, D. M. (1985). Issues of level in organizational research: Multi-level and cross-level perspectives. In: L. L. Cummings & B. M. Staw (Eds), Research in organizational behavior (Vol. 7, pp. 1–37). Greenwich, CT: JAI Press. Rowe, W., & Wright, P. (1997). Related and unrelated diversiﬁcation and their effect on human resource management controls. Strategic Management Journal, 18, 329–338. Rumelt, R., Schendel, D., & Teece, D. (1991). Strategic management and economics. Strategic Management Journal, 12, 5–29.

Bridging the Micro–Macro Divide in the Organizational Sciences

67

Rynes, S. L. (2006). ‘‘Getting on board’’ with AMJ: Balancing quality and innovation in the review process. Academy of Management Journal, 49, 1097–1102. Sanders, W., & Carpenter, M. (1998). Internationalization and ﬁrm governance: The roles of CEO compensation, top team composition, and board structure. Academy of Management Journal, 41, 158–178. Seibert, S. E., Kraimer, M. L., & Liden, R. C. (2001). A social capital theory of career success. Academy of Management Journal, 44, 219–237. Short, J., Palmer, T., & Ketchen, D. (2003). Multi-level inﬂuences on ﬁrm performance: Insights from the resource-based view and strategic groups research. In: F. Dansereau & F. J. Yammarino (Eds), Research in multi-level issues (Vol. 2, pp. 155–187). Greenwich, CT: JAI Press. Simons, T., & Roberts, P. (2008). Local and non-local pre-founding experience and new organizational form penetration: The case of the Israeli wine industry. Administrative Science Quarterly, 53, 235–265. Simsek, Z., Veiga, J. F., Lubatkin, M. H., & Dino, R. N. (2005). Modeling the multilevel determinants of top management team behavioral integration. Academy of Management Journal, 48, 69–84. Smith, K. G., Smith, K. A., Sims, H. P., O’Bannon, D. P., Scully, J. A., & Olian, J. D. (1994). Top management team demography and process: The role of social integration and communication. Administrative Science Quarterly, 39, 412–438. Snell, S. A. (1992). Control theory in strategic human resource management: The mediating effect of administrative information. Academy of Management Journal, 35, 292–327. Soo Min, T., Morgeson, F. P., & Campion, M. A. (2008). Human resource conﬁgurations: Investigating ﬁt with the organizational context. Journal of Applied Psychology, 93, 864–882. Staw, B. (1991). Dressing up like an organization: When psychological theories can explain organizational action. Journal of Management, 17, 805–819. Taras, V., Steel, P., & Kirkman, B. (2010). Negative practice–value correlations in the GLOBE data: Unexpected ﬁndings, questionnaire limitations and research directions. Journal of International Business Studies, 41, 1330–1338. Tasa, K., Seijts, G. H., & Taggar, S. (2007). The development of collective efﬁcacy in teams: A multilevel and longitudinal perspective. Journal of Applied Psychology, 92, 17–27. Taylor, S., Beechler, S., & Napier, N. (1996). Toward an integrative model of strategic international human resource management. Academy of Management Review, 21, 959–985. Tsai, W., & Ghoshal, S. (1998). Social capital and value creation: The role of intraﬁrm networks. Academy of Management Journal, 41, 464–476. Van Dyne, L., & Ellis, J. B. (2004). Job creep: A reactance theory perspective on organizational citizenship behavior as overfulﬁllment of obligations. In: J. A. M. Coyle-Shapiro, L. M. Shore, M. S. Taylor & L. E. Tetrick (Eds), The employment relationship: Examining psychological and contextual perspectives (pp. 181–205). Oxford: Oxford University Press. Van Dyne, L., Graham, J. W., & Dienesch, R. M. (1994). Organizational citizenship behavior: Construct redeﬁnition, measurement, and validation. Academy of Management Journal, 37, 765–802. Weaver, G. R., Trevino, L. K., & Cochran, P. L. (1999). Integrated and decoupled corporate social performance: Management commitments, external pressures, and corporate ethics practices. Academy of Management Review, 42, 539–552.

68

M. RONALD BUCKLEY ET AL.

Weick, K. E. (1993). The collapse of sensemaking in organizations: The Mann Gulch disaster. Administrative Science Quarterly, 38, 628–652. Wenspin, T. (2000). Social capital, strategic relatedness and the formation of intraorganizational linkages. Strategic Management Journal, 21, 925–939. Westphal, J. (1998). Board games: How CEOs adapt to increases in structural board independence from management. Administrative Science Quarterly, 43, 511–513. Wiersema, M., & Bantel, K. (1992). Top management team demography and corporate strategic change. Academy of Management Journal, 35, 91–121. Williams, L. J., & Anderson, S. E. (1991). Job satisfaction and organizational commitment as predictors of organizational citizenship and in-role behaviors. Journal of Management, 17, 601–617. Wright, P., Dunford, B., & Snell, S. (2001). Human resources and the resource based view of the ﬁrm. Journal of Management, 27, 701–721. Zajac, E. J., & Westphal, J. D. (1995). Accounting for the explanations of CEO compensation: Substance and symbolism. Administrative Science Quarterly, 40, 283–308. Zammuto, R. F., & Connolly, T. (1984). Coping with disciplinary fragmentation. Organizational Behavior Teaching Review, 9, 30–32. Zeitz, G. (1980). Interorganizational dialectics. Administrative Science Quarterly, 20, 72–80. Zhao, Z., & Anand, J. (2009). A multilevel perspective on knowledge transfer: Evidence from the Chinese automotive industry. Strategic Management Journal, 30, 959–983. Zohar, D. (2000). A group-level model of safety climate: Testing the effect of group climate on microaccidents in manufacturing jobs. Journal of Applied Psychology, 85, 587–596. Zyphur, M. J., Narayanan, J., Koh, G., & Koh, D. (2009). Testosterone–status mismatch lowers collective efﬁcacy in groups: Evidence from a slope-as-predictor multilevel structural equation model. Organizational Behavior & Human Decision Processes, 110, 70–79.

STRATEGIC HUMAN RESOURCE STAFFING AND ORGANIZATION RESEARCH: ARE THEY ONE-SIZE-FITS-ALL ENDEAVORS? Philip Bobko and Denise Potosky ABSTRACT Purpose – We show that, although most private employer establishments are small, much reported research (and subsequent suggestions for practice) in management comes from large ﬁrms. In turn, we wanted to explore if organizational knowledge gained from studying one or more large ﬁrms is necessarily applicable to numerous smaller ﬁrms. Design/methodology/approach – We computed ﬁrm size in the United States using existing databases, and we then considered published literature in human resources and strategy to see if the large sample results logically applied to smaller ﬁrms. Findings – At the job-analytic level, it is suggested that jobs might be deﬁned differently and more broadly in smaller establishments than in large organizations. Also, the feasibility of best corporate strategies may be moderated by the size of the ﬁrm. In addition, we noted that the underlying model of selection utility in human resource management

Building Methodological Bridges Research Methodology in Strategy and Management, Volume 6, 69–87 Copyright r 2011 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-8387/doi:10.1108/S1479-8387(2011)0000006011

69

70

PHILIP BOBKO AND DENISE POTOSKY

(HRM), and several factors in its numerical estimation, might need to be modiﬁed as a function of ﬁrm size. Originality/value – We hope that this chapter inspires HRM and strategy researchers by helping to focus future evidence-based efforts, creating new initiatives, and providing results that are useful (or scalable) to the large number of small, private-sector U.S. ﬁrms. Keywords: Best practices; ﬁrm size; job analysis; selection utility; strategy

INTRODUCTION Most of the data generated, reported on, and analyzed in the organizational sciences are based on samples from (or across) large ﬁrms and corporations. This focus on large ﬁrms is somewhat understandable, as relatively large sample sizes enable within-study quantitative analyses. For example, large sample sizes provide statistical power to detect signiﬁcant results and allow researchers to engage in sophisticated multivariate techniques (e.g., factor analysis or structural equation modeling). In other words, the pursuit of quantitative information from/about large organizations (to the omission of data about small ﬁrms) might seem to be a reasonable strategy for organizational researchers. Unfortunately, such a research strategy may have resulted in a gap in our knowledge about typical organizations. There is an underlying, counterpoint theme that motivates this chapter – that is, most private employer establishments are small. Indeed, most people work in settings with fewer than 100 other employees. Large ﬁrms are exceptions to this norm, and knowledge gained from studying one or more large ﬁrms may not be automatically applicable to numerous smaller ﬁrms. This potential mismatch seems particularly noteworthy in light of the ﬁeld’s recent embracement of the concept of evidence-based management (EBM), which advocates that practicing managers make decisions that are informed by the ‘‘best available scientiﬁc evidence’’ (Rousseau, 2006, p. 256). We do note that several authors have called attention to the lack of research focused on smaller establishments, particularly research within academic journals (e.g., Barber, Wesson, Roberson, & Taylor, 1999; Duberly & Walley, 1995; Gatewood, Feild, & Barrick, 2008; Hendry, Arthur, & Jones, 1995; Heneman, Tansky, & Camp, 2000b; McElwee & Warren, 2000;

Firm Size and Strategic Human Resources

71

Vickerstaff, 1993; Williamson, 2001). For example, Williamson reviewed the 207 articles focused on recruitment, selection, and other hiring issues published in three top-tier journals between 1988 and 1998 and found only seven that either focused on small businesses or used small ﬁrms in their samples (see also Barber et al., 1999). Similarly, Heneman, Tansky, and Camp reviewed both the academic and practitioner research literature between 1984 and 1999 and found only 17 empirical articles focused on small- and mediumsized enterprises. After illustrating how many U.S. ﬁrms fall into the overlooked/understudied ‘‘small’’ categorization, we brieﬂy discuss the potential lack of ﬁt (between ﬁrm size and research-generated organizational knowledge) with regard to (a) strategic human resource management (HRM) with a focus on selection and (b) strategic management in general. We offer some potential areas where current organizational knowledge might not match the needs of most U.S. ﬁrms, and we identify methodological concerns and considerations to provide some direction for future research to address this gap.

SOME STATISTICS ON FIRM SIZE (THE ELEPHANT IN THE ROOM?) The Data for ‘‘Establishments’’ As deﬁned by the Quarterly Census of Employment and Wages (QCEW), an ‘‘establishment’’ is akin to a ‘‘workplace’’ (or an economic unit focused on a speciﬁc activity), whereas a ‘‘ﬁrm’’ refers to a corporate parent organization that can consist of several establishments. Similarly, the Bureau of Labor Statistics (BLS) deﬁnes an establishment as an economic unit that produces goods or provides services (usually at a single physical location) and engages in predominantly one activity, whereas a ﬁrm is deﬁned as an aggregation of establishments that are under common ownership by a corporate parent.1 Figure 1 graphically depicts the total number of U.S. establishments according to the number of individuals employed in each establishment. Note that of the 8,737,209 establishments listed by the QCEW2 in March 2008, the most recent data year available, more than half (5,347,059 establishments) listed fewer than 5 workers (including establishments with no employees but with payroll at some time during the year), and 2,996,241 establishments employed between 5 and 50 workers (U.S. Bureau of Labor Statistics, 2008). In contrast, 221,242 establishments employed between 50 and 99 workers, and

72

PHILIP BOBKO AND DENISE POTOSKY

Number of Employees in Private-Sector U.S. Establishments 3% 2% 0.2%

<5 5-49

34%

50-99 100-499 61%

500+

Fig. 1. Number of Establishments According to the Number of People Employed. Note. Data from Quarterly Census of Employment and Wages (QCEW), March 2008. Available at http://www.bls.gov/cew/ew08table3.pdf.

156,331 establishments employed between 100 and 499 workers. As of the ﬁrst quarter March 2008, there were 16,336 private U.S. establishments with more than 500 employees. One implication from this data is that researchers who study companies with more than 500 employees are studying companies that represent only 0.2 percent of U.S. private-sector establishments, the smallest sliver of the pie wedges shown in Fig. 1. Another way of looking at these data is in terms of the size of the work setting in which people are employed; that is, organizational size can be described in terms of how many other employees a person works with. For example, using a cut-point of 500 employees, of the 112,661,107 people employed in the U.S. private sector in March 2008, 16.7 percent (18,853,263 people) worked in establishments of 500 or more employees, whereas 83.3 percent of private-sector employees (93,807,744 people) were employed in establishments of less than 500 workers. As of March 2008, 50 percent of all private-sector U.S. employees (56,820, 767 employees) worked in establishments comprised of 5–99 employees (U.S. Bureau of Labor Statistics, 2008). Thirty-three percent of U.S. employees were employed in micro-businesses, which the U.S. Small Business Administration deﬁnes as businesses that employ fewer than ﬁve total employees.

Firm Size and Strategic Human Resources

73

Micro-businesses may include home-based and start-up businesses, and as noted later may be less likely than other small businesses to have demarcated jobs or work roles. Also, because the QCEW data used here excludes the selfemployed, the ‘‘fewer than ﬁve employees’’ category does not adequately represent all micro-businesses in the United States, the majority of which are ‘‘nonemployer’’ organizations.3 In either of these two related ways of looking at the values, the data indicate that the majority of private-sector establishments have a relatively small number of employees. In turn, and as outlined later, there may be theoretical and operational differences in maximally efﬁcient organizational processes used by small and large companies.

The Data for ‘‘Firms’’ Many establishments are not stand-alone business entities, but belong to a parent company or ‘‘ﬁrm.’’ As noted above, a ﬁrm is deﬁned as an aggregation of establishments that are under common ownership by a corporate parent. However, it is worth noting that according to the BLS, most U.S. ﬁrms (indexed using IRS-issued employer identiﬁcation numbers) are comprised of a single establishment. For example, in March 2005 (the latest year with ﬁrm-level data), 95.2 percent of U.S. ﬁrms had a single establishment, covering 44.6 percent of employment in the United States (see http://www.bls.gov/bdm/sizeclassqanda.htm#q21). Given the data discussed above (and Fig. 1), it is likely that ﬁrms reﬂected in the ﬁeld’s published research endeavors represent unusual (atypical) employment situations. At the end of the ﬁrst quarter of 2005 (the most recent ﬁrm-level data available), there were 4.9 million ﬁrms in the private sector of the U.S. economy. Firms with fewer than 500 employees accounted for 99.6 percent of all ﬁrms and 55.8 percent of total employment. Firms with fewer than 100 employees accounted for 97.6 percent of all ﬁrms and 38.1 percent of total U.S. employment. Of the ﬁrms included in the Business Employment Dynamics (BED) data (U.S. Bureau of Labor Statistics, December 8, 2005), 54.4 percent had 1–4 employees, representing 5.2 percent of total employment. In contrast, ﬁrms with 1,000 or more employees represented 0.2 percent of ﬁrms and accounted for 37.4 percent of total employment in the United States. Hence, even at the aggregated ﬁrm-level of analysis, a substantial majority of workplaces in the United States employ fewer than 100 employees.

74

PHILIP BOBKO AND DENISE POTOSKY

It is also potentially useful to note that private small and large ﬁrms are not evenly represented across industry sectors. For example, Kobe (2007) analyzed the contribution of small businesses4 in nonfarm industries to the gross domestic product (GDP), and she showed that small businesses and large businesses do not necessarily dominate the same industry sectors in the United States. Small businesses represent more than 50 percent of real estate and leasing, professional and technical services, health and social services, arts and entertainment, accommodation and food services, construction, and other services industry sectors. Large ﬁrms dominate mining and manufacturing, utilities, transportation and warehousing, information, ﬁnance and insurance, and education services industry sectors. About the same number of small and large ﬁrms is represented in the trade (wholesale and retail), administrative services, waste management services, and holding companies. Kobe reports that in the period from 1998 to 2004, small businesses produced half of the private nonfarm GDP. Once again, the data indicate that the majority of private-sector ﬁrms have a relatively small number of employees. In turn, questions arise as to the adequacy of theories and research based almost exclusively on large companies (that represent a relatively small proportion of ﬁrms in the private sector). We now attempt to delineate some general implications of these ﬁndings within more speciﬁc management subdisciplines.

HUMAN RESOURCE (HR) MANAGEMENT, STRATEGIC HR, AND SELECTION/UTILITY In a subsequent section, we consider a small case analysis and a paper by Huselid (1995) to help address our points. However, as further context, we ﬁrst address the notion of job analysis, in general, and the potential role of ﬁrm size in the use of job analysis and personnel selection. Job Analysis Job analysis may be brieﬂy described as the process of analyzing the components and characteristics of each job in an organization; environmental contexts and job requirements can also be taken into account in this process (Sanchez & Levine, 2003; Simola, Taggar, & Smith, 2007). One traditional way of designing jobs has been to precisely deﬁne tasks associated with each job and then develop selection devices from these speciﬁcations (Heneman,

Firm Size and Strategic Human Resources

75

Judge, & Heneman, 2000a). Functional job analysis (FJA), for example, assumes that all work activity involves the way employees relate to data, people, and things within the work environment (Fine & Getkate, 1995; Fine & Wiley, 1974). As with many other job analysis methods (JAMs), conducting FJA entails having a trained job analyst collect and compare information about jobs, interview job incumbents (subject matter experts or SMEs), write task statements, and rate jobs on dimensions or ‘‘functions’’ such as comparing, analyzing, or synthesizing data, and so forth (cf. Fine & Getkate, 1995). Other JAMs exist and are used. For example, the Position Analysis Questionnaire (PAQ; McCormick, Jeanneret, & Mecham, 1972) asks SMEs to rate jobs on a set of predetermined questions that represent six dimensions/ factors that distinguish jobs from one another (e.g., the nature of input involved in the job, type of communication, and so forth). Thus, use of the PAQ depends upon collecting standardized and comparable information on a set of distinct jobs. Brannick, Levine, and Morgeson (2007) describe a combination job analysis method (C-JAM), which focuses on tasks (to determine what gets done and how), as well as the human attributes (knowledges, skills, abilities, other characteristics; KSAOs) that are most relevant to task performance. This method reﬂects a common approach in job analysis – the generation of a taxonomy of tasks and a taxonomy of KSAOs for each job. In C-JAM (and several other approaches), different groups of SMEs are used to generate tasks and KSAOs, reﬁne these lists, and rate each item on the list for various characteristics (e.g., importance, needed at entry, etc.).

Job Analysis in Small Firms Organizations that do not have distinct jobs to analyze, that have few or no incumbents, or use jobs that are not very differentiated may ﬁnd it particularly difﬁcult to apply such job analysis approaches. Again, we suspect this concern tends to apply to smaller (more typically sized) ﬁrms than larger ﬁrms. Efforts to ‘‘scale down’’ existing JAMs for use in smaller ﬁrms without any adjustment in fundamental assumptions about work in these organizations could likely have limited theoretical usefulness and low practical appeal (for a different view – that much knowledge developed on large ﬁrms can be directly applied to small ﬁrms – see Gatewood et al., 2008, p. 25; Gatewood & Feild, 1987).5 Again, research that is meaningful to these more frequent settings

76

PHILIP BOBKO AND DENISE POTOSKY

must not only incorporate methodologies suitable for small sample sizes, but must carefully consider underlying assumptions of traditional job analysis.

A Brief Qualitative Example By way of example of the potential differences between large and smaller ﬁrms, consider the case of a catering business (as described to one of us by the owner of the business). What began as a small family business grew to a very successful small business. The owner is the executive chef, but the kitchen employs several others who prepare food and menus. In addition to preparing hundreds of sandwiches, salads, etc., kitchen workers might also load catering trucks and deliver catered meals to clients. The company maintains a small ﬂeet of delivery trucks, with drivers who can also set up and take down catered spreads. The tasks associated with the kitchen and delivery jobs might initially be considered as relatively straightforward. However, delivery personnel must also present a good face to customers on behalf of the caterer, and drivers can also set up and take down catered spreads, so that other KSAOs (food design, sociability, etc.) are useful. Conversely, the kitchen workers perform the usual tasks (chopping, preparation, etc.) but they also help load trucks and deliver catered meals to clients. The point is that smaller ﬁrms may require less distinction between who does what at work. That is, the structure of work and job positions in smaller organizations may not be highly specialized or differentiated, relative to work and jobs in larger ﬁrms. In particular, there is theoretical as well as empirical support for the notion that larger ﬁrms need more formalized procedures, speciﬁcation (and differentiation) of procedures, and are more focused on efﬁciencies and achieving economies of scale than smaller ﬁrms (Barber et al., 1999; Chandler, 1990; Williamson, 1975). In support of this logic, a survey study in the United Kingdom (UK) reported that systematic job analysis was not conducted in small ﬁrms (Carroll, Marchington, Earnshaw, & Taylor, 1999). Also, Tanova (2003) (citing Torrington, 1991) noted that small ﬁrms may not need detailed job analysis, job descriptions, or job speciﬁcation processes, but could more simply determine an employee’s main tasks and recruit new employees using informal methods. Stines and Kleiner (2003) recommended that small business owners themselves generate job descriptions. Note that this job description would not be derived from traditional job analysis, but would simply reﬂect the owner’s judgment at the time. In turn, and consistent with

Firm Size and Strategic Human Resources

77

our comments below about Huselid (1995), Simola et al. (2007, p. 39) noted that basing job interview questions solely upon job descriptions without the beneﬁt of systematically collected information about the requirements of the position to be ﬁlled ‘‘would be counter to research-based ‘best practices’.’’

Other Possibilities The above case example is consistent with the deﬁnition of ‘‘ﬂexible jobs’’ provided by Heneman et al. (2000a): Many small business owners, general managers of start-up strategic business units, and top management members perform such ﬂexible jobs. These jobs are ‘‘loose cannon’’ ones, characterized by broad job titles (e.g., administrator, general manager, director, scientist) and job descriptions with only cursory statements about tasks and duties (e.g., ‘‘manages budget planning, human resources, and marketing processes’’). (p. 158)

A different set of job analysis options, possibly useful for smaller ﬁrms, might be ‘‘work analysis’’ (cf. Sanchez & Levine, 1999, 2003) and competency modeling (cf. Shippmann et al., 2000; Spencer & Spencer, 1993). Unlike traditional job analysis efforts, which tend to focus on identifying job content and related KSAOs, competency modeling tends to focus on the longer-term organizational ﬁt (Shippmann et al., 2000), which may make competency modeling seem feasible and appealing to the smaller ﬁrm. Indeed, competency modeling is associated with ﬂexible jobs, as competencies can span multiple jobs and are understood as contributing not only to job performance but also to organizational success (Heneman et al., 2000a). Whether or how smaller organizations actually use competency modeling is not well known, however. Organizations that choose to implement competency modeling will do so in relatively ‘‘uncharted legal waters’’ (Heneman et al., 2000a, p. 189). Because they are perceived as less rigorous than traditional job analyses (Sanchez & Levine, 2003; Shippmann et al., 2000), researchers have encouraged efforts that improve the psychometric characteristics of competency modeling (Lievens, Sanchez, & DeCorte, 2004; Shippmann et al., 2000). In sum, rather than precisely deﬁning job titles and/or tasks, smaller ﬁrms more likely need to embrace and deﬁne broad job types. Creative methods might be needed that will address concerns about psychometric precision. Competency modeling (or similar processes) may be required because job ﬂexibility (overlapping job roles) may be the norm in small ﬁrms.

78

PHILIP BOBKO AND DENISE POTOSKY

AN EXAMPLE FROM THE LITERATURE A recent issue of the Academy of Management Journal provided a crossarticle citation analysis (Tsai & Wu, 2010), which placed prevailing management research into subcategories. One of the subcategories in this analysis was a small, closed system of research on strategic HRM, and one of the most oft-cited articles in this category is a paper by Huselid (1995). This paper reports on recommended high-performance management practices. We consider this paper here because it is an inﬂuential paper and we believe it can serve as an example of many articles in strategy and management in terms of its studied ﬁrms. The reported mean ﬁrm size (labeled ‘‘total employment’’) in this study is 4,412.80,6 although, as noted above, ﬁrms with fewer than 1,000 employees represent 99.8 percent of U.S. ﬁrms. Firms that had fewer than 100 employees were completely excluded from analysis in the Huselid study (p. 644). However, as noted above, ﬁrms with fewer than 100 employees accounted for 97.6 percent of all ﬁrms and 38.1 percent of total U.S. employment. Thus, it is not clear how the highperformance management practices recommended in this study, and subsequently endorsed in the ﬁeld, might apply to the majority of ﬁrms in the United States. For example, some of the 13 factors in Huselid (1995) included a ‘‘formal job analysis,’’ ‘‘attitude surveys,’’ a newsletter or other formal information sharing processes, programs such as quality circles or quality of work-life, formal grievance procedures, administration of preemployment tests, and formal performance appraisals. It is unclear how such practices might efﬁciently be accomplished by management in smaller (i.e., the majority of ) ﬁrms, or whether it is even realistic and/or feasible to do so. To recap, recommendations for best practices in strategic HRM have been based almost exclusively on the study of large ﬁrms, and scholars have made little effort to consider and explain how most ﬁrms might scale these recommendations to their situations.

SOME IMPLICATIONS FOR PERSONNEL SELECTION AND UTILITY ANALYSIS Maurer and Fay (1986) called attention to the notion that smaller U.S. ﬁrms might need to comply with legal expectations that base valid selection methods upon job analyses. We have already pointed out some issues regarding the ﬁeld’s dependence on large ﬁrms in developing what deﬁnes a

Firm Size and Strategic Human Resources

79

good job analysis. Also, regarding selection, we noted that small ﬁrms might best focus on a large constellation of competencies when assessing applicants for a single, open position. In contrast, larger ﬁrms can often focus on a well-deﬁned job (and well-deﬁned tasks), because the large ﬁrm has more narrowly deﬁned jobs that exploit efﬁciencies from specialization. Further yet, large companies might be able to afford a streaming of hiring dates (e.g., the military’s delayed entry program) because of their large size even without the new hire(s), whereas small ﬁrms might have more ﬁxed, immediate hiring needs. Regarding utility analysis (i.e., the assignment of dollar value to selection systems), we wonder if the underlying assumptions need to change as a function of ﬁrm size. The usual formula for selection test utility involves a variety of terms, including validity (r), the selection ratio (s.r.), and the number of individuals selected (ns) using the test (for a review, see Roth, Bobko, & Mabon, 2001). In particular, the validity and number of individuals selected are multiplicative factors in the calculation, whereas the selection ratio appears in the denominator and is thus inversely related to utility. However, the functional role of these factors, as well as their point estimates, may be different for smaller ﬁrms. For example, it is possible that estimates of test validity (r) are lower than assumed in smaller ﬁrms.7 That is, meta-analyses of test validities in the literature are inﬂuenced by large ﬁrms/samples, and these validities are based on individuals who will be doing a relatively speciﬁc job (with relatively welldeﬁned criteria). However, as noted, in smaller ﬁrms individuals take on varied organizational roles/jobs/tasks. As such, the validity of any single test is lowered, as the ultimate criterion for such an individual becomes quite heterogeneous. As another example, the selection ratio (s.r.) appears in the denominator of the usual utility formula. That is, as an organization becomes more selective (i.e., s.r. is smaller), the estimated test utility rises. However, large ﬁrms might be able to engage in substantial, national/global recruiting efforts. In turn, the number of applicants is larger, s.r. becomes smaller, and estimated utility rises. Smaller ﬁrms, however, might recruit more locally and their selection ratios might be relatively large. Thus, differences in estimated test utility might be artifactual and current formulas may need rethinking if they are applied to smaller ﬁrms. As yet another example, the number of individuals selected (ns) is usually used as a direct multiplicative factor in overall estimated utility; that is, if the number of individuals hired is increased by a factor of ﬁve, then organization utility estimated to be gained from the selection process increases by a factor

80

PHILIP BOBKO AND DENISE POTOSKY

of ﬁve. However, in smaller ﬁrms, increasing hiring ﬁvefold for a particular job may be unrealistic. On the one hand, the formula for utility might be logical because the larger ﬁrm gains more utility from its larger number of hires. However, for the smaller ﬁrms, a single hire might be crucial – so selecting one good person is more important to this organization than selecting ﬁve individuals for the large organization. To put this another way, the larger ﬁrm might be able to absorb ‘‘mis-hires’’ but the smaller organization cannot. This aspect (the percentage of the ﬁrm’s workforce represented by one individual) is not taken into account in current utility estimates.

STRATEGIC MANAGEMENT IN GENERAL Some Additional Facets of Large Firms Large ﬁrms can exploit advantages of both scale and scope (Chandler, 1990). They can invest in production facilities large enough to take advantage of economies of scale. They can invest in marketing and distribution networks to keep pace with increased production. They can consider strategies that involve mass production and mass marketing. Large ﬁrms can also invest in management – by recruiting and training managers for administration and oversight. Such economies of scale and scope are frequently noted. For example, consider the General Foods Corporation. As noted by D’Antonio (2006), its roots are the Postum Cereal Company that started in a small barn in Battle Creek, MI. By consolidating many different brands, the larger company was able to achieve economies of scale by using common shipping, sales, and distribution networks. The use of complementary skills and knowledge across brands and establishments was also involved. Additionally, as Chandler (1990) notes, such massive systems and complexities result in the separation of management and ownership because investors do not have adequate time or experience or information to run these large companies. That is, he notes that the ‘‘enlarged enterprises came to be operated by teams of salaried managers who had little or no equity in the ﬁrm’’ (p. 1). An implication of this thinking is that selection systems might systematically differ across small and large ﬁrms because, for example, owner, manager, and employee work roles would be less differentiated in smaller organizations. This reduced differentiation of work roles seems likely to affect the dynamics of organizational selection. For example, in large ﬁrms,

Firm Size and Strategic Human Resources

81

the possible hiring of professional managers, devoid of much experience with organizationally speciﬁc tasks, might require a stronger focus on selection systems that explicitly assess levels of intrinsic motivation to encompass the company’s values and culture – because these professional managers have less experience, or potentially less vested interest, in the ﬁrm’s services/products. However, there would seem to be some additional inherent/intrinsic motivation levels for applicants to smaller ﬁrms (because the goals/values/tasks within the small organizations are more strongly tied to the overall organizational purpose and structure). Thus, although facets such as culture ﬁt, person-job ﬁt, and motivational commitment can be important personnel attributes in most ﬁrms, it may be that these non-job-speciﬁc KSAOs are already present in applicants to small ﬁrms. On the other hand, the relatively increased importance of these non-job-speciﬁc KSAOs to smaller ﬁrms may require increased attention at the point of formal selection. Research on these two competing hypotheses could be fruitful.

Some Strategy Types and the Firm Size Issue A classic set of organization strategies involves the Porter typology (cf. Porter, 1985, as described in Hitt, Ireland, & Hoskisson, 1999). In a nutshell, four strategic types arise from a two-by-two consideration of the factors of (i) cost versus differentiation strategies and (ii) broad or narrow/focused target strategies. Within this typology, we again suggest that ﬁrm size matters: large ﬁrms have the capacity to realistically consider all of the four resulting strategies, whereas small ﬁrms do not. For example, consider a broad cost strategy. This is sometimes labeled a cost leadership strategy. Hitt et al. (1999) note it ‘‘is an integrated set of actions designed to produce products at the lowest cost, y with features that are acceptable to customers’’ (p. 136). Such ﬁrms might need (a) signiﬁcant access to ﬁnancial resources in order to invest in production and (b) a scope that allows sharing of resources across different product lines. Such resources/economies might include having a common network of suppliers, common production facilities, and/or common distribution networks – the very things we noted that were more typically associated with larger ﬁrms. Thus, we suggest that, all else equal, small ﬁrms are less likely to be able to feasibly consider the cost leadership strategy. Focused low cost might be possible, but economies of scale seem to make such strategies less competitive for small ﬁrms.

82

PHILIP BOBKO AND DENISE POTOSKY

In contrast, logic suggests that relatively small ﬁrms could consider differentiation, and particularly focused differentiation, strategies. Differentiation strategies, by deﬁnition, energize ﬁrms to produce products or provide services that are unique. Uniqueness might be obtained by unusualness or by generation of a commonly available product/service but with a focus on high quality. We suggest these strategic options are indeed available to smaller ﬁrms, despite their size (or possibly because of it). Smaller ﬁrms can potentially more readily build strong corporate culture (e.g., recall owners and managers are less distinct). Smaller ﬁrms (and each of their employees) might also be ‘‘closer to the knitting’’ and thus more readily recognize high quality (i.e., recognition of high quality might be a precondition to providing high quality). Our point is that ﬁrm size might inﬂuence what is regarded as useful strategic practice. In particular, we suggest that large ﬁrms have available to them all of the strategic possibilities indicated by Porter. (This includes, of course, success at differentiation strategies; e.g., Starbucks, Lexus, and so forth.) In contrast, small ﬁrms may be logically constrained in their strategic options. What is best for them is not necessarily best for large ﬁrms. Further, not all small ﬁrms expect or hope to be acquired by larger ﬁrms in order to achieve economies of scale. A recent paper (Cao, Gedajlovic, & Zhang, 2009) supports our contention that small ﬁrms face different strategic options and constraints than large ﬁrms. Cao et al. were interested in the strategic concept of organizational ambidexterity. They studied the concept of balance between exploratory and exploitative organizational activities. They also looked at the simultaneous embracement of exploratory and exploitative activities. They found that resource availability (measured as ﬁrm size and environmental muniﬁcence) inﬂuences the number and types of activities a ﬁrm can effectively pursue. More speciﬁcally, smaller, more resource-constrained ﬁrms need to balance their ‘‘exploration’’ and ‘‘exploitation’’ activities. In contrast, larger, less resource-constrained ﬁrms can pursue both activities effectively, and such pursuits enhance ﬁrm performance if the organization is large. Strategic management research might move forward with new assumptions about the ‘‘strategies of small ﬁrms’’ and new methods for studying them. For example, ﬁeld research might consider analyzing a set of smaller establishments within larger parent organizations in order to build conﬁdence in the models obtained. An example is an analysis of corporate entrepreneurship strategies conducted by Kuratko, Ireland, and Hornsby (2001). These researchers studied 50 companies of 42–200 employees each that were part of a larger umbrella ﬁrm. They also examined one of these smaller companies

Firm Size and Strategic Human Resources

83

(initially 97 employees) in more detail in order to focus in-depth on corporate entrepreneurship activities. The point is that researchers need to be more proactive in seeking new methodologies and reﬁning existing approaches related to topics such as job analysis, personnel selection, and strategy.

SUMMARY AND CONCLUSIONS To summarize, we have noted in this chapter that most private employer establishments are small; most individuals work in settings with fewer than 100 other employees. Large ﬁrms are exceptions to this norm, and organizational knowledge generated from the study of large ﬁrms may not be automatically applicable to numerous smaller ﬁrms. For example, jobs might be deﬁned differently and more broadly in smaller establishments than in large organizations. In addition, the feasibility of best corporate strategies may be moderated by the size of the ﬁrm. A related issue raised in this chapter is that most job analysis methods assume that work is differentiated, and that distinct jobs can be deﬁned primarily according to tasks performed (and associated KSAOs). Work analysis or competency modeling approaches, pending additional research and validation studies, may be most appropriate for organizations that must deﬁne broad job categories and work roles along with, or in place of, differentiated tasks. We also suggested that selection in large ﬁrms is more likely to be characterized by greater focus on KSAOs relevant to task performance. In contrast, selection in small ﬁrms may be more likely to focus on ‘‘other characteristics’’ such as ﬁt with the work setting, schedule availability, experience, and ﬂexibility alongside requisite KSAOs or occupational qualiﬁcations for the speciﬁc types of work. This does not mean to imply that larger ﬁrms do not care about organizational ‘‘ﬁt’’ during selection, but that ﬁt characteristics are generally not the primary basis for selection into large, differentiated organizational systems. Synthetic validity approaches may also prove useful, but it is important to keep in mind that job analysis is likely needed to deﬁne the key components of each job no matter what validation approach is applied. In addition, we noted that the underlying model of selection utility, and several factors in its numerical estimation, might need to be modiﬁed as a function of ﬁrm size. We also suggested that large ﬁrms have available to them all of the strategic possibilities indicated by Porter. In contrast, small ﬁrms may be logically constrained in their strategic options. Indeed, as noted to us by the

84

PHILIP BOBKO AND DENISE POTOSKY

series editors, perhaps future researchers can more systematically juxtapose ‘‘best practices’’ recommendations with ‘‘ﬁrm size ‘‘(i.e., delineate which practices might increase the probability of success for small ﬁrms and which practices might increase the probability of success for large ﬁrms). This chapter may also inspire HRM and strategy researchers by helping to focus future research. In her recommendations for bridging the organizational practice and science gap, Rousseau (2007) suggested forming an evidence-based management collaboration (EBMC) that would leverage practice-informed research, create new initiatives, and provide results that are useful or scalable across many different disciplines. It appears to be essential to keep the large number of small, private-sector U.S. ﬁrms in mind in this effort.

NOTES 1. Although the QCEW and BLS demarcate establishments as embedded within ﬁrms, the academic research literature on small organizations rarely makes this distinction. Aside from the precision used when referring to BLS data, this chapter follows the convention of using the word ‘‘ﬁrm’’ as broadly deﬁned and synonymous with the terms organization, workplace, company, business, etc. 2. The QCEW program produces data on establishments, employment, and wage by size of establishment for the ﬁrst quarter of each year. Tables are available in the QCEW’s annual bulletin, Employment and Wages, and are also listed by the Bureau of Labor Statistics on their website (www.bls.gov). 3. The QCEW does not distinguish between establishments with zero, one, two, three, or four employees, but groups ﬁrms with less than ﬁve employees together in one category. Note that ﬁrms with zero employees in this dataset are ﬁrms that paid employees at some time during the year, but listed zero paid employees during the ﬁrst quarter/mid-March pay period. 4. In Kobe’s (2007) analyses, small businesses were ﬁrms with fewer than 500 employees. 5. Gatewood and Feild (1987) outlined a three-step approach that small businesses could use. Step 1 involves conducting a systematic job analysis that produces a list of task statements (rated on time spent and importance). Step 2 involves identifying knowledge and skill levels of workers. Step 3 involves developing selection instruments that assess these worker characteristics in terms of the critical job tasks. This approach mirrors traditional large-scale procedures. However, if a job title has no incumbents or if differentiation between jobs is low, a small business owner might have difﬁculty getting past step 1. 6. The standard deviation of ﬁrm size is reported to be 18,967, so the mean is likely inﬂuenced by several large ﬁrms (outliers). Hence, the median is likely less than the 4,412.80 value, but the point remains.

Firm Size and Strategic Human Resources

85

7. Indeed, small ﬁrms might have difﬁculty empirically validating their tests and having enough statistical power to ﬁnd a signiﬁcant correlation, if one exists. ‘‘For a predictive validity study to have statistical power of .80, to detect a correlation of .30 at the .05 level of statistical signiﬁcance, would require hiring 84 people for one job category over some relatively short time period’’ (Hollenbeck & Whitener, 1988, p. 536). Other techniques, such as synthetic validity, might need to be considered (see Scherbaum, 2005, for a review or a recent, 2010 issue of The Industrial Organizational Psychologist that featured papers and comments on synthetic validity).

REFERENCES Barber, A., Wesson, M., Roberson, Q., & Taylor, M. (1999). A tale of two job markets: Organizational size and its effects of hiring practices and job search behavior. Personnel Psychology, 52, 841–867. Brannick, M., Levine, E., & Morgeson, F. (2007). Job and work analysis: Methods, research, and applications for human resource management (2nd ed.). Cambridge, MA: Sage. Cao, Q., Gedajlovic, E., & Zhang, H. (2009). Unpacking organizational ambidexterity: Dimensions, contingencies, and synergistic effects. Organization Science, 20, 781–796. Carroll, M., Marchington, M., Earnshaw, J., & Taylor, S. (1999). Recruitment in small ﬁrms: Processes, methods, and problems. Employee Relations, 21, 236–250. Chandler, A. (1990). Scale and scope: The dynamics of industrial capitalism. Cambridge, MA: Belknap Press. D’Antonio, M. (2006). Hershey. New York: Simon and Schuster. Duberly, J., & Walley, P. (1995). Assessing the adoption of HRM by small and medium-sized manufacturing organizations. The International Journal of Human Resource Management, 4, 891–909. Fine, S., & Getkate, M. (1995). Benchmark tasks for job analysis. Mahwah, NJ: Lawrence Erlbaum. Fine, S., & Wiley, W. (1974). An introduction to functional job analysis. In: E. Fleishman & A. Bass (Eds), Studies in personnel and industrial psychology (pp. 6–13). Homewood, IL: Irwin. Gatewood, R., & Feild, H. (1987). A personnel selection program for small business. Journal of Small Business Management, 25, 16–24. Gatewood, R., Feild, H., & Barrick, M. (2008). Human Resource Selection (6th ed.). Mason, OH: Thomson Higher Education of Thomson Southwestern. Hendry, C., Arthur, M., & Jones, A. (1995). Strategy through people: Adaptation and learning in the small and medium sized enterprise. London: Routledge. Heneman, H., Judge, T., & Heneman, R. (2000a). Stafﬁng organizations. Middleton, WI: Irwin McGraw-Hill. Heneman, R., Tansky, J., & Camp, S. (2000b). Human resource management practices in small and medium-sized enterprises: Unanswered questions and future research perspectives. Entrepreneurship Theory and Practice, Fall, 11–26. Hitt, M., Ireland, D., & Hoskisson, R. (1999). Strategic management: Competitiveness and globalization (3rd ed.). Cincinnati, OH: South-Western College Publishing.

86

PHILIP BOBKO AND DENISE POTOSKY

Hollenbeck, J., & Whitener, E. (1988). Criterion-related validation for small sample contexts: An integrated approach to synthetic validity. Journal of Applied Psychology, 73, 536–544. Huselid, M. (1995). The impact of human resource management practices on turnover, productivity, and corporate ﬁnancial performance. Academy of Management Journal, 38, 635–672. Kobe, K. (2007). The small business share of GDP, 1998–2004. Small Business Research Summary no. 299. Economic Consulting Services, LLC, Washington, DC. Kuratko, D., Ireland, R., & Hornsby, J. (2001). Improving ﬁrm performance through entrepreneurial actions: Acordia’s corporate entrepreneurship strategy. Academy of Management Executive, 15, 60–71. Lievens, F., Sanchez, J., & DeCorte, W. (2004). Easing the inferential leap in competency modeling: The effects of task-related information and subject matter expertise. Personnel Psychology, 57, 881–904. Maurer, S., & Fay, C. (1986). Legally fair hiring practices in the small business environment. Journal of Small Business Management, 24, 47–54. McCormick, E., Jeanneret, P., & Mecham, R. (1972). A study of job dimensions based on the Position Analysis Questionnaire. Journal of Applied Psychology, 56, 347–368. McElwee, G., & Warren, L. (2000). The relationship between total quality management and human resource management in small and medium-sized enterprises. Strategic Change, 9, 427–435. Porter, M. (1985). Competitive advantage: Creating and sustaining superior performance. New York: Free Press. Roth, P., Bobko, P., & Mabon, H. (2001). Utility analysis: A review and analysis at the turn of the century. In: N. Anderson, D. Ones, H. Sinangil & C. Viswesvaran (Eds), Handbook of industrial, work and organizational psychology (Vol. 1, pp. 363–384). Thousand Oaks, CA: Sage. Rousseau, D. (2006). Is there such a thing as ‘‘evidence-based management’’? Academy of Management Review, 31, 256–269. Rousseau, D. (2007). A sticky, leveraging, and scalable strategy for high-quality connections between organizational practice and science. Academy of Management Journal, 50, 1037–1042. Sanchez, J., & Levine, E. (1999). Is job analysis dead, misunderstood, or both? New forms of work analysis and design. In: A. Kraut & A. Dorman (Eds), Evolving practices in human resource management (pp. 43–68). San Francisco: Jossey-Bass. Sanchez, J., & Levine, E. (2003). The analysis of work in the 20th and 21st centuries. In: N. Anderson, D. Ones, H. Sinagil & C. Viswesvaran (Eds), Handbook of industrial, work, and organizational psychology: Personnel psychology (Vol. 1, pp. 71–89). New York: Sage. Scherbaum, C. (2005). Synthetic validity: Past, present, and future. Personnel Psychology, 58, 481–515. Shippmann, J., Ash, R., Battista, M., Carr, L., Eyde, L., Hesketh, B., et al. (2000). The practice of competency modeling. Personnel Psychology, 53, 703–740. Simola, S., Taggar, S., & Smith, G. (2007). The employment selection interview: Disparity among research-based recommendations, current practices and what matters to human rights tribunals. Canadian Journal of Administrative Sciences, 24, 30–44.

Firm Size and Strategic Human Resources

87

Spencer, L., & Spencer, S. (1993). Competence at work: Models for superior performance. New York: Wiley. Stines, S., & Kleiner, B. (2003). Keys to hiring employees effectively in a small business. Management Research News, 26, 170–180. Tanova, C. (2003). Firm size and recruitment: Stafﬁng practices in small and large organizations in north Cyprus. Career Development International, 8, 107–114. Torrington, D. (1991). Employee resourcing. London: Institute of Personnel Management. Tsai, W., & Wu, C. (2010). Knowledge combination: A co-citation analysis. Academy of Management Journal, 53, 441–450. U.S. Bureau of Labor Statistics. (2005). Newsletter: New quarterly data from BLS on business employment dynamics by size of ﬁrm, December 8. Available online at http://www. bls.gov/bdm/ U.S. Bureau of Labor Statistics. (2008). Employment and wages, annual averages 2008. Table 3: Private industry by supersector and size of establishment: Establishments and employment, ﬁrst quarter 2008. Available at http://www.bls.gov/cew/ew08table3.pdf Vickerstaff, S. (1993). The training needs of small ﬁrms. Human Resource Management Journal, 2, 1–15. Williamson, I. (2001). Employer legitimacy and recruitment success in small businesses. Entrepreneurship, Theory and Practice, 25, 27–42. Williamson, O. (1975). Markets and hierarchies. New York: Free Press.

LEADER–MEMBER EXCHANGE (LMX) MEASUREMENT: EVIDENCE FOR CONSENSUS, CONSTRUCT BREADTH, AND DISCRIMINANT VALIDITY Dana L. Joseph, Daniel A. Newman and Hock-Peng Sin ABSTRACT Purpose – This chapter (a) summarizes leader–member exchange (LMX) measurement practices since the inﬂuential reviews by Schriesheim, Castro, and Cogliser (1999) and Gerstner and Day (1997), (b) clariﬁes the status of LMX as a broad construct from a hierarchical factor model, (c) conducts multitrait-multimethod (MTMM) analyses on leader and follower reports of multidimensional LMX, and (d) investigates discriminant validity between Member LMX and satisfaction with supervisor. Methodology/Approach – We used (a) a literature search of LMX measurement practices, (b) a combination of meta-analysis and factor analysis to specify the broad LMX construct underlying Liden and Maslyn’s (1998) (LMX-MDM) multidimensional instrument, (c) MTMM analyses

Building Methodological Bridges Research Methodology in Strategy and Management, Volume 6, 89–135 Copyright r 2011 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-8387/doi:10.1108/S1479-8387(2011)0000006012

89

90

DANA L. JOSEPH ET AL.

of leader and member ratings of the LMX-MDM, and (d) a combination of meta-analysis and multiple regression to assess incremental validity of Member LMX beyond satisfaction with supervisor. Findings – Since 1999, 85% of LMX studies now use one of two dominant LMX scales (LMX-7, Scandura, & Graen, 1984; LMX-MDM, Liden & Maslyn, 1998). These two measures are correlated (rcorrected ¼ .9), suggesting the LMX-7 and the LMX-MDM are alternate forms of the same instrument. 94% of studies that used these two measures treat LMX as a single, broad construct rather than as a multidimensional set of constructs. MTMM analyses suggest Leader LMX and Member LMX are two, separate-but-related constructs (i.e., conﬁrming two source factors and no lower-order trait factors). Last, Member LMX meta-analytically correlates with satisfaction with supervisor at rcorrected ¼ .8. There is some incremental validity of LMX, but the pattern is inconsistent across samples. Social Implications – We point out that LMX researchers have now moved toward standard measurement of LMX – as a broad, higher-order factor that varies between leader and follower. By doing so, we reveal that the stage is set for cumulative and replicable research on leadership as a dyadic, follower-speciﬁc phenomenon. Originality/Value of Paper – Our chapter is the ﬁrst to reveal consensus in LMX measurement across studies; to summarize the standard treatment of LMX as a single, broad factor; and to apply MTMM analyses to demonstrate separate Leader LMX and Member LMX source factors. Keywords: Bandwidth-ﬁdelity; factor analysis; leader–member exchange; LMX; measurement; meta-analysis; multitraitmultimethod; satisfaction with supervisor Leader–member exchange (LMX) is a construct that has been deﬁned in various ways. It has been treated as member perceptions of leader characteristics, as leader perceptions of member characteristics, and/or as a dyadic construct representing a property of the leader–member unit (i.e., leader and member perceptions of their relationship, at the dyadic level of analysis). LMX has been employed as a unidimensional concept (Graen & Schiemann, 1978) and also as a multidimensional set of concepts (Dienesch & Liden, 1986; Liden & Maslyn, 1998; Schriesheim, Neider, Scandura, & Tepper, 1992).

Leader–Member Exchange (LMX) Measurement

91

Similarly, the content of LMX has variously included the leader’s ﬂexibility, supportiveness, use of feedback, decision making, likability, loyalty, performance, reputation, and trustworthiness (see Ferris et al., 2009; Graen & Scandura, 1987), as well as the generic effectiveness of the leader–member relationship itself (Graen, Liden, & Hoel, 1982). Thus it appears the LMX label has a history of being liberally applied to a wide variety of concepts (Schriesheim, Castro, & Cogliser, 1999), which are all at least loosely associated with how leaders and members interact. Despite its sometimes slippery construct validity, LMX remains one of the most studied concepts in the ﬁeld of leadership (Barling, Christie, & Hoption, 2010). In this chapter, we forward the position that the LMX concept has evolved from a vague notion (i.e., deﬁned imprecisely) to now being most accurately described as an abstract notion (i.e., deﬁned broadly, as a general tendency to behave in relationally supportive ways). Along the way toward substantiating this position, we illustrate several analytic techniques that can be used to assess the construct validity of dyadic and interpersonal constructs in the organizational sciences. This chapter attempts to make four contributions to the study of LMX conceptualization and measurement. First, we review the coevolution of deﬁnitions and measures of LMX and attempt to quantify the effects of previous reviews (e.g., Schriesheim et al., 1999) on the current practice of LMX measurement. We demonstrate that LMX researchers have recently converged on two particular measures – the LMX-7 (a unidimensional measure of LMX; Scandura & Graen, 1984) and the LMX-MDM (a multidimensional measure of LMX; Liden & Maslyn, 1998). Second, we note that whereas the LMX-MDM is ostensibly a multidimensional measure of LMX, both factor-analytic evidence and common research practice emphasize the broad, higher-order factor of the LMX-MDM. This single factor from the LMX-MDM exhibits a disattenuated correlation of r ¼ .9 with the LMX-7, suggesting the two dominant measures of LMX do indeed tap into the same, broad LMX construct. Third, we address previous metaanalytic ﬁndings that suggest member perceptions of LMX are very different from leader perceptions of LMX (Gerstner & Day, 1997; Sin, Nahrgang, & Morgeson, 2009) by applying a multitrait-multimethod (MTMM) analysis to two multisource LMX-MDM datasets. MTMM results reveal that Leader LMX and Member LMX are two, separate-but-correlated source constructs, and facet-level LMX provisions from one partner do not imply reciprocation-in-kind from the other partner (e.g., member loyalty does not speciﬁcally enhance the probability of leader loyalty, and vice-versa). Fourth and ﬁnally, we summarize the strong overlap between the broad

92

DANA L. JOSEPH ET AL.

Member LMX factor and satisfaction with supervisor (meta-analytic corrected r ¼ .8), and then estimate incremental validity of LMX over satisfaction with supervisor in predicting attitudinal and behavioral outcomes. Although LMX does sometimes exhibit meaningful incremental validity beyond satisfaction with supervisor, these results seem inconsistent across contexts.

HOW HAS LMX BEEN MEASURED? Current LMX Measurement Practices In the late 1990s, three events occurred that would have a major impact on future measurement of LMX. First and foremost, Schriesheim and colleagues (1999) published a comprehensive review of LMX measurement, which revealed that researchers were using many, different measures (and different versions of measures) of LMX. They noted the general state of LMX measurement was ad hoc and summarized that scales used to measure LMX were generally not based on explicit construct validation. They remarked, ‘‘exactly what these [LMX] scales are measuring is unknown and attempting to substantively synthesize the existing literature may therefore not make much sense’’ (p. 101). Second, around the same time, Gerstner and Day (1997) provided a comprehensive and inﬂuential meta-analytic review of the LMX literature, in which they recommended that future researchers interested in unidimensional LMX should use the LMX-7 measure (p. 837). This recommendation came on two grounds: (a) the LMX-7 exhibited higher internal consistency reliability than other LMX measures, on average, and (b) the LMX-7 showed stronger criterion validity correlations with job performance, job attitudes, role conﬂict, and role clarity than did other LMX measures, on average. Third, Liden and Maslyn (1998) published a new measure of LMX (known as the LMX-MDM) designed to capture the multiple dimensions of LMX conceptualized by Dienesch and Liden (1986) and extended the measurement work of Dienesch (1986) and Schriesheim et al. (1992). The Liden and Maslyn (1998) instrument development was exemplary in many ways and followed state-of-the-art scale development procedures (see Hinkin, 1995), including content validity analyses similar to those described by Schriesheim, Powers, Scandura, Gardiner, and Lankau (1993). The new multidimensional LMX-MDM scale was brief and appeared to have unusually strong construct validity evidence supporting its use (Liden & Maslyn, 1998).

93

Leader–Member Exchange (LMX) Measurement

What was the cumulative result of these three events in the late-1990s [i.e., the Gerstner-Day (1997) meta-analysis, the new Liden and Maslyn (1998) LMX-MDM scale, and the Schriesheim et al. (1999) expose´ of LMX measurement practices]? Our answer to this question appears in Fig. 1. In Fig. 1, we present a review of LMX measurement practices both before and since 1999. Figure 1 indicates that the combined effect of the above three events created a virtual consensus among LMX researchers that LMX should be measured through either the LMX-7 or the LMX-MDM. That is, before 1999, there was little consensus in how to measure LMX, with only 22% of researchers employing the LMX-7, and over 56% using a hodgepodge of other LMX measures. But after Schriesheim et al.’s (1999) review, Gerstner and

70

% of LMX studies

60 50 40 30 20 10 0 LBDQ

Neg. Lat., LMX-4 & LMX-5

LMX-7

Before 1999

Other

LMX-7

LMX-MDM

Other

After 1999

Fig. 1. Usage of Various LMX Measures, both Before and After Schriesheim et al. (1999). Note: Before 1999, N ¼ 137 empirical studies of LMX taken from the review by Schriesheim et al. (1999; note the original article misnumbered these studies from 1 to 147). After 1999 (through 2010), N ¼ 241 empirical studies of LMX. LBDQ ¼ leader behavior description questionnaire (Stogdill & Coons, 1957); Neg. Lat. ¼ negotiating latitude scale of Dansereau et al. (1975); LMX-4 ¼ negotiating latitude scale of Graen & Cashman (1975); LMX-5 ¼ Graen et al., 1982a; LMX7 ¼ versions of the Scandura and Graen (1984) 7-item instrument (also see Graen et al., 1982b; Graen & Uhl-Bien, 1995); LMX-MDM ¼ multidimensional LMX instrument (Liden & Maslyn, 1998). See Fig. 2 for description of measures.

94

DANA L. JOSEPH ET AL.

Day’s (1997) meta-analysis, and the introduction of the LMX-MDM (Liden & Maslyn, 1998), we see 66% of LMX studies using the LMX-7, and most of the others (19%) using the LMX-MDM (i.e., 85% of LMX studies since 1999 have used either the LMX-7 or LMX-MDM; see Fig. 1). In retrospect, it appears that the late 1990s were an important transition period toward consensus in LMX measurement, when the LMX-7 and the LMX-MDM were established as the de facto standards for assessing the construct.

A Brief History We next brieﬂy describe the history of LMX measurement by focusing on the dominant LMX measurement practices since 1999, and we refer the reader to Schriesheim et al.’s (1999) comprehensive review for a more thorough discussion of LMX measurement before 1999. Before 1999, there was little consensus in how to measure LMX, and many LMX measures proliferated. Amidst the multitude of measures used during the early decades of LMX research, one can identify a single, cohesive sequence of LMX measurement that would ultimately evolve into the dominant measures used today. To see the cohesive stream, we present a genealogy of LMX measurement in Fig. 2 (i.e., the LMX timeline). To be concise, the development of the dominant measures of LMX to date has essentially progressed in four, temporally overlapping stages: (1) the Leader Behavior Description Questionnaire (LBDQ; Stogdill & Coons, 1957), (2) precursors to the LMX-7, (3) the LMX-7, and (4) the LMXMDM. First, LMX research originated using the LBDQ from the Ohio State studies (Stogdill & Coons, 1957) to measure leader consideration and initiating structure (at the level of the individual follower; e.g., Dansereau, Cashman, & Graen, 1973; Graen, Dansereau, & Minami, 1972a, 1972b; Katerberg & Hom, 1981). Second, this practice of using the LBDQ was quickly replaced using much shorter measures (two to ﬁve items) that focused on leader helpfulness, leader ﬂexibility in supporting follower job changes, leader supportiveness, and openness to suggestions from the follower [see Fig. 2; these second-stage measures include the Negotiating Latitude scale (Dansereau, Graen, & Haga, 1975), the LMX-4 (Graen & Cashman, 1975) and the LMX-5 (Graen, Liden, et al., 1982)]. This second stage of LMX measurement can be considered ‘‘precursors to the LMX-7’’ (see item-level review by Bernerth, Armenakis, Field, Giles, & Walker, 2007, p. 984). Third, the LMX-7 was presented (Scandura & Graen, 1984) as a measure that contained three items from the LMX-5 (i.e., two leader

95

Leader–Member Exchange (LMX) Measurement

The Vertical Dyad Linkage approach is formally introduced and contrasted with the traditional leadership styles approach (Dansereau, Cashman, & Graen, 1973; Dansereau, Graen, & Haga, 1975; Graen, Dansereau, & Minami, 1972a, 1972b)

Dansereau and colleagues (Dansereau et al., 1973; Graen et al., 1972a, 1972b) use scores on the Leader Behavior Description Questionnaire (Stogdill & Coons, 1957) to investigate Vertical Dyad Linkage Dansereau et al. (1975) create the 2item Negotiating Latitude measure

Item intercorrelations, test-retest reliability

Graen & Cashman (1975) create the Negotiating Latitude/LMX-4 by adding 2 items to Dansereau et al.’s (1975) Negotiating Latitude measure

Internal consistency, Heise test-retest stability coefficients, MTMM analysis

Liden & Graen (1980) introduce a 4point response scale for the LMX-4

Leader-member agreement, test-retest reliability

1980

Graen (1976) proposes a theory of LMX development, including role-taking, role-making, and roleroutinization Graen & Schiemann (1978) relabel the construct LMX, which includes “reciprocal influence, extra-contractual behavior exchange, mutual trust, respect and liking, and common fate” (p. 206)

Methodology Used in LMX Measurement

1970

LMX Measurement

1975

LMX Theory

Graen, Liden, & Hoel (1982) add a “centroid” item (“How would you characterize your working relationship with your supervisor?”) to the LMX-4 to create the LMX-5 Graen, Novak, & Sommerkamp (1982) add items to Liden & Graen (1980), to get LMX-7 (items not reported)

Dienesch & Liden (1986) propose 3 dimensions of LMX: affect, contribution, and loyalty

1985

Scandura & Graen (1984) report the LMX-7, which includes 3 items from the LMX-5. Graen & Scandura (1985) add 10 items to Scandura & Graen’s (1984) LMX-7 measure to create the LMX-17. Wakabayashi, Graen, Graen, & Graen (1988) introduce a 12-item LMX scale (written in Japanese)

Cronbach’s alpha, testretest reliability

Kozlowski & Doherty (1989) introduce an 8-item measure of Information Exchange

Cronbach’s alpha, itemtotal correlations, PCA, convergent validity with LMX-7

Seers (1989) develops a 10 item measure of team-member exchange (TMX)

Fig. 2.

Cronbach’s alpha, testretest reliability

LMX Measurement Timeline.

96

DANA L. JOSEPH ET AL.

LMX Measurement

1990

LMX Theory

Wayne & Ferris (1990) re-word Scandura & Graen’s (1984) LMX-7 to accommodate a 5-point disagree-agree response scale

Methodology Used in LMX Measurement

Cronbach’s alpha

1995

Wakabayashi, Graen, & Uhl-Bien (1990) add 2 items to Wakabayashi et al.’s (1988) 12 item LMX measure (written in Japanese)

Gerstner & Day (199) metaanalytically show low agreement between leader and member perceptions of LMX, implying two distinct LMX constructs Wayne, Shore, & Liden (1997) theoretically integrate LMX theory with social exchange theory

Schriesheim, Neider, Scandura, &Tepper (1992) introduce the LMX-6, which measures Dienesch & Liden’s (1986) LMX dimensions of contribution, loyalty, and affect. [The 2 items measuring affect were taken from the Satisfaction with Supervisor subscale of the Minnesota Satisfaction Questionnaire (Weiss, Dawis, England, & Lofquist, 1967).]

Multiple samples, CFA, Cronbach’s alpha, testretest reliability, criterionrelated validity, convergent validity with LMX-7, item intercorrelations

Liden, Wayne, & Stilwell (1993) reword Scandura & Graen’s (1984) LMX-7 to accommodate a 7-point disagree-agree response scale, and also create a leader-LMX scale which includes “mirror” items.

Cronbach’s al pha, PCA

Graen &Uhl-Bien (1995) modify the wording of Scandura & Graen’s (1984) LMX-7 and change the response scale to a 5-point Likert

Bauer and Green (1996) split the first item of Liden et al.’s (1993) version of the LMX-7 into 2 items to create the LMX-8 Bhal & Ansari (1996) develop the 10 item Quality of Interaction (QI) scale, which is designed to measure Dienesch & Liden’s (1986) LMX dimensions of contribution, loyalty, and affect.

Fig. 2.

(Continued)

Cronbach’s alpha Item generation, content validity analysis, multiple samples, convergent validity with LMX-5 and Negotiating Latitude, PCA, Cronbach’s alpha, item intercorrelations, item-total correlations, criterion-related validity, WABA

97

Leader–Member Exchange (LMX) Measurement

LMX Theory

LMX Measurement

Liden & Maslyn (1998) add professional respect to Dienesch & Liden’s (1986) multidimensional LMX theory

2000

Schriesheim et al. (1999) review the LMX literature to-date and highlight the lack of consensus in LMX theory, measurement, and analyses

Liden & Maslyn (1998) create a 12item, multidimensional measure of LMX (LMX-MDM) that measures the dimensions of affect, loyalty, contribution, and professional respect

Keller & Dansereau (2001) propose Negotiating Latitude is a separate construct, and not part of LMX

Keller &Dansereau (2001) create the LMX-3 by deleting the Negotiating Latitude items from the LMX-5

Methodology Used in LMX Measurement Item generation, content validity analysis, multiple samples, convergent validity with LMX-7, PCA, CFA, Cronbach’s alpha, test-retest reliability, correlations with social desirability, item intercorrelations, criterion-related validity, discriminant validity Compared LMX-3, LMX4, LMX-5 on leadermember agreement, Cronbach’s alpha, and criterion-related validity

CFA, Cronbach’s alpha

Maslyn & Uhl-Bien (2001) develop a “mirror version” of Graen & UhlBien’s (1995) LMX-7 scale to assess supervisor perceptions of members.

Cronbach’s alpha, EFA, leader-member agreement

Raabe & Beehr (2003) re-word the LMX-MDM to create a measure of coworker-member exchange (CMX)

Cronbach’s alpha

2005

Wang, Law, Wang, & Chen (2001; see also Wang, Law, & Chen, 2008) modify the LMX-MDM scale to include 4 items per dimension (16 items total)

The construct of LMX differentiation, or the withingroup variance on LMX, is presented (Ford & Seers, 2006; Liden, Erdogan, Wayne, & Sparrowe, 2006; Schyns, 2006)

Greguras & Ford (2006) modify the LMX-MDM and LMX-7 scales to create 2 new scales ( SLMX-MDM, SLMX-7) that are designed to assess supervisor perceptions of LMX

Fig. 2.

(Continued)

Cronbach’s alpha, convergent validity with LMX-MDM and LMX-7, criterion-related validity, CFA, hierarchical CFA

98

DANA L. JOSEPH ET AL.

LMX Theory

LMX Measurement

Cogliser, Schriesheim, Scandura, & Gardner (2009) introduce the idea of LMX balance between supervisor and subordinate

Sluss, Klimchak, & Holmes (2008) reduce the LMX-MDM to 8 items

Cronbach’s alpha

Harris, Harris, & Harvey (2008) reduce the LMX-MDM to 3 items

Cronbach’s alpha

Cronbach’s alpha

2010

Sin, Nahrgang, & Morgeson (2009) attempt to explain low leader-member agreement on LMX, concluding low agreement is consistent with LMX theory

Bernerth et al. (2007) develop a measure of leader -member social exchange ( LMSX) to assess two-way exchange in the leader-member dyad

Item generation, content validity analysis, EFA, measurement equivalence, multiple samples, CFA, convergent validity with LMX-7 and LMX-MDM,

Mardanov, Heischmidt, & Henson (2008) add a 13th item (“Overall, I’m very satisfied with my supervisor”) to the LMX-MDM

Ferris et al. (2009) introduce a multidimensional theory of dyadic work relationships that includes dimensions of trust, support, affect, loyalty, accountability, instrumentality, respect, and flexibility

Methodology Used in LMX Measurement

Vidyarthi, Liden, Anand, Erdogan, & Ghosh (2010) introduce measures of LMX social comparison (LMXSC) and relative LMX (RLMX)

Fig. 2.

Multiple samples, PCA, CFA, Cronbach’s alpha, incremental validity

(Continued)

helpfulness items and one relationship effectiveness item were retained) and four additional items, including one item each about the leader’s provision of performance feedback, leader understanding of follower needs, leader recognition of follower potential, and follower conﬁdence in leader decisions (see Fig. 2, and LMX-7 items in Table 1). The LMX-7 then became the most-used measure of LMX (Gerstner & Day, 1997; Schriesheim et al., 1999). The LMX-7 has undergone slight modiﬁcations, such as altering the

.73

.81

.77

.76

Leader understanding

Leader recognizes your potential

Leader help

Factor Loadings from Collins, (2007) (N ¼ 1,016)

Leader feedback

Item Content

.70

.70

.60

.50

Factor Loadings from Schriesheim and Cogliser (2009) (EFA; N ¼ 379)

.70

.78

.47

.54

Factor Loadings from Schriesheim and Cogliser (2009) (CFA; N ¼ 350)

LMX-7 Items (Scandura & Graen, 1984) and Illustrative Factor Loadings.

1. Do you usually feel you know where you stand y Do you usually know how satisﬁed your supervisor is with what you do? (1) Never know where I stand, (2) seldom know where I stand, (3) usually know where I stand, (4) always know where I stand 2. How well do you feel that your immediate supervisor understands your problems and needs? (1) Not at all, (2) some but not enough, (3) as much as the next person, (4) fully 3. How well do you feel that your immediate supervisor recognizes your potential? (1) Not at all, (2) some but not enough, (3) as much as the next person, (4) fully 4. Regardless of how much formal authority your immediate supervisor has built into his or her position, what are the chances that he or she would be personally inclined to use power to help you solve problems in your work? (1) No chance, (2) might or might not, (3) probably would, (4) certainly would

LMX-7 Item

Table 1.

Leader–Member Exchange (LMX) Measurement 99

.69

.63

.72

.72

Leader decisionmaking/member loyalty

Leader–member relationship, ‘‘centroid item’’

.72

Factor Loadings from Schriesheim and Cogliser (2009) (EFA; N ¼ 379)

.67

Factor Loadings from Collins, (2007) (N ¼ 1,016)

Leader help

Item Content

.60

.80

.76

Factor Loadings from Schriesheim and Cogliser (2009) (CFA; N ¼ 350)

Notes: Schriesheim and Cogliser (2009) CFA results are obtained from a model that allows items 2 and 7 to double-load onto a Satisfaction with Supervisor factor. Collins (2007) loadings are obtained from a larger CFA model.

5. Again, regardless of the amount of formal authority your immediate supervisor has, to what extent can you count on him or her to ‘‘bail you out’’ at his or her expense when you really need it? (1) No chance, (2) might or might not, (3) probably would, (4) certainly would 6. I have enough conﬁdence in my immediate supervisor that I would defend and justify his or her decisions if he or she were not present to do so. (1) Probably not, (2) maybe, (3) probably would, (4) certainly would 7. How would you characterize your working relationship with your immediate supervisor? (1) Less than average, (2) about average, (3) better than average, (4) extremely effective

LMX-7 Item

Table 1. (Continued )

100 DANA L. JOSEPH ET AL.

Leader–Member Exchange (LMX) Measurement

101

response scale to a Likert format, and occasionally revising item 1 (which has unusual wording) to split it into two items (see Fig. 2 timeline). In the fourth and ﬁnal stage of LMX measurement, the LMX-MDM (Liden & Maslyn, 1998; items in Table 2) rose to prominence as an alternative measure of LMX, second only to the LMX-7 in popularity (see Fig. 1). The ancestry of present-day LMX measurement, advancing from the LBDQ to the LMX-7 to the LMX-MDM, may appear haphazard (see Fig. 2 timeline). Indeed, the use of different measures to asses the same construct can, at times, greatly impede scientiﬁc progress. But at the bottom line, one may wonder just exactly how different the LMX-7 actually is from its predecessor, the LBDQ, or from the more recent LMX-MDM instrument. Schriesheim and Cogliser (2009) have recently reported that the correlation between the LMX-7 and distantﬃ precursor LBDQ (consideration facet) is pits ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ r ¼ :75 ½rcorrected ¼ :75= ð:89Þð:85Þ ¼ :86. This implies the LMX-7 is in large part a re-invention of the LBDQ consideration facet from the Ohio State studies (measured at the individual member-level). Furthermore (as we review below), the LMX-7 also appears to overlap very strongly with the LMXMDM measure (i.e., they have a near-perfect correlation). In light of this evidence, it now seems that the progression of LMX research (Fig. 2) starting with the LBDQ (Stogdill, 1963) and transitioning to the LMX-7 (Scandura & Graen, 1984) and the LMX-MDM (Liden & Maslyn, 1998), in retrospect, may have been largely a process of ‘‘changing the nametags’’ of leader consideration research, rather than actual evolution of the LMX construct.

Comparing the LMX-7 to the LMX-MDM In light of Fig. 1, the problems with nonstandard LMX measurement decried by Schriesheim and colleagues (1999) appear to have now been resolved through the adoption of two consensus measures of LMX (Fig. 1 and Tables 1 and 2). However, is having two dominant measures of LMX one too many? In general, having two measures that carry the same construct label can engender confusion and yield noncomparable results across studies. For instance, if two studies conducted in two different contexts (e.g., a blue collar and a white collar context) report different results for the effects of LMX, yet the two studies employed two different measures of LMX, then the variation in results across studies has two, confounded explanations: (a) the blue-collar vs. white-collar context, and (b) the choice of LMX measure. Comparing two different results might be more straightforward if both studies had used the same LMX measure.

.91 .80

.72 .91

.74

.70

.83

.79 .35

.69

.87

Factor Loadings from 4-Factor Liden and Maslyn (1998) (PCA; N ¼ 302)

.90

Factor Loadings from 1-Factor Eisenberger et al. (2010) (CFA; N ¼ 251)

.86

.78

.55

.88

.86

.85

Factor Loadings from 4-Factor Liden and Maslyn (1998) (CFA; N ¼ 249)

LMX-MDM Items (Liden & Maslyn, 1998) and Illustrative Factor Loadings.

1. I like my supervisor very much as a person. (affect 1) 2. My supervisor is the kind of person one would like to have as a friend. (affect 2) 3. My supervisor is a lot of fun to work with. (affect 3) 4. My supervisor defends my work actions to a superior, even without complete knowledge of the issue in question. (loyalty 1) 5. My supervisor would come to my defense if I were ‘‘attacked’’ by others. (loyalty 2) 6. My supervisor would defend me to others in the organization if I made an honest mistake. (loyalty 3)

LMX-MDM Item

Table 2.

102 DANA L. JOSEPH ET AL.

.86

.81 (item revised)

– .97

.91

.79

.28

.47

.78 .85

.88

.77

.87

.87

.85

–

.55 (item revised)

.74

Notes: Eisenberger et al. (2010) CFA results are obtained from a larger model that included other factors (contribution item loadings are likely small because the larger CFA model included extrarole performance). Liden and Maslyn (1998) loadings are from 11-item version of the LMX-MDM, before the contribution scale was expanded to three items.

7. I do work for my supervisor that goes beyond what is speciﬁed in my job description. (contribution 1) 8. I am willing to apply extra efforts, beyond those normally required, to meet my supervisor’s work goals. (contribution 2) 9. I do not mind working my hardest for my supervisor. (contribution 3) 10. I am impressed with my supervisor’s knowledge of his or her job. (professional respect 1) 11. I respect my supervisor’s knowledge of and competence on the job. (professional respect 2) 12. I admire my supervisor’s professional skills. (professional respect 3)

Leader–Member Exchange (LMX) Measurement 103

104

DANA L. JOSEPH ET AL.

On the contrary, the magnitude of this problem (using two measures of the same construct) is mitigated to the extent that the two ‘‘different’’ measures of LMX are correlated close to unity. That is, if the LMX-7 and the LMX-MDM are correlated close to r ¼ 1.0, then these two measures can be treated as alternate forms of the same instrument. [That is, the correlation between two measures of LMX is a type of reliability index (Gulliksen, 1950).] To assess this possibility, we searched the LMX literature for studies that employed both the LMX-7 and the LMX-MDM and were able to locate two such reports. Maslyn and Uhl-Bien (2001) reported an observed correlation between LMXpﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ﬃ 7 and the LMX-MDM total of r ¼ :80 ½rcorrected ¼ :80= ð:90Þð:92Þ ¼ :88; N ¼ 182. Greguras and Ford (2006) replicated this estimate with an observed correlation p between LMX-7 and the LMX-MDM total of ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ﬃ r ¼ :77 ½rcorrected ¼ :77= ð:90Þð:92Þ ¼ :85; N ¼ 420. In other words, the LMX-7 instrument (Scandura & Graen, 1984) and the composite total scores of the LMX-MDM instrument (Liden & Maslyn, 1998) appear to be alternate forms of the same test, based on their latent correlation of r ¼ .8–.9. [To restate this conclusion, we point out that according to Fornell and Larcker’s (1981, p. 46) test of discriminant validity, two measures assess the same construct when their latent correlation is stronger than the square root of the average indicator (observed measure) variance extracted by the latent constructs. Using the LMX factor loadings reported in the ﬁrst columns of Tables 1 and 2, Fornell and Larcker’s (1981) critical discriminant validity correlation is .74. Thus, because the latent correlation between LMX-7 and LMX-MDM is larger than .74, we conclude there is no evidence for discriminant validity between the LMX-7 and the LMX-MDM, and excellent convergent validity between the two.] Noting that the LMX-7 and the LMX-MDM may be largely interchangeable (rcorrected X .85) and these are the two dominant measures of LMX (Fig. 1), it appears that the ﬁeld of LMX research has now reached excellent consensus in how the construct is measured [at least during the decade following Schriesheim et al.’s (1999) review]. We next turn to the question: What is this construct that LMX researchers have been measuring?

CONSTRUCT VALIDITY AND HIERARCHICAL CONFIRMATORY FACTOR ANALYSIS Construct Validity Given the summary of current LMX measurement choices presented in Fig. 1, we must now ask: What construct (or mix of constructs) do the LMX-7

Leader–Member Exchange (LMX) Measurement

105

and the LMX-MDM actually measure? This is the question of construct validity. The notion of construct validity involves the demonstration that measures of a concept occupy expected positions within the nomological network of theoretically related concepts and operations (Cronbach & Meehl, 1955). Construct validity evidence involves both convergent validity (i.e., when multiple measures of the same concept are strongly correlated) and discriminant validity (i.e., when multiple measures of different concepts are not too strongly correlated; Campbell & Fiske, 1959; see also Edwards, 2003, for a review of factor-analytic methods for establishing construct validity). In the language of factor analysis, convergent validity is obtained when two indicators (observed measures) of the same construct exhibit large factor loadings onto their intended latent factor, whereas discriminant validity is obtained when two latent factors are intercorrelated less than perfectly (Widaman, 1985; Fornell & Larcker, 1981).

Hierarchical CFA In discussing the construct validity of LMX measurement, this chapter emphasizes a technique known as hierarchical conﬁrmatory factor analysis (hierarchical CFA). This technique allows the analyst to estimate a factor model that operates simultaneously at multiple levels of abstraction. An example of a hierarchical CFA model appears in Fig. 3. As seen in Fig. 3, the Liden and Maslyn (1998) LMX-MDM measure can be modeled hierarchically – including a set of four lower-order factors (affect, loyalty, contribution, and respect), which simultaneously reﬂect (or load onto) a single, higher-order factor of overall LMX. We note that Fig. 3 speciﬁes LMX at three, distinct levels of generality (i.e., three measurement strata): (a) the lowest level of generality, with 12 speciﬁc LMX-MDM items that serve as concrete, observable instantiations of LMX, (b) the middle stratum, which is occupied by four, lower-order LMX latent constructs, and (c) the highest level of generality, which holds a single, broad factor of LMX. The idea of modeling constructs hierarchically has been exempliﬁed in research on mental abilities (Carroll, 1993; Johnson & Bouchard, 2005), personality (Digman, 1997; Judge, Locke, & Durham, 1997), withdrawal behavior (Hanisch, Hulin, & Roznowski, 1998), job attitudes (Harrison, Newman, & Roth, 2006; Newman, Joseph, & Hulin, 2010), and work performance (Fisher, 1980; Harrison et al., 2006; Viswesvaran, Schmidt, & Ones, 2005). The question of whether researchers should focus their attention on constructs at a broad, higher-level of abstraction vs. a narrow, lower- or

106

DANA L. JOSEPH ET AL.

LMX (higher-order factor)

.86

.80

Affect

A1

A2

.67

Loyalty

A3

L1

L2

.77

Contribution

L3

C1

C2

C3

Respect

R1

R2

R3

Fig. 3. Hierarchical Factor Model of LMX Based Upon the LMX-MDM Measure (Liden & Maslyn, 1998). Note: Second-order factor loadings are based on a metaanalysis of the LMX-MDM facets, reported in Table 3.

facet-level of abstraction, has been contentious. Cronbach (1956) described psychologists as falling into two camps: lumpers (who prefer to combine psychological variables and treat them as indicating broad constructs), and splitters (who prefer to identify ﬁne distinctions between increasingly speciﬁc, narrow constructs). Valuable discussions of the empirical, theoretical, and analytical implications of these two approaches to construct conceptualization have been provided by Ones and Viswesvaran (1996), Johns (1998), Edwards (2001), and numerous others.

LMX Researchers as Lumpers In this chapter, we conclude that LMX has essentially been measured as a broad, higher-order construct (Fig. 3). That is, LMX researchers appear to be lumpers [i.e., both the LMX-7 and the LMX-MDM (Fig. 1) seem to measure the broad, higher-order LMX factor depicted at the top of Fig. 3]. We present evidence of LMX lumperism in three forms, all related to the use of the LMX-MDM instrument (Liden & Maslyn, 1998). First, we note that

107

Leader–Member Exchange (LMX) Measurement

Table 3. Meta-Analytic Correlation Matrix among the Liden and Maslyn (1998) LMX-MDM Subscales, and Loadings onto the HigherOrder LMX Factor. LMX Facet 1. 2. 3. 4.

Affect Loyalty Contribution Respect

1

2

1.00 .68 .58 .67

1.00 .56 .61

3

Loadings onto LMX Factor

CFA Fit Indices

1.00 .49

.86 .80 .67 .77

RMSEA ¼ .053 TLI ¼ .99 CFI ¼ 1.00 SRMR ¼ .012

Notes: N ¼ 1,358; k ¼ 6 samples; all samples involved English translations of the LMXMDM scale; all meta-analytic correlations are sample size-weighted average Pearson’s r. For the single-factor conﬁrmatory factor analysis (CFA) based on the correlation matrix, w2(df) ¼ 9.62 (2).

several LMX researchers have explicitly found adequate empirical ﬁt for hierarchical CFA models of the LMX-MDM measure [i.e., a model with four lower-order LMX factors and one higher-order LMX factor (Fig. 3); Erdogan & Enders, 2007; Erdogan & Liden, 2006; Erdogan, Kraimer, & Liden, 2004; Erdogan, Liden, & Kraimer, 2006; Greguras & Ford, 2006; Sparrowe, Soetjipto, & Kraimer, 2006; Wang, Law, Hackett, Wang, & Chen, 2005]. Second, we observe an overwhelming tendency for LMX researchers to analyze LMX as a single factor, even when they are using Liden and Maslyn’s (1998) ostensibly multidimensional LMX-MDM scale. To quantify this assertion, we conducted a survey of the LMX-MDM literature in PsycINFO (1887 – December 2010). Of 147 papers that cited Liden and Maslyn (1998), 23 were book chapters or journal articles that were not accessible or written in a foreign language. Of the remaining 124 studies, 46 used Liden and Maslyn’s scale to measure LMX (this does not include two studies that each employed only one dimension of the LMXMDM scale). From these 46 empirical studies that used the LMX-MDM scale, 33 treated it exclusively as a single factor (i.e., they conducted analyses based upon the LMX-MDM composite total scores only), 11 treated it exclusively as four factors, two studies treated it both as one factor and as four factors, and two studies conducted only item-level analyses. The notable result is that 33 of 46, or 72% (nearly three-quarters) of empirical studies using the multidimensional LMX-MDM measure (Liden & Maslyn, 1998) treated it exclusively as a single factor in the analyses. Third and ﬁnally, we conducted our own meta-analysis of the facet-level correlations

108

DANA L. JOSEPH ET AL.

among the lower-order LMX-MDM factors (i.e., affect, loyalty, contribution, and respect; Liden & Maslyn, 1998; see Table 2), to estimate a metaanalytic CFA to conﬁrm the broad, higher-order LMX factor. Of the 46 empirical studies that used the LMX-MDM instrument (Liden & Maslyn, 1998), only 10 reported the matrix of correlations among the four LMXMDM subscales. In an attempt to summarize these LMX lower-order facet correlations, we meta-analyzed the correlation matrices from 6 of the 10 studies (the four studies that were excluded involved non-English translations of the LMX-MDM [i.e., Vandenberghe, Bentein, & Stinglhamber, 2004 (French); Schyns & Wolfram, 2008 (German), and Wang, Law, & Chen, 2008 (Chinese)], and a nonindependent sample [i.e., Lee, 2005; Lee, 2008]). In sum, the total sample size across the six meta-analyzed samples was 1,358, and the sample-weighted average Pearson correlation coefﬁcients appear in Table 3. As seen in Table 3, all the lower-order facets from the Liden–Maslyn LMX-MDM appear to be strongly intercorrelated, and a CFA of this meta-analytic facet-level correlation matrix conﬁrms a onefactor solution (RMSEA ¼ .053; CFI ¼ 1.00). In short, the LMX-MDM, despite being labeled ‘‘multidimensional,’’ reﬂects a very prominent higherorder factor, and 72% of studies that employ this measure have used it exclusively to assess the higher-order factor (essentially ignoring the lowerorder structure of LMX). On the basis of these results, it appears LMX researchers are indeed lumpers. They tend to focus overwhelmingly on the broad, higher-order factor of LMX and to ignore lower-order dimensions. To summarize, we ﬁrst noted 66% of LMX studies have relied on the LMX-7 measure (Fig. 1), which is strictly unidimensional. Second, from the 19% of LMX studies that have used the multidimensional LMX-MDM measure (Liden & Maslyn, 1998), we found that three-quarters of these ignore the multidimensional nature of the construct. Finally, we remind the reader of the above-reported disattenuated correlations between LMX-7 and LMX-MDM of rcorrected ¼ .85 and .88. In sum, both the LMX-MDM and the LMX-7 appear to measure the same concept, and that concept is the broad, higherorder factor of LMX. So, unlike the conclusions from a previous review of LMX research (Schriesheim et al., 1999), we conclude here that there is a strong consensus in LMX measurement – 85% of researchers are using either the LMX-7 or the LMX-MDM, and 94% of these are treating LMX as a single, broad dimension [i.e., (66% LMX-7 þ 19% LMX-MDM72% treating LMX-MDM as unidimensional)/(66% LMX-7 þ 19% LMXMDM) ¼ 94% of LMX-7 and LMX-MDM studies are treating LMX as unidimensional].

Leader–Member Exchange (LMX) Measurement

109

Content Validity At this point, we direct the reader’s attention to the content validity of the two dominant LMX measures. Content validity (which is an aspect of construct validity; Landy, 1986) refers to the subjective ‘‘adequacy with which a measure [is judged to assess] the domain of interest’’ (Hinkin, 1995, p. 968; Schriesheim et al., 1993). Tables 1 and 2 present the items from the LMX-7 (Scandura & Graen, 1984) and the LMX-MDM (Liden & Maslyn, 1998), respectively. As noted by Schriesheim and Cogliser (2009), the LMX7 appears to reﬂect supportive leadership (items 2 through 5) and performance feedback (item 1). This subjective observation is further conﬁrmed by Schriesheim and Cogliser’s (2009) demonstration that the LMX-7 correlates r ¼ .74 with Schriesheim’s (1978) leader support scale, and r ¼ .78 with House and Dessler’s (1974) supportive leadership scale. So, the LMX-7 appears to primarily measure supportive leadership (see the items in Table 1). It is also worth mentioning that item 7 of the LMX-7 – the so-called centroid item (‘‘characterize your working relationship with your immediate supervisor y,’’ Graen, Liden et al., 1982) – is the only item in the LMX-7 that uses an explicitly dyad-level referent (i.e., it assesses the ‘‘relationship’’; see Chan, 1998). The other LMX-7 items use the supervisor as the referent, rather than the dyadic relationship. Furthermore, by inspection of the factor loadings in Table 1, it does not appear that the centroid item is empirically more ‘‘central’’ than the other LMX-7 items. If anything, either the LMX-7 items appear to be essentially tau-equivalent (i.e., all have similar loadings) or the leader support items (items 2 through 5) appear to be most central, rather than the centroid item. In sum, whereas the leader–member relationship may be theoretically key to the deﬁnition of LMX, the LMX7 items appear to over-sample leader supportiveness from the content domain and focus very little on the dyadic relationship [cf. Bernerth et al.’s (2007) LMX measure, which focuses on the two-way exchange relationship in LMX]. Regarding the content validity of the LMX-MDM instrument (Liden & Maslyn, 1998), we provide the 12 LMX-MDM items in Table 2. In contrast to the LMX-7, when the LMX-MDM was developed, Liden and Maslyn (1998) followed content validation procedures described by Hinkin (1995, 1998). In Liden and Maslyn’s (1998) content analysis, two sets of content experts attempted to classify 120 LMX items along an a priori set of construct deﬁnitions, resulting in modiﬁcation of the construct deﬁnitions and the retention of only 38 items from the initial item pool (items were

110

DANA L. JOSEPH ET AL.

retained if ﬁve out of six judges agreed that an item belonged to a particular dimension). These content analysis procedures ultimately yielded a set of items that performed well in subsequent factor analyses. Our only reservation about the content/construct validity of the LMX-MDM involves the contribution dimension, which appears to assess organizational citizenship behavior (OCB) enacted in service of the supervisor’s goals (e.g., item 8: ‘‘I am willing to apply extra efforts, beyond those normally required, to meet my supervisor’s work goals.’’). Relatively poor psychometric performance of the LMX contribution items in the presence of OCB items (Eisenberger et al., 2010; see Table 2) afﬁrms our concern. Next, we look at content validity of the LMX-7 and the LMX-MDM together, under the perspective that these two instruments sample the same, broadly deﬁned LMX content domain. Recalling that the composite scores of the LMX-7 and the LMX-MDM correlate at rcorrectedZ.85 (see above), we now compare the items in Tables 1 and 2. The interesting conclusion from comparing Tables 1 and 2 is that LMX-7 items (Table 1) and LMXMDM items (Table 2) do not appear very similar, in terms of their speciﬁc content. For example, the LMX-MDM items (Table 2) appear to measure the following perceptions about a supervisor: (a) I like her (affect), (b) she defends me (loyalty), (c) I would work extra hard for her (contribution), and (d) I like her competence (professional respect). The only obvious overlaps with the LMX-7 items are between the LMX-MDM loyalty dimension (she defends me) and LMX-7 item 5 (‘‘count on him or her to ‘bail you out’’’), and between the LMX-MDM respect dimension (I like her competence) and LMX-7 item 6 (‘‘I have enough conﬁdence in my immediate supervisor that I would defend and justify his or her decisions y ’’). Otherwise, the individual LMX-7 items and LMX-MDM items do not appear to be the same (cf. Bernerth et al., 2007). Interestingly, we assert that this state of affairs (i.e., near-perfect correlation between two instruments, but very little overlap between their individual items) is normal, at least when measuring broadly deﬁned concepts. That is, the broader the content domain, the more likely it is that two, parallel measures of that domain will contain non-content-overlapping individual items. This occurs because broader content domains are reﬂected in increasingly many plausible manifestations. For example, imagine we are measuring the narrow personality content domain of assertiveness. If we randomly sample two sets of seven trait adjectives each from this domain, then when we compare the two lists of seven adjectives, several of these sampled adjectives are likely to be synonyms between the two lists. On the contrary, if we randomly sample from the much broader content domain of

Leader–Member Exchange (LMX) Measurement

111

extraversion (which includes assertiveness, warmth, gregariousness, positive emotion, and activity), we are much more likely to obtain two lists of seven trait adjectives for which very few synonyms can be found between the two lists. (Despite having very few synonymous items across lists, the correlation between composite scores based on the two lists is still very high, because both seven-adjective composites are representative reﬂections of the broad domain.) This example is similar to measuring the broad LMX construct (Fig. 3), by content sampling seven items to compose the LMX-7 and 12 items to compose the LMX-MDM. Because the content domain is broad (as opposed to narrow and speciﬁc), it is increasingly possible to develop two good, brief measures of the construct for which the item contents barely overlap. This point can be subtle, but it explains why subjective content validity analyses may fail to reveal measurement equivalence at the same time that convergent validity methods (Campbell & Fiske, 1959) demonstrate strong evidence of equivalence across measures. In summary, the LMX-7 (Scandura & Graen, 1984) and the LMX-MDM (Liden & Maslyn, 1998) both measure the broad, higher-order LMX construct (Fig. 3), but these two measures sample different items from the broad content domain when doing so.

MEMBER LMX VERSUS LEADER LMX: CONSIDERING BOTH SIDES Small Correlation between Leader LMX and Member LMX We have shown that LMX is broadly measured (i.e., it can be speciﬁed as a single, higher-order factor) and that the two dominant methods of measuring the broad LMX construct (LMX-7 and LMX-MDM) are roughly equivalent (i.e., the two measures of broad LMX correlate rcorrected ¼ .85–.88). In this section, we now turn to address a central – but often hidden – aspect of LMX conceptualization and measurement: Leader LMX is not the same as Member LMX. This key observation was highlighted by Gerstner and Day’s (1997) meta-analysis of the LMX literature, in which they reported an average correlation of r ¼ .29 (rcorrected ¼ .37) between leader perceptions of LMX and member perceptions of LMX (k ¼ 24 studies, N ¼ 3,460). A correlation of this modest magnitude suggests strong discriminant validity between Leader LMX and Member LMX – meaning that when a leader rates LMX with the member,

112

DANA L. JOSEPH ET AL.

s/he is rating a different construct from when the member rates LMX with the leader. More recently, Sin et al. (2009) replicated Gerstner and Day’s (1997) meta-analytic ﬁnding (k ¼ 64 studies, N ¼ 10,884; leader–member r ¼ .32; rcorrected ¼ .37), and went further to demonstrate that leader–member agreement gets stronger with longer relationship tenure and increased dyadic interaction at work. Before moving further, we feel an important distinction must be drawn. The respective roles of leader and member in the measurement of LMX can be arrayed in a 2 2 matrix, with two axes: (a) leader-reported vs. member-reported LMX, and (b) leader provisions vs. member provisions (Fig. 4). To clarify, leader-reported implies that the LMX survey was ﬁlled in by the leader, while member-reported implies the survey was ﬁlled in by the member. Leader provisions indicates that the leader is offering a contribution (in the sense of social exchange), to the presumed beneﬁt of the member, while member provisions involve a contribution offered by the member. The meaning of leader–member LMX agreement depends on which quadrants of Fig. 4 are being compared. Two sorts of leader–member LMX comparisons are most prevalent: (a) cross-report LMX comparisons – for example, member-report of leader provisions vs. leader-report of leader provisions (Fig. 4, quadrant I vs. quadrant II, or quadrant III vs. quadrant IV); and (b) cross-diagonal LMX comparisons – for example, member-report of member provisions vs. leader-report of leader provisions (Fig. 4, quadrant I vs. quadrant IV, or quadrant II vs. quadrant III). These distinctions have been noted by past researchers; Sin et al. (2009, p. 1052) meta-analyzed both cross-report LMX agreement rcorrected ¼ .38 (k ¼ 34 studies; N ¼ 5,551) and cross-diagonal LMX agreement rcorrected ¼ .38

Leader Provisions

Member Provisions

Leader Reported I. Leader Report of Leader Provisions (e.g., I do not mind working my hardest for my subordinate.)

Member Reported II. Member Report of Leader Provisions (e.g., My supervisor does not mind working her/his hardest for me.)

III. Leader Report of Member Provisions (e.g., My subordinate does not mind working her/his hardest for me.)

IV. Member Report of Member Provisions (e.g., I do not mind working my hardest for my supervisor.)

Greguras & Ford (2006) compared leader reports of leader provisions vs. member reports of member provisions (Quadrant I vs. Quadrant IV; cross-diagonal comparison)

Sin, Nahrgang, & Morgeson (2009) compared leader vs. member reports of leader provisions (Quadrant I vs. Quadrant II; cross-report comparison)

Fig. 4.

Four Types of LMX Measures (reports provisions)

Leader–Member Exchange (LMX) Measurement

113

(k ¼ 23 studies; N ¼ 4,147), showing very similar magnitudes of agreement for the two types of leader–member comparisons shown in Fig. 4. We here notice that the LMX-7 and the LMX-MDM both draw items from multiple quadrants of Fig. 4. For the LMX-7 instrument in Table 1, items 1 through 5 belong in quadrant II of Fig. 4 (member-report of leader provisions), item 6 belongs in quadrant IV (member-report of member provisions), and item 7 does not belong in any quadrant (i.e., effectiveness of the dyadic relationship). For the LMX-MDM instrument (Liden & Maslyn, 1998; see Table 2), the affect, contribution, and respect items all belong in quadrant IV (member-report of member provisions), and the loyalty items belong in quadrant II (member-report of leader provisions). Ultimately, comparisons of LMX measures involve implicit use of the quadrants in Fig. 4, and comparisons of leader vs. member reports of LMX depend critically on the notions in Fig. 4. This issue arises in the following section, where we discuss a particular analysis that can be applied to multi-source (leader perspective and member perspective) LMX data.

Multitrait-Multimethod (MTMM) Analysis One interesting technique for looking at the construct validity of Leader LMX and Member LMX concepts is MTMM analysis ( Campbell & Fiske, 1959; Widaman, 1985). MTMM analyses can be performed whenever a set of constructs (i.e., ‘‘traits’’) is assessed through multiple instruments or rating sources (i.e., ‘‘methods’’). Campbell and Fiske (1959) discussed how monotrait-heteromethod correlations (i.e., when two methods are used to assess the same trait construct) could be used to help establish convergent validity, and comparisons within the MTMM correlation matrix could also be used to establish discriminant validity among the traits, buttressing the inference of construct validity. Widaman (1985) formalized Campbell and Fiske’s (1959) MTMM approach using CFA. To explain Widaman’s (1985) speciﬁcation, we refer to the MTMM-CFA model in Fig. 5a. Notice that this model is ﬁt to data from multiple traits (i.e., LMX affect, loyalty, contribution, and respect), each of which was measured via two methods (i.e., leader perspective and member perspective). To ﬁt the model, each measure is allowed to double-load – once onto a trait factor, and once onto a method factor. For example, leader-reported affect LMX double-loads onto an affect trait factor and also onto a leader method factor (Fig. 5a). A sequence of models can then be used to test whether the data best represent a model with four traits vs. one trait (discriminant validity), and whether

114

DANA L. JOSEPH ET AL.

Member Affect

.78 (.77)

Member Loyalty

.89 (.90)

Member Contribution

.54 (.54)

.25 Affect .25 .00

Member

Loyalty .00 .20 Contribution

Member Respect

Respect .20

.20 (.21)

Leader Affect

.86 (.85)

Leader Loyalty

.83 (.84)

.20 .20

.72 (.71)

Leader Leader Contribution Leader Respect

.73 (.79) .79 (.72)

Fig. 5a. Multitrait-Multimethod Analysis of the Liden and Maslyn (1998) LMXMDM Measure (Sin et al., 2009, data). Note: Parameter estimates are from model XIII, and from model IV (in parentheses). Results show two strong source factors (leader factor and member factor), but very weak (or nonexistent) trait dimension factors (affect, loyalty, contribution, and respect). For model IV, RMSEA ¼ .027, CFI ¼ .99, TLI ¼ .99, and SRMR ¼ .041. Additional model ﬁt indices appear in Table 4a.

trait measures from two different methods load onto the same trait factor (convergent validity). For other examples of MTMM-CFA, see Lim and Ployhart (2006) and Joseph and Newman (2010). To illustrate this technique, we ﬁt a series of MTMM models to two multisource LMX datasets: (a) the two-source (leader and member) crossreport data of LMX-MDM from Sin et al. (2009) (Fig. 4), and (b) the twosource (leader and member) cross-diagonal data of LMX-MDM from Greguras and Ford (2006) (Fig. 4). Results appear in Tables 4a and 4b. In short, after ﬁtting various MTMM models (e.g., 4 traits-2 methods, 4 traitsno methods, 1 trait-no methods, no traits-1 method, etc.), we found three

115

Leader–Member Exchange (LMX) Measurement

Table 4a. Multitrait-Multimethod Results for LMX-MDM (4 Facets), from Reanalyzing Correlation Matrix of Sin et al. (2009). MTMM Model Model I Model II Model III Model IV Model V Model VI Model VIIa Model VIII Model IX Model X Model XI Model XIIa Model XIIIa Model XIV Model XVb Model XVIa,b Model XVII

Trait Factors

Method Factors

w2

None None None None 1, general trait 1, general trait 1, general trait 1, general trait 1, general trait 4, uncorrelated 4, uncorrelated 4, uncorrelated 4, uncorrelated 4, correlated 4, correlated 4, correlated 4, correlated

None 1, general method 2, uncorrelated 2, correlated None 1, general method 1, leader method 2, uncorrelated 2, correlated None 1, general method 2, uncorrelated 2, correlated None 1, general method 2, uncorrelated 2, correlated

444.13 251.93 23.38 20.35 251.93 – 18.87 – – 409.60 – 21.41 18.56 193.51 12.22 18.00 –

df RMSEA (90% CI) TLI CFI SRMR

28 20 20 19 20 – 16 – – 24 – 17 16 18 10 14 –

.391 .346 .042 .027 .346 .043

.407 .052 .041 .317 .048 .054

(.360, (.308, (.000, (.000, (.308, – (.000, – – (.373, – (.000, (.000, (.278, (.000, (.000, –

.424) .385) .101) .095) .385) .108)

.442) .111) .106) .358) .126) .119)

.167 .352 .980 .987 .352 – .979 – – .046 – .978 .984 .435 .981 .977 –

.167 .537 .986 .991 .537 – .988 – – .182 – .986 .991 .637 .993 .988 –

.357 .242 .098 .041 .242 – .035 – – .353 – .092 .039 .200 .040 .105 –

Notes: Best-ﬁtting models are underlined. N ¼ 98 leader–member dyads from Sin et al. (2009, p. 1053). Several models from Widaman’s (1985) MTMM typology did not converge (see cells with ‘‘–’’), likely due to trying to ﬁt trait models onto a methods-only true structure. In Models X through XVII, the two trait loadings were constrained to equality for each LMX-MDM subdimension. a To achieve empirical identiﬁcation, the following post hoc constraints were required: Model VII – Method loadings onto member factor were constrained to zero; Models XII, XIII, & XVI – Trait loadings onto loyalty factor were constrained to zero; b Estimates include nonsense correlations outside the [1.0 to 1.0] range.

results: (a) a model with no traits (affect, loyalty, contribution, and respect) and two correlated methods (Leader LMX and Member LMX) provided excellent ﬁt to the Sin et al. data (RMSEA ¼ .027; CFI ¼ .99), (b) many of the more sophisticated models that speciﬁed trait factors had problems converging, unless one or more of the trait factors’ loadings were constrained to zero (Table 4a), and (c) when the models with trait loadings did converge, the loadings onto the trait factors were very small, whereas the loadings onto the method factors were large (Fig. 5a). These same three ﬁndings (from Sin et al.’s cross-report data) were then replicated in the analysis of Greguras and Ford’s (2006) cross-diagonal data (RMSEA ¼ .067; CFI ¼ 98; see Table 4b and Fig. 5b). Together, these results are consistent with the interpretation

116

DANA L. JOSEPH ET AL.

Table 4b. Multitrait-Multimethod Results for LMX-MDM (4 facets), from Reanalyzing Correlation Matrix of Greguras and Ford (2006). MTMM Model Model I Model II Model III Model IV Model V Model VI Model VIIa Model VIII Model IX Model X Model XI Model XIIa Model XIIIa Model XIVb Model XV Model XVIa,b Model XVII

Trait Factors

Method Factors

w2

None None None None 1, general trait 1, general trait 1, general trait 1, general trait 1, general trait 4, uncorrelated 4, uncorrelated 4, uncorrelated 4, uncorrelated 4, correlated 4, correlated 4, correlated 4, correlated

None 1, general method 2, uncorrelated 2, correlated None 1, general method 1, leader method 2, uncorrelated 2, correlated None 1, general method 2, uncorrelated 2, correlated None 1, general method 2, uncorrelated 2, correlated

2052.5 477.22 142.09 54.71 477.22 – 50.64 25.76 24.32 1394.2 – 136.62 49.91 387.19 – 134.21 –

df RMSEA (90% CI) TLI CFI SRMR

28 20 20 19 20 – 16 12 11 24 – 18 17 18 – 17 –

.414 .233 .120 .067 .233 .072 .052 .054 .368 .125 .068 .221 .128

(.399, (.215, (.102, (.047, (.215, – (.050, (.024, (.024, (.352, – (.106, (.047, (.202, – (.108, –

.430) .251) .139) .088) .251) .094) .080) .083) .385) .145) .090) .240) .149)

.331 .756 .907 .974 .756 – .970 .986 .985 .313 – .903 .974 .760 – .898 –

.331 .826 .934 .983 .826 – .983 .994 .994 .411 – .937 .984 .846 – .938 –

.368 .126 .194 .033 .126 – .031 .021 .023 .354 – .191 .032 .117 – .193 –

Notes: Best-ﬁtting models are underlined. N ¼ 422 leader-member dyads from Greguras and Ford (2006, p. 449). Several models from Widaman’s (1985) MTMM typology did not converge (see cells with ‘‘–’’), likely due to trying to ﬁt trait models onto a methods-only true structure. In Models X through XVII, the two trait loadings were constrained to equality for each LMXMDM subdimension. a To achieve empirical identiﬁcation, the following post hoc constraints were required: Model VII – Method loadings onto member factor were constrained to zero; Models XII, XIII, & XVI – Trait loadings onto loyalty and respect factors were constrained to zero; b Estimates include nonsense correlations outside the [1.0 to 1.0] range. Model VIII exhibits good ﬁt, but the loadings onto the leader factor are unusual (i.e., .61, .18, .26, and .03, respectively) and thus difﬁcult to interpret.

that leader and member responses to the LMX-MDM (Liden & Maslyn, 1998) do not reﬂect any content-speciﬁc LMX traits (affect, loyalty, contribution, and respect), but rather should be interpreted as reﬂecting two, distinct LMX source factors (Leader LMX and Member LMX). What do these MTMM results mean? When leaders and members each rate their LMX perceptions, it appears that they are not rating speciﬁc dyadic content (i.e., a leader’s report of leader respect LMX does not imply a member’s report of leader respect LMX; nor does it imply a member’s report of member respect LMX). Instead, Leader LMX and Member LMX emerge as two, distinct source constructs. Furthermore, both Leader LMX

117

Leader–Member Exchange (LMX) Measurement

Member Affect

.90 (.91)

Member Loyalty

.77 (.76)

Member Contribution

.63 (.63)

.22 Affect .22 .00

Member

Loyalty .00 .14 Contribution

Member Respect

Respect .00

.53 (.54)

Leader Affect

.73 (.74)

Leader Loyalty

.83 (.83)

.14 .00

.74 (.74)

Leader Leader Contribution Leader Respect

.63 (.63) .62 (.62)

Fig. 5b. Multitrait-Multimethod Analysis of the Liden and Maslyn (1998) LMXMDM Measure (Greguras & Ford, 2006, data). Note: Parameter estimates are from model XIII, and from model IV (in parentheses). Results show two strong source factors (leader factor and member factor), but very weak (or nonexistent) trait dimension factors (affect, loyalty, contribution, and respect). For model IV, RMSEA ¼ .067, CFI ¼ .98, TLI ¼ .97, and SRMR ¼ .033. Additional model ﬁt indices appear in Table 4b.

and Member LMX can be conceptualized at the highest level of abstraction: as the broad, higher-order LMX factor from Fig. 3. This result implies the need for two, separate positions in the nomological network (and perhaps two separate theories) to explain the antecedents and consequences of LMX – one for the Leader LMX construct and one for the Member LMX construct (see Gerstner & Day, 1997). Before going further, it is important for us to return to Fig. 4 to interpret the meaning of these particular MTMM data. The Sin et al. (2009) data involved cross-report comparisons [e.g., comparing leader-reported leader provisions (quadrant I) vs. member-reported leader provisions (quadrant II), where leader report and member report are the method factors], whereas

118

DANA L. JOSEPH ET AL.

the Greguras and Ford (2006) data involved cross-diagonal comparisons of the quadrants from Fig. 4 [e.g., comparing leader-reported leader provisions (quadrant I) vs. member-reported member provisions (quadrant IV)]. The type of comparison one conducts (i.e., leader vs. member reports of leader provisions, in contrast to leader reports of leader provisions vs. member reports of member provisions) has direct implications for the meaning of the MTMM ‘‘method’’ factors. When conducting cross-report comparisons (Fig. 4, Sin et al., looking at two different reports of the same provisions), then method factors refer to some aspects of the reporting process (e.g., rater’s opportunity to observe, rater bias). On the contrary, when applying MTMM analyses to cross-diagonal comparisons (i.e., Fig. 4, Greguras & Ford, looking at two different reports of different provisions), then the interpretation of the ‘‘method’’ factors is that they reﬂect a combination of different methods (reports) and different basic content (leader provisions vs. member provisions) – so the ‘‘methods’’ are somewhat confounded with the trait content in these data (e.g., Greguras & Ford, 2006). Thus, a much more informative design for future research will be to include measures of all four quadrants from Fig. 4, enabling researchers to fully partition reporting factors from provision factors. The interesting conclusion from the current data is that cross-report comparisons yield the same MTMM results as cross-diagonal comparisons: ‘‘Leader perceptions of Leader provisions’’ are distinct from both ‘‘Member perceptions of Leader provisions’’ and from ‘‘Member perceptions of Member provisions’’ (i.e., different rater implies different LMX construct). Another ﬁnding from the MTMM analysis with these data was that Leader LMX and Member LMX sources were correlated (r ¼ .21 in the Sin et al. data, and r ¼ .54 in the Greguras and Ford data; see Figs. 5a and 5b). Correlated source factors, in this case, index the extent to which the broad LMX factor from the leader’s perspective corresponds to the broad LMX factor from the member’s perspective. In the Sin et al. (cross-report) data, this method correlation is interpreted as perceptual similarity between leader and member (i.e., leader and member share the same frame of reference when rating Leader LMX provisions). In the Greguras and Ford (crossdiagonal) data, this method correlation can be interpreted as evidence of reciprocity in the LMX. That is, at a broad (higher-order factor) level of abstraction, having more leader provisions corresponds to having more member provisions. The absence of MTMM trait factors (affect, loyalty, contribution, and respect) suggests that Leader LMX and Member LMX are not reciprocated at the level of the lower-order LMX dimensions. Instead, any leader–member reciprocation occurs at the level of the broad

Leader–Member Exchange (LMX) Measurement

119

LMX factor [e.g., leader affect/liking is not necessarily reciprocated with member affect/liking, but leader broad LMX is reciprocated with member broad LMX, at least to some extent (r ¼ .54)]. As another example of this, it is possible for affect/liking from one party to be reciprocated in various ways (through loyalty, respect) by the other partner in the dyad. Thus the magnitude of reciprocation of the LMX relationship appears to be moderate (r ¼ .54), and it is generic reciprocation rather than reciprocation-in-kind – for example, loyalty is not met with loyalty. In other words, the four subdimensions of LMX are suitable substitutes for one another in the exchange process (cf. Wilson, Sin, & Conlon, 2010).

CAN MEMBER LMX BE DISTINGUISHED FROM SATISFACTION WITH SUPERVISOR? Lastly, we address the notion that Member LMX is potentially redundant with a construct known as satisfaction with the supervisor. The theoretical issue here is that ones’ liking of the supervisor/supervision (i.e., a facet of job satisfaction; Smith, Kendall, & Hulin, 1969) is conceptually redundant with the affect dimension of LMX (Dienesch & Liden, 1986), which is also deﬁned as liking of the supervisor. Indeed, satisfaction with supervisor is so closely conceptually aligned with LMX affect that Schriesheim et al. (1992) excerpted items directly from the satisfaction with supervisor section of the Minnesota satisfaction questionnaire (MSQ; Weiss, Dawis, England, & Lofquist, 1967) to use them as measures of LMX affect. Empirically speaking, Schriesheim and Cogliser (2009) have provided a systematic factor-analytic study of the discriminant validity between the LMX-7 measure (Scandura & Graen, 1984) and satisfaction with supervisor (as measured via the MSQ). Results from this study revealed that, from an exploratory factor analysis (EFA) of 379 bakery employees: (a) factor loadings onto a satisfaction with supervisor factor were uniformly at or above .64 for the MSQ items, and cross-loadings of LMX-7 items onto the satisfaction with supervisor factor were uniformly at or below .34 (with the exception of LMX-7 item 7, which loaded at .42 onto satisfaction with supervisor), and (b) factor loadings onto the LMX factor were at or above .50 for the LMX-7 items, and cross-loadings of MSQ items onto the LMX factor were all at or below .35. A CFA performed on an independent sample of 350 MBA students conﬁrmed adequate ﬁt of the two-factor model (NFI ¼ .90), although a model that allowed cross-loadings of LMX-7 items

120

DANA L. JOSEPH ET AL.

2 and 7 (which had cross-loadings of only .38 and .33) showed slightly better ﬁt (NFI ¼ .92). Notably, in the best-ﬁtting two-factor CFA model, the latent correlation between satisfaction with supervisor and LMX was a whopping .86. [According to Fornell and Larcker’s (1981) discriminant validity criterion (described earlier), and using the factor loadings from Schriesheim and Cogliser (2009, p. 729), the critical correlation above which the discriminant validity test fails for LMX and satisfaction with supervisor is r ¼ .75. Thus, LMX does not exhibit strong discriminant validity from satisfaction with supervisor (i.e., r ¼ .86W.75).] In some ways, this result is unsurprising. In 1997, Gerstner and Day reported a meta-analytic uncorrected correlation between the LMX-7 and satisfaction with supervision of r ¼ .74 (p. 834). By correcting Gerstner and Days’ (1997) p meta-analytic correlation for unreliability, we estimate ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rcorrected ¼ :74= ð:89Þð:88Þ ¼ :84. In other words, the LMX-7 and satisfaction with supervisor seem to largely measure the same construct, because they are correlated at .84. Next, we naturally wanted to assess whether the other dominant measure of LMX – the LMX-MDM (Liden & Maslyn, 1998) – was redundant with satisfaction with supervisor. To determine how strongly the LMX-MDM overlaps with satisfaction with supervisor, we searched the LMX literature for primary estimates of this relationship. In total, only three studies could be located that measured both the LMX-MDM and satisfaction with supervisor (Liden & Maslyn, 1998; Mardanov, Heischmidt, & Henson, 2008; Greguras & Ford, 2006). Furthermore, only the last of these studies reported the zero-order correlation between LMX-MDM and satisfaction with supervisor. Mardanov et al. (2008) did not report any information that can be used to estimate the LMX-MDM correlation with satisfaction with supervisor. Furthermore, Liden and Maslyn (1998) only reported a multiple R2 when satisfaction with supervisor was regressed onto all four LMX-MDM dimensions (R2 ¼ .65). Using this estimate, pﬃﬃﬃﬃﬃﬃﬃ we can calculate multiple R ¼ :65 ¼ :81, and Wherry’s shrunken R ¼ .80 (N ¼ 249), which is the population correlation between an optimally weighted LMX-MDM composite and satisfaction with supervisor. The most informative estimate, however, comes from the Greguras and Ford (2006) sample, because we can more directly calculate the correlation between the LMX-MDM broad LMX factor and satisfaction with supervisor. From the Greguras and Ford (2006) data, the correlation between the LMX-MDM composite pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ and satisfaction with supervisor is r ¼ .59 (rcorrected ¼ :59= ð:80Þð:92Þ ¼ :69). [The SEM latent correlation between the LMX-MDM higher-order factor (estimated from the correlations among

121

Leader–Member Exchange (LMX) Measurement

the lower-order LMX-MDM facets) and satisfaction with supervisor is latent r ¼ .75 (N ¼ 421).] Altogether, the results discussed above suggest: (a) the LMX-7 metaanalytically correlates above .8 with satisfaction with supervisor, and (b) the LMX-MDM total correlates .7 with satisfaction with supervisor. Given the magnitudes of these disattenuated correlations, we next wanted to assess whether the LMX measures could display incremental validity for predicting work outcomes, beyond the effects of satisfaction with supervisor. In other words, does LMX produce any added value (in the sense of criterion validity) over and above satisfaction with supervisor? We attempted to answer this question in two ways. First, we calculated the meta-analytic attenuation-corrected correlations between the LMX-7 and several individual-level outcomes (Gerstner & Day, 1997, pp. 834 & 832), and compiled the attenuation-corrected meta-analytic correlations between these same outcomes and satisfaction with supervisor from the meta-analysis of Kinicki, McKee-Ryan, Schriesheim, and Carson (2002; pp. 17–19). All these correlations are reported in Table 5a. We then estimated the incremental validity (change in R2) when using LMX-7 to predict each outcome, after controlling for satisfaction with supervisor (results are in the ﬁnal column of Table 5a). As shown in Table 5a, the LMX-7 does not

Table 5a. Outcome Variable

Organizational commitment Turnover intentions Performance ratings Role conﬂict Role ambiguity

Meta-Analytic Estimates of Incremental Validity of LMX-7 over Satisfaction with Supervisor. Correlation with LMX7a

Correlation with Satisfaction with Supervisorb

Multiple R2 for Incremental R2 R2 Satisfaction with Change for Supervisor LMX-7

.43

.42

.20

.18

.02

.31

.31

.10

.10

.01

.35

.23

.13

.05

.08

.47 .54

.40 .43

.22 .30

.16 .18

.06 .11

Note: Incremental changes in R2 that are greater than 2%. All correlations are corrected for unreliability in the predictor and the criterion. Correlation between LMX-7 and Satisfaction with Supervisor is rcorrected ¼ .84. a Correlations calculated from Gerstner and Day (1997, pp. 834 & 832). b Correlations from Kinicki et al. (2002, pp. 17–19). Owing to rounding after calculation, the ﬁnal column is not always exactly equal to the difference between the 4th and 5th columns.

122

DANA L. JOSEPH ET AL.

appear to meaningfully predict organizational commitment and withdrawal/ turnover intentions after controlling for satisfaction with supervisor (DR2 ¼ .02; .01). However, the LMX-7 does seem to exhibit meaningful incremental prediction of role ambiguity (DR2 ¼ .11), role conﬂict (DR2 ¼ .06), and performance ratings (DR2 ¼ .08), after controlling for satisfaction with supervisor, based on meta-analytic data. These incremental validity results are interesting, given the rcorrected ¼ .84 correlation between LMX-7 and satisfaction with supervisor. This raises a key question: What is the speciﬁc content of the LMX-7 that can predict role perceptions and performance ratings beyond simple satisfaction/liking for the leader? To answer this question, we default to a subjective inspection of the items in Table 1. Speciﬁcally, it appears that LMX-7 item 1 (the leader feedback ‘‘you know where you stand’’ item) reﬂects LMX domain content that taps role clarity/ambiguity. Schriesheim and Cogliser (2009) also suggest this LMX-7 item may tap into performance feedback from the leader, which can potentially explain the meta-analytic incremental validity of the LMX-7 beyond satisfaction with supervisor when predicting job performance. Next, we set out to estimate incremental validity of the LMX-MDM (Liden & Maslyn, 1998) measure, above satisfaction with supervisor. Owing to limited availability of studies that report correlations between the LMXMDM and satisfaction with supervisor, our estimates here were based on a single, large-sample study (Greguras & Ford, 2006; N ¼ 421). Greguras and Ford (2006) measured satisfaction with supervisor using a 4-item subscale from Spector’s (1985) job satisfaction survey, as well as self-reports of affective organizational commitment, job involvement, job performance, and OCB. Incremental validity results in Table 5b are somewhat different from the metaanalytic results reported above. Speciﬁcally, incremental validity of both the LMX-7 and the LMX-MDM (over satisfaction with supervisor) was observed for organizational commitment and job involvement attitudinal outcomes, but no incremental validity was observed for the job performance and citizenship behavior outcomes (i.e., these Table 5b incremental validity results are the opposite of the meta-analytic ﬁndings in Table 5a). Why do the metaanalytic data suggest LMX-7 incrementally predicts performance, but not commitment; whereas the Greguras and Ford (2006) data suggest both LMX7 and LMX-MDM incrementally predict commitment, but not performance? Because there are inﬁnitely many reasons why results from two samples may differ [and Gerstner and Day (1997) similarly reported signiﬁcant variability in LMX-7 criterion validities across their primary study samples], the most we can conﬁdently say is that incremental validity of LMX measures beyond satisfaction with supervisor is inconsistent.

123

Leader–Member Exchange (LMX) Measurement

Table 5b. Estimates of Incremental Validity of the LMX-7 and LMXMDM over Satisfaction with Supervisor (Based on Greguras & Ford, 2006) (LMX-7 results; LMX-MDM results). Outcome Variable

Correlation with LMX [LMX-7; LMX-MDM]

Organizational commitment Job involvement Performance ratings Citizenship behavior

Correlation with Multiple R2 for Incremental Satisfaction with R2 Satisfaction R2 Change for Supervisor with LMX Supervisor

.61; .54

.50

.39; .32

.25

.13; .07

.40; .32

.23

.16; .10

.05

.11; .05

.24; .29

.28

.08; .10

.08

.00; .02

.24; .27

.29

.09; .09

.08

.00; .01

Notes: Based on correlation matrix from Greguras and Ford (2006; p. 449; N ¼ 421). LMX-7 results are presented ﬁrst, followed by LMX-MDM results. Incremental changes in R2 that are greater than 2%. All correlations are corrected for unreliability in the predictor and the criterion. Correlation between LMX-7 and satisfaction with supervisor is rcorrected ¼ .70; correlation between LMX-MDM and satisfaction with supervisor is rcorrected ¼ .69. Owing to rounding after calculation, the ﬁnal column need not always be exactly equal to the difference between the 4th and 5th columns.

CONCLUSION AND FUTURE DIRECTIONS This chapter focused on describing and interpreting past measurement practices in research on LMX. We extended previous work (Schriesheim et al., 1999) by noting that LMX measurement has now converged onto two, dominant measures [the LMX-7 (Scandura & Graen, 1984) and the LMXMDM (Liden & Maslyn, 1998)]. When using these measures, LMX is commonly conceptualized as a broad, higher-order construct (not as a multidimensional construct), which is captured by the composite total score on the LMX-MDM, and by the LMX-7 (which correlates rcorrected ¼ .9 with the LMX-MDM total – suggesting the LMX-7 and LMX-MDM are alternate forms of the same instrument). Thus, there now appears to be great consensus in how LMX is measured: as the broad, higher-order LMX factor (Fig. 3). Next, we analyzed two MTMM datasets to demonstrate that Member LMX and Leader LMX are two, distinct constructs, in need of two, distinct locations in the nomological network (cf. Dienesch & Liden, 1986). Finally, we used meta-analytic data to show that the two dominant

124

DANA L. JOSEPH ET AL.

LMX measures both correlate very strongly with satisfaction with supervisor, but each shows some incremental validity beyond satisfaction with supervisor (although the pattern of incremental validity results seems to vary meaningfully across contexts). To restate: the overwhelming majority of LMX researchers are now consistently measuring the broad LMX construct using standard instruments, Member LMX and Leader LMX are two distinct constructs, and Member LMX is quite similar (though perhaps not identical) to the satisfaction with supervisor facet of job satisfaction. These ﬁndings are summarized in Table 6. As for future directions in the measurement of LMX, we have a few ideas (Table 7). First and foremost, we notice that most LMX measurement has evaluated the LMX construct using parts of the framework depicted in Fig. 4 (e.g., as member reports of leader contributions, leader reports of member contributions, etc.). Such data structures can naturally lend themselves to dyadic modeling (Kenny, Kashy, & Cook, 2006). But there are at least three

Table 6.

Summary of Findings.

1. Consensus. Since Schriesheim et al.’s (1999) review, 85% of LMX studies have used either the LMX-7 (Scandura & Graen, 1984) or the LMX-MDM (Liden & Maslyn, 1998). Total scores on these two measures are correlated rcorrected ¼ .9, suggesting the LMX-7 and the LMXMDM are alternate forms of the same instrument. There is thus great consensus in how LMX is currently measured. 2. Construct breadth. Of the studies that use the LMX-7 or the LMX-MDM to measure LMX, 94% treat LMX as a single dimension. In support of this practice, a half-dozen published papers have reported good model ﬁt for hierarchical CFA models of the LMX-MDM, conﬁrming the higher-order, broad LMX construct. Our own meta-analysis of LMX-MDM subdimensions shows the lower-order facets have factor loadings from .7 to .9 onto the broad LMX factor. The preponderance of LMX studies now consistently conceptualizes LMX at a broad level of abstraction. 3. Leader vs. Member LMX. Leader LMX is not the same construct as Member LMX. This conclusion generalizes for cross-report comparisons (leader reports vs. member reports of leader provisions) and for cross-diagonal comparisons (leader reports of leader provisions vs. member reports of member provisions). Both issues – who is supporting the LMX relationship and who is reporting the LMX relationship – can be used to meaningfully parse the LMX construct space. 4. Discriminant validity. Member LMX exhibits very weak discriminant validity from satisfaction with supervisor (meta-analytic rcorrected ¼ .8). LMX exhibits incremental validity over satisfaction with supervisor (predicting attitudes and behavior) in some contexts, but the pattern is inconsistent. Future research may beneﬁt from merging the LMX literature and the satisfaction with supervisor literature.

Same as above

Leader

Leader

LBDQ-consideration, LMX-7, LMX-MDM

Same as above

Dyad

2. Leader differential provisions Leader across followers (LMX differentiation,

Same as above

Dyad

Instrument(s)

LBDQ-consideration, LMX-7 (e.g., items 1–5), LMX-MDM (e.g., loyalty subscale) Same as above

Dyad

Level of Analysis

Member(s)

Leader

Member(s)

Leader & member

Leader

Member

Rater

Standard deviation of member-reports of leader provisions across

For each dyad, use leader report of leader provisions For each dyad, use the average of leader report and member report of leader provisions; calculate interrater reliability and interrater agreement to justify aggregation of leader with member ratings If more than one member gives rating of the same leader, use the average across raters; calculate ICCs and rWG to justify aggregation (Bliese, 2000) If leader rates her/his provisions to more than one member, use the average across members; calculate ICCs and rWG to justify aggregation (Bliese, 2000)

For each dyad, use member report of leader provisions

Data Analysis

Measuring the Broad LMX Factor: Summary of Constructs, Levels of Analysis, and Raters.

1. Leader provisions

LMX Construct

Table 7.

Leader–Member Exchange (LMX) Measurement 125

3. Member provisions

individualized leadership, favoritism)

LMX Construct

Same as above

Same as above

Leader

Leader

Leader

Member(s)

Leader & member

Same as above

Dyad

Leader

Rater

Dyad

Same as above

Instrument(s)

LMX-7 (e.g., item 6), LMX- Member MDM (e.g., affect, contribution, & respect subscales) Same as above Leader

Dyad

Leader

Level of Analysis

Table 7. (Continued )

For each dyad, use leader report of member provisions For each dyad, use the average of leader report and member report of member provisions; calculate interrater reliability and interrater agreement to justify aggregation of leader with member ratings If more than one member gives rating of provisions to the same leader, use the average across raters; calculate ICCs and rWG to justify aggregation (Bliese, 2000) If leader rates provisions received from more than one member, use the average across members; calculate ICCs and rWG to justify aggregation (Bliese, 2000)

For each dyad, use member report of member provisions

members, within-leader (see Henderson et al., 2009) Standard deviation of leader-reports of leader provisions across members, within-leader (see Henderson et al., 2009)

Data Analysis

126 DANA L. JOSEPH ET AL.

5. Balance between leader provisions and member provisions (Reciprocity)

4. Relationship quality

LMSX (Bernerth, Armenakis, Field, Giles, & Walker, 2007)

Same as above Same as above Same as above

Dyad

Dyad

Dyad

Leader

Same as above

Dyad

LMX-7, LMX-MDM

Same as above

Dyad

Dyad

LMX-7 (e.g., item 7)

Dyad

Member(s)

Member

Leader

Leader & member

Leader & member

Leader & Member

Leader

Member

For each dyad, use the interaction term between reports of leader provisions member provisions (Edwards, 2002) For each dyad, use the average of leader report and member report of LMSX; calculate interrater reliability and interrater agreement to justify aggregation of leader with member ratings For each dyad, use leader report of LMSX For each dyad, use member report of LMSX If more than one member gives rating of LMSX with the same leader, use the average across raters; calculate ICCs and rWG to justify aggregation (Bliese, 2000)

For each dyad, use member report of relationship quality For each dyad, use leader report of relationship quality. For each dyad, use the average of leader report and member report of relationship quality; calculate interrater reliability and interrater agreement to justify aggregation of leader with member ratings

Leader–Member Exchange (LMX) Measurement 127

Dyad

Social network

7. Network reciprocity and higher-order network effects (e.g., tendency for reciprocation between giving and receiving LMX provisions; pay-it-forward triadic exchange effects; etc., across the entire network)

Leader

Level of Analysis

6. Balance between leader perceptions and member perceptions of LMX (agreement)

LMX Construct

LMX-7, LMX-MDM

LBDQ-consideration, LMX-7, LMX-MDM

Same as above

Instrument(s)

Rater

All leaders & members in the network

Leader & member

Leader

Table 7. (Continued )

Exponential random graph modeling (Snijders et al., 2006)

For each dyad, use the interaction term between leader perceptions of LMX member perceptions of LMX (Edwards, 2002; cf. Cogliser, Schriesheim, Scandura, & Gardner, 2009; Sin et al., 2009)

If leader rates LMSX with more than one member, use the average across members; calculate ICCs and rWG to justify aggregation (Bliese, 2000)

Data Analysis

128 DANA L. JOSEPH ET AL.

Leader–Member Exchange (LMX) Measurement

129

alternative (yet intuitive) ways for future researchers to index the notion of exchange between a leader and follower. First, researchers could directly measure perceptions of the dyadic relationship [e.g., Bernerth et al., 2007; also see Chan’s (1998) discussion of referent shift composition models, and the description of methods for validating group-level constructs by Bliese (2000); Dyer, Hanges, & Hall (2005); and Yammarino, Dionne, Chun, & Dansereau (2005)]. Second, researchers could assess the ﬁt between leader provisions and member provisions, using methods articulated by Edwards (2002), to evaluate the effects of congruence between leader provisions and follower provisions (i.e., good congruence between the contributions of leaders and the contributions of followers should produce perceptions of relational fairness). A similar concept was advanced by Cogliser, Schriesheim, Scandura, & Gardner (2009; i.e., LMX balance), with regard to agreement between leader reports and member reports of leader provisions (i.e., perceptual ﬁt regarding leader provisions). Such indices focus more directly on the notion of interpersonal exchange. Third, the concept of exchange relationships can be captured at the team- or organizational-level of analysis using social networks of exchange (cf. Sparrowe & Liden, 1997), by assessing tendencies for dyadic reciprocity (and also triadic exchange effects) at the level of the entire interpersonal network, as well as assessing the correspondence between the occupation of formal leader roles and symmetry/asymmetry of dyadic LMX provisions (ties) (see review by Contractor, Wasserman, & Faust, 2006; Snijders, Pattison, Robins, & Handcock, 2006). Such analyses and measures will take much more advantage of the dyadic conceptualization of LMX, helping to pinpoint the effects of leadership when it is viewed as a relational phenomenon that varies across followers, within each leader.

REFERENCES Barling, J., Christie, A., & Hoption, A. (2010). Leadership. In: S. Zedeck (Ed.), Handbook of industrial and organizational psychology (pp. 183–240). Washington, DC: American Psychological Association. Bauer, T. N., & Green, S. G. (1996). Development of leader–member exchange: A longitudinal test. Academy of Management Journal, 39, 1538–1567. Bernerth, J. B., Armenakis, A. A., Field, H. S., Giles, W. F., & Walker, H. J. (2007). Leadermember social exchange (LMSX): Development and validation of a scale. Journal of Organizational Behavior, 28, 979–1003. Bhal, K. T., & Ansari, M. A. (1996). Measuring quality of interaction between leaders and members. Journal of Applied Social Psychology, 26, 945–972.

130

DANA L. JOSEPH ET AL.

Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and analysis. In: K. J. Klein & S. W. Kozlowski (Eds), Multilevel theory, research, and methods in organizations (pp. 512–553). San Francisco: JosseyBass, Inc. Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait–multimethod matrix. Psychological Bulletin, 56, 81–105. Carroll, J. B. (1993). Human cognitive abilities: A survey of factor analytic studies. New York: Cambridge University Press. Chan, D. (1998). Functional relations among constructs in the same content domain at different levels of analysis: A typology of composition models. Journal of Applied Psychology, 83, 234–246. Cogliser, C. C., Schriesheim, C. A., Scandura, T. A., & Gardner, W. L. (2009). Balance in leader and follower perceptions of leader-member exchange: Relationships with performance and work attitudes. The Leadership Quarterly, 20, 452–465. Collins, M. D. (2007). Understanding the relationship between leader-member exchange (LMX), psychological empowerment, job satisfaction, and turnover intent in a limited-service restaurant environment. Unpublished doctoral dissertation, The Ohio State University. Contractor, N. S., Wasserman, S., & Faust, K. (2006). Testing multi-theoretical multilevel hypotheses about organizational networks: An analytic framework and empirical example. Academy of Management Review, 31, 681–703. Cronbach, L. J. (1956). Assessment of individual differences. Annual Review of Psychology, 7, 173–196. Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52, 281–302. Dansereau, F., Jr., Cashman, J., & Graen, G. B. (1973). Instrumentality theory and equity theory as complementary approaches in predicting the relationship of leadership and turnover among mangers. Organizational Behavior and Human Performance, 10, 184–220. Dansereau, F., Jr., Graen, G., & Haga, W. J. (1975). A vertical dyad linkage approach to leadership within formal organizations: A longitudinal investigation of the role-making process. Organizational Behavior and Human Performance, 13, 46–78. Dienesch, R. M. (1986). A three dimensional model of leader-member exchange: An empirical test. Presented at the meetings of the Academy of Management, Chicago, IL. Dienesch, R. M., & Liden, R. C. (1986). Leader–member exchange model of leadership: A critique and further development. Academy of Management Review, 11, 618–634. Digman, J. M. (1997). Higher-order factors of the big ﬁve. Journal of Personality and Social Psychology, 73, 1246–1256. Dyer, N. G., Hanges, P. J., & Hall, R. (2005). Applying multilevel conﬁrmatory factor analysis techniques to the study of leadership. The Leadership Quarterly, 16, 149–167. Edwards, J. R. (2001). Multidimensional constructs in organizational behavior research: An integrative analytical framework. Organizational Research Methods, 4, 144–192. Edwards, J. R. (2002). Alternatives to difference scores: Polynomial regression analysis and response surface methodology. In: F. Drasgow & N. W. Schmitt (Eds), Advances in measurement and data analysis (pp. 350–400). San Francisco: Jossey-Bass. Edwards, J. R. (2003). Construct validation in organizational behavior research. In: J. Greenberg (Ed.), Organizational behavior: The state of the science (pp. 327–371). Mahwah, NJ: Erlbaum.

Leader–Member Exchange (LMX) Measurement

131

Eisenberger, R., Karagonlar, G., Stinglhamber, F., Neves, P., Becker, T. E., Gonzalez-Morales, M. G., & Steiger-Mueller, M. (2010). Leader–member exchange and affective organizational commitment: The contribution of supervisor’s organizational embodiment. Journal of Applied Psychology, 95, 1085–1103. Erdogan, B., & Enders, J. (2007). Support from the top: Supervisors’ perceived organizational support as a moderator of leader-member exchange to satisfaction and performance relationships. Journal of Applied Psychology, 92, 321–330. Erdogan, B., & Liden, R. C. (2006). Collectivism as a moderator of responses to organizational justice: Implications for leader-member exchange and ingratiation. Journal of Organizational Behavior, 27, 1–17. Erdogan, B., Kraimer, M. L., & Liden, R. C. (2004). Work value congruence and intrinsic career success: The compensatory role of leader-member exchange and perceived organizational support. Personnel Psychology, 57, 305–332. Erdogan, B., Liden, R. C., & Kraimer, M. L. (2006). Justice and leader-member exchange: The moderating role of organizational culture. Academy of Management Journal, 49, 395–406. Ferris, G. R., Liden, R. C., Munyon, T. P., Summers, J. K., Basik, K. J., & Buckley, M. R. (2009). Relationships at work: Toward a multidimensional conceptualization of dyadic work relationships. Journal of Management, 25, 1379–1403. Fisher, C. D. (1980). On the dubious wisdom of expecting job satisfaction to correlate with performance. Academy of Management Review, 5, 607–612. Ford, L. R., & Seers, A. (2006). Relational leadership and team climates: Pitting differentiation versus agreement. Leadership Quarterly, 17, 258–270. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39–50. Gerstner, C. R., & Day, D. V. (1997). Meta-analytic review of leader–member exchange theory: Correlates and construct issues. Journal of Applied Psychology, 82, 827–844. Graen, G. (1976). Role-making processes within complex organizations. In: M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 1201–1245). Chicago: Rand-McNally. Graen, G., & Cashman, J. F. (1975). A role-making model of leadership in formal organizations: A developmental approach. In: J. G. Hunt & L. L. Larson (Eds), Leadership frontiers (pp. 143–165). Kent, OH: Kent State University Press. Graen, G., Dansereau, G., Jr., & Minami, T. (1972a). Dysfunctional leadership styles. Organizational Behavior and Human Performance, 7, 216–236. Graen, G., Dansereau, G., Jr., & Minami, T. (1972b). An empirical test of the main-in-themiddle hypothesis among executives in a hierarchical organization employing a unit-set analysis. Organizational Behavior and Human Performance, 8, 262–285. Graen, G., Liden, R. C., & Hoel, W. (1982a). The role of leadership in the employee withdrawal process. Journal of Applied Psychology, 67, 868–872. Graen, G. B., Novak, M. A., & Sommerkamp, P. (1982b). The effects of leader-member exchange and job design on productivity and satisfaction: Testing a dual attachment model. Organizational Behavior and Human Performance, 30, 109–131. Graen, G., & Scandura, T. A. (1985). Leader-member exchange scale-17. Cincinnati, OH: University of Cincinnati. Graen, G. B., & Scandura, T. A. (1987). Toward a psychology of dyadic organizing. Research in Organizational Behavior, 9, 175–208.

132

DANA L. JOSEPH ET AL.

Graen, G. B., & Schiemann, W. (1978). Leader-member agreement: A vertical dyad linkage approach. Journal of Applied Psychology, 63, 206–212. Graen, G. B., & Uhl-Bien, M. (1995). Relationship-based approach to leadership: Development of leader–member exchange (LMX) theory of leadership over 25 years: Applying a multi-level multi-domain perspective. Leadership Quarterly, 6, 219–247. Greguras, G. J., & Ford, J. M. (2006). An examination of the multidimensionality of supervisor and subordinate perceptions of leader-member exchange. Journal of Occupational and Organizational Psychology, 79, 433–465. Gulliksen, H. (1950). Theory of mental tests. Hillsdale, NJ: Lawrence Erlbaum. Hanisch, K. A., Hulin, C. L., & Roznowski, M. (1998). The importance of individuals’ repertoires of behaviors: The scientiﬁc appropriateness of studying multiple behaviors and general attitudes. Journal of Organizational Behavior, 19, 463–480. Harris, R. B., Harris, K. J., & Harvey, P. (2008). An examination of the impact of supervisor on the relationship between job strains and turnover intention for computer workers. Journal of Applied Social Psychology, 38, 2108–2131. Harrison, D. A., Newman, D. A., & Roth, P. L. (2006). How important are job attitudes? Metaanalytic comparisons of integrative behavioral outcomes and time sequences. Academy of Management Journal, 49, 305–325. Henderson, D. J., Liden, R. C., Glibowski, B. C., & Chaudry, A. (2009). LMX differentiation: A multilevel review and examination of its antecedents and outcomes. Leadership Quarterly, 20, 517–534. Hinkin, T. (1995). A review of scale development practices in the study of organizations. Journal of Management, 21, 967–988. Hinkin, T. (1998). A brief tutorial on the development of measures for use in survey questionnaires. Organizational Research Methods, 1, 104–121. House, R. J., & Dessler, G. (1974). The path-goal theory of leadership: Some post hoc and a priori tests. In: J. G. Hunt & L. L. Larson (Eds), Contingency approaches to leadership. Carbondale, IL: Southern Illinois University Press. Johns, G. (1998). Aggregation or aggravation? The relative merits of abroad withdrawal construct. Journal of Organizational Behavior, 19, 453–462. Johnson, W., & Bouchard, T. J., Jr. (2005). The structure of human intelligence: It is verbal, perceptual, and image rotation (VPR), not ﬂuid and crystallized. Intelligence, 33, 393–416. Joseph, D. L., & Newman, D. A. (2010). Discriminant validity of self-reported emotional intelligence: A multitrait-multisource study. Educational and Psychological Measurement, 70, 672–694. Judge, T. A., Locke, E. A., & Durham, C. C. (1997). The dispositional causes of job satisfaction: A core evaluations approach. Research in Organizational Behavior, 19, 151–188. Katerberg, R., & Hom, P. W. (1981). Effects of within-group and between-group variation in leadership. Journal of Applied Psychology, 66, 218–223. Keller, T., & Dansereau, F. (2001). The effect of adding items to scales: An illustrative case of LMX. Organizational Research Methods, 4, 131–143. Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic data analysis. New York, NY: Guilford Press. Kinicki, A. J., McKee-Ryan, F. M., Schriesheim, C. A., & Carson, K. P. (2002). Assessing the construct validity of the Job Descriptive Index: A review and meta-analysis. Journal of Applied Psychology, 87, 14–32.

Leader–Member Exchange (LMX) Measurement

133

Kozlowski, S. W. J., & Doherty, M. L. (1989). Integration of climate and leadership: Examination of a neglected issue. Journal of Applied Psychology, 74, 546–553. Landy, F. J. (1986). Stamp collecting versus science: Validation as hypothesis testing. American Psychologist, 41, 1183–1192. Lee, J. (2005). Effects of leadership and leader-member exchange on commitment. Leadership & Organization Development Journal, 26, 655–672. Lee, J. (2008). Effects of leadership and leader-member exchange on innovativeness. Journal of Managerial Psychology, 23, 670–687. Liden, R. C., Erdogan, B., Wayne, S. J., & Sparrowe, R. T. (2006). Leader-member exchange, differentiation, and task interdependence: Implications for individual and group performance. Journal of Organizational Behavior, 27, 723–746. Liden, R. C., & Graen, G. B. (1980). Generalizability of the vertical dyad linkage model of leadership. Academy of Management Journal, 23, 451–465. Liden, R. C., & Maslyn, J. M. (1998). Multi-dimensionality of leader–member exchange: An empirical assessment through scale development. Journal of Management, 24, 43–72. Liden, R. C., Wayne, S. J., & Stilwell, D. (1993). A longitudinal study on the early development of leader-member exchanges. Journal of Applied Psychology, 78, 662–674. Lim, B. C., & Ployhart, R. E. (2006). Assessing the convergent and discriminant validity of Goldberg’s international personality item pool. Organizational Research Methods, 9, 29–54. Mardanov, I. T., Heischmidt, K., & Henson, A. (2008). Leader-member exchange and job satisfaction bond and predicted employee turnover. Journal of Leadership and Organizational Studies, 15, 159–175. Maslyn, J. M., & Uhl-Bien, M. (2001). Leader-member exchange and its dimensions: Effects of self-effort and other’s effort ton relationship quality. Journal of Applied Psychology, 86, 697–708. Newman, D. A., Joseph, D. L., & Hulin, C. L. (2010). Job attitudes and employee engagement: Considering the attitude ‘‘A-factor’’. In: S. Albrecht (Ed.), Handbook of employee engagement. Cheltenham, UK: Edwin Elgar Publishing. Ones, D. S., & Viswesvaran, C. (1996). Bandwidth-ﬁdelity dilemma in personality measurement for personnel selection. Journal of Organizational Behavior, 17, 609–626. Raabe, B., & Beehr, T. A. (2003). Formal mentoring versus supervisor and coworker relationship: Differences in perceptions and impact. Journal of Organizational Behavior, 24, 271–293. Scandura, T. A., & Graen, G. B. (1984). Moderating effects of initial leader-member exchange status on the effects of a leadership intervention. Journal of Applied Psychology, 69, 428–436. Schriesheim, C. A. (1978). Development and empirical veriﬁcation of new leadership and expectancy research instruments. Unpublished doctoral dissertation, Ohio State University. Schriesheim, C. A., Castro, S. L., & Cogliser, C. C. (1999). Leader-member exchange (LMX) research: A comprehensive review of theory, measurement, and data-analytic practices. Leadership Quarterly, 10, 63–113. Schriesheim, C. A., & Cogliser, C. C. (2009). Construct validation in leadership research: Explication and illustration. The Leadership Quarterly, 20, 725–736. Schriesheim, C. A., Neider, L. L., Scandura, T. A., & Tepper, B. J. (1992). Development and preliminary validation of a new scale (LMX-6) to measure leader-member exchange in organizations. Educational and Psychological Measurement, 52, 135–147.

134

DANA L. JOSEPH ET AL.

Schriesheim, C. A., Powers, K. J., Scandura, T. A., Gardiner, C. C., & Lankau, M. J. (1993). Improving construct measurement in management research: Comments and a quantitative approach for assessing the theoretical content adequacy of paper-andpencil survey-type instruments. Journal of Management, 19, 385–417. Schyns, B. (2006). Are group consensus in leader–member exchange (LMX) and shared work values related to organizational outcomes? Small Group Research, 37, 20–35. Schyns, B., & Wolfram, H. J. (2008). The relationship between leader-member exchange and outcomes as rated by leaders and followers. Leadership & Organization Development Journal, 29, 631–646. Seers, A. (1989). Team-member exchange quality: A new construct for role-making research. Organizational Behavior and Human Decision Processes, 43, 118–135. Sin, H. P., Nahrgang, J. D., & Morgeson, F. P. (2009). Understanding why they don’t see eye to eye: An examination of leader-member exchange (LMX) Agreement. Journal of Applied Psychology, 94, 1048–1057. Sluss, D. M., Klimchak, M., & Holmes, J. J. (2008). Perceived organizational support as a mediator between relational exchange and organizational identiﬁcation. Journal of Vocational Behavior, 73, 457–464. Smith, P. C., Kendall, L. M., & Hulin, C. L. (1969). The measurement of satisfaction in work and retirement: A strategy for the study of attitudes. Chicago: Rand McNally. Snijders, T. A. B., Pattison, P. E., Robins, G. L., & Handcock, M. S. (2006). New speciﬁcations for exponential random graph models. Sociological Methodology, 36, 99–153. Sparrowe, R. T., & Liden, R. C. (1997). Process and structure in leader-member exchange. Academy of Management Review, 22, 522–552. Sparrowe, R. T., Soetjipto, B. W., & Kraimer, M. L. (2006). Do leaders’ inﬂuence tactics relate to members’ helping behavior? It depends on the quality of the relationship. Academy of Management Journal, 49, 1194–1208. Spector, P. E. (1985). Measurement of human service staff satisfaction: Development of the Job Satisfaction Survey. American Journal of Community Psychology, 13, 693–713. Stogdill, R. M. (1963). Manual for the leader behavior description questionnaire – Form XII. Columbus, OH: Ohio State University, Bureau of Business Research. Stogdill, R. N., & Coons, A. E. (1957). Leader behavior: Its description and measurement. Columbus, OH: Bureau of Business Research, Ohio State University. Vandenberghe, C., Bentein, K., & Stinglhamber, F. (2004). Affective commitment to the organization, supervisor, and work group: Antecedents and outcomes. Journal of Vocational Behavior, 64, 47–71. Vidyarthi, P. R., Liden, R. C., Anand, S., Erdogan, B., & Ghosh, S. (2010). Where do I stand? Examining the effects of leader-member exchange social comparison on employee work behaviors. Journal of Applied Psychology, 95, 849–861. Viswesvaran, C., Schmidt, F. L., & Ones, D. S. (2005). Is there a general factor in ratings of job performance? A meta-analytic framework for disentangling substantive and error inﬂuences. Journal of Applied Psychology, 90, 108–131. Wakabayashi, M., Graen, G. B., Graen, M. R., & Graen, M. G. (1988). Japanese management progress: Mobility into middle management. Journal of Applied Psychology, 73, 217–227. Wakabayashi, M., Graen, G., & Uhl-Bien, M. (1990). The generalizability of the hidden investment hypothesis in leading Japanese corporations. Human Relations, 90, 1100– 1116.

Leader–Member Exchange (LMX) Measurement

135

Wang, H., Law, K. S., & Chen, Z. X. (2008). Leader-member exchange, employee performance, and work outcomes: An empirical study in the Chinese context. The International Journal of Human Resource Management, 19, 1809–1824. Wang, H., Law, K. S., Hackett, R. D., Wang, D., & Chen, Z. X. (2005). Leader-member exchange as a mediator of the relationship between transformational leadership and followers’ performance and organizational citizenship behavior. Academy of Management Journal, 48, 420–432. Wang, H., Law, K. S., Wang, D., & Chen, Z. (2001). The linkage role of LMX: A mediating effect of LMX on the relationship between transformational leadership and followers’ performance and OCB. Paper presented at the Annual Academy of Management Meeting. Washington, DC, August 3–8. Wayne, S. J., & Ferris, G. R. (1990). Inﬂuence tactics, affect, and exchange quality in supervisor-subordinate interactions: A laboratory experiment and ﬁeld study. Journal of Applied Psychology, 75, 487–499. Wayne, S. J., Shore, L. M., & Liden, R. C. (1997). Perceived organizational support and leadermember exchange: A social exchange perspective. Academy of Management Journal, 40, 82–111. Weiss, D. J., Dawis, R. V., England, G. W., & Lofquist, L. H. (1967). Manual for the Minnesota satisfaction questionnaire. Minneapolis, MN: Industrial Relations Center, University of Minnesota. Widaman, K. F. (1985). Hierarchically nested covariance structure models for multitraitmultimethod data. Applied Psychological Measurement, 9, 1–26. Wilson, K. S., Sin, H. P., & Conlon, D. E. (2010). What about the leader in leader-member exchange? The impact of resource exchanges and substitutability on the leader. Academy of Management Review, 35, 358–372. Yammarino, F. J., Dionne, S. D., Chun, J. U., & Dansereau, F. (2005). Leadership and levels of analysis: A state-of-the-science review. The Leadership Quarterly, 16, 879–919.

ENTREPRENEURIAL MOBILITY Mike Wright ABSTRACT Purpose – Although there is extensive work on labor mobility, research on entrepreneurial mobility is fragmented and many aspects are largely neglected. We develop a framework for analysis that integrates different perspectives on entrepreneurial mobility to provide a broad agenda for future research. Design/methodology – We build upon the strategic entrepreneurship, entrepreneurial behaviour theory, resource-based theory and other literatures, to distinguish four quadrants involving high and low geographical mobility and high and low organizational mobility. Findings – Within each quadrant we identify different types of entrepreneurial mobility, speciﬁcally habitual entrepreneurs, management buyouts, university spin-offs, returnee entrepreneurs and transnational entrepreneurs. Issues concerning the development of research programs and methods, with particular emphasis on datasets, are discussed. Originality/value – It is hoped that this chapter will spur entrepreneurship and strategy scholars to recognize that the scope of the entrepreneurial mobility concept is considerably greater than hitherto appreciated, providing interesting new avenues for theoretical and methodological development in this area.

Building Methodological Bridges Research Methodology in Strategy and Management, Volume 6, 137–159 Copyright r 2011 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-8387/doi:10.1108/S1479-8387(2011)0000006008

137

138

MIKE WRIGHT

Keywords: Academic entrepreneurship; cognition; management buyouts; micro-foundations; habitual entrepreneurs; resource-based view; returnee entrepreneurs; transnational entrepreneurs; strategic entrepreneurship

INTRODUCTION Strategic entrepreneurship is concerned with the resources and capabilities required to recognize and exploit entrepreneurial opportunities (Ireland, Hitt, & Sirman, 2003). Increasing attention is being directed towards the hitherto-neglected processes of where resources come from and how they are coordinated in the design of organizations (Sirmon, Hitt, & Ireland, 2007). These processes depend heavily on the partially understood decision-making processes of entrepreneurs. However, it is recognized that it is not possible to study individual entrepreneurs independently of the context in which they ﬁnd themselves (Shane & Venkataraman, 2000). How and why entrepreneurs ﬁnd themselves in particular contexts raises major issues concerning entrepreneurial mobility. Understanding entrepreneurial mobility helps identify the boundaries of the strategic entrepreneurship construct. In this chapter, we adopt a broad perspective on entrepreneurship that focuses on the recognition and exploitation of opportunities by entrepreneurs through different organizational modes. As such, we include organizational modes such as start-ups, spin-offs, buyouts and the transfer of family ﬁrms to successor generations where these involve the pursuit of new entrepreneurial opportunities. Adopting this perspective on entrepreneurship, entrepreneurial mobility is deﬁned broadly as entrepreneurial activity that involves the movement of an entrepreneur from a current context to another one. Our focus is distinct from analyses of international trading activities of entrepreneurial ﬁrms that remain located in their home economy (Cumming, Sapienza, Siegel, & Wright, 2009b). The aim of this chapter is to develop a framework of the dimensions of entrepreneurial mobility. This chapter makes contributions to several literatures. First, we extend the emerging strategic entrepreneurship literature by focusing on the inﬂuence of mobility across different contexts on the nature of opportunity recognition and the types of resources needed in these different contexts. Second, by focusing on individual behaviour we contribute to the emerging literature on the micro-foundations of strategy and entrepreneurship. Third,

Entrepreneurial Mobility

139

we extend the literature on human capital mobility across national boundaries. This literature has tended to focus on the mobility of employees in multinational corporations but we draw attention to the role of individual entrepreneurial mobility. Fourth, we have extended the limited literature on entrepreneurial mobility by providing a systematic framework for analysis that enriches the extent and types of entrepreneurial mobility. In the following section, we outline a deﬁnition of entrepreneurial mobility before proceeding to consider entrepreneurial mobility in the context of entrepreneurial behaviour and resource-based theory. Building on these concepts, the subsequent section develops a typology of entrepreneurial mobility based on high and low aspects of geographical and organizational mobility. The discussion section then considers methodological issues relating to developing theory and empirical analysing the entrepreneurial mobility framework.

ENTREPRENEURIAL MOBILITY There is an extensive literature on employee mobility in the form of expatriates working for multinational enterprises (Harvey & Moeller, 2009) and in terms of the migration of individuals to ﬁnd work in more developed industrial centers (Go¨rg & Strobl, 2005); Song, Almeida, & Wu, 2003). However, the existing literature has taken a limited and fragmented view of entrepreneurial mobility. The focus of the entrepreneurship literature has mainly been on entrepreneurs starting businesses in their local environments because of their knowledge of resources, social capital, etc. (Iyer, Kitson, & Toh, 2005; Nijkamp, 2003; Stuart & Sorenson, 2003; Pe’er & Vertinsky, 2008). The limited attention to mobility mainly concerns the geographical mobility involved in immigrant entrepreneurship (Levie, 2007; Kalnins & Chung, 2006). Yet, geographical mobility is complex and multi-dimensional. Entrepreneurs may move in different geographic directions, such as away from their home country or return there after a period spent abroad. Entrepreneurial mobility also involves organizational mobility. Entrepreneurs may shift from their organization or shift part of the organization in which they are employed or which they own. There is some literature on employees leaving their employers to start ventures (Agarwal, Echambadi, Franco, & Sarkar, 2004; Groysberg, Nanda, & Prats, 2007) and on corporate spin-offs (Phan, Wright, Ucbasaran, & Tan, 2009), but the processes by which both these aspects of entrepreneurial mobility occur have tended to be neglected. Recognition of the multi-faceted nature of geographical and

140

MIKE WRIGHT

organizational mobility introduces a greater landscape in which to study entrepreneurial mobility than hitherto recognized. To help understand the processes of entrepreneurial mobility we draw upon strategic entrepreneurship, entrepreneurial behaviour theory and resource based theory. Mobility and Entrepreneurial Behaviour The emerging research on the micro-foundations of strategy and entrepreneurship stresses the importance of individual behaviour and the interaction of this behaviour with ﬁrm attributes (Foss, 2011; Coff & Kryscynski, 2011). Cognitive aspects of entrepreneurship are being analysed (Baron, 2004) and the inﬂuence of heuristics and biases on internal information acquisition and processing challenges faced by managers and entrepreneurs is recognized (King, Garbuio, & Lovallo, 2011). This literature has considered experience but generally in a very limited sense. For example, it has not paid much attention to the difference between the nature and extent of experience (Ucbasaran, Westhead, & Wright, 2009; Ucbasaran, Westhead, Wright, & Flores, 2010). This literature also has a limited view of the geographical and organizational boundaries of entrepreneurial cognition and learning. Studies have mainly involved owners of independent businesses, with little attention to the contexts of corporate entrepreneurship (Wiklund & Shepherd, 2008) or the purchase of businesses, as in management buyouts (Wright, Hoskisson, Busenitz, & Dial, 2000). There is also a need to understand the interaction between these cognitive aspects of entrepreneurial behaviour and social capital resources (Davidsson & Honig, 2003). With respect to entrepreneurial mobility, therefore, entrepreneurial behaviour raises important questions that are as yet unanswered concerning how entrepreneurs’ cognitive processes work when they move to different contexts and whether and how entrepreneurs adapt their cognitive processes to the new environment. The question of whether only entrepreneurs with certain types of cognitive attributes move their context also remains unanswered. Mobility and Resource-Based Theory Three particular aspects of resources and capabilities provide conceptual insights of relevance to entrepreneurial mobility: the development of human capital resources and capabilities, social capital, and corporate governance.

Entrepreneurial Mobility

141

Developing Human Capital Resources and Capabilities. The strategic entrepreneurship perspective (Sirmon et al., 2007) emphasizes the need to link opportunity recognition to the resources needed to exploit an opportunity. It follows that how entrepreneurial ﬁrms access the resources and capabilities needed to develop new opportunities raises major challenges that are only now beginning to be understood. This literature has so far remained quite general and has not considered how these resources and capabilities are developed when entrepreneurs move to new geographical locations or organizational contexts. It is also unclear how this process unfolds over time. A major challenge concerns how entrepreneurs and ﬁrms shift from inherited path dependencies as they move contexts and locations. For example, family ﬁrms that are transferred to the next generation may ﬁnd it difﬁcult to move away from ways of doing things that have been ingrained by the founder. In the case of spin-offs and divisional buy-outs from corporations, there may also need to be a shift from the inherited path dependencies of the parent if the newly independent management and ﬁrm are to survive (Ahuja & Katila, 2004). Similarly, in the case of spin-offs from traditionally non-commercial universities, academic entrepreneurs need to develop resources and capabilities that are more commercially oriented (Rasmussen, Mosey, & Wright, 2011).

Mobility and Social Capital. Social capital and networks represents an important resource for entrepreneurship, with a long research tradition (Hoang & Antoncic, 2003). Different aspects of social capital have been recognized as important in different institutional contexts (Hoskisson, Eden, Lau, & Wright, 2000). The importance and paradox of embedded relationships have also been identiﬁed (Uzzi, 1997; Meuleman, Lockett, Manigart, & Wright, 2010). Embedded relationships may pose limitations in facilitating the mobility of entrepreneurs across geographical and organizational contexts. Mobile entrepreneurs may need to access new knowledge through weak rather than strong ties (Sequiera & Rasheed, 2006) with new social capital partners as they shift contexts and locations. The challenges for mobile entrepreneurs in developing this new social capital have not been explored. For example, the importance of organizational social capital has been recognized in the context of family ﬁrms (Arregle, Hitt, Sirmon, & Very, 2007; Zahra, 2010). However, as family ﬁrms move from ﬁrst to subsequent generations, it may be important to stimulate entrepreneurship through accessing new knowledge, otherwise the family

142

MIKE WRIGHT

ﬁrm may be too inward looking in the face of changing environmental conditions and the loss of the tacit knowledge embodied in founders. Mobility and Corporate Governance as a Resource. Corporate governance is recognized as a resource for ﬁrms in general (Barney, Wright, & Ketchen, 2001). The role of corporate governance is also coming to be seen as important in entrepreneurial ﬁrms. In particular, boards of directors assume particularly central roles. Although it has sometimes been thought that agency issues are absent in owner-managed ﬁrms, new insights not only contest this view but also suggest that agency issues may be quite complex and intertwined with other aspects. For example, family ﬁrms may experience agency issues between family owner-managers and family owner-non-managers (Schulze, Lubatkin, & Dino, 2003). In such cases there may be conﬂicts between members focused on the stewardship of the ethos of the family business and those seeking to focus on proﬁtability. These perspectives may be complementary over the life-cycle of the business (Schulze & Gedajlovic, 2010) but they may also pose challenges in a context of entrepreneurial mobility. Boards may play an advisory role focused on adding value to the venture and not just a monitoring role. Whether boards play a monitoring or an advisory role may depend on different context. One aspect of context that has received attention concerns the life-cycle stage of the venture. For example, Zahra, Filatotchev, and Wright (2009) have examined the role of boards in the context of threshold ﬁrms, that is those ﬁrms at the stage between start-up and being established companies. They suggest that boards can play a key role in accessing knowledge that facilitates entrepreneurship and may complement or substitute for the role of entrepreneurs depending on the absorptive capacity of the entrepreneur. These insights suggest that the nature of governance structures and processes may facilitate or constrain entrepreneurial mobility.

ORGANIZATIONAL AND GEOGRAPHICAL MOBILITY – A TYPOLOGY Table 1 presents a typology of entrepreneurial mobility. For expositional purposes, a distinction is made between high and low geographical mobility and high and low organizational mobility. Each quadrant is discussed using the conceptual insights provided by the previous discussion on

143

Entrepreneurial Mobility

entrepreneurial behaviour and the development of different aspects of resources and capabilities.

Quadrant 1: Low Geographical and Low Organizational Mobility Entrepreneurs are traditionally viewed as starting and owning one business. Geographical and organizational mobility is typically low. However, signiﬁcant numbers of businesses are owned by habitual entrepreneurs and their sub-categories of serial entrepreneurs (who own multiple ventures sequentially) and portfolio entrepreneurs who own multiple ventures concurrently (Ucbasaran, Alsos, Westhead, & Wright, 2008a). Low organizational and geographical mobility is generally involved as few habitual entrepreneurs change the geographical location of their subsequent ventures or their industrial sectors. Table 1.

Entrepreneurial Mobility Typology. Geographical Mobility Low

Organizational mobility Low Quadrant 1: creation of new ﬁrm in local context Traditional: nascents; start-ups, family ﬁrm succession Nature of previous organization not considered New: habitual entrepreneurs Learn from previous ventures/generation High Quadrant 2: move out of established organization Traditionally: Corp. spin-offs

High

Quadrant 3: individual mobility Traditional: immigrant entrepreneurs Either established locally or moved to a foreign country New: returnee entrepreneurs Moved back to home country Quadrant 4: move established ﬁrm Traditionally: ﬁrm relocation to another region or country Relocate to new region or new country

Redraw organizational boundaries by creating new organization by taking IP outside established organization New: academic spin-outs; divisional New: transnational entrepreneurs MBOs Redraw non-commercial to commercial; Retain home but establish new transfer entities to independence location overseas ownership

144

MIKE WRIGHT

Increasing interest has been focused on the cognitive factors that inﬂuence how habitual entrepreneurs create new ventures. In particular, attention has been devoted to examining the heuristics and biases used by habitual entrepreneurs in processing information relation to opportunities. Authors have also linked prior experience from previous entrepreneurial ventures to the literature on expert cognition. Entrepreneurial experience is found to enable habitual entrepreneurs to direct their attention in a more focused manner and to be able to interpret information more efﬁciently so as to better facilitate the generation of ideas. As experts have more developed knowledge structures and schemas this allows them to unify superﬁcially disparate information and to process more fragmentary information. Empirical evidence shows that habitual entrepreneurs identify more opportunities in a given period than novices and more innovative opportunities (Ucbasaran, Westhead, & Wright, 2008b). Not all opportunities that are identiﬁed are pursued. However, prior ownership experience may increase the likelihood of pursuing opportunities. Experience and learning reduce the costs of exploitation. Entrepreneurs may increase their expected return from the pursuit of opportunities and reduce the associated uncertainties due to the skills and knowledge they gain from prior experience. Experience may also allow the identiﬁcation of better quality opportunities. It may also enable better and cheaper access to resources required. The skills accumulated from prior experience better able entrepreneurs to seek out resources and to have better developed social capital that enables them to access these resources. Evidence shows that habitual entrepreneurs do indeed pursue more opportunities in a given period than novice entrepreneurs (Ucbasaran et al., 2008a). Entrepreneurial experience also contributes to the development of knowledge structures. As knowledge structures become richer with experience, entrepreneurs are able to engage in faster and more effective information processing. In turn, this reduces the burden on cognitive processing and allows for greater attention to novel and unique information. Alternatively, as a result of prior experience, entrepreneurs may become more risk averse and be less inclined to pursue innovative ventures. This relationship may be non-linear. Habitual entrepreneurs do indeed report higher levels of innovation than novice entrepreneurs. It is not just the extent of prior entrepreneurial experience that matters, however. Rather, the nature of prior experience may inﬂuence learning. Learning from failure may be different from learning from prior success. Whether entrepreneurs previously succeeded or failed may inﬂuence their optimism which in turn inﬂuences not only whether they decide to continue

Entrepreneurial Mobility

145

with entrepreneurship but also what kind of entrepreneurship if they do continue. The nature of the sequence of previous failures and successes may also be inﬂuential. Failure may lead to risk aversion and less innovativeness but if it is punctuated by success (or vice versa) the effect may be different. A further dimension concerns whether prior experience was conducted as a serial entrepreneur or as a portfolio entrepreneur (Ucbasaran et al., 2009, 2010). For portfolio entrepreneurs, a failure within a larger portfolio may be less concerning than it is for serials and therefore they learn and adapt from it. For serial entrepreneurs, prior failure is likely to be more emotionally important as they had only one current venture. As a result, serial entrepreneurs may adopt coping strategies and rationalize failure as being due to external factors (attribution theory). The recent studies that have examined these issues have provided interesting insights into habitual entrepreneurship behaviour but a number of important issues remain unanswered. Although habitual entrepreneurs are sometimes viewed as being cognitively unable to help themselves in terms of pursuing new opportunities, there is little direct analysis of what determines whether habitual entrepreneurs continue to exploit new opportunities. In particular, the juxtaposition of cognitive aspects and resource constraints has not been explored. We also know little about the processes by which habitual entrepreneurs acquire or develop resources and whether these processes change between earlier and later ventures. Relatedly, the mode adopted to pursue ventures may change as between acquisition of an existing venture versus creation/start-up (Ucbasaran, Wright, & Westhead, 2003b). It is recognized that much entrepreneurship takes place in teams and that the nature of teams may change as a venture develops (Ucbasaran, Lockett, Wright, & Westhead, 2003a). What is not known is whether and how habitual entrepreneurs change their founding teams between ventures to provide different mixes of human and social capital resources that they perceive are needed. Further, there has been some qualitative examination of the genaeologies of habitual entrepreneurs’ portfolios (Rosa, 1998). However, there is little systematic evidence on how habitual entrepreneurs conﬁgure their portfolio of ventures, including how governance mechanisms are developed to oversee these ventures.

Quadrant 2: Low Geographical and High Organizational Mobility Traditional research that ﬁts in this quadrant has focused on corporate spinoffs (Phan et al., 2009) and employees who leave their current employer

146

MIKE WRIGHT

(Agarwal et al., 2004; Klepper & Sleeper, 2005) to establish a venture. In this quadrant we identify two new forms of entrepreneurial mobility, management buyouts of divisions of larger corporations and spin-offs from universities. Management buyouts (MBOs) of a divested division or subsidiary of a corporation involve the acquisition by a newly created company in which the existing management takes a substantial proportion of the equity. As such, the individuals involved move away from the protection or constraints afforded by the parental corporate umbrella. A related form involves management buyins (MBIs) where a new management team with experience in other organizations buys into the venture that is being divested. MBOs and MBIs usually, but not always, require the support of a private equity ﬁrm and extensive borrowings. The former parent may retain an equity stake, perhaps to support a continuing trading relationship. MBOs typically involve a small group of senior managers as equity-holders, who in smaller transactions, likely obtain the majority of the voting equity (Wright & Bruining, 2008). Wright et al. (2000) integrate insights from agency theory and cognition perspectives to present a conceptual typology of buyout types. In efﬁciency buyouts, executives with a managerial mindset are expected to respond positively to enhanced monetary incentives and control mechanisms designed to reduce agency costs by improving efﬁciencies in mature ﬁrms with stable cash ﬂows. In revitalization buyouts, the change of ownership creates the discretionary power for the newly independent management team to decide what is best for the business and to engage in catch-up innovations which have been frustrated by the former corporate owner. Meuleman, Amess, Wright, and Scholes (2009) argue that a limitation of the agency perspective is that it under-emphasizes the upside potential of MBOs. An entrepreneurial perspective recognizes that incumbent management has capabilities with respect to identifying and exploiting growth opportunities. Before an MBO, managers might be unable or unwilling to utilize their knowledge to achieve value creating growth. Management in buyouts may not merely respond to greater incentives. Rather, managers may have an entrepreneurial mindset that enables them to perceive entrepreneurial opportunities; however, such opportunities cannot be realized within the existing ownership and control structure (Wright et al., 2000). In entrepreneurial buyouts, management with an entrepreneurial cognition or mindset, who adopt heuristic-based decisionmaking may be able to pursue entrepreneurial opportunities that they

Entrepreneurial Mobility

147

have identiﬁed, but which could not previously be pursued within a larger group. Evidence from divisional buyouts shows signiﬁcant increases in corporate entrepreneurship, including new product development, better use of research and development, and increased patent citations (Zahra, 1995; Lerner, Stro¨mberg, & Sørensen, 2008). These Post-MBO ﬁrms do achieve growth and the experience of private equity ﬁrms aids this achievement (Meuleman et al., 2009). The heterogeneity of private equity ﬁrms involved in MBOs is beginning to be examined, in terms of prior sector experience and intensity of monitoring models. However, the processes by which management and their private equity ﬁrm backers identify and exploit the new opportunities facilitated by the buyout are little understood. Similarly, while the Wright et al. (2000) typology has provided a conceptual distinction between managers in buyouts who have a managerial or an entrepreneurial mindset, empirical evidence exploring this distinction is lacking. Spin-offs from universities represent a second aspect of entrepreneurial mobility in this quadrant. University spin-offs involve the creation of a new venture involving intellectual property generated by a university scientist in which the scientist takes an equity stake and an active role (Wright, Clarysse, Mustar, & Lockett, 2007). In contrast to start-ups and corporate spin-offs, university spin-offs involve mobility from a traditionally noncommercial context and the academic may or may not leave their academic post to become the entrepreneur. These academic entrepreneurs need to access the human and social capital they need to recognize and exploit opportunities but face major challenges because while they possess high levels of scientiﬁc knowledge and social capital, the commercial dimensions of these are typically absent. Academic entrepreneurs therefore need to acquire and develop commercial knowledge and social capital and shift away from the path dependencies of the non-commercial university environment. External surrogate entrepreneurs with commercial knowledge and capabilities to engage in the startup process may be recruited, if the university has the network of contacts to identify such individuals (Franklin, Wright, & Lockett, 2001). Also, some academic entrepreneurs do have prior entrepreneurial experience and have built up commercial social capital and so may be better able to identify opportunities (Mosey & Wright, 2007). However, at present, we know little about the cognitive challenges that arise in creating and developing successful university spin-offs. There is also a need for analysis of the capabilities and competencies that are required of academic entrepreneurs and how these are acquired. How academic

148

MIKE WRIGHT

entrepreneurs build commercial legitimacy and the social capital they need is also an area that is little understood. For example, how do they recruit team and board members and how do they develop alliances with corporate trading partners?

Quadrant 3: High Geographical and Low Organizational Mobility Traditional research that ﬁts this quadrant has focused on immigrant and ethnic entrepreneurs who have moved from their home country to a host country. Returnee entrepreneurs as a new form of mobility ﬁt this quadrant and are scientists and engineers returning to start up a new venture in their native countries, after several years of business experience and/or education in OECD countries (Filatotchev, Liu, Lu, & Wright, 2011; Saxenian, 2006; Wright, Liu, Buck, & Filatotchev, 2008). Returnee entrepreneurs have acquired knowledge from their experience in OECD countries that can be transferred to an emerging economy. This knowledge may be academic knowledge from general education. Alternatively, it may involve scientiﬁc and technical training together with practical business human capital as a result of working in a commercial environment or from starting a business in a developed economy. Through being exposed to education and commercial experience, returnee entrepreneurs have likely also acquired social capital in the host country which enables them to obtain access to diverse sources of knowledge when they become a returnee entrepreneur. This knowledge, which is embodied in individual returnee entrepreneurs, adds to strategic entrepreneurship insights into how resources and capabilities are accessed. Returnee entrepreneurs in emerging economies are expected to be more innovative than non-returnee ﬁrms (Liu, Lu, Filatotchev, Buck, & Wright, 2010). They may also affect the innovative activities of non-returnee ﬁrms through knowledge spillovers that help to enhance the technological capabilities of local ﬁrms (Audretsch & Lehmann, 2005). These innovative activities are likely to occur as long as the non-returnee ﬁrm has the absorptive capacity to assimilate it. The greater the technology gap experienced by the non-returnee ﬁrm in relation to international developed economy industry standards, the greater the effect of returnees. Although interest in returnee entrepreneurs has been developing, much of this research has been focused at the ﬁrm level. We know little about the cognitive processes engaged in by returnee entrepreneurs in terms of identifying opportunities in their home countries. We also know little about

Entrepreneurial Mobility

149

how they assemble the resources and capabilities to create their ventures in their home countries and how and whether this differs from the entrepreneurial behaviour of non-returnee entrepreneurs in their home country. Further, analysis is needed of whether returnee entrepreneurs set out explicitly to gain experience abroad before returning to their home country or whether they decide to return home after failing to establish themselves in the host country. A potentially important difference concerns the role of social and political capital, which are recognized as key to doing business in emerging economies (Hoskisson et al., 2000). Returnee ﬁrms may be able to transfer knowledge effectively due to their embeddedness in both environments of home and host countries which enable them to avoid cultural incompetence and lacking local networks (Lin, 2010). However, a downside of having been abroad is that returnee entrepreneurs may have they seen their critical political social capital depreciated. If this is the case, how is this problem addressed? Are returnee entrepreneurs able to rebuild this capital or do they seek substitutes and, if so, what form do they take? For example, given the importance of extended families in many emerging economies, returnee entrepreneurs may be able to use the channels provided by this aspect of social capital. This in itself may not be straightforward, particularly where returnees bring approaches that are in conﬂict with traditional family values. An alternative approach concerns the extent to which returnee entrepreneurs develop founding teams comprising local managers with the requisite social capital. This raises important issues relating to the processes by which returning entrepreneurs identify and integrate potential team members which need to be explored.

Quadrant 4: High Geographical and High Organizational Mobility Traditional research in this quadrant relates to studies that have examined the relocation of entrepreneurial ﬁrms to other locations (Cumming, Fleming, & Schwienbacher, 2009a). Transnational entrepreneurship involves the formation and maintenance of business ﬁrms by entrepreneurs whose activities span home and host countries (Drori, Honig, & Wright, 2009). Although much of the attention in international entrepreneurship research has focused on the ﬁrm-level, the characteristics and behaviours of the individual entrepreneur play a key role IE (Coviello & Jones, 2004; Cumming et al., 2009b). There is, therefore a need for further examination of individual transnational entrepreneurs.

150

MIKE WRIGHT

Transnational entrepreneurs establish businesses utilizing social capital and networks to exploit opportunities in their new and former homes (Light, 2007). These actions require embeddedness in the economic and social life of at least two culturally and geographically dispersed environments (Yeung, 2002). How transnational entrepreneurs are able to access resources and capabilities by embedding themselves in these multiple settings and actively shaping, modifying and reinforcing those domains is little understood (Patel & Conklin, 2009). Further research is needed to examine the degree of embeddedness of transnational entrepreneurs in different social spaces. Little is known about the aptitude and cognitive attributes for exploring and exploiting business opportunities in international spaces. How transnational entrepreneurs conﬁgure their home and host country activities is little studied. In designing their organizations, transnational entrepreneurs may envision that each serves a different function. For example, one location may serve as exploration activity while the other provides an exploitation role. The determinants of this choice need to be examined. For example, to be effective, transnational entrepreneurs may need to develop the ability to minimize or exploit geographical, social, cultural, economic and political disparities in these international spaces. How they do this, and the challenges they face in so doing, have not been addressed. The inﬂuence of competitive international forces on their practices also needs to be considered.

DISCUSSION Drawing on strategic entrepreneurship, entrepreneurial behaviour theory and resource based theory, this chapter has set out a systematic framework for the analysis of entrepreneurial mobility. This chapter makes contributions to several literatures. First, we extend the emerging strategic entrepreneurship literature. By focusing on the inﬂuence of movement across different contexts on the nature of opportunity recognition and the types of resources needed in these different contexts, we add dynamic and contextual aspects that help delineate the boundaries of this perspective. Second, by focusing on individual behaviour we contribute to the emerging literature on the micro-foundations of strategy and entrepreneurship (Barney, Ketchen, & Wright, 2011). Third, we extend the literature on human capital mobility across national boundaries. This literature has tended to focus on the mobility of employees in multinational corporations but we draw attention to the role of individual

151

Entrepreneurial Mobility

entrepreneurial mobility. Fourth, we have extended the limited literature on entrepreneurial mobility by providing a systematic framework for analysis that enriches the extent and types of entrepreneurial mobility. The analysis presented in this chapter has a number of implications for research methods.

Datasets Archival datasets may be particularly important for examining entrepreneurial mobility. As an over-arching implication, it is clear that studies using general datasets of entrepreneurial ventures and owner-managed ﬁrms may have failed to account for unobserved heterogeneity of the owners of these ventures that includes mobility. For example, these datasets likely contain ﬁrst time (novice) entrepreneurs, who may have acquired, inherited or created their venture, as well as experienced entrepreneurs who have owned multiple ventures either sequentially or concurrently. These datasets may also contain ﬁrms and founders involved in spin-offs from corporations and universities. Similarly, datasets of entrepreneurs in emerging economies may contain both local entrepreneurs and returnee entrepreneurs. Studies that have focused on more speciﬁc types of entrepreneurs and entrepreneurial ventures may also contain hitherto unobserved heterogeneity. For example, datasets of immigrant and ethnic entrepreneurs may contain some entrepreneurs who are transnational entrepreneurs. As such, studies of immigrant and ethnic entrepreneurs may be misspeciﬁed (Drori et al., 2009). More general studies of international entrepreneurship that consider the international experience of the entrepreneurs involved may need to include more ﬁne-grained measures that distinguish not just prior international experience as employees in multinational corporations or as entrepreneurs, but also distinguish whether these entrepreneurs have moved from or to a particular country context. Datasets involving university spin-offs also need to recognize the heterogeneity of spin-offs. Some spin-offs may include formal intellectual property (IP) generated by the university, such as patents. However, more informal knowledge may also be transferred into a university spin-off. Studies that omit the latter may seriously under-estimate the extent of mobility through academic entrepreneurship. Further, it is important to consider whether the academic has actually moved out of the university or

152

MIKE WRIGHT

whether they retain their position (full or part time) and have recruited a surrogate entrepreneur to run the venture (Franklin et al., 2001). With respect to analyses of buyouts, datasets need to recognize the vendor source of the transaction since not all buyouts arise on divestment. More importantly, there is a need for datasets to recognize the distinction between MBOs and MBIs. The different corporate organizational backgrounds from which these managers emerge provide different knowledge bases from their experience. Further, MBI managers likely experience important asymmetric information challenges not faced by managers in MBOs who have already been running the business, albeit within the constraints imposed by the corporate headquarters. There is also a need to distinguish MBOs and MBIs that are private equity backed from those which may only be ﬁnanced by debt, since the governance mechanisms may differ signiﬁcantly. Although commercial private equity datasets are accessible, at a price, they typically do not include non-private equity deals, have patchy and limited coverage of earlier deals especially outside the United States, and have limited distinction between the heterogeneous types of buyout. Similarly, some proprietary databases provided, say by particular private equity ﬁrms or funds of funds may be suffer from selection biases as they are skewed towards larger ﬁrms and may be accessible only on condition that ﬁrms and individuals are anonymized. However, an example of a longitudinal dataset that does address these issues is that compiled by the Centre for Management Buyout Research (CMBOR) that stretches back to the late 1970s (www.cmbor.com). Datasets relating to entrepreneurial mobility may be restricted to mobility between particular contexts. For example, Filatotchev et al. (2010) consider returnee entrepreneurs in one part of China. Alternatively, analyses using more general, global datasets covering many countries may not take into account differences between institutional contexts. There is a need for further theory development and accompanying dataset development that considers other contexts where the relationships may differ. For example, datasets might be developed to examine returning entrepreneurs in Latin America and India. Datasets might also be developed that enable comparator analysis of transnational entrepreneurs operating across emerging economies and those operating across emerging and developed economies (Wright, Filatotchev, Hoskisson, & Peng, 2005; Drori et al., 2009). Datasets of private ﬁrms may not include details of whether owners have or have had stakes in multiple ventures, their education, or previous work experience. Further, such datasets may not enable the identiﬁcation of an

153

Entrepreneurial Mobility

entrepreneur’s portfolio of ventures, whether they are habitual or transnational entrepreneurs. An alternative approach to compiling appropriate data is to access datasets of directors of private ﬁrms. In some jurisdictions, for example the United Kingdom, private company registration documents are accessible and contain information on founding directors, changes in directors, and director characteristics and their experience (Wilson & Altanlar, 2011). As a second example, in Sweden the LISA database maintained by Statistics Sweden provides data on individual founders, including annual data on education, employment, changes in employment. This dataset also enables data on whether individuals have transferred directly from employment in a university or from a private corporation to enable university and corporate spin-offs to be identiﬁed (Wennberg, Wiklund, & Wright, 2010). In both these examples from the United Kingdom and Sweden, individual data can be matched with ﬁrm-level datasets. Research design that seeks to compare different forms of entrepreneurial mobility within a single dataset, therefore, may need to consider the development of novel datasets that integrate data from different sources to enable the different forms of entrepreneurial mobility to be identiﬁed. For example, Wilson, Wright, and Altanlar (2010) combine UK Companies House data with the CMBOR data on management buyouts and Harris, Siegel, and Wright (2005) combine the UK government’s establishment level data with the CMBOR data to enable panel analysis of the population of buyouts in the context of the national population. Clearly there is a need to be able to identify ﬁrms within each dataset to enable them to be matched, itself a signiﬁcant challenge and not always feasible. Constructing and maintaining such proprietary databases over time is a major and costly task, for which some form of sponsorship may be required but which may place constraints on what data can be collected and what can be done with it.

Constructs Further research is needed to reﬁne and sharpen the entrepreneurial mobility construct developed in this chapter and address issues of construct validity. The multi-dimensional aspects of the construct need to be identiﬁed with respect to such items as the range from local to cross-national mobility, the organizational mode of mobility in terms of start-up, spin-off and acquisition, the range of entrepreneurial opportunities from local to foreign, and the range of resources from local to foreign.

154

MIKE WRIGHT

Our discussion has highlighted cognitive aspects of entrepreneurial behaviour within each of the four quadrants. Further research needs to compare the cognitive dimensions of entrepreneurs in each quadrant to identify differences and similarities. For example, to what extent are there cognitive differences in opportunity identiﬁcation and in learning? To what extent do entrepreneurs engaging in different forms of mobility adopt different heuristics? Studies could be designed to examine these issues using well-established measures (Baron, 2004). For expositional purposes, geographical and organizational mobility were dichotomized into high and low levels. However, within each category there may be degrees of mobility which need to be explored. Also for expositional purposes, each quadrant was treated as independent. Yet, there may be mobility between quadrants and there may also be interactions. For example, habitual entrepreneurs may be returnees or transnational entrepreneurs. Entrepreneurs that spin-off or buyout from other organizations may become transnational entrepreneurs if, say, they acquire their venture from a multinational corporation and decide to internationalize. Further studies could usefully be designed to analyse the extent and determinants of these interactions. The boundary conditions for entrepreneurial mobility and the determinants of those boundaries also need to be explored both theoretically and empirically. To what extent do entrepreneurs continue to engage in entrepreneurial mobility? Analyses also need to examine whether mobility is a linear or non-linear process of development? Such analysis might also consider the factors inﬂuencing exit or retrenchment from mobility.

Quantitative Analysis Methods Analysing entrepreneurial mobility using quantitative analysis presents considerable opportunities and challenges. We need to know more about the different proﬁle of the entrepreneurs engaging in different forms of entrepreneurial mobility and the ﬁrms they develop in terms of their demographic and human capital backgrounds, such as prior managerial and entrepreneurial experience. Multivariate analyses of, say, ﬁrm performance could simply adopt dummy variables to denote different forms of entrepreneurial mobility. Multinomial logistic estimation methods, or discriminant analysis, could be used to examine differences between variables relating to each type

155

Entrepreneurial Mobility

of entrepreneurial mobility. More sophisticated analyses could adopt methods such as canonical correlation analysis to examine the differences between variables associated with each type of entrepreneurial mobility.

Theory Building Understanding of the processes by which entrepreneurial occurs remains fragmented. To elaborate insights regarding the processes of different types of entrepreneurial mobility, we envision considerable scope for theory building using comparative case studies from each quadrant. We identiﬁed human capital resources and capabilities, social capital and corporate governance as key dimensions of a resource-based theory approach to understanding entrepreneurial mobility. Theory building research could usefully help shed light on how these resources and capabilities are accessed and evolved during the processes involved in the different forms of entrepreneurial mobility.

CONCLUSIONS The framework developed in this chapter provides a basis for the development of a new research agenda that integrates and extends the hitherto fragmented research on entrepreneurial mobility. The research program embodied in the framework makes major data demands that are unlikely to be accomplished in one all-encompassing dataset but a number of areas have been identiﬁed where datasets can be reﬁned to avoid misspeciﬁcation and under-reporting problems. It is to be hoped that other scholars will be able to build on the conceptual and methodological insights presented here to enhance our understanding of the increasingly important phenomenon of entrepreneurial mobility.

ACKNOWLEDGMENTS Thanks to Don Bergh and Dave Ketchen and to participants at seminars at Syracuse University, the 2010 RENT conference in Maastricht and at the Micro-Foundations of Organizational Design Conference, Copenhagen Business School, 2010 for comments on earlier versions.

156

MIKE WRIGHT

REFERENCES Agarwal, R., Echambadi, R., Franco, A., & Sarkar, M. (2004). Knowledge transfer through inheritance: Spin-out generation, development and survival. Academy of Management Journal, 47, 501–522. Ahuja, G., & Katila, R. (2004). Where do resources come from? The role of idiosyncratic situations. Strategic Management Journal, 25, 887–907. Arregle, J.-L., Hitt, M. A., Sirmon, D. G., & Very, P. (2007). The development of organizational social capital: Attributes of family ﬁrms. Journal of Management Studies, 44, 72–95. Audretsch, D., & Lehmann, E. (2005). Does the knowledge spillover theory of entrepreneurship hold for regions? Research Policy, 34(8), 1191–1202. Barney, J. B., Ketchen, D., & Wright, M. (2011). The future of resource-based theory: Revitalization or decline? Journal of Management, forthcoming. Barney, J. B., Wright, M., & Ketchen, D. J. (2001). The resource based view of the ﬁrm: ten years after 1991. Journal of Management, 27, 625–643. Baron, R. A. (2004). The cognitive perspective: A valuable tool for answering entrepreneurship’s basic ‘Why’ questions. Journal of Business Venturing, 19(2), 221–239. Coff, R., & Kryscynski, D. (2011). Drilling for micro-foundations of human capital based competitive advantages. Journal of Management, forthcoming. Coviello, N. E., & Jones, M. (2004). Methodological issues in international entrepreneurship research. Journal of Business Venturing, 19, 485–508. Cumming, D., Fleming, G., & Schwienbacher, A. (2009a). Corporate relocation in venture capital ﬁnance. Entrepreneurship Theory & Practice, 33, 1121–1154. Cumming, D., Sapienza, H., Siegel, D., & Wright, M. (2009b). International entrepreneurship: Managerial and policy implications. Strategic Entrepreneurship Journal, 3, 283–296. Davidsson, P., & Honig, B. (2003). The role of social and human capital among nascent entrepreneurs. Journal of Business Venturing, 18, 301–331. Drori, I., Honig, B., & Wright, M. (2009). Transnational entrepreneurs. Entrepreneurship Theory and Practice, 33, 1001–1022. Filatotchev, I., Liu, X., Lu, J., & Wright, M. (2011). Knowledge spillovers through human mobility across national borders: Evidence from Zhongguancun Science Park in China. Research Policy, forthcoming. Foss, N. (2011). Micro-foundations for the resource-based view? Journal of Management, forthcoming. Franklin, S., Wright, M., & Lockett, A. (2001). Academic and surrogate entrepreneurs in university spin-out companies. Journal of Technology Transfer, 26, 127–141. Go¨rg, H., & Strobl, E. (2005). Spillovers from foreign ﬁrms through worker mobility: An empirical investigation. Scandinavian Journal of Economics, 107(4), 693–709. Groysberg, B., Nanda, A., & Prats, M. J. (2007). Does individual performance affect entrepreneurial mobility? Empirical evidence from the ﬁnancial analysis market, NBER Working Paper 13633. Harris, R., Siegel, D. S., & Wright, M. (2005). Assessing the impact of management buyouts on economic efﬁciency: plant-level evidence from the United Kingdom. Review of Economics and Statistics, 87, 148–153. Harvey, M., & Moeller, M. (2009). Expatriate managers: A historical review. International Journal of Management Reviews, 11(3), 275–296.

Entrepreneurial Mobility

157

Hoang, H., & Antoncic, B. (2003). Network-based research in entrepreneurship: A critical review. Journal of Business Venturing, 18, 165–187. Hoskisson, R., Eden, L., Lau, C. M., & Wright, M. (2000). Strategies in emerging economies. Academy of Management Journal, 43(3), 249–267. Ireland, R. D., Hitt, M. A., & Sirman, D. G. (2003). A model of strategic entrepreneurship: The construct and its dimensions. Journal of Management, 29(6), 963–989. Iyer, S., Kitson, M., & Toh, B. (2005). Social capital, economic growth and regional development. Regional Studies, 39, 1015–1040. Kalnins, A., & Chung, W. (2006). Social capital, geography, and survival: Gujarati immigrant entrepreneurs in the U.S. lodging industry. Management Science, 52(2), 233–247. King, A. W., Garbuio, M., & Lovallo, D. (2011). Looking inside: Psychological inﬂuences on structuring a ﬁrm’s portfolio of resources. Journal of Management, forthcoming. Klepper, S., & Sleeper, S. (2005). Entry by spinoffs. Management Science, 51, 1291–1306. Lerner, J., Stro¨mberg, P., & Sørensen, M. (2008). Private equity and long-run investment: the case of innovation. In: Lerner, J., & Gurung, A. (Eds.), The Global Impact of Private Equity Report 2008, globalization of alternative investments, Working Papers (Vol. 1., pp. 27–42). Davos, Switzerland: World Economic Forum. Levie, J. (2007). Immigration, in-migration, ethnicity and entrepreneurship: Insights from the GEM UK database. Small Business Economics, 28(2), 143–169. Light, I. (2007). Global entrepreneurship and transnationalism. In: L. P. Dana (Ed.), Handbook of Research in Ethnic Minority Entrepreneurship (pp. 3–15). Cheltenham: Edward Elgar. Lin, X. (2010). The diaspora solution to innovation capacity development: Immigrant entrepreneurs in the contemporary world. Thunderbird International Business Review, 52(2), 124–136. Liu, X., Lu, J., Filatotchev, I., Buck, T., & Wright, M. (2010). Returnee entrepreneurs, knowledge spillovers and innovation in high-tech ﬁrms in emerging economies. Journal of International Business Studies, 41, 1183–1197. Meuleman, M., Amess, K., Wright, M., & Scholes, L. (2009). Agency, strategic entrepreneurship, and the performance of private equity-backed buyouts. Entrepreneurship Theory and Practice, 33(1), 213–239. Meuleman, M., Lockett, A., Manigart, S., & Wright, M. (2010). Partner selection decisions in interﬁrm collaborations: The paradox of relational embeddedness. Journal of Management Studies, 47, 995–1019. Mosey, S., & Wright, M. (2007). From human capital to social capital: A longitudinal study of technology-based academic entrepreneurs. Entrepreneurship Theory and Practice, 31, 909–935. Nijkamp, P. (2003). Entrepreneurship in a modern network economy. Regional Studies, 37, 395–405. Patel, P., & Conklin, B. (2009). The balancing act: The role of transnational habitus and social networks in balancing transnational entrepreneurial activities. Entrepreneurship Theory & Practice, 33, 1045–1078. Pe’er, A., & Vertinsky, I. (2008). Firm exits as a determinant of new entry: Is there evidence of local creative destruction? Journal of Business Venturing, 23, 280–306. Phan, P., Wright, M., Ucbasaran, D., & Tan, W-L. (2009). Corporate entrepreneurship: Current research and future directions. Journal of Business Venturing, 24, 197–205.

158

MIKE WRIGHT

Rasmussen, E., Mosey, S., & Wright, M. (2011). The evolution of entrepreneurial competencies: A longitudinal study of university spin-off venture emergence. Journal of Management Studies, forthcoming. Rosa, P. (1998). Entrepreneurial processes of business cluster formation and growth by ‘‘habitual’’ entrepreneurs. Entrepreneurship: Theory and Practice, 22, 43–61. Saxenian, A. (2006). The new Argonauts: Regional Advantage in a Global Economy. Cambridge, MA: Harvard University Press. Schulze, W., & Gedajlovic, E. R. (2010). Whither family business? Journal of Management Studies, 47(1), 191–204. Schulze, W. S., Lubatkin, M. H., & Dino, R. N. (2003). Exploring the agency consequences of ownership dispersion among the directors of private family ﬁrms. Academy of Management Journal, 46(2), 179–194. Sequiera, J. M., & Rasheed, A. (2006). Start-up and growth of immigrant small businesses: The impact of social and human capital. Journal of Developmental Entrepreneurship, 11(4), 357–375. Shane, S. A., & Venkataraman, S. (2000). The promise of entrepreneurship as a ﬁeld of research. Academy of Management Review, 26, 217–226. Sirmon, D. G., Hitt, M. A., & Ireland, R. D. (2007). Managing ﬁrm resources in dynamic environments to create value: Looking inside the black box. Academy of Management Review, 32(1), 273–292. Song, J., Almeida, P., & Wu, G. (2003). Learning by hiring: When is mobility more likely to facilitate inter-ﬁrm knowledge transfer? Management Science, 49(4), 351–365. Stuart, T., & Sorenson, O. (2003). The geography of opportunity: Spatial heterogeneity in founding rates and the performance of biotechnology ﬁrms. Research Policy, 32, 229–253. Ucbasaran, D., Alsos, G., Westhead, P., & Wright, M. (2008a). Habitual Entrepreneurs. Foundations and Trends in Entrepreneurship, 4(4), 1–93. Ucbasaran, D., Lockett, A., Wright, M., & Westhead, P. (2003a). Entrepreneurial founder teams: Factors associated with team member entry and exit. Entrepreneurship Theory and Practice, 28(2), 107–128. Ucbasaran, D., Westhead, P., & Wright, M. (2008b). Opportunity identiﬁcation and pursuit: Does an entrepreneur’s human capital matter? Small Business Economics, 30(2), 153–173. Ucbasaran, D., Westhead, P., & Wright, M. (2009). The extent and nature of opportunity identiﬁcation by experienced entrepreneurs. Journal of Business Venturing, 24(2), 99–115. Ucbasaran, D., Westhead, P., Wright, M., & Flores, M. (2010). The nature of entrepreneurial experience, business failure and comparative optimism. Journal of Business Venturing, 25, 541–555. Ucbasaran, D., Wright, M., & Westhead, P. (2003b). A longitudinal study of habitual entrepreneurs: Starters and acquirers. Entrepreneurship and Regional Development, 15, 207–228. Uzzi, B. (1997). Social structure and competition in interﬁrm networks: The paradox of embeddedness. Administrative Science Quarterly, 42, 35–67. Wennberg, K., Wiklund, J., & Wright, M. (2010). The effectiveness of university knowledge spillovers: Performance differences between university spin-offs and corporate spin-offs. Paper presented at the Bayh-Dole Act 30 Years On Conference, Washington, DC, October.

Entrepreneurial Mobility

159

Wiklund, J., & Shepherd, D. (2008). Portfolio entrepreneurship, habitual and novice founders, new entry and mode of organizing. Entrepreneurship Theory and Practice, 32, 701–725. Wilson, N., & Altanlar, A. (2011). The survival of newly incorporated companies and the impact of director (board) characteristics. Leeds: CMRC, Leeds University Business School. Wilson, N., Wright, M., & Altanlar, A. (2010). Private equity, buyouts, leverage and failure. CMBOR Working Paper. Wright, M., & Bruining, J. (2008). Private equity and management buyouts. Cheltenham: Edward Elgar. Wright, M., Clarysse, B., Mustar, P., & Lockett, A. (2007). Academic entrepreneurship in Europe. Cheltenham: Edward Elgar. Wright, M., Filatotchev, I., Hoskisson, R., & Peng, M. (2005). Strategy research in emerging economies: Challenging the conventional wisdom. Journal of Management Studies, 42, 1–34. Wright, M., Hoskisson, R., Busenitz, L., & Dial, J. (2000). Entrepreneurial growth through privatization: The upside of management buy-outs. Academy of Management Review, 25(3), 591–601. Wright, M., Liu, X., Buck, T., & Filatotchev, I. (2008). Returnee entrepreneurs, science park location choice and performance: An analysis of high-technology SMEs in China. Entrepreneurship Theory and Practice, 32(1), 131–156. Yeung, H. (2002). Entrepreneurship in international business: An institutional perspective. Asia Paciﬁc Journal of Management, 19, 29–61. Zahra, S. (2010). Harvesting family ﬁrms’ organizational social capital: A relational perspective. Journal of Management Studies, 47, 345–366. Zahra, S. A. (1995). Corporate entrepreneurship and ﬁnancial performance: The case of management leveraged buyouts. Journal of Business Venturing, 10, 225–247. Zahra, S., Filatotchev, I., & Wright, M. (2009). How do threshold ﬁrms sustain corporate entrepreneurship? The role of boards and absorptive capacity. Journal of Business Venturing, 24, 248–260.

PART II BRIDGES TO STRONGER DESIGNS AND ANALYSIS

MULTI-STUDY PACKAGES IN ORGANIZATIONAL SCIENCE RESEARCH Wayne A. Hochwarter, Gerald R. Ferris and T. Johnston Hanes ABSTRACT Purpose – The purpose of this chapter is to examine the frequency of multi-study research packages in the organizational sciences and advocate for their use by detailing strengths and recognizing limitations. Methodology/approach – Philosophy of science research, focusing on multi-study research packages, is discussed followed by a 20-year review of incidence of these packages in top organizational sciences journals. Findings – The publication of multi-study research packages have increased over the past 10 years, most notably in micro-level journals. Social implications – For reasons of validity and generalizability, society beneﬁts if scholars adopt multi-study approaches to knowledge generation and disseminate. Originality/value of the chapter – This chapter provides the most comprehensive review of multiple-study research packages in the organizational sciences to date, examining publication trends in eight leading micro-and Building Methodological Bridges Research Methodology in Strategy and Management, Volume 6, 163–199 Copyright r 2011 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-8387/doi:10.1108/S1479-8387(2011)0000006005

163

164

WAYNE A. HOCHWARTER ET AL.

macro-level journals. We also summarize the use of multi-study packages in our own research and offer recommendations for improving the science of replication. Keywords: Falsiﬁcation; multidisciplinary; multi-study; replication; theory development; validity Contributing meaningfully to the organizational sciences requires sound theory as well as the inclusion of well-designed and executed empirical research (Ferris, Hall, Royle, & Martocchio, 2004). Seemingly, the objective is to build a more comprehensive and informed understanding of workplace phenomena as well as explicate the inherently complex interrelationships that occur in dynamic contexts. The development of collective knowledge, however, is contingent upon the belief that the results of prior empirical studies are valid, replicable, and thus generalizable to other samples, settings, and construct operationalizations (Hendrick, 1990; Sidman, 1960). Unfortunately, this ‘‘belief’’ is taken almost completely on faith and assumption rather than scholarly conﬁrmation. In reality, replications and veriﬁcations of research ﬁndings, across disciplines, rarely are published in inﬂuential journals (Easley, Madden, & Dunn, 2000; Mezias & Regnier, 2007; Neuliep & Crandall, 1991), preventing empirical generalizability and knowledge creation (Galtung, 1967; Lindsay & Ehrenberg, 1993). Instead, editor (and reviewer) often-documented preferences for ‘‘path-breaking’’ ﬁndings have rendered replications and multi-study investigations undeserving of competitively sought-after journal space (Evanschitzky, Baumgarth, Hubbard, & Armstrong, 2007; Singh, Ang, & Leong, 2003). In many cases, what was originally viewed as path-breaking assumes ‘‘ﬂavor of the volume’’ status, virtually ignored after its initially presentation to the ﬁeld. In view of this reality, it becomes obvious that only a few of the widely accepted tenets guiding research activity over the past 50 years have been comprehensively endorsed through replication (i.e., falsiﬁcation) (McKinley, 2010; Tsang & Kwan, 1999). As a result, the organizational sciences have been built by a knowledge base consisting largely of single-study investigations (Allen & Priess, 1993). This state of affairs stands in stark contrast to basic prescriptions for establishing knowledge in the hard sciences (Hedges, 1987; Tweney, 2008), which mandates that scholars conduct multiple replications with carefully chosen samples before generalizations are presented and accepted in their respective ﬁelds (Williams, Frankel, Campbell, & Deci, 2003). Contrary to the objectionable qualities of replications held by journal monitors, scholars maintain that non-veriﬁed phenomena generate little

Multi-Study Research

165

substantive conﬁdence, even when robust results are demonstrated. Stated more emphatically, Lindsay and Ehrenberg (1993) argued that isolated social science investigations are, for all practical purposes, meaningless in terms of advancing knowledge. Instead, these studies should be subjected to consternation given their inability to consider the myriad of extraneous factors that extend beyond the scope of the initial investigation (Guttman, 1985). As a result of this artiﬁcial limitation, the checks and balances associated with replication/extension studies remain largely dormant in most research protocols (Gravetter & Forzano, 2008). We propose multi-study research packages as an antidote to the overreliance on single-study research designs (McKinley, 2010). By deﬁnition, such research packages involve more than one study arranged within a single article, often employing multiple methodologies and theoretical explanations (e.g., Munyon, Hochwarter, Perrewe´, & Ferris, 2010). In terms of beneﬁts to science, adopting this approach allows for heightened awareness of the interplay between theory and method (Van Maanen, Sorensen, & Mitchell, 2007), as replications/extensions mandate both within and across explanations of organizational phenomena (Srivastava, Locke, Judge, & Adams, 2010). In addition, this approach addresses many of the methodological and philosophical concerns that have limited the inclusion of verifying studies in past research (Mezias & Regnier, 2007). As an example of a complementary approach that we expound, Rosenthal (1990) proposed the use of a ‘‘replication battery,’’ which consists of multiple verifying studies, each reﬂecting varying degrees of dissimilarity to the original focal investigation. Ostensibly, this collective chain of studies would strengthen conﬁdence in the validity of the initial results if substantiated and offer the opportunity for thoughtful discussions if unanticipated (or disconﬁrming) results were to surface. This chapter, in which we hope to develop support for the consideration of multi-package research studies, unfolds as follows. First, replication/ extension studies are deﬁned, and their long-standing role in the organizational sciences research is explicated. Second, we identify documented limitations that have encumbered the use of multi-study investigations. Third, we describe the critical role of theory when developing multi-study investigation, brieﬂy note well-established replication taxonomies present across scholarly domains, and discuss the role of replication in meta-analysis (and vice versa). Fourth, we consider prior studies that have documented the incidence of replication activity. Fifth, we follow with a review of leading organizational science journals to determine the publication frequency of multi-study investigations; following which, we describe our own programs of

166

WAYNE A. HOCHWARTER ET AL.

research that has adopted this approach. The role of multi-study research packages in addressing past criticisms of replication or extension research is discussed, as are concerns that warrant scholars’ attention. Caveats require brief discussion, as it is important to establish the parameters of our review and proposed recommendations. First, there are several topics relevant to replication studies that perhaps merit consideration in the chapter but are not addressed. For example, it is not our intention to include a discussion of best statistical approaches when undertaking replication/extension studies. The intricacies of comparing and contrasting results generated by potentially dissimilar methodologies warrants a chapter dedicated speciﬁcally to this topic (see, e.g., Iverson, Wagenmakers, & Lee, 2010; Killeen, 2007). Second, it is impractical to outline the various combinations of replication studies available to researchers (e.g., experiment/experiment/ﬁeld study; literal/operational/ constructive/Type 1–Type 4), nor is it pragmatic to describe unique strengths and weaknesses characteristic of each grouping. What’s best is a moving target, predicted by factors too many to list. Currently, numerous taxonomies of replication/extension exist (e.g., Kelly, Chase, & Tucker, 1979; Leavitt, Mitchell, & Peterson, 2010; Lindsay & Ehrenberg, 1993; Tsang & Kwan, 1999), ranging from largely imitative to widely divergent. Instead, we brieﬂy describe the types exhibiting the most inﬂuence on scholarly research as well as deemed most practically viable for those conducting replication/extension studies (Easley et al., 2000). Moreover, our discussion focuses predominantly on the initiation, development, and submission of multi-study research designs in the organizational sciences (i.e., behavioral and business policy/strategy domains). Other treatments have examined a wider array of disciplines, in both organizational and non-organizational settings. For example, Hubbard and Vetter (1996) documented replication incidence in accounting, ﬁnance, economics, marketing, and management journals, documenting less than modest inclusion. Moreover, Ioannidis (2003) examined the prevalence of replication in molecular medicine research settings, whereas Kelly (2006) described both the infrequency of replication and its underappreciated value in behavioral ecology research. Finally, we use the terms replication, extension, multi-sample, and conﬁrmation reciprocally, which may be viewed as improper, given that each possesses elements that provide differentiation (Toncar & Munch, 2010; Tsang & Kwan, 1999). In this chapter, however, we focus on the general similarities rather than discrete dissimilarities. Speciﬁcally, we use these terms to correspond to the systematic evaluation of prior ﬁndings to determine their

Multi-Study Research

167

legitimacy and to promote science by expanding the current domain of knowledge (Kline, 2008; Short, McKelvie, Ketchen, & Chandler, 2009).

IF AT FIRST YOU ‘‘DO’’ (OR DON’T) SUCCEED: TRY, TRY, AGAIN Replications and multi-study research packages have a long-standing and distinguished history in virtually all scientiﬁc disciplines. As an umbrella term for the scholarly development of multiple evaluative studies, deﬁnitions frequently possess similar rudiments. As selected examples, Gould and Kolb (1964) discussed replication in terms of a methodology designed to verify or conﬁrm the accuracy of documented results. Similarly, Cooper and Rosenthal’s (1980) deﬁnition describes a supplementary assessment of a previously tested hypothesis, whereas Singh et al. (2003) maintained that replication’s objective is to determine the reproducibility of an alreadypublished study. Hubbard and Armstrong (1994) classiﬁed a replication as ‘‘a duplication of a previously – published empirical study that is concerned with assessing whether similar ﬁndings can be obtained upon repeating the study’’ (p. 236). As a point of clariﬁcation, Kerlinger and Lee (2000) cautioned against the interchangeable use of the terms repetition and replication. Speciﬁcally, replication implies potential variations in subjects, contexts, and variables, whereas replica does not. Hence, positioning a study as a ‘‘replica’’ (i.e., repetition), when dissimilarities do, in fact, exist is potentially deceptive. In terms of its inﬂuence on science, Tsang and Kwan (1999) argued that replication and knowledge generation are inextricable. Also, Lamal (1991) noted that replication and multi-sample designs epitomize the core of the scientiﬁc method, critical because ‘‘our knowledge is corrigible’’ (p. 31). Despite the voluminous support for the development of multiple, authenticating (redirecting) studies (Rosenthal, 1991), replications reportedly remain rare (Umphress, Labianca, Brass, Kass, & Scholten, 2003). In fact, it can be argued that scholars have spent equal (or more) time demarcating the reasons for a lack of replication involvement as in the participation itself. As an example, Smith’s (1970) review uncovered a myriad of factors including a lack of ﬁnancial resources, time considerations, an inability to secure a comparable group of research subjects, a lack of novel research interests, researcher’s diminished interest in publishing, and ego connection with the data. Without question, many of these constraints remain

168

WAYNE A. HOCHWARTER ET AL.

inﬂuential in contemporary research environments (Ferris & Hochwarter, 2011; Toncar & Munch, 2010). However, this list is far from exhaustive (Ferris et al., 2006; Mezias & Regnier, 2007). For example, we strongly contend that many scholars ﬁnd participation in replication/extension studies bothersome due to the policies of their university’s Institutional Review Board (IRB) (Zywicki, 2007). Both internally and externally, we are aware of situations where approval of a relatively simple survey design took in excess of one year to secure. For most scholarly endeavors, organizational science research assumes organizational participation at some level (e.g., macro/micro continuum), often leaving the investigator powerless to initiate the research. Assume that an investigator is contacted by an industry colleague who seeks assistance in dealing with attitudes for employees participating in an upcoming downsizing. Given the complexities of layoff occurrences, opportunities to replicate prior work are exceedingly rare requiring painstaking supervision and execution. Furthermore, the tumult of most organizations involved in such change implies that windows for data collection may close with little forewarning or explanation. Hence, ‘‘striking while the iron is hot’’ has become an important ingredient for publishing work of scholarly and practical value, and structural impediments of IRB’s may hasten the cooling of the iron (Oakes, 2002). Of course, there is no way to verify how many important replication/extension studies have failed to take place because of the time disconnect between the researcher’s window of opportunity and the institution’s review policies. However, we argue that the number is considerably greater than what most researchers would speculate (Liddle & Brazelton, 1996). Acknowledging the barriers noted earlier, and others noted in recent research (Leavitt et al., 2010; McKinley, 2010), we choose to discuss what we (and many others) view as the constraint most strongly inﬂuencing the development and publication of multi-study research programs. Unequivocally, replication/extension studies are not viewed by research gatekeepers (i.e., editors and reviewers) and other evaluative audiences (i.e., department heads and promotion/tenure committees) as valuable, often abruptly dismissed as stale, boring, and redundant (Lindsay & Ehrenberg, 1993). In support of such contempt, replications have been further described as substandard (Umapathy, 1987), irrelevant (Campbell, 1986), unoriginal (Hubbard & Armstrong, 1994), denigrating (Hubbard & Lindsay, 2002), and unable to offer science ‘‘cutting edge stuff’’ (Neuliep & Crandall, 1991, p. 88). With these labels innately linked to replication/extension research, it is of little surprise that scholars often dissociate themselves with the practice. As an illustration, it has been suggested that replication studies indeed exist at a level signiﬁcantly noted in prior reviews. However, these studies often go

Multi-Study Research

169

unadvertised when being developed and upon submission by scholars who fear that such positioning would be viewed with disdain even before externally reviewed (Eden, 2002). Ferris et al. (2006) found this relationship paradoxical and offered evidence for an opposite perspective. Speciﬁcally, they argued that the most informative and complex conceptual approaches would stimulate the most theory testing and replication/expansion interest from scholars. Regrettably, this view is not uniformly help by the majority of organizational science scholars. Moreover, the perspective held by those holding the ‘‘keys to publishing kingdom,’’ notably that replication has no place in the organizational sciences, is contradictory with teaching doctrine. First, as a means to demonstrate ‘‘best approaches,’’ these academicians often encourage doctoral students to develop multi-study research packages in their philosophy of science seminars (Kerlinger & Lee, 2000; Jick, 1979) as well as when crafting a dissertation (e.g., lab study ﬁrst, then ﬁeld research). Second, it is frequently these same scholars who criticize single-study investigations in their reviewing capacities, lamenting the fact that results are largely context-speciﬁc. Replication, then, serves as a vehicle of rejection (i.e., single studies need veriﬁcation, and veriﬁcation does not warrant publication) regardless of whether it is included in submitted research or not. Despite this intractable conundrum, we agree with both past (Mack, 1951; Popper, 1959) and more contemporary scholars (McKinley, 2010; Mezias & Regnier, 2007; Schmidt, 2009) who view replications, extensions, and validation as important and necessary mechanisms for scientiﬁc advancement. In support, Miller and Contay (1980) argued that a theory passing a test provides only limited evidence that it will do so comparably under different conditions and in unlike contexts. As self-correctness is important in building a body of knowledge (Rosenthal, 1979), theory veriﬁcation or falsiﬁcation across contexts is essential (Popper, 1968; Tsang & Kwan, 1999). We discuss the interplay between theory and context in supporting the need for multistudy research below and develop our own proposal for their increased use as a strategy to broaden existing research.

THE ROLE OF THEORY AND CONTEXT IN REPLICATIONS The term ‘‘organizational sciences’’ has been used in the literature to refer to the collective body of knowledge formed through the investigation of behavior in and of formal entities (i.e., proﬁt and nonproﬁt). With its roots

170

WAYNE A. HOCHWARTER ET AL.

in psychology, sociology, political science, and economics (among others; Cummings, 1978), the ﬁeld has come to encapsulate well-identiﬁed subareas (Heath & Sitkin, 2001), including human resources management, organizational behavior, organizational theory, and strategic management. As a result of its inherent multidisciplinary rudiments, a collective body of knowledge can be formed only when theoretical explanation developed to drive empirical research appropriately joins or separates phenomenon (Strong, Jeanneret, McPhail, Blakley, & D’Egidio, 1999). Thus, theory development is fundamental for any scientiﬁc ﬁeld of inquiry, as it represents the basis on which ideas are tested and new knowledge and insights are gained. From this exercise, or preferably a series of exercises, validation or disconﬁrmation is established. Bacharach (1989, p. 496) deﬁned theory as ‘‘a statement of relations among concepts within a set of boundary assumptions and constraints. It is no more than a linguistic device used to organize a complex empirical world.’’ Dubin (1976) described theory as simply the ‘‘attempt of man to model some theoretical aspect of the real world’’ (p. 26). Scholars maintain theory is the basis upon which new ideas are tested and knowledge gained (Jaccard & Jacoby, 2010) by addressing the questions of how, when, and why (King, Felin, & Whetten, 2010; Whetten, 1989). Extending this discussion, Weick (1989) argued that theories often are developed because they lend themselves to methodological validation (e.g., replication/extension), and as such, their expansion is critical for the advancement of organizational science research. Finally, Bacharach argued that the two main criteria upon which a theory should be evaluated are falsiﬁability (i.e., whether a theory is constructed in such a manner that it is subject to empirical refutation) and utility (i.e., the usefulness of the theory), both of which possess implications for the development of multi-study research programs we describe later. Despite the vigorous support for theory-developing replication/extension studies, reality suggests a ubiquitous detachment. For example, McKinley (2010) recently argued that theory development and theory testing have become unbalanced in recent years. Speciﬁcally, the development of theory has become the ultimate goal in the organizational sciences; a view that is shared by both micro- and macro-level scholars (e.g., Hambrick, 2007). Commenting on the trend in the strategic management literature, Mezias and Regnier (2007, p. 287) argued, ‘‘Strategic management as a ﬁeld, in part because of its close ties to economics, may be particularly hard hit by the perception that empirical work is not as valid as theoretical work.’’

Multi-Study Research

171

As a result, theory testing and empirical replications of proposed relationships that are activities typical of traditional science advancement (Kuhn, 1970) have become preventable casualties. This perspective has subsequently exiled two important features of science, namely, instrument standardization and replication, to outsider status, well behind the pursuit of theoretical development (McKinley, 2007; Neuliep & Crandall, 1991). An obvious, but often overlooked, consequence is that the cycle of theory development has replication as a critical link (Parkhe, 1993), and without its inclusion, theory cannot be accepted and further cultivated. In essence, it dies on the vine. In support of the replication-theory development link, Helfat (2007, p. 187) noted that ‘‘an accumulation of evidence that points to empirical regularities provides us with a much broader and more generalized understanding of the world.’’ Historically, conceptual tests and the subsequent replications (i.e., sometimes several) were critically important components of theory conﬁrmation in the 1960s and 1970s (McKinley, 2010). During this time, the inﬂuential Aston studies were developed, tested, and replicated to validate the original theoretical statements (e.g., Pugh, Hickson, Hinings, & Turner, 1968, 1969). In this multi-study research stream, inconsistencies often were uncovered (Child, 1972), leading to additional thought regarding the interplay between underlying theory and potentially inﬂuential contextual factors (McKinley, 2010) including size, dependency, and technology (Pugh & Hickson, 1972). It is important to note that this program extended over the course of several years, which reﬂects time rarely allowed in contemporary scholarly settings for the research development and dissemination of ﬁndings. For example, how likely is it for a department head to advise a newly minted Ph.D. to undertake a program of research that imitates the Aston studies? Presumably, not very likely, and in fact, many would term this approach ‘‘career suicide.’’ To a great extent, meaningful has been replaced by publishable, as junior scholars have been trained to carve rather than cultivate their research. Moreover, replications and reexaminations of research like these were considered important enough (and submission was actually encouraged by editors) to be published in the top academic journals at the time (e.g., Administrative Science Quarterly and Human Relations). As testament to their enduring impact, these studies remain frequently cited in contemporary organizational theory research (Fiedler & Welpe, 2010; Samra-Fredericks, 2010), most notably for their attention to contextual inﬂuences (Pugh, Hickson, & Hinings, 2006; Wong & Birnbaum-More, 1994). Advocating future contributions similar in scope and magnitude, McKinley, Mone, and

172

WAYNE A. HOCHWARTER ET AL.

Moon (1999) argued neither internal nor external conﬁrmation can be realized devoid of repeated operationalizations varying with respect to research approach, samples, and contexts.

Theory Extensions and the Role of Context Largely because of its in situ inﬂuence on role perceptions and construal of workplace dynamics (Ferris, Munyon, Basik, & Buckley, 2008b), scholars have vigorously appealed for increased sensitivity of contextual effects in organizational research (Grifﬁn, 2007; Whetten, 2009) in recent years. This request, however, is not new to the organizational sciences. Over 30 years ago, Katz and Kahn’s (1978) discussion of role development conﬁrmed that ‘‘enactment does not occur in isolation; it is shaped by additional or contextual factors’’ (p. 195). More recently, Dierdorff and Morgeson (2007) argued that elements of task and social contexts shape both role cultivation and the stability of perceived role expectations. Requests for increased attention to contextual features permeate other content areas as well including strategic decision effectiveness (Elbanna & Child, 2007) and leadership (Liden & Antonakis, 2009). Appeals notwithstanding, researchers have habitually disregarded the effect of context when developing scholarly investigations (Dierdorff, Rubin, & Morgeson, 2009; Johns, 2006). Whetten (2009) identiﬁed two reasons for such neglect. First, he argued that scholars knowingly have avoided environmental considerations, assuming that context-free studies (i.e., isolation) are more informative than those considered context-rich (i.e., inclusion). Second, he suggested that scholars neglect contextual effects because ‘‘they simply don’t see them’’ (Whetten, 2009, p. 31), as is likely the case when a standard survey design is the tool of choice. We add to this list the assumption that research evaluating the role of context is particularly difﬁcult to conduct given the tacit volatility associated with most work environments (Nadkarni & Barr, 2008; Schmidt, Dolis, & Tolli, 2009). In sum, scholars may conclude that considering context is impractical because laboratory experiments often fail to fulﬁll requirements associated with external validity, and ﬁeld research settings often possess temporal and access constraints, (Johns, 2001). Regardless of the rationale, a failure to consider context in organizational research comes with considerable costs. The most notable limitation is the promotion of conclusions that are incomplete and thus vulnerable to heightened skepticism. Moreover, empirical results that support a proposed

Multi-Study Research

173

theory, in the absence of contextual recognition, signal to the ﬁeld that ‘‘this is the way it is’’ (Helfat, 2007). Replications, often undertaken to further establish the veracity of the initial ﬁndings, often implement comparable research approaches (i.e., same thing leads to the same outcomes). As the stream of context-neglected research builds consensus, the conclusions likely drift farther and farther away from reality. Furthermore, the inability to more strongly consider the environment is disconcerting given the recent development of contextual taxonomies (Morgeson, Dierdorff, & Hmurovic, 2010; Rousseau & Fried, 2001). For example, Johns (2006) described context in terms of two dimensions: omnibus and discrete. Omnibus context takes into account more expansive environmental factors (e.g., society, organization, and culture), which help explain the ‘‘who,’’ ‘‘what,’’ ‘‘when,’’ and ‘‘why’’ of a particular phenomenon (Whetten, 1989). On the contrary, discrete context refers to identiﬁable environmental characteristics (e.g., task, social, and physical) that impact behavior directly, or moderate/mediate relationships between substantive omnibus variables (Johns, 2006). More speciﬁcally, the task context includes informational elements, the social context includes interpersonal features, whereas the physical context includes structural/building dimensions, of the discrete environment (Dierdorff et al., 2009; Johns, 2006). These typologies have been proven very effective in allowing for a balanced view of person–situation relationships (Johns, 2010; Morgeson et al., 2010). Finally, given the pervasive state of ﬂux in which organizational environments reside (Dierdorff et al., 2009), contextual features uniquely inﬂuence the interpretation of the particular phenomena under investigation. As such, research needs to be repeated across multiple environments to determine whether legitimate effects exist, if ﬁndings are context-speciﬁc (Blair & Hunt, 1986; Liden & Antonakis, 2009), or some effect-context combination is present. Indeed, by replicating/extending a research study, it can be determined that results transcend context (Eden, 2004). More directly, the stuff of science, or the demonstration of what Lykken (1968) referred to as an ‘‘empirical fact,’’ requires sound demonstration of replicability results. The following distinguish different types of replications/multi-study designs. Literal Replications Lykken’s (1968) often-cited deﬁnition describes a literal replication as ‘‘exact duplication of the ﬁrst investigations’ sampling procedure, experimental conditions, measuring techniques, and methods of analysis’’ (p. 155). Johnson and Bouchard (2005) suggested that literal replications serve largely as a check to ensure the correctness of the initial study, rather than offer directives for

174

WAYNE A. HOCHWARTER ET AL.

augmentation. Schmidt (2009) referred to studies that closely mirror a prior investigation as ‘‘direct replications,’’ whereas others synonymously use the term ‘‘exact replication’’ (Singh et al., 2003; Tsang & Kwan, 1999). Others warn against the use of the term ‘‘exact,’’ due to the inability to control for all incidental factors (Lykken, 1968) across studies. Eden (2002) argued that the strict replication of an original work may be viewed favorably. However, he warns that ‘‘strict replications or strict replications’’ would likely be considered of only minimal value. Operational Replications An operational replication involves the duplication of a previous study’s sampling and experimental procedures (Lykken, 1968). The main objective of this approach is to determine whether the methodological strategies used in the initial study produce comparable results. Unlike a strict or literal replication, an operational approach allows for relevant variation (Bedeian, Mossholder, Kemery, & Armenakis, 1992). Kelly et al. (1979) contended that an operational replication exists, for example, by adding or modifying a criterion outcome, while maintaining other processes/procedures. Constructive Replications Constructive replications differ from literal and operational tactics in that the researcher deliberately avoids duplication of procedures used in the target study (Lykken, 1968). Moreover, the use of different methodologies promotes a riskier test of substantive predictions. In terms of beneﬁts, the use of research tactics far removed from the initial study may mitigate concerns associated with common method factors (Lykken, 1984). Moreover, Hunter and Schmidt (2004) argued that constructive replications have the potential to contribute greatly to external validity, both in absolute terms and when compared with other conﬁrmation strategies. However, their value can be realized only if they are considered an important component of scholarly development. Furthermore, Schmidt (2009) maintained that a conceptual replication offers science signiﬁcantly greater insight into a phenomenon than a direct approach. Speciﬁcally, successfully replicating a proposed relationship not only corroborates the hypothesis but also supports its theoretical underpinnings (Amir & Sharon, 1990). In the end, this hypothesis has been tested by two different experimental strategies, which is highly preferable, if the objective is to build theory through a consensus of ﬁndings (Schmidt & Hunter, 2001). In sum, a direct replication generates facts for further contemplation, whereas a conceptual replication offers heightened understanding (Schmidt, 2009).

Multi-Study Research

175

Of course, understanding increases when a body of replication studies is simultaneously evaluated, often in the form of meta-analysis. Replications and Meta-Analysis Over 50 years ago, Miller (1956) acknowledged that more accurate generalizations come to light when comprehensive patterns of results emerge. Conversely, theoretical progress is thwarted when investigators are limited in terms of variables and streams of results to base resulting arguments (Cronbach, 1957). Eden (2002) suggested that ‘‘Such generalizations are best based on meta-analyses, and meta-analyses are enriched by encompassing a large number of high-quality replication studies’’ (p. 841). As a result, scholars conducting primary research must supply data, adequate in amount and composition, for meta-analytic studies to prosper. ‘‘Otherwise, once the supply of data to be mined is depleted, the meta-analysts may be out of business’’ (Eden, 2002, pp. 843–844). Replication and meta-analysis are inextricable linked, largely because they both serve as instruments to evaluate an amalgam of single-study investigations reporting results that are ‘‘usually not as straightforward as they might ﬁrst appear’’ (Hofer & Picinnin, 2009, p. 152). By extension, meta-analyses cannot occur without dedicated replication research (e.g., evaluation of primary research) (Aguinas, Pierce, Bosco, Dalton, & Dalton, 2011; Leavitt et al., 2010). Allen and Priess (1993) strongly argue for the importance of replicating given that meta-analyses are unable to substantiate their own existence without the contributions of primary research (Eden, 2002). Stated differently, meta-analysis is not a viable tool to assess consistency across studies when few commonalities across investigations exist. Hence, metaanalysis is largely reliant on replication research, and this dependence is especially important when protocol is stringently required to demonstrate comparability (Leavitt et al., 2010; Higgins & Thompson, 2002). An equally compelling case argues for the committed replication of metaanalyses (Allen & Priess, 1993). Such an approach is especially needed when rapid growth offers discrepancy evidence from initial ﬁndings. An as example, Feingold (1994) replicated and extended several prior meta-analytic studies, which examined the relationship between gender and personality. More recently, Humphrey, Nahrgang, and Morgeson (2007) replicated the Fried and Ferris (1987) meta-analytic summary of particular motivational characteristics. In further support, Allen and Burrell (2002) argued ‘‘Metaanalysis does not represent a truly objective method of analysis; but it does provide a method of literature summary that permits an assessment of the literature that others can replicate’’ (p. 132).

176

WAYNE A. HOCHWARTER ET AL.

In sum, we share the view of Aguinas et al. (2011) who note that metaanalysis is the near-perfect conduit for discovery and conﬁrmation (i.e., replication) with associated scholarly and practical beneﬁts that are sizeable. For example, Schmidt (1992) argued that meta-analysis can mitigate the improper use of scholar’s effort and research funds by determining when additional replication research is unwarranted. In addition to conserving resources, these ﬁndings can direct new theory development and, as a result, cultivate a need for further replication research. Taken in their entirety, these realities conﬁrm Eden’s (2002) view that meta-analysis and replication reside in a constant state of ‘‘symbiosis’’ (p. 844).

INCIDENCE OF MULTI-STUDY PACKAGES Scholars have examined the publication trends associated with replication/ multi-study research designs. As expected, replication studies represent a small percentage of published research in the organizational sciences. In other disciplines, which we brieﬂy review later, tendencies differ. These conclusions are supported, largely, by Blalock’s (1961) assertion that replication in the social sciences is considerably more difﬁcult than the physical sciences (Hedges, 1987). Treatments in Strategic Management Literature Hubbard, Vetter, and Little’s (1998) review found that 4.6% of articles published in ﬁrst-tier journals were replication with extension studies during the 1976–1985 period. This number increased to 6% during the1986–1995 period. For the entire 1976–1995 period, 5.3% of all 701 reviewed journal articles represented replication with extension studies. Not a single strict replication was identiﬁed during the 20 years of reviewed research. Third-tier journals were found to be the most receptive to replication with extension studies (6.9%), followed by ﬁrst-tier (5.3%) and second-tier (2.8%) publications. Review of the Marketing Literature The scarcity of replication submissions, advanced by recent publishing realities, has lead to strong encouragement by marketing journals for multistudy submissions. Evanschitzky et al. (2007), building on prior research (Hubbard & Armstrong, 1994), documented the incidence of published

177

Multi-Study Research

replication studies in this domain. Results indicated that the percentage of replications published over the 1990–2004 period dropped by 50% over the earlier 1974–1989 period. The authors concluded by advising practitioners to view scholarly research with skepticism and advocated that teachers consider one-shot studies as uninformative.

Review of Other Sciences Reviews of replication incidence exist across virtually all scientiﬁc disciplines, including those considered the ‘‘hard sciences.’’ For example, Ioannidis (2005) found that 44% of highly referenced medical studies published between 1900 and 2003 were replicated. Sullivan (2007) reported that the majority of the published genetic association studies have been subjected to replications, most of which subsequently went unconﬁrmed (Lohmueller, Pearce, Pike, Lander, & Hirschhorn, 2003). The complexity of human disease represents an often-cited reason for a lack of corroboration in medical science settings (Hirschhorn, Lohmueller, Byrne, & Hirschhorn, 2002). Finally, Jones, Derby, and Schmidlin (2010) found that at least onehalf of chosen human factors investigations were replicated. Given that the sample was neither large (N ¼ 8), nor randomly chosen, these results should be viewed with prudence.

Twenty-Year Review of Eight Top Journals To identify tendencies and chart trends, we conducted an extensive review of organizational science journals focusing speciﬁcally on multi-study research packages. We selected the 20-year period of 1990–2009, evaluating articles published in the following eight organizational science journals that represent the top micro and macro publications in the ﬁeld: Academy of Management Journal, Administrative Science Quarterly, Journal of Applied Psychology, Journal of Management, Organizational Behavior and Human Decision Processes, Organization Science, Personnel Psychology, and Strategic Management Journal. We acknowledge the potential for inclusion of other journals that are perceived as comparable or superior to those we have chosen (especially by scholars prone to frequent journals considered more international than domestic). For each identiﬁed multi-study article published, we noted several pieces of information: the number published in each journal, what type of

178

WAYNE A. HOCHWARTER ET AL.

multi-study package (e.g., replication, and if so, what type, and theoretical extension), the number of studies or samples, and the stated rationale for the number and sequencing of samples or studies. The results of this review yielded a document in excess of 30 pages. Because it was not possible to include such a sizeable table as part of this chapter, we summarize the results in Table 1, and we make the complete table available upon request from the authors. The review identiﬁed a total of 1,072 multi-study publications. We assumed that the more micro-oriented journals would account for the greatest number of multi-study package publications, with the more macro-oriented journals producing far fewer. Consistent with our expectations, the more behavioral journals yielded the greatest incidence, with Organizational Behavior and Human Decision Processes accounting for 470 such packages, followed by the Journal of Applied Psychology with 360, and Personnel Psychology with 65. In terms of macro journals, Organization Science published 31 multistudy articles, Strategic Management Journal accounted for 29, whereas 24 articles were found in Administrative Science Quarterly. Finally, there are arguably two journals that publish a mix of both micro and macro studies, and it is expected that their multi-study publishing tendencies would fall in between the micro and macro journals. In support, the Academy of Management Journal and Journal of Management produced 51 and 42 multistudy articles, respectively. Also, we suggested that we would see a substantial increase in the publication of such multi-study packages over time, with considerably more articles identiﬁed in current research than conﬁrmed 20 years ago. Indeed, that was the case, as the 1990–1999 period yielded 457 multi-study publications. In comparison, the frequency rose to 614 for the 2000–2009 period. The incidence over these time periods by journal showed considerable increases for the Journal of Applied Psychology (i.e., 144 for 1990–1999, and 216 for 2000– 2010), Personnel Psychology (i.e., 23 and 42, respectively), Journal of Management (i.e., 17 and 24, respectively), and Organization Science (i.e., 11 and 20, respectively). Interestingly, the incidence of multi-study articles declined over the two decade time periods for the Academy of Management Journal (i.e., 26 and 25, respectively), Administrative Science Quarterly (i.e., 16 and 8, respectively), and the Strategic Management Journal (i.e., 15 and 14, respectively). Because journals often change in the absolute number of published articles (e.g., from four to six volumes per year), we thought it appropriate to examine the percentage of multi-study packages appearing in each outlet, rather than the absolute number. For clarity of interpretation, the results for

179

Multi-Study Research

Table 1.

Frequency Assessment of Multi-Study Packages in Selected Journals.

Year

AMJ

JAP

PPSYCH

OBHDP

JOM

ASQ

SMJ

ORG SCI

Total

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Total

1 1 1 2 1 11 2 3 2 2 4 – 4 1 – 4 2 7 – 3 51

8 16 12 19 25 16 9 10 13 16 25 21 23 12 15 21 24 26 24 25 360

3 – 3 5 1 1 3 2 3 2 2 6 4 3 7 6 3 4 1 6 65

7 12 15 14 23 33 43 27 17 14 19 18 26 38 24 27 25 31 26 31 470

1 2 2 – 2 1 1 4 1 3 2 – 3 3 1 2 5 1 3 5 42

2 2 3 2 2 – – 2 1 2 2 1 1 – 1 – 2 1 – – 24

1 – 1 2 2 1 1 3 3 1 – 1 3 2 1 3 2 – 1 1 29

1 – – 1 3 1 1 – 4 – 2 2 3 – 4 2 1 5 1 – 31

24 33 37 45 59 64 60 51 44 40 56 49 67 59 53 65 64 75 56 71 1072

Notes: AMJ, Academy of Management Journal; ASQ, Administrative Science Quarterly; JAP, Journal of Applied Psychology; JOM, Journal of Management; OBHDP, Organizational Behavior and Human Decision Processes; ORG SCI, Organization Science; PPSYCH, Personnel Psychology; SMJ, Strategic Management Journal

micro, macro, and mixed foci journals are shown in Figs. 1–3. As shown, Organizational Behavior and Human Decision Processes (mean % ¼ 42.7%, range 16–71%), the Journal of Applied Psychology (mean % ¼ 18.9%, range 12.6–29.4%), and Personnel Psychology (mean % ¼ 10.7%, range 0–25.0%) most frequently included multi-study research designs. Multi-study packages in the macro- and mixed-orientation journals were Organization Science (mean % ¼ 3.7%, range 0–8.7%), Strategic Management Journal (mean % ¼ 2.4%, range 0–5.1%), Administrative Sciences Quarterly (mean % ¼ 5.3%, range 0–12.5%), Academy of Management Journal (mean % ¼ 3.9%, range 0–5.2%), and the Journal of Management (mean % ¼ 5.3%, range 0–13.1%).

180

WAYNE A. HOCHWARTER ET AL. 80% 70%

Journal of Applied Psychology

Replication %

60% 50%

Personnel Psychology

40% 30%

Organizational Behaviour and Human Decision Processes

20% 10%

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

0%

Year

Fig. 1.

Prevalence of Replication Studies in Journals with a Micro-Level Emphasis (1990–2009).

20% 18% 16%

Replication %

14% 12%

Organization Science

10% 8%

Strategic Management Journal

6% 4% 2% 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

0%

Year

Fig. 2.

Prevalence of Replication Studies in Journals with a Macro-Level Emphasis (1990–2009).

181

Multi-Study Research 16% 14%

Replication %

12% Journal of Management

10% 8%

Academy of Management Journal

6%

Administrative Science Quarterly

4% 2%

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

0%

Year

Fig. 3.

Prevalence of Replication Studies in Journals with Possessing both Microand Macro-Level Status (1990–2009).

In general, results are consistent with results provided in prior organizational research; notably, multi-study investigations remain largely underrepresented relative to single-study designs. However, by analyzing speciﬁc journals uniquely, we were able to show that incidents were far from homogeneous. We acknowledge that the objective of conducting this exercise was not to offer deﬁnitive statements. Rather, our goal was to stimulate thought and further assessment focusing on the ‘‘why’’ and ‘‘how’’ of multi-study publishing. In completing our review, we identiﬁed several stellar multi-study investigations, which are shown and reviewed in Table 2.

Multi-Study Packages in Our Own Research We have embraced the importance of the multi-study research strategies dating back over 20 years (i.e., Wayne & Ferris, 1990). Building on this philosophy, our more contemporary research reﬂects a commitment to science through the development of multiple sample investigations in social inﬂuence processes/politics (Breaux, Munyon, Hochwarter, & Ferris, 2009; Frink & Ferris, 1998; Hochwarter, Ferris, Laird, Treadway, & Gallagher, 2010;

Replicating and extending Fried and Ferris (1987) metaanalytic study (259 studies). Fourteen work characteristics explained, on average, 43% of variance in 19 worker attitudes and behaviors

Control was more strongly related to commitment when coping with layoff threat than

Summary of meta-analyses and extension of theoretical extension using multistep mediation

Examination of two unique layoff settings (some witnessed the layoff, some did not –

Humphrey et al. (2007). Integrating motivational, social, and contextual work design features: A metaanalytic summary and theoretical extension of the work design literature. Journal of Applied Psychology, 92, pp. 1332–1356. Brockner et al. (2004). Perceived control as an antidote to the negative effects of layoffs on

CEO commitment to the status quo is a multilevel (e.g., individual, ﬁrm, and industry) phenomenon

Summary of Findings

Replication and extension using content analysis incorporating content analysis

Unique Design Strengths

Value of studying a comparable theoretical argument in environments that possessed

Content analysis can be used as a less intrusive method of data collection. Including theoretically relevant variables is important. The CSQ – performance relationship is more complicated than initially described Importance of considering mediation when conducting meta-analysis. Expanding theoretical discussions when amalgamating/replicating studies beyond simply addressing past relationships

Implications for Research

Exemplary Multi-Study Research Investigations in Recent Years.

McClelland, Liang, and Barker (2010). CEO commitment to the status quo: Replication and extension using content analysis. Journal of Management

Author(s)/Year

Table 2.

182 WAYNE A. HOCHWARTER ET AL.

Felps et al. (2009). Turnover contagion: How coworkers’ job embeddedness and job search behaviors inﬂuence quitting. Academy of Management Journal, 52, pp. 545–581

Tang, Kacmar, and Busenitz (in press). Entrepreneurial alertness in the pursuit of new opportunities. Journal of Business Venturing

survivors’ organizational commitment and job performance. Administrative Science Quarterly, 49, pp. 76–100

Study 1; all survived the layoff but experienced different levels of distress as a result – Study 2). Outcomes differed as well (commitment – Study 1; performance – Study 2. Study 1 generated entrepreneurial alertness scale items and determined adequacy. Study 2 factor structure as indicated by South Korean CEOs. Study 3 two data sources were used to established nomological network Study 1 consisted of over 8,000 workers in of 1,037 departments. Supplemental qualitative analyses were conducted. Sample 2, which consisted of over 230 employees in 45 branches, both replicated and extended Study 1 ﬁndings Job embeddedness and job search behavior explained variance in voluntary turnover beyond the individual and group-level predictors conﬁrmed in prior research

Extending previous research (Kirzner, 1973), a 13-item scale was developed, which demonstrated appropriate validity

in low-risk settings. Furthermore, performance increased when high control was coupled with high threat to well-being

Merit in developing replication/ extension studies in diverse settings, utility in mixing analysis methods to include quantitative and qualitative approaches, and acknowledging that mesolevel phenomena are amenable to multiple conﬁrmation studies

wide-ranging procedural and operational differences. Also, theoretical expansion as a result of documenting relationships across unique outcomes. Utility in capturing data from multiple contexts (US and Korea) to validate initial scale items. Also, wide ranges of industries and settings are preferable to narrow ones

Multi-Study Research 183

184

WAYNE A. HOCHWARTER ET AL.

Hochwarter, Ferris, Zinko, Arnell, & James, 2007; Kiewitz, Restubog, Zagenczyk, & Hochwarter, 2009; Treadway et al., 2005) and interpersonal stressors and strain (Ferris, Rogers, Blass, & Hochwarter, 2009; Hall et al., 2006; Hochwarter, Laird, & Brouer, 2008; Hochwarter, Perrewe´, Hall, & Ferris, 2005; Hochwarter et al., 2009; Lanivich, Brees, Hochwarter, & Ferris, 2010) content domains. These studies have spanned the full spectrum of replication typologies, from closely exact to constructive replications with extensions. For example, Hochwarter et al.’s (2005) two-study investigation, which examined the role of negative affectivity on accountability – tension relationship is best described as a literal replication. With the exception of unique contexts (e.g., wide range of occupations and administrative health care employees), the studies mirrored one another in terms of methodology (Lykken, 1968). Recently, Hochwarter et al. (2008) investigated the hurricane stress – resource availability relationship in ﬁve organizations. Data were collected over the course of two volatile hurricane seasons. This study is considered a replication with extension because of the inclusion of work effort as a dependent variable in studies 4 and 5 and the use of a different data collection tactic. Also, some organizations experienced hurricanes in 2004 (i.e., Charley, Frances, Ivan, and Jeanne), whereas others faced hurricanes in 2005 (i.e., Katrina). All other methodological components were consistent across samples. Finally, Treadway et al.’s (2005) constructive replication examined the interactive relationship of age and politics perceptions on job performance across three unique environments. In addition, alternative measures of politics perceptions and job performance were used, including second-source data of work contribution (e.g., subordinates direct supervisor) in one sample. The research program that has initiated the greatest incidence of multistudy research has been in the area of political skill (Blickle et al., in press; Blickle, Kramer, Schneider, Meurs, Ferris, Mierke, Witzki, & Momm, 2011; Blickle et al., 2009; Ferris et al., 2008a; Ferris et al., 2005; Jawahar, Meurs, Ferris, & Hochwarter, 2008; Kapoutsis, Papalexandris, Nikolopoulos, Hochwarter, & Ferris, 2011; Kolodinsky, Hochwarter, & Ferris, 2004; Liu et al., 2007; Witt & Ferris, 2003). Because political skill represents a relatively new area of inquiry (Lindsay & Ehrenberg, 1993; Tsang & Kwan, 1999), initial expectations were higher relative to constructs with more lenghty histories (e.g., to establish that it is unique and does not simply duplicate other established constructs). Thus, providing even stronger validity evidence across studies was critical. In response, the scale development report contained three studies and seven

Multi-Study Research

185

samples (Ferris et al., 2005). Replication/extension strategies initially lent support for the scales’ factor structure and then was critical in forging support for its construct and predictive validity.

THE CONTRIBUTIONS OF MULTI-STUDY RESEARCH PACKAGES Our objectives in designing and executing multi-study research packages are many. Addressing potentially extraneous error variance and respecting the ‘‘ﬁle drawer problem’’ associated with much of the existing research (Rosenthal, 1990) are obvious considerations, as is the potential for method bias to be introduced into the interpretation of results (Rosenthal & Rosnow, 1991). However, our pursuit of theoretical validation has been the goal most fervently guiding this research (Kurke & Aldrich, 1983). In support, Hubbard, Vetter, and Little (1998) maintained that extensions broaden the capacity of initial ﬁndings most effectively when results are generalizeable to other contexts, time periods, and populations. Hypothetically, we ﬁnd it incredibly naı¨ ve (and equally self-serving) to assume that a phenomena identiﬁed by a sample of 118 insurance agents (with less than a 40% survey response rate no less) is a decree requiring serious observance by both scholars and practitioners. In terms of execution, we sought maximum differentiation between studies, whenever possible, to provide validity evidence (Singh et al., 2003). Despite our advocacy, this approach has often been the subject to reprimand by evaluators of our work. For example, editors have all too frequently requested eliminating studies or have outright rejected manuscripts because of perceived dilution or a personal preference for single-study designs. As noted, multi-sample research studies range widely in potential differentiation, all of which have merit when used appropriately. On one hand, a package with minimal differentiation may have only one variable or factor manipulated, as is often the case in scale development studies. As example, the inﬂuence of time may be considered when conducting the traditional Time 1-Time 2 research design within one context or ﬁrm. Minimally, other study ingredients with the potential for manipulation include population, context, size of ﬁrm, location of activity, onset of event, measures of constructs, inclusion/deletion of substantive predictor or outcome, demographic control factor, individual difference factors, company classiﬁcation, change in leadership, data collection approach, data source,

186

WAYNE A. HOCHWARTER ET AL.

analytic strategy, and countless more (Evanschitzky & Armstrong, 2010; Schmidt, 2009; Tsang & Kwan, 1999). On the other hand, studies within a multi-study package may be differentiated across a number of factors. Rosenthal (1991) maintained that an initial study should be evaluated by at least two replications/extension, one that is nearly exact and one that deviates substantially. Lindsay and Ehrenberg (1993) outlined three reasons for differentiation. First, differentiation addresses concerns regarding convergent validity. Speciﬁcally, it increases conﬁdence that the conceptual variables, rather than study artifacts (i.e., research, instrument), are the actual source of results (Rosenthal, 1976). Second, processes and measurement associated with differentiation promote acceptance by a wider range of audiences (i.e., industries, scholarly domains, and corporate levels) who are better able to glean appreciation for both study tools and outcomes. Third, differentiation increases the likelihood that inconsistencies or deviations will surface, contributing to reﬁnements in both empiricism and theory (Ehrenberg & Bound, 1993). Without the potential for falsiﬁcation, meaningful contributions are simply not possible (Lykken, 1968; McKelvey, 2003). An example of a differentiated replication was recently conducted by Kapoutsis et al. (2011), who examined the politics perceptions – political skill relationship on job performance. Variations, both numerous and signiﬁcant, were embedded in the research design including context (United States – Study 1; Greece – Study 2), vocational focus (insurance employees – Study 1; Graduate School Alumni – Study 2), measure of politics perceptions (Hochwarter, Kacmar, Perrewe´, & Johnson, 2003 – Study 1; Kacmar & Carlson, 1997 – Study 2), measure of job performance (Wright, Kacmar, McMahan, & Deleeuw, 1995 – Study 1; Williams & Anderson, 1991 – Study 2), source of performance data (self – Study 1; supervisor – Study 2), and measurement of individual differences (Watson, Clark, & Tellegen, 1988 – Study 1; John, Donahue, Ernst, & Kentle, 1991 – Study 2). Despite these methodological deviations, results across studies ran largely parallel. Notably, political skill predicted performance most strongly in settings perceived as reﬂecting minimal political activity, as theorized.

Addressing a Long History of Disdain – Beneﬁts of Multi-Study Research Packages Many of the criticisms of replication research are potentially ameliorated by the thoughtful use of multi-study research packages. Given the sheer number of condemning statements proposed over the past 50 years, it is

Multi-Study Research

187

implausible to adequately address each with an appropriate level of substantiation. However, we see considerable value in speaking to those that are perhaps wider in scope, and thus, more prominent in inﬂuencing scholarship and practice. By extension, we contend that developing a multistudy mindset has both scientiﬁc and career-related beneﬁts. Limitations exist as well, and we discuss each later. First, it has been suggested that replication research has little value, is unimportant, and rarely original (Evanschitzky & Armstrong, 2010; Hubbard & Armstrong, 1994). Unquestionably, replication is important, especially when novel, complex, or intuitively discrepant results surface during the evolution process (Rosenthal & Rosnow, 1991). We take this perspective one step further by arguing that there is no science without studies that conﬁrm and extend existing knowledge. In essence, Eden (2002) argued that all studies are replications to varying degrees even when not publicized as such. We challenge scholars to identify a study that is 100% original, with absolutely no consideration of prior theory or reported ﬁndings. Even in our own work in the area of organizational politics, it is obvious that we are replicating Ferris, Russ, and Fandt (1989) despite our intention to offer unique insights into the phenomenon. In sum, we are conﬁdent when we argue that ‘‘all studies are replication studies.’’ In terms of value added to the ﬁeld, multi-study packages not only have the potential to verify/disconﬁrm original tenets, they are able to offer additional explanation, and empirical assessment of alternative models. Even if the original results are conﬁrmed in the ﬁrst study, a second and third study can be used to reestablish its parameters and advance its theoretical underpinnings in ways that are creative and valuable to science. It has also been noted that replication research is undeserving of journal space (Hubbard et al., 1998), given its often-recognized potential to offer the ﬁeld only ‘‘marginal contributions’’ (Singh et al., 2003, p. 544). We empathize with the editor and reviewer who must consider an exact replication study for publication. If the ﬁndings are consistent with the initial research, ‘‘so what?’’ is a plausible reaction. If the ﬁndings are decidedly different, especially when a well-acknowledged program of research is scrutinized, a ‘‘you better show me this again to cover myself’’ response is to be expected. Either way, the manuscript has little chance of publication. A multiple-study research package is an antidote because of its potential to offer augmented levels of insight while only adding a modest level of pages to the ﬁnal document. In our experiences, additional studies have increased manuscript length by 1–3 pages (i.e., mostly when describing methods and results), which is an approximation we have conﬁrmed in other disciplines.

188

WAYNE A. HOCHWARTER ET AL.

Moreover, methodological issues have hindered scholar’s perceptions of replication research as well as the legitimacy of publication. Much of this consternation has arisen as a result of inappropriate labeling. For example, many studies are touted as exact replications despite the fact that they are virtually impossible to conduct (we would argue for impossible rather than virtually impossible). In response, Leavitt et al. (2010) noted the infrequency of exact replications as well as the detrimental career consequences for pursuing this approach. We advocate emphasizing the story rather than forcing a label that distracts evaluators at the outset. Problems also arise when researchers are unable to gather the detailed information (e.g., scales, sample characteristics, and research protocol) required to judge results of prior studies. The sharing of information between scholars also may be a hindrance. For example, Reid, Rotfeld, and Wimmer (1982) found that the probability of would-be replicators securing the needed information from the original study’s authors was approximately 50% (36% of which simply failed to respond even after a second request). Page restrictions often ration written speciﬁcation of methodologies, leading scholars to exhaust resources pursuing prior protocol rather than execution of the research. Unfortunately, the cost associated with the purchase of copyrighted instruments used in the original research represents a limitation as well. Other information-gathering factors include the passage of time, changes in scholarly interests, lost instrumentation, retirement, and even mortality. Developing multi-study investigations resolves much of this confusion, principally due to scholar’s control over the research process. For example, this approach allows for more accurate classiﬁcation (e.g., conceptual replication versus differentiated replication versus systematic replication – Kelly et al., 1979; Finifter, 1972). Moreover, the execution of subsequent studies within the package is less complicated as the researcher develops a better understanding of what works and what procedural reﬁnements require ﬁne-tuning. The jurisprudence associated with a self-directed research program affords the ﬂexibility to do what is best instead of adhering strictly to the methodologies of a study potentially conducted half a century ago. Finally, inﬂuence over the use of measures, across measures, limits confusion with respect to scale acquisition and manipulation.

Limitations of a Multi-Study Approach The greatest beneﬁt associated with multi-study research packages may also be its most telling limitation. Speciﬁcally, scholars have argued that

189

Multi-Study Research

replication/extension research has more validity if performed by someone other than the author of the original study. As an example, Rosenthal (1991) argued to an investigator gradient consisting of low (i.e., the researcher of the initial study), intermediate (i.e., a student or member within the research team), and high (i.e., a researcher completely disassociated with the initial study’s author). Hubbard and Armstrong (1994) contended that replication studies conducted by independent researchers often contain fewer biases associated with experimenter expectations. Rosenthal and Rosnow (1991) discussed the inﬂuence of correlated replicators, questioning the value of 10 replications conducted by a single investigator as compared to 10 carried out by scholars not associated with the initial research. As a result, Monroe (1992) argued that replication research originated by independent scholars is preferred to ‘‘reduce concerns about the interaction of the research and study being a reasonable explanation for the original results’’ (p. 1). In response to these potential biases, the editorial policy for the Quarterly Journal of Finance and Accounting clearly states, ‘‘Extreme dependence (such as an author replicating his own work or a student replicating the work of a professor) destroys the econometric audit function y The policy of this journal will be to reject such replications.’’ This view, which has elements of legitimacy, is shared by editors across social science domains. Finally, a concern that cannot be fully discounted is time. Conducting three or four data collections, regardless of their format, will deplete a scholar’s resources more rapidly than a single-study approach. Leavitt et al. (2010) aptly concluded that developing multiple studies to judge theories is time-consuming and expensive. In an age of shrinking university budgets, resources for data collection represent targets of reduction. Adding to this stress is the never-spoken, but often-endorsed, view that data need to come from multiple sources to warrant publication in the top behavioral science journals. In reality, scholars early in their careers may not have acquired the contacts in their new setting to accomplish the feat of collecting multi-source data once, let alone three times. Regardless of whether three data collections result in three articles with one sample each, or one article with three samples, we advocate that scholars develop a data collection mentality that is both consistent and creative.

CONCLUSION Cartwright (1949) acknowledged the difﬁculty of constructing replication studies, and their resulting absence in the social sciences, stating ‘‘It is hard

190

WAYNE A. HOCHWARTER ET AL.

to sell the need of replication to men of action, but it can be sold’’ (p. 202). In this chapter, we have proposed the increased use of multi-study research packages as a supplement to single-study designs that permeate virtually all of the literature. Multi-study research packages involve more than one study, and the objectives can vary from testing the hypotheses in the ﬁrst study and then conducting literal or constructive replications, or conceptual extensions, of those initial study results. Such multi-study packages thus serve to increase conﬁdence in the validity of the obtained ﬁndings as reﬂecting valid variance and thus theory conﬁrmation support, versus leaving speculation about whether the ﬁndings (i.e., of single-study designs) are merely artifactual to sample, settings, or measures, and are not replicable ﬁndings that deserve serious attention. Also, Schmidt (2009) argued in favor of packaging studies together that make an integrative statement regarding some phenomenon of interest noting ‘‘systematic connection between the replication condition and the new hypothesis condition is the important link that demonstrates that the underlying assumptions are working and that they can be transferred to a similar design testing for new ideas’’ (Schmidt, 2009, p. 99). Campbell and Stanley (1963) argued, ‘‘The experiments we do today, if successful, will need replication and cross-validation at other times under other conditions before they can become an established part of science, before they can be theoretically interpreted with conﬁdence’’ (p. 3). We concur and strongly advocate for the increased use of multi-study research packages to ultimately advance the organizational sciences. We hope the ideas presented in this chapter stimulate further interest in this important area.

REFERENCES Aguinas, H., Pierce, C., Bosco, F., Dalton, D., & Dalton, C. (2011). Debunking myths and urban legends about meta-analysis. Organizational Research Methods, 14, 306–311. Allen, M., & Burrell, N. (2002). Sexual orientation of the parent: The impact on the child. In: M. Allen, R. Preiss, B. Gayle & N. Burrell (Eds), Interpersonal communication: Advances through meta-analysis (pp. 125–143). Mahwah, NJ: Lawrence Erlbaum. Allen, M., & Priess, R. (1993). Replication and meta-analysis: A necessary connection. Journal of Social Behavior and Personality, 8, 9–20. Amir, Y., & Sharon, I. (1990). Replication research: A ‘‘must’’ for the scientiﬁc advancement of psychology. Journal of Social Behavior and Personality, 5, 51–69. Bacharach, S. (1989). Organizational theories: Some criteria for evaluation. Academy of Management Review, 14, 496–515.

Multi-Study Research

191

Bedeian, A., Mossholder, K., Kemery, E., & Armenakis, A. (1992). Replication requisites: A second look at Klenke-Hamel and Mathieu (1990). Human Relations, 45, 1093– 1105. Blair, J., & Hunt, J. (1986). Getting inside the head of the management researcher one more time: Context-free and context-speciﬁc orientations in research. Journal of Management, 12, 147–166. Blalock, H. (1961). Theory, measurement, and replication in the social sciences. American Journal of Sociology, 66, 342–347. Blickle, G., Ferris, G., Munyon, T., Momm, T., Zettler, I., Schneider, P., & Buckley, M. (in press). A multi-source, multi-study investigation of job performance prediction by political skill. Applied Psychology: An International Review. Blickle, G., Kramer, J., Schneider, P. B., Meurs, J. A., Ferris, G. R., & Mierke, J. (2011). Role of political skill in job performance prediction beyond general mental ability and personality in cross-sectional and predictive studies. Journal of Applied Social Psychology, 41, 239–265. Blickle, G., Momm, T., Kramer, J., Mierke, J., Liu, Y., & Ferris, G. (2009). Construct and criterion-related validation of a measure of emotional reasoning skills: A two-study investigation. International Journal of Selection and Assessment, 17, 101–118. Breaux, D., Munyon, T., Hochwarter, W., & Ferris, G. (2009). Politics as a moderator of the accountability-job satisfaction relationship: Evidence across three studies. Journal of Management, 35, 307–326. Brockner, J., Spreitzer, G., Mishra, A., Hochwarter, W., Pepper, L., & Weinberg, J. (2004). Perceived control as an antidote to the negative effects of layoffs on survivors’ organizational commitment and job performance. Administrative Science Quarterly, 49, 76–100. Campbell, D. (1986). Science’s social system of validity-enhancing collective belief change and the problems of the social sciences. In: D. Fiske & R. Shweder (Eds), Metatheory in social science: Pluralisms and subjectivities (pp. 108–175). Chicago: University of Chicago Press. Campbell, D., & Stanley, J. (1963). Experimental and quasi-experimental designs for research on teaching. In: N. Gage (Ed.), Handbook of research on teaching (pp. 171–246). Chicago: Rand McNally. Cartwright, D. (1949). Some principles of mass persuasion. Human Relations, 2, 253–267. Child, J. (1972). Organization structure and strategies of control: A replication of the Aston study. Administrative Science Quarterly, 17, 163–177. Cooper, H., & Rosenthal, R. (1980). Statistical versus traditional procedures for summarizing research ﬁndings. Psychological Bulletin, 87, 442–449. Cronbach, L. (1957). The two disciplines of scientiﬁc psychology. American Psychologist, 12, 671–684. Cummings, L. (1978). Toward organizational behavior. Academy of Management Review, 3, 90–98. Dierdorff, E., & Morgeson, F. (2007). Consensus in work role requirements: The inﬂuence of discrete occupational context on role expectations. Journal of Applied Psychology, 92, 1228–1241. Dierdorff, E., Rubin, R., & Morgeson, F. (2009). The milieu of managerial work: An integrative framework linking work context to role requirements. Journal of Applied Psychology, 94, 972–988.

192

WAYNE A. HOCHWARTER ET AL.

Dubin, R. (1976). Theory building in applied areas. In: M. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 17–39). Skokie, IL: Rand McNally. Easley, R., Madden, C., & Dunn, M. (2000). Conducting marketing science: The role of replication in the research process. Journal of Business Research, 48, 83–92. Eden, D. (2002). From the editors. Academy of Management Journal, 45, 841–846. Eden, D. (2004). Reﬂections on the AMJ associate editor role. Academy of Management Journal, 47, 167–173. Ehrenberg, A., & Bound, J. (1993). Predictability and prediction. Journal of the Royal Statistical Society Series, 156, 167–206. Elbanna, S., & Child, J. (2007). Inﬂuences on strategic decision effectiveness: Development and test of an integrative model. Strategic Management Journal, 28, 431–453. Evanschitzky, H., & Armstrong, J. (2010). Replication of forecasting research. International Journal of Forecasting, 26, 4–8. Evanschitzky, H., Baumgarth, C., Hubbard, R., & Armstrong, J. (2007). Replication research’s disturbing trend. Journal of Business Research, 60, 411–415. Feingold, A. (1994). Gender differences in personality: A meta-analysis. Psychological Bulletin, 116, 429–456. Felps, W., Mitchell, T., Hekman, D., Lee, T., Holtom, B., & Harman, W. (2009). Turnover contagion: How coworkers’ job embeddedness and job search behaviors inﬂuence quitting. Academy of Management Journal, 52, 545–561. Ferris, G., Blickle, G., Schneider, P., Kramer, J., Zettler, I., & Solga, J. (2008a). Political skill construct and criterion-related validation: A two-study investigation. Journal of Managerial Psychology, 23, 744–771. Ferris, G., Bowen, M., Treadway, D., Hochwarter, W., Hall, A., & Perrewe´, P. (2006). The assumed linearity of organizational phenomena: Implications for occupational stress and well-being. In: P. Perrewe´ & D. Ganster (Eds), Research in occupational stress and wellbeing (pp. 205–232). Oxford, UK: Elsevier Science Ltd. Ferris, G., Hall, A., Royle, M., & Martocchio, J. (2004). Theoretical development in the ﬁeld of human resources management: Issues and challenges for the future. Organizational Analysis, 12, 231–254. Ferris, G., & Hochwarter, W. (2011). Organizational politics. In: S. Zedeck (Ed.), APA handbook of industrial and organizational psychology (Vol. 3, pp. 435–459). Washington, DC: American Psychological Association. Ferris, G., Munyon, T., Basik, K., & Buckley, M. (2008b). The performance evaluation context: Social, emotional, cognitive, political, and relationship components. Human Resource Management Review, 18, 146–163. Ferris, G., Rogers, L., Blass, F., & Hochwarter, W. (2009). The interaction of job-limiting pain and political skill on job satisfaction and organizational citizenship behavior. Journal of Managerial Psychology, 24, 584–608. Ferris, G., Russ, G., & Fandt, P. (1989). Politics in organizations. In: R. Giacalone & P. Rosenfeld (Eds), Impression management in the organization (pp. 143–170). Hillsdale, NJ: Lawrence Erlbaum. Ferris, G., Treadway, D., Kolodinsky, R., Hochwarter, W., Kacmar, C., & Douglas, C. (2005). Development and validation of the political skill inventory. Journal of Management, 31, 126–152. Fiedler, M., & Welpe, I. (2010). How do organizations remember? The inﬂuence of organizational structure on organizational memory. Organizational Studies, 31, 381–407.

Multi-Study Research

193

Finifter, B. (1972). The generation of conﬁdence: Evaluating research ﬁndings by random subsample replication. Sociological Methodology, 4, 112–175. Fried, Y., & Ferris, G. (1987). The validity of the job characteristics model: A review and metaanalysis. Personnel Psychology, 40, 287–322. Frink, D., & Ferris, G. (1998). Accountability, impression management, and goal setting in the performance evaluation process. Human Relations, 51, 1259–1283. Galtung, J. (1967). Theory and methods of social research. New York: Columbia University Press. Gould, J., & Kolb, W. (Eds). (1964). A dictionary of social sciences. London: Tavistock Publications. Gravetter, F., & Forzano, L. (2008). Research methods for the behavioral sciences. Belmont, CA: Wadsworth, Cenage Learning. Grifﬁn, M. (2007). Specifying organizational contexts: Systematic links between contexts and processes in organizational behavior. Journal of Organizational Behavior, 28, 859–863. Guttman, L. (1985). The illogic of statistical inference for cumulative science. Applied Stochastic Models and Data Analysis, 1, 3–10. Hall, A., Royle, M., Brymer, R., Perrewe´, P., Ferris, G., & Hochwarter, W. (2006). Relationship between felt accountability as a stressor and strain reactions: The neutralizing role of autonomy across two studies. Journal of Occupational Health Psychology, 11, 87–99. Hambrick, D. (2007). The ﬁeld of management’s devotion to theory: Too much of a good thing? Academy of Management Journal, 50, 1346–1352. Heath, C., & Sitkin, S. (2001). Big-B versus big-O: What is organizational about organizational behavior? Journal of Organizational Behavior, 22, 43–48. Hedges, L. (1987). How hard is hard science, how soft is soft science? American Psychologist, 42, 443–455. Helfat, C. (2007). Stylized facts, empirical research and theory development in management. Strategic Organization, 5, 185–192. Hendrick, C. (1990). Replications, strict replications, and conceptual replications: Are they important? Journal of Social Behavior and Personality, 5, 41–50. Hirschhorn, J., Lohmueller, K., Byrne, E., & Hirschhorn, K. (2002). A comprehensive review of genetic association studies. Genetic Medicine, 4, 45–61. Hochwarter, W., Ferris, G., Laird, M., Treadway, D., & Gallagher, V. (2010). Non-linear politics perceptions–work outcomes relationships: A three-study, ﬁve-sample investigation. Journal of Management, 36, 740–763. Hochwarter, W., Ferris, G., Zinko, R., Arnell, B., & James, M. (2007). Reputation as a moderator of political behavior-work outcomes relationships: A two-study investigation with convergent results. Journal of Applied Psychology, 92, 567–576. Hochwarter, W., Kacmar, C., Perrewe´, & Johnson, D. (2003). Perceived organizational support as a mediator of the relationship between politics perceptions and work outcomes. Journal of Vocational Behavior, 63, 438–465. Hochwarter, W., Laird, M., & Brouer, R. (2008). Board up the windows: The interactive effects of hurricane-induced job stress and perceived resources on work outcomes. Journal of Management, 34, 263–289. Hochwarter, W., Perrewe´, P., Hall, A., & Ferris, G. (2005). Negative affectivity as a moderator of the form and magnitude of the relationship between felt accountability and job tension. Journal of Organizational Behavior, 26, 517–535.

194

WAYNE A. HOCHWARTER ET AL.

Hochwarter, W., Rogers, L., Summers, J., Meurs, J., Perrewe´, P., & Ferris, G. (2009). Personal control antidotes to the strain consequences of generational conﬂict as a stressor: A twostudy constructive replication and extension. Career Development International, 14, 465–486. Hofer, S., & Picinnin, A. (2009). Integrative data analysis through coordination of measurement and analysis protocol across independent longitudinal studies. Psychological Methods, 14, 150–164. Hubbard, R., & Armstrong, J. S. (1994). Replications and extensions in marketing: Rarely published but quite contrary. International Journal of Research in Marketing, 11, 233–248. Hubbard, R., & Lindsay, R. (2002). How the emphasis on ‘original’ empirical marketing research impedes knowledge development. Marketing Theory, 2, 381–402. Hubbard, R., & Vetter, D. (1996). An empirical comparison of published replication research in accounting, economics, ﬁnance, management, and marketing. Journal of Business Research, 35, 153–164. Hubbard, R., Vetter, D., & Little, E. (1998). Replication in strategic management: Scientiﬁc testing for validity, generalizability, and usefulness. Strategic Management Journal, 19, 243–254. Hunter, J., & Schmidt, F. (2004). Methods of meta-analysis. Thousand Oaks, CA: Sage. Higgins, J., & Thompson, S. (2002). Quantifying heterogeneity in a meta-analysis. Statistics in Medicine, 21, 1539–1558. Humphrey, S., Nahrgang, J., & Morgeson, F. (2007). Integrating motivational, social, and contextual work design features: A meta-analytic summary and theoretical extension of the work design literature. Journal of Applied Psychology, 92, 1332–1356. Ioannidis, J. (2003). Genetic associations: False or true? Trends in Molecular Medicine, 9, 135–138. Ioannidis, J. (2005). Contradicted and initially stronger effects in highly cited clinical research. Journal of the American Medical Association, 294, 218–228. Iverson, G., Wagenmakers, E., & Lee, M. (2010). A model-averaging approach to replication: The case of prep. Psychological Methods, 15, 172–181. Jaccard, J., & Jacoby, J. (2010). Theory construction and model-building skills. New York: Guilford Press. Jawahar, I., Meurs, J., Ferris, G., & Hochwarter, W. (2008). Self-efﬁcacy and political skill as comparative predictors of task and contextual performance: A two-study constructive replication. Human Performance, 21, 1–20. Jick, T. (1979). Mixing qualitative and quantitative methods: Triangulation in action. Administrative Science Quarterly, 24, 602–611. John, O., Donahue, E., Ernst, D., & Kentle, R. (1991). The big ﬁve inventory-versions 4a and 5a. Berkley, CA: University of California, Berkeley Institute of Personality and Social Research. Johns, G. (2001). In praise of context. Journal of Organizational Behavior, 22, 31–42. Johns, G. (2006). The essential impact of context on organizational behavior. Academy of Management Review, 31, 385–408. Johns, G. (2010). Some unintended consequences of job design. Journal of Organizational Behavior, 31, 361–369. Johnson, W., & Bouchard, T. (2005). Constructive replication of the visual-perceptual-image rotation model in Thurstone’s (1941) battery of 60 tests of mental ability. Intelligence, 33, 417–430.

Multi-Study Research

195

Jones, K., Derby, P., & Schmidlin, E. (2010). An investigation of the prevalence of replication research in human factors. Human Factors: The Journal of the Human Factors and Ergonomics Society, 52, 586–595. Kacmar, K., & Carlson, D. (1997). Further validation of the perceptions of politics scale (POPS): A multiple sample investigation. Journal of Management, 23, 627–658. Kapoutsis, I., Papalexandris, A., Nikolopoulos, A., Hochwarter, W., & Ferris, G. (2011). Politics perceptions as moderator of the political skill-job performance relationship: A two-study, cross-national, constructive replication. Journal of Vocational Behavior, 78, 123–135. Katz, D., & Kahn, R. (1978). The social psychology of organizations. New York: Wiley. Kiewitz, C., Restubog, S., Zagenczyk, T., & Hochwarter, W. (2009). The interactive effects of psychological contract breach and organizational politics on perceived organizational support: Evidence from two longitudinal studies. Journal of Management Studies, 46, 806–834. Kirzner, I. (1973). Competition and entrepreneurship. Chicago: The University of Chicago Press. Kelly, C. (2006). Replicating empirical research in behavioral ecology: How and why it should be done but rarely ever is. Quarterly Review of Biology, 81, 221–236. Kelly, C., Chase, L., & Tucker, R. (1979). Replication in experimental communication research: An analysis. Human Communication Research, 5, 338–342. Kerlinger, F., & Lee, H. (2000). Foundations of behavioral research. Orlando, FL: Harcourt College Publishers. Killeen, P. (2007). Replication statistics. In: J. Osborne (Ed.), Best practices in quantitative methods (pp. 103–125). Thousand Oaks, CA: Sage. King, B., Felin, T., & Whetten, D. (2010). Finding the organization in organization theory: A meta-theory of the organization as social actor. Organization Science, 21, 290–305. Kline, R. (2008). Becoming a behavioral science researcher: A guide to producing research that matters. New York: Guilford. Kolodinsky, R., Hochwarter, W., & Ferris, G. (2004). Nonlinearity in the relationship between political skill and work outcomes: Convergent evidence from three studies. Journal of Vocational Behavior, 65, 294–308. Kuhn, T. (1970). The structure of scientiﬁc revolutions. Chicago: University of Chicago Press. Kurke, L., & Aldrich, H. (1983). Mintzberg was right!: A replication and extension of the nature of managerial work. Management Science, 29, 975–984. Lamal, P. (1991). On the importance of replication. In: J. Neuliep (Ed.), Replication in the social sciences (pp. 31–36). Newbury Park, CA: Sage. Lanivich, S. E., Brees, J. R., Hochwarter, W. A., & Ferris, G. R. (2010). P-E ﬁt as moderator of the accountability-employee reactions relationships: Convergent results across two samples. Journal of Vocational Behavior, 77, 425–436. Leavitt, K., Mitchell, T., & Peterson, J. (2010). Theory pruning: Strategies to reduce our dense theoretical landscape. Organizational Research Methods, 13, 644–667. Liddle, B., & Brazelton, E. (1996). Psychology faculty satisfaction and compliance with IRB procedures. IRB: A Review of Human Subjects Research, 18, 4–6. Liden, R., & Antonakis, J. (2009). Considering context in psychological leadership research. Human Relations, 62, 1587–1605. Lindsay, R., & Ehrenberg, A. (1993). The design of replicated studies. The American Statistician, 47, 217–228.

196

WAYNE A. HOCHWARTER ET AL.

Liu, Y., Ferris, G., Zinko, R., Perrewe´, P., Weitz, B., & Xu, J. (2007). Dispositional antecedents and outcomes of political skill in organizations: A four-study investigation with convergence. Journal of Vocational Behavior, 71, 146–165. Lohmueller, K., Pearce, C., Pike, M., Lander, E., & Hirschhorn, J. (2003). Meta-analysis of genetic association studies supports a contribution of common variants to susceptibility to common disease. Nature Genetics, 33, 177–182. Lykken, D. (1968). Statistical signiﬁcance in psychological research. Psychological Bulletin, 70, 151–159. Lykken, D. (1984). Statistical signiﬁcance in psychological research. In: T. Bateman & G. Ferris (Eds), Method and analysis in organizational research (pp. 54–64). Reston, VA: Reston Publishing. Mack, R. (1951). The need for replication research in sociology. American Sociological Review, 16, 93–94. McClelland, P., Liang, X., & Barker, V. (2010). CEO commitment to the status quo: Replication and extension using content analysis. Journal of Management, 36, 1251–1277. McKelvey, B. (2003). Postmodernism vs. truth in management theory. In: E. Locke (Ed.), Post modernism and management: Pros, cons, and the alternative. Research in the sociology of organizations (pp. 115–170). Amsterdam: Elsevier. McKinley, W. (2007). Managing knowledge in organization studies through instrumentation. Organization, 14, 123–146. McKinley, W. (2010). Organizational theory development: Displacement of ends. Organization Studies, 31, 47–68. McKinley, W., Mone, M., & Moon, G. (1999). Determinants and development of schools in organization theory. Academy of Management Review, 24, 634–648. Mezias, S., & Regnier, M. (2007). Walking the walk as well as talking the talk: Replication and the normal science paradigm in strategic management research. Strategic Organization, 5, 283–296. Miller, N. (1956). Effects of drugs on motivation: The value of using a variety of measures. Annals of the New York Academy of Sciences, 65, 318–333. Miller, G., & Contay, J. (1980). Differentiation in organizations: Replication and cumulation. Social Forces, 59, 265–274. Monroe, K. (1992). On replications in consumer research: Part I. Journal of Consumer Research, 19, 1. Morgeson, F., Dierdorff, E., & Hmurovic, J. (2010). Work design in situ: Understanding the role of occupational and organizational context. Journal of Organizational Behavior, 31, 351–360. Munyon, T., Hochwarter, W., Perrewe´, P., & Ferris, G. (2010). Optimism and the nonlinear citizenship behavior – job satisfaction relationship in three studies. Journal of Management, 36, 1505–1528. Nadkarni, S., & Barr, P. (2008). Environmental context, managerial cognition, and strategic action: An integrated view. Strategic Management Journal, 29, 1395–1427. Neuliep, J., & Crandall, R. (1991). Editorial bias against replication research. In: J. Neuliep (Ed.), Replication research in the social sciences (pp. 85–90). London: Sage. Oakes, J. (2002). Risks and wrongs in social science research: An evaluator’s guide to the IRB. Evaluation Review, 24, 443–478. Parkhe, A. (1993). ‘‘Messy’’ research, methodological predispositions, and theory development in international joint ventures. Academy of Management Review, 18, 227–268.

Multi-Study Research

197

Popper, K. (1959). The logic of scientiﬁc discovery. London: Hutchison. Popper, K. (1968). Conjectures and refutations. New York: Harper & Row. Pugh, D., & Hickson, D. (1972). Causal inference and the Aston Studies. Administrative Science Quarterly, 17, 273–276. Pugh, D., Hickson, D., & Hinings, C. (2006). Theory undergirding the Aston studies. In: J. Miner (Ed.), Organizational behavior 2: Essential theories of process and structure (pp. 267–288). Armonk, NY: M.E. Sharpe. Pugh, D., Hickson, D., Hinings, C., & Turner, C. (1968). Dimensions of organization structure. Administrative Science Quarterly, 13, 65–105. Pugh, D., Hickson, D., Hinings, C., & Turner, C. (1969). The context of organization structures. Administrative Science Quarterly, 14, 91–114. Reid, L., Rotfeld, H., & Wimmer, R. (1982). How researchers respond to replication requests. The Journal of Consumer Research, 9, 216–218. Rosenthal, R. (1976). Experimenter effects in behavioral research. New York: Irvington. Rosenthal, R. (1979). The ‘‘ﬁle drawer problem’’ and tolerance for null results. Psychological Bulletin, 86, 638–641. Rosenthal, R. (1990). Replication in behavioral research. Journal of Social Behavior and Personality, 5, 1–30. Rosenthal, R. (1991). Cumulating psychology: An appreciation of Donald T. Campbell. Psychological Science, 2, 217–221. Rosenthal, R., & Rosnow, R. L. (1991). Essentials of behavioral research: Methods and data analysis. New York: McGraw-Hill. Rousseau, D., & Fried, Y. (2001). Location, location, location: Contextualizing organizational research. Journal of Organizational Behavior, 22, 1–13. Samra-Fredericks, D. (2010). Where is the ‘I’? One silence in strategy research. In: B. Silverman (Ed.), The globalization of strategy research-advances in strategic management (pp. 411–444). London: Emerald Group Publishing Limited. Schmidt, F. (1992). What do data really mean? Research ﬁndings, meta-analysis, and cumulative knowledge in psychology. American Psychologist, 47, 1173–1181. Schmidt, S. (2009). Shall we really do it again? The powerful concept of replication is neglected in the social sciences. Review of General Psychology, 13, 90–100. Schmidt, A., Dolis, C., & Tolli, A. (2009). A matter of time: Individual differences, contextual dynamics, and goal progress effects on multiple-goal self-regulation. Journal of Applied Psychology, 94, 692–709. Schmidt, F., & Hunter, J. (2001). Meta-analysis. In: N. Anderson, D. Ones, H. Sinangil & C. Viswesvaran (Eds), Handbook of industrial, work, and organizational psychology (Vol. 1, pp. 51–70). London: Sage. Short, J., McKelvie, A., Ketchen, D., & Chandler, G. (2009). Firm and industry effects on ﬁrm performance: A generalization and extension for new ventures. Strategic Entrepreneurship Journal, 3, 47–65. Sidman, M. (1960). Tactics of scientiﬁc research. New York: Basic Books. Singh, K., Ang, S., & Leong, S. (2003). Increasing replication for knowledge accumulation in strategy research. Journal of Management, 29, 533–549. Smith, N. (1970). Replication study: A neglected aspect of psychological research. American Psychologist, 25, 970–975. Srivastava, A., Locke, E., Judge, T., & Adams, J. (2010). Core self-evaluations as causes of job satisfaction: The mediating role of seeking task complexity. Journal of Vocational Behavior, 77, 255–265.

198

WAYNE A. HOCHWARTER ET AL.

Strong, M., Jeanneret, P., McPhail, S., Blakley, B., & D’Egidio, E. (1999). Work context: Taxonomy and measurement of the work environment. In: N. Peterson, M. Mumford, W. Borman, P. Jeanneret & E. Fleishman (Eds), An occupational information system for the 21st century: The development of the O NET (pp. 127–145). Washington, DC: American Psychological Association. Sullivan, P. (2007). Spurious genetic associations. Biological Psychiatry, 61, 1121–1126. Tang, J., Kacmar, M., & Busenitz, L. (in press). Alertness in the pursuit of new opportunities. Journal of Business Venturing. Toncar, M., & Munch, J. (2010). Meaningful replication: When is a replication no longer a replication? A rejoinder to Stella and Adam (2008). Journal of Marketing, Theory, and Practice, 18, 71–80. Treadway, D., Ferris, G., Hochwarter, W., Perrewe´, P., Witt, L., & Goodman, J. (2005). The role of age in the perceptions of politics–job performance relationship: A three-study constructive replication. Journal of Applied Psychology, 90, 872–881. Tsang, E., & Kwan, K. (1999). Replication and theory development in the organizational sciences: A critical realistic perspective. Academy of Management Review, 24, 759–780. Tweney, R. (2008). Studies in historical replication in Psychology I: Introduction. Science and Education, 17, 467–475. Umapathy, S. (1987). Unfavorable variances in budgeting: Analysis and recommendations. In: K. Ferris & J. Livingstone (Eds), Management planning and control (pp. 163–179). Beavercreek, OH: Century VII. Umphress, E., Labianca, G., Brass, D., Kass, E., & Scholten, L. (2003). The role of instrumental and expressive social ties in employees’ perceptions of organizational justice. Organization Science, 14, 738–753. Van Maanen, J., Sorensen, J., & Mitchell, T. (2007). The interplay between theory and method. Academy of Management Review, 32, 1145–1154. Watson, D., Clark, L., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scale. Journal of Personality and Social Psychology, 54, 1063–1070. Wayne, S., & Ferris, G. (1990). Inﬂuence tactics, affect, and exchange quality in supervisorsubordinate interactions: A laboratory experiment and ﬁeld study. Journal of Applied Psychology, 75, 487–499. Weick, K. (1989). Theory construction as disciplined imagination. Academy of Management Review, 14, 516–531. Whetten, D. (1989). What constitutes a theoretical contribution? Academy of Management Review, 14, 490–495. Whetten, D. (2009). An examination of the interface between context and theory applied to the study of Chinese organizations. Management and Organization Review, 5, 29–55. Williams, G., Frankel, R., Campbell, T., & Deci, E. (2003). The science of the art of medicine: Research on the biopsychosocial model of health care. In: R. Frankel, T. Quill & S. McDaniel (Eds), The biopsychosocial approach: Past, present, and future (pp. 109–121). Rochester, NY: University of Rochester Press. Williams, L., & Anderson, S. (1991). Job satisfaction and organizational commitment as predictors of organizational citizenship and in-role behaviors. Journal of Management, 17, 601–617.

Multi-Study Research

199

Witt, L., & Ferris, G. (2003). Social skill as moderator of the conscientiousness-performance relationship: Convergent results across four studies. Journal of Applied Psychology, 88, 809–821. Wong, G., & Birnbaum-More, P. (1994). Culture, context and structure: A test of Hong Kong banks. Organizational Studies, 15, 99–124. Wright, P., Kacmar, K., McMahan, G., & Deleeuw, K. (1995). P ¼ f (M X A): Cognitive ability as a moderator of the relationship between personality and job performance. Journal of Management, 21, 1129–1139. Zywicki, T. (2007). Institutional review boards as academic bureaucracies: An economic and experiential analysis. Northwestern University Law Review, 101, 861–895.

TEMPLATES AND TURNS IN QUALITATIVE STUDIES OF STRATEGY AND MANAGEMENT Ann Langley and Chahrazad Abdallah ABSTRACT Purpose – This chapter presents four different approaches to doing and writing qualitative research in strategy and management based on different epistemological foundations. It describes two well-established ‘‘templates’’ for doing such work, and introduces two more recent ‘‘turns’’ that merit greater attention. Design/Methodology/Approach – The chapter draws on methodological texts and a detailed analysis of successful empirical exemplars from the strategy and organization literature to show how qualitative research on strategy processes can be effectively carried out and written up. Findings – The two ‘‘templates’’ are based on different logics and modes of writing. The ﬁrst is based on a positivist epistemology and aims to develop nomothetic theoretical propositions, while the second is interpretive and more concerned to capture and gain insight from the meanings given to organizational phenomena. The two ‘‘turns’’ (the practice turn and the discursive turn) are not as well deﬁned but are generating innovative contributions based on new ways of considering the social world. Building Methodological Bridges Research Methodology in Strategy and Management, Volume 6, 201–235 Copyright r 2011 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-8387/doi:10.1108/S1479-8387(2011)0000006007

201

202

ANN LANGLEY AND CHAHRAZAD ABDALLAH

Originality/Value – The chapter should be helpful to researchers considering qualitative methods for the study of strategy processes. It contributes by comparing different approaches and by recognizing that part of the challenge of doing qualitative research lies in writing it up to communicate its insights in a credible way. Thus while describing the different methods, the chapter also draws attention to effective forms of writing. In addition, it introduces and assesses two more recent ‘‘turns’’ that offer promising routes to novel insight as well as having particular ontological and epistemological afﬁnities with qualitative research methods. Keywords: case studies; qualitative research; strategy process; strategy as practice; discourse This chapter discusses a range of ways in which qualitative methods may be used to study and theorize about strategy processes, that is, to examine the how questions of strategic management that deal with phenomena such as decision making, learning, strategizing, planning, innovating, and changing (Van de Ven, 1992). Qualitative data have particular strengths for understanding processes because of their capacity to capture temporally evolving phenomena in rich detail, something that is hard to do with methodologies based on quantitative surveys or archival databases that are coarse-grained and tend to ‘‘skim the surface of processes rather than plunging into them directly’’ (Langley, 1999, p. 705). Our focus will thus be on the study of strategy processes taken as an empirical phenomenon drawing on qualitative data that examines these processes over time, that is, using what has been called ‘‘process data’’ (Langley, 1999). Process data tend to incorporate a mix of in vivo observations (meetings, conversations, events, shadowing, etc.), memories and interpretations (real time or retrospective interviews, focus groups, questionnaires, diaries, etc.) and artifacts (minutes, plans, reports, archival records, etc.). However, the key challenge of doing qualitative research on organizational processes lies not so much in collecting these data but in making sense of them to generate a valuable theoretical contribution. The data tend to be complex, messy, eclectic, and with varying degrees of temporal embeddedness. In a previous paper, the ﬁrst author proposed seven strategies for addressing this challenge include composing case narratives, quantiﬁcation of incidents, using alternate theoretical templates, grounded theorizing, visual mapping, temporal decomposition, and case comparisons (Langley, 1999).

Templates and Turns in Qualitative Studies of Strategy and Management

203

In this chapter, while building on previous work, we take a somewhat different perspective on the mobilization of qualitative data to analyze strategy processes. First, the chapter recognizes that qualitative methods are associated with a range of different epistemological assumptions and that these may have important implications for the way in which data are interpreted as well as for the theoretical products generated by the analysis. Second, the chapter also recognizes that part of the challenge of doing qualitative research lies in writing it up to communicate its insights in a credible way. Thus while describing methods, we also draw attention to effective forms of writing. Third, we focus the chapter around two rather well-established ‘‘templates’’ for doing qualitative studies of strategy processes and contrast these with two more recent ‘‘turns’’ that offer promising routes to novel insight as well as having particular ontological and epistemological afﬁnities with qualitative research methods. We begin by describing the two ‘‘templates’’ that have each given rise to a body of work where it seems that the norms of presentation and methodological process have become to a degree standardized and institutionalized among a set of scholars. These templates are far from exhaustive of approaches for qualitative research on strategy processes. However, we believe that they are particularly instructive. Then we consider the implications of two nascent ‘‘turns’’ (the practice turn and the discursive turn) in qualitative analysis of strategy processes that we argue merit greater attention.

TWO TEMPLATES One of the common complaints (but for some of us, the rather attractive qualities) about qualitative research is that unlike quantitative studies, the rules, formats, and norms for doing, writing, and publishing it are not uniform or well-established. It is not for nothing that Michael Pratt titled a recent editorial in Academy of Management Journal about writing qualitative research for the journal ‘‘For the lack of a boilerplate’’ (Pratt, 2009). We do however see the emergence of at least two templates for qualitative studies that have achieved some penetration in the North American management journals, that are each based on different epistemological assumptions, and that are sometimes being used as yardsticks by others. In honor of their originators, we label these the Eisenhardt method and the Gioia method. Both of these have given rise to some highly inﬂuential contributions to strategy process research.

204

ANN LANGLEY AND CHAHRAZAD ABDALLAH

As mentioned earlier, in describing these approaches, we focus not only on the logical structure of the method itself but also on the rhetorical structure that is used to support it in published articles. These two dimensions seem to us to be inextricably linked and indeed contribute to constituting the template. Since Golden-Biddle and Locke (2006, 1993) drew our attention to the way in which skillful writers of qualitative research convince their readers, there is increasing realization that writing and rhetoric matter. Thus, the two approaches each have their own internal logics and rhetorical power that we describe below and summarize in Table 1. Note that our accounts of these approaches are based for the most part on a close reading of published papers by key authors, but include also ideas gleaned from conference presentations and in the second case from personal communication.1

The Eisenhardt Template: Credibly Novel Nomothetic Theory from Case Comparisons Kathleen Eisenhardt’s (1989a) article on ‘‘Building theories from case study research’’ is now a classic methodological reference both within the ﬁeld of management and beyond (Ravenswood, forthcoming), with over 11,000 citations on Google scholar at time of writing. Even more impressive perhaps, Eisenhardt and her colleagues have published a continuous stream of exemplars of the approach that while innovating in their substantive topic foci, replicate both the logic of the method and the rhetoric underpinning its ﬁrst empirical applications (Eisenhardt & Bourgeois, 1988; Eisenhardt, 1989b). For example, papers coauthored by Eisenhardt or her students and collaborators have examined factors associated with fast decision making (Eisenhardt, 1989b), successful approaches to continuous innovation (Brown & Eisenhardt, 1997), charter changes in multi-divisional businesses (Galunic, 2001; Galunic & Eisenhardt, 1996), how entrepreneurs successfully shape organizational boundaries and markets in their favor (Santos & Eisenhardt, 2009), networking strategies associated with successful industry positioning (Ozcan & Eisenhardt, 2009), the role of seller perspectives and trust in acquisitions (Graebner & Eisenhardt, 2004; Graebner, 2004, 2009), patterns of planning and improvisation in successful internationalization (Bingham, 2009), the origins of success in cross-business collaboration (Martin & Eisenhardt, 2010), and the strategies used by entrepreneurs to build relationships with venture capitalists (Hallen & Eisenhardt, 2009).

Templates and Turns in Qualitative Studies of Strategy and Management

Table 1.

Two Templates for Qualitative Studies of Strategy and Management. The ‘‘Eisenhardt Method’’

Key methodological reference

Eisenhardt (1989a)

Exemplar empirical articles

Eisenhardt (1989b), Brown and Eisenhardt (1997), Martin and Eisenhardt (2010) Yin (2009) on case study research, but see also Miles and Huberman (1994) Post-positivist assumptions – Purpose: developing theory in the form of testable propositions – Search for facts (e.g., emphasis on court-room style interviewing) – Product: nomothetic theory

Central methodological inspirations Epistemological foundations and purposes

Logic of the method

Design to maximize credible novelty – Multiple cases (4–10) chosen to be sharply distinct on one key dimension (e.g., performance) while similar on others – Interview data with diverse informants – Identify elements that distinguish high and low performing cases building on cross-case comparison

Rhetoric of the writing

205

The ‘‘Gioia Method’’ None, but see Gioia (2004) for personal reﬂections on research philosophy Gioia and Chittipeddi (1991), Corley and Gioia (2004), Gioia et al. (2010) Glaser and Strauss (1967); Strauss and Corbin (1990) on grounded theory Interpretive assumptions – Purpose: capturing and modeling of informant meanings – Search for informants’ understandings of organizational events. – Product: process model/ novel concept Design for revelation, richness and trustworthiness – Single case chosen for its revelatory potential and richness of data

– Validity and reliability from multiple researchers, triangulation of data

– Real-time interviews and observation – Build ‘‘data structure’’ by progressive abstraction starting with informant ﬁrstorder codes and building to second-order themes and aggregate dimensions – Trustworthiness from insideroutsider roles, member checks, triangulation

Establishing novelty: Contrasting ﬁndings with previous research; Providing evidence: Data presentation in two steps: (a) data tables; (b) narrative examples of high and low cases

Establishing the gap: Show how this study ﬁlls a major gap Distilling the essence: Present the data structure emphasizing second-order themes and overarching dimensions

206

ANN LANGLEY AND CHAHRAZAD ABDALLAH

Table 1. (Continued )

Examples of other authors using similar approaches

The ‘‘Eisenhardt Method’’

The ‘‘Gioia Method’’

Offering explanation: Ask why for every proposition. Reasons offered building on data and literature; Integrating contribution: Link separate propositions together to build theory

Elaborating the story: Elaborate the model in two ways; (a) present the narrative; (b) additional quotes in tables Reafﬁrm contribution: Return to opening gap to show novel insight.

Zott and Huy (2007), Gilbert (2005), Maitlis (2005)

Maguire and Phillips (2008), Anand, Gardner, and Morris (2007), Anand et al. (2007), Rindova et al. (2011)

In another sign of the inﬂuence of this approach, in the late 1990s, the ﬁrst author received a review on a submission to a journal in which the reviewer used Eisenhardt’s (1989a) eight-step method as a framework to guide the review. Every one of the eight steps was analyzed in detail and the submission was matched up against its standards. For better or worse, the method had already acquired something of the character of a template. Epistemological Foundations and Purposes: Toward Testable Propositions Eisenhardt (1989a, p. 546) establishes her method as positivist in orientation, aimed at ‘‘the development of testable hypotheses and theory which are generalizable across settings.’’ The method is oriented toward induction, that is, generating sets of formal propositions from case study evidence, and is presented as suitable for situations where little is known about a phenomenon or where current perspectives are conﬂicting or confusing, and where case study evidence can therefore be seen to contribute novel insight. At the same time, the method draws inspiration from Yin’s (2009 [1984]) discussion of case study research, emphasizing a logic of replication in which different cases are considered (much like different experiments) as occasions for verifying and elaborating theoretical relationships developed from previous cases. Overall, after reading many of the articles produced with this approach, its power seems to lie in its ability to generate ﬁndings that are claimed as novel-even ‘‘surprising,’’ and yet at the same time to render these ﬁndings highly credible, something that appears

Templates and Turns in Qualitative Studies of Strategy and Management

207

paradoxical at ﬁrst sight. The need for both defamiliarization and plausibility in qualitative research is probably universal and has been noted before (e.g., Golden-Biddle & Locke, 1993). However, it seems to be a particularly strong leitmotiv underlying this particular approach, and both the logic of the method and the rhetoric of the writing in empirical articles seem designed to achieve it. Logic of the Method: Designing to Maximize the Chances of Credible Novelty The replication logic proposed by Eisenhardt requires a substantial number of comparative units of analysis or cases [Eisenhardt (1989a) suggests from four to ten] because the objective is to abstract from these cases common constructs that can then be used to describe and compare generic process components across all the cases (usually in terms of categorical or ordinal scales), and ultimately to relate these to outcome constructs representing some kind of performance. Although the speciﬁcs of individual cases contribute importantly to the nature of the constructs induced from the data, it is their common dimensions across cases and not their idiosyncratic features that are emphasized. Thus, the processes examined using this approach are taken as wholes synthesized into a limited number of descriptive dimensions (constructs), rather than being elaborated idiographically. However, to make this logic work, and to optimize the chances of credible but novel insight, the cases cannot be and are not chosen arbitrarily. Key elements of design include choosing and gaining access to promising phenomena where new knowledge is likely to emerge, setting up comparisons to maximize differences on one dimension while controlling for differences on others, and ensuring coverage of perspectives within each case. Planning for novel insight of course begins with the research questions and empirical phenomena studied. Thus, Eisenhardt and her colleagues have studied phenomena that have often been subject to quantitative research previously (e.g., acquisitions, alliances, new technology ventures), but where prior process-oriented research has been limited, and particularly so in the dynamic fast-paced technological settings they have favored. Additionally, recent studies demonstrate an impressive level of access to complex situations that few have been able to obtain previously, enhancing the probability of novel ﬁndings. For example, Ozcan and Eisenhardt (2009) accessed six new entrants to the wireless video-gaming industry (of which two turned out to be the top players) conducting three waves of interviews with multiple organization members over time as well as interviews with their main partner ﬁrms as they constructed their alliance portfolios. One might speculate that the potential for such good access to novel situations

208

ANN LANGLEY AND CHAHRAZAD ABDALLAH

might be enhanced by previous successful research that has had practical impact (as evidenced in this case by several Harvard Business Review articles). While controlling for secondary sources of variation (such as size, industry, etc.), cases are also carefully selected to represent what Pettigrew (1990) labeled ‘‘polar types,’’ thus emphasizing comparisons between extremes so that, for example, the distinguishing features of highperforming and low-performing cases have the strongest possible chance of emerging clearly. As Eisenhardt and Graebner (2007, p. 27) explain, ‘‘Although such an approach can surprise reviewers because the resulting theory is so consistently supported by the empirical evidence, this sampling leads to very clear pattern recognition of the central constructs, relationships, and logic of the focal phenomenon.’’ Sometimes, the authors have collected data on more cases than they actually used in the analysis to preserve the sharpness of the contrast (e.g., Brown & Eisenhardt, 1997). One might ask what is missing from our understanding by removing consideration of average run-of-the-mill ﬁrms. However, the sharpness in contrast is clearly helpful in enhancing the clarity of insights. The credibility of those insights is further enhanced by sampling multiple perspectives within each case. For example, Graebner (2004, 2009) interviewed both buyers and sellers in her study of acquisitions, Martin and Eisenhardt (2010) interviewed managers at corporate and business unit levels in their study of cross-divisional collaboration. While interviews tend to be the main source of information with all their inherent limitations, strong emphasis is also placed on collecting several kinds of data (e.g., quantitative scales embedded in interview protocols to triangulate responses; archival sources), as well as on obtaining factual accounts through techniques such as ‘‘courtroom style questioning’’ (mentioned in the methods sections of most published articles). Finally, tandem interviewing, electronic recording, and rapid transcription are cited as further means of enhancing validity and reliability. A good research question, a strong design and excellent data are clearly helpful for developing novel and credible insight, but it is in the analysis that this all comes together. Eisenhardt and her colleagues describe data analysis as essentially a two-stage process, beginning ﬁrst with the construction of complete within-case narratives and followed by iterative processes of case comparison that continues until a set of constructs that might explain similarities and differences in outcomes begins to emerge (Eisenhardt, 1989a). The fashioning of these constructs is a creative moment of the method because it involves bringing together pieces of case evidence to reﬁne

Templates and Turns in Qualitative Studies of Strategy and Management

209

emerging measures of constructs by tabulating data, as well as elaborating understanding of how and why emerging relationships might make sense. Clearly without being there, it is hard to experience the process of analysis itself. However, its products can be appreciated more easily, which brings us to the rhetorical dimension of the template. Rhetoric of the Writing: Establishing Novelty, Providing Evidence, and Offering Explanation In addition to a methodological approach that maximizes the chances of offering a novel but credible contribution, Eisenhardt and her colleagues have perfected a distinct mode of writing case study articles that establishes this value. We will use Eisenhardt’s (1989b) article on the speed of organizational decision making and a more recent study by Martin and Eisenhardt (2010) to illustrate the approach. The most interesting rhetorical feature concerns how each individual ﬁnding or proposition is argued in three key moves. The ﬁrst move involves establishing novelty. Here, for each ﬁnding, a contrast is explicitly drawn between what previous literature and theory would lead one to expect and the current ﬁnding. For example, Eisenhardt (1989b) uses expressions such as, ‘‘The data from this research indicate a different view’’ (p. 549), ‘‘In contrast,’’ (pp. 555, 559, 562). Martin and Eisenhardt (2010) use expressions such as ‘‘unexpectedly’’ (p. 271) and ‘‘However we observed the opposite’’ (p. 283). The sharply constructed contrast serves to introduce an unexpected or novel ﬁnding but also sets up a tension that then has to be resolved – if this is so surprising, can we believe it? The resolution begins with the second move involving the presentation of the evidence. In most of this stream of work, this occurs in two steps. The ﬁrst step involves presenting an overall semi-quantitative portrait of the evidence supporting the proposed relationship in a table in which cases are ordered vertically from more to less high performing. The columns of the table draw together evidence from various sources. For example, Martin and Eisenhardt (2010) argued that engaging in deliberate learning activities contributes to successful cross-divisional collaboration and tabulated evidence on this that included both counts of the number of activities engaged in and two or three quotes from different sources in each ﬁrm. As is typical, their chapter includes one table for each proposition (ﬁve in this case; with from four to seven columns) plus an additional table documenting evidence of performance (including multiple columns for different quantitative assessments as well as quotes). Some writers might stop the

210

ANN LANGLEY AND CHAHRAZAD ABDALLAH

presentation of the data here, since the tabulations generally provide unambiguous support for the propositions and extracts from the data on all the cases.2 However, the authors generally elaborate on the ﬁndings by offering more qualitative narrative examples of typically two highperforming and two low-performing units that add depth to the information provided in tables. Eisenhardt and colleagues then always engage in an important ﬁnal move before closing the presentation of their propositions. This is to ask themselves why the observed relationships might hold, that is, offering not just evidence but explanation. Usually two or three reasons are offered for each proposition. To present these, the authors draw on both the data themselves and on prior theory and research in an attempt to deepen understanding, and thus further enhance the credibility of the relationships discovered. This may also be an occasion to reconcile the ﬁndings of the research with prior literature (see, e.g., Eisenhardt, 1989b). The importance of offering explanation is sometimes forgotten in qualitative research, but it is particularly important, because it is here that a mere observed empirical regularity is transformed into the beginnings of a theoretical contribution. Extending this theme, a theory-building multiple case study will offer a strong contribution to knowledge if its atomistic propositions can further be integrated together into a coherent theoretical story that reaches beyond the individual components. This ﬁnal step is also important and can be quite challenging because the need for novelty and credibility must also be maintained. For example, after presenting a series of propositions about factors that seemed associated with successful continuous innovation, it is at this stage that Brown and Eisenhardt (1997) began to draw on complexity theory as a metaphor to tie their ﬁndings together, noting that a persistent theme in their work was the simultaneous need for structure but also for ﬂexibility. Assessing the Template: Limitations and Variations Overall, the ‘‘Eisenhardt method’’ has emerged as a very successful approach to strategy process research as shown by the multiple publications of the author and her collaborators. Although its logical and rhetorical structure have not been quite so sharply replicated by other authors, many have drawn inspiration from it while adapting it to their distinctive research problems and contexts and mobilizing other sources of methodological inspiration. For example, Zott and Huy (2007) used a comparative case method with similar features to examine how more or less successful entrepreneurial startups used symbolic management approaches, including a

Templates and Turns in Qualitative Studies of Strategy and Management

211

focus on extreme cases to sharpen insights. In a prize-winning paper, Gilbert (2005) used a similar method to explore patterns of inertia and modes of overcoming them in the newspaper industry. Others have used multiple case study methods that although not necessarily directly inspired by Eisenhardt’s work share methodological and rhetorical elements. For example, Maitlis (2005) used multiple cases to generate a model of different forms of leader and stakeholder sensemaking and their relationships with outcomes using extensive tabulated data to add credibility to the relationships she identiﬁed. The template has however its boundary conditions and limitations. First, while empirical processes are analyzed and interesting new process ‘‘constructs’’ emerge from these studies, the approach often tends to lead to ‘‘variance’’ rather than ‘‘process’’ theorizations, that is, the emphasis in most applications is on explaining variation in outcomes rather than on understanding patterns of evolution over time (Mohr, 1982; Langley, 1999, 2009). Variance models have their own value but they compress time, limit attention to temporal ordering, and assume that there is such a thing as a ﬁnal outcome, something that can be questionable in many cases. For example, ﬁrm performance evolves over time – it is not ﬁxed once and for all. Performance ‘‘outcomes’’ are just way-stations in ongoing processes. Indeed, they might sometimes better be seen as inputs to ongoing processes since evaluations and interpretations of performance can have important effects on subsequent actions (Langley, 2007). There is however actually no inherent reason why multiple case analyses cannot be used to develop process models and elements of ordering do appear in a few studies (e.g., Bingham, 2009; Galunic & Eisenhardt, 1996). Yet, when this is the objective, the logic is different from the dominant pattern described above. Rather than seeking explanations for differences between cases, a process theoretical analysis requires looking for regularities in temporal patterns across cases. One study that does this rather well using multiple cases is Ambos and Birkinshaw’s (2010) recent paper on the developmental patterns and transitions of new science-based ventures. This study indeed demonstrates how the outcomes of one phase of development become stimuli for change for the next. Nevertheless, the retrospective interview methodology used in multiple case studies often limits the depth of evolutionary process detail that can be captured in these studies. A second issue concerns the degree to which the ﬁndings emerging from such studies are indeed as theoretically novel and surprising as often claimed. However interesting the studies are, the subsequent capacity of the authors to explain their results drawing on other literature suggests that the

212

ANN LANGLEY AND CHAHRAZAD ABDALLAH

rhetoric of surprise might sometimes be overemphasized. Several authors have mitigated such claims while still legitimating their research efforts and methods by referring to them as ‘‘theory elaboration’’ rather than ‘‘theory development’’ (Lee, Mitchell, & Sablynski, 1999). In most cases, this would seem to be a more realistic and yet valuable research enterprise, because it involves explicitly building on previous work while developing it in new directions. Finally, as we noted at the beginning of this section, the Eisenhardt multiple case method is positivist in orientation [or more precisely, what Guba and Lincoln (1994) would label post-positivist]. It attempts to access ‘‘factual’’ data about what happened in a sample of relevant processes, and it aims to develop generalizable nomothetic causal laws about objectively observable phenomena in the real world. There are other ways of conceiving the research enterprise with qualitative research, one of which we shall consider in the next section.

The Gioia Method: Interpretive Modeling of Informant Understandings over Time Ever since Kathleen Eisenhardt published her ﬁrst papers using the distinctive comparative case method described above, the approach has been both a source of admiration and emulation for many, yet a source of some discomfort to certain other qualitative researchers who have seen in it a distortion of the principles of the traditional interpretive case method that emphasizes depth of understanding of unique situations (Dyer & Wilkins, 1991; Ahrens & Dent, 1998). Yet, cross-case comparative studies and single case analyses have very different objectives and make different kinds of theoretical contributions, valued for different reasons (Langley, 1999). One group of scholars who appear to have perfected an approach for both doing and successfully publishing single in-depth interpretive case studies is Dennis Gioia and his colleagues and students. Their qualitative work has a distinctive ﬂavor that has given rise to numerous empirical studies, beginning with a series on strategic sensemaking and sensegiving in the 1990s (Gioia & Chittipeddi, 1991; Gioia, Thomas, Clark, & Chittipeddi, 1994; Gioia & Thomas, 1996) and following up with another impressive series of papers on organizational identity change in different settings with or by colleagues and students (e.g., Corley & Gioia, 2004; Corley, 2004; Nag, Corley, & Gioia, 2007; Clark, Gioia, Ketchen, & Thomas, 2010; Gioia, Price, Hamilton, & Thomas, 2010). The paper by Corley and Gioia (2004)

Templates and Turns in Qualitative Studies of Strategy and Management

213

dealing with identity ambiguity during a spinoff (based on Kevin Corley’s Ph.D. thesis) received the ASQ Scholarly Contribution Award for the most signiﬁcant paper published ﬁve years earlier and has been frequently cited not only as a strong contribution to organizational identity theory but also as a methodological exemplar by other authors (e.g., Pratt, 2009; Rindova, Dalpiaz, & Ravasi, 2011; Maguire & Phillips, 2008). From our personal observations, it is frequently mentioned by reviewers. There is evidence that we have here the elements of another emergent template. Epistemological Foundations and Purposes: Toward Interpretive Understanding Unlike Kathleen Eisenhardt, Dennis Gioia has never published a paper explicitly describing step by step his methodology. However, in a reﬂexive piece about his career as an organizational scholar, he noted: In my research life, I am a grounded theorist. I pick people’s brains for a living, trying to ﬁgure out how they make sense of their organizational experience. I then write descriptive, analytical narratives that try to capture what I think they know. Those narratives are usually written around salient themes that represent their experience to other interested readers. (Gioia, 2004, p. 101)

This quotation neatly sums up the interpretive philosophy driving the approach described here. The data Gioia and his colleagues are interested in concern how people understand the changes they are both instigating and dealing with, and how those meanings evolve. The key methodological references the authors build on are the original grounded theorists (Glaser & Strauss, 1967; Strauss & Corbin, 1990). The theoretical products they generate are narratives that attempt at the same time to provide closeness to so-called ‘‘ﬁrst order’’ participant perspectives, and yet to add the authors’ ‘‘second-order’’ interpretations of these perspectives distilled into a set of inter-related overarching categories or themes that resonate with both participants and readers, and yet communicate new insight. Of course, as in the previous case, there remains a certain tension between novelty and plausibility. We now brieﬂy summarize the logic of the method and the rhetoric of the writing that contribute to achieve both. The Logic of the Method: Designing for Revelation, Richness, and Trustworthiness When studying one case at a time in the hope of offering distinctive insights, it would seem important to choose the right site. Yin (2009) suggests that three different logics can be used to select sites for holistic case studies: choose ‘‘critical’’ cases for the ‘‘test’’ of a particular theory, choose ‘‘extreme’’ cases where something exceptional seems to be occurring, or choose ‘‘revelatory’’ cases that offer high potential for developing new

214

ANN LANGLEY AND CHAHRAZAD ABDALLAH

insight into an understudied phenomenon. Gioia and colleagues’ recent contributions seem to have been designed to build successively on a developing body of cognitively oriented theories of sensemaking and identity change, each study adding new identity-critical situations in a kind of sequential revelatory case logic. For example, while Corley and Gioia (2004) examined the dynamics of identity change during a spinoff, Nag et al. (2007) looked at identity change in the context of the addition of new forms of knowledge, Clark et al. (2010) focused on evolving identity dynamics during a merger, and Gioia et al.’s (2010) study investigated the emergence of identity in a new organization. The timing of these studies has been such that although others have worked in the area organizational identity, each individual study was able to lay claim to a novel context and related set of insights and the whole series of studies takes on a programmatic character. Beyond the technical criterion of selecting cases for their revelatory potential, in-depth ethnographic studies of change require organizations that provide good access to ensure data richness. Thus, Gioia and colleagues have not hesitated to study organizations close to home: ‘‘No organization is more salient or more important to me than my own organization, so that helps to explain why I sometimes study my own university’’ (Gioia, 2004, p. 102). For several articles, Gioia and colleagues have also developed a rather innovative insider-outsider perspective that truly optimizes access to richness, in which one member of the research team has been an active participant in the events studied (e.g., Gioia et al., 1994, 2010; Gioia & Chittipeddi, 1991). The authors argue that the combination of insider and outsider perspectives both enriches the research and can contribute to its trustworthiness as long as precautions are taken to ensure conﬁdentiality and independence (Gioia et al., 2010). In terms of data collection more generally, the researchers have made extensive use of interviews, often carried out in multiple rounds and at multiple levels and positions, but also of observational data (Clark et al., 2010; Gioia & Chittipeddi, 1991; Corley & Gioia, 2004). Following Strauss and Corbin (1990), the methods sections of these articles generally describe a highly disciplined coding and analysis process whose central artifact, a hierarchical ‘‘data structure’’ is presented as a key output of the research, usually in the form of a horizontal tree-shaped ﬁgure (see, e.g., Corley & Gioia, 2004, p. 184). To arrive at this, the authors ﬁrst develop in vivo codes through ‘‘open coding’’ of data extracts using the words of participants, and then group these into ‘‘ﬁrst order’’ (participant-based) concepts through ‘‘constant comparison’’ (Strauss & Corbin, 1990) between different extracts. Linkages between ﬁrst-order concepts are then sought through ‘‘axial coding’’ leading to so-called

Templates and Turns in Qualitative Studies of Strategy and Management

215

second-order themes situated at a higher level of abstraction. Through further comparisons of the data, the researchers generally arrive at a limited number of ‘‘aggregate dimensions’’ or ‘‘core categories’’ that serve to summarize the elements of an emerging theoretical model. For example, the ideas of ‘‘sensemaking’’ and ‘‘sensegiving’’ emerged as the key explanatory concepts from the study of the initiation of strategic change in a university (Gioia & Chittipeddi, 1991); the notion of ‘‘identity ambiguity’’ along with its triggers and consequences emerged as central in the study of identity change following a corporate spinoff (Corley & Gioia, 2004). Each of these concepts is linked to others and underpinned by the ﬁrst-order and second-order themes that successively and in tree-like fashion gave rise to it. All this takes place iteratively, with constant moving back and forth between codes and data, and with emerging ideas leading to additional data collection to ﬁll out the framework as the research progresses. Instead of terms like validity and reliability, the authors use Lincoln and Guba’s (1985) set of criteria for naturalistic inquiry to assess the quality of their research method. In particular, their claims for the ‘‘trustworthiness’’ of their data are supported by the involvement of multiple researchers and by member-checking (i.e., gaining feedback from insiders on emerging interpretations). Again, the simple description of the design and procedures does not do justice to the uncertainties involved in generating these outputs. Finding the twist that will pull all the ideas together is of course necessarily a creative act. As Suddaby (2006) has noted, grounded theory is not easy, although when examining its products, it sometimes looks easy, since at least in the case of these researchers, the emerging models tend to be neatly parsimonious despite the mass of data that generated them. This brings us to the question of rhetoric. Rhetoric of the Writing: Establishing the Gap, ‘‘Distilling the Essence,’’ Elaborating the Story My awareness of my cognitive limitations helps me empathize with the poor reader trying to understand the point(s) I am trying to make in a given article. For that reason, I work hard at trying to distil ﬁndings to their essences and to communicate them in simple compelling ways. Although I once disdained it, I have developed a great appreciation for ‘‘sound-bite’’ research reporting. (y) A well-constructed sound-bite has a certain memorability about it-what I like to call a ‘‘cognitive stickiness’’ that allows readers to remember the most important points you are trying to make. (Gioia, 2004)

The rhetorical structure of the articles by Gioia and colleagues that we have reviewed here is perhaps not as uniform as that described above for

216

ANN LANGLEY AND CHAHRAZAD ABDALLAH

Eisenhardt and colleagues’ work. However, there are some very instructive commonalities that are worthy of note. First, the positioning of the contribution is more often in the nature of establishing a gap in understanding of important processes than of establishing a contradiction with previous research as we saw above. However, perhaps the most striking and powerful rhetorical pattern lies in the presentation of the ﬁndings. This begins with the overall ‘‘data structure’’ diagram we described in the previous section. For example, Corley and Gioia’s (2004) data structure diagram has 24 ‘‘ﬁrst-order’’ concepts grouped into 9 ‘‘second order themes,’’ which are in turn grouped again into three ‘‘aggregate dimensions’’ that form the core of the theoretical contribution. Gioia et al.’s (2010) study of the creation of a new identity in a university department has 16 ‘‘ﬁrst order categories’’ grouped into 8 ‘‘second order themes.’’ In both these papers and others, another ﬁgure that shows how the second order themes are related with each other over time is also provided. These ﬁgures, accompanied by a short verbal description, provide an upfront distillation of the paper’s central message (see Gioia’s remarks at the beginning of this section). All that remains then is to elaborate on each of the main themes. This is done in two ways that together provide compelling support for the emerging model. First each of the themes is elaborated as part of a narrative account in the body of the paper, with multiple references to speciﬁc incidents and quotations from informants or documents. Second, additional quotations for each theme are displayed in a large accompanying table (with very little overlap in content with the textual narrative). This data presentation strategy, very obvious in the Corley and Gioia (2004) paper and followed through in subsequent writings, builds strong credibility around the ﬁndings. In a recent Academy of Management Journal editorial, Pratt (2009) noted the value of this approach, suggesting that writers might keep their most striking ‘‘power quotes’’ (Gioia’s sound-bites?) for the narrative, but place additional ‘‘proof quotes’’ in tables to solidify their arguments. Finally, after the presentation of the ﬁndings, the authors return to a description of the overall model, and elaborate on the contribution of the paper, often though not always in a series of propositions. Assessing the Template: Limitations and Variations Again, the ‘‘Gioia method’’ has been very successful on its own terms in generating knowledge about strategic and identity change in various situations. Several of its elements have also been taken up by others, especially but not only by researchers in the area of organizational identity.

Templates and Turns in Qualitative Studies of Strategy and Management

217

Speciﬁcally, the authors’ approach to summarizing the derivation of their emergent grounded conceptual framework in the form of a data structure diagram has become increasingly common. For example, Maguire and Phillips (2008) used this device in a study of identity change at Citigroup, Anand et al. (2007) used it for a study of the development of new practices in consulting ﬁrms, and Rindova et al. (2011) used it in their study of Alessi’s incorporation of new cultural resources into their strategy. This template has limitations too. One potential limitation that seems, however, not to have hindered these researchers concerns the challenge of convincing readers about the transferability and relevance of the ﬁndings given the propensity to study single cases. In interpretive research, it is argued that it is the depth of contextual detail in a case study that provides the understanding necessary for a reader to judge whether the theoretical elements might apply to their own situation. Also, one might expect that cases (of for example mergers) might have certain generic qualities that could make some types of ﬁndings relevant almost anywhere. And yet, working with a single idiographic case considered holistically is, in our own experience, often more challenging than working with some form of comparative design where similarities and differences more naturally stimulate theorization (Langley, 1999). With a single case, it is easy to fall into the trap of having nothing but a boring sequential narrative to tell, with no insightful plot or any hope of catching readers’ minds and imaginations with the ‘‘cognitive stickiness’’ that Gioia (2004) was referring to. The ability to generate theoretical insights that have obvious value beyond the speciﬁc context of their development is a crucial skill for this type of research. Finally, although the Gioia method does lead to process models of how people make sense over time, these models sometimes seem to describe phenomena at rather a high level of aggregation (as described in the secondorder themes) so that a complete understanding of how and why things occur in the everyday from one moment to the next is to a degree glossed over. This may be partly a consequence of the grounded theory methodology where the coding and categorizing process may generate a certain decontextualization; to achieve generality, the chaining and interplay of particular events may sometimes become lost in this process. In addition, despite their interpretive roots, these studies usually produce singular narratives where differences in perspective are subsumed as ‘‘tensions’’ but are not elaborated in depth (Buchanan & Dawson, 2007). As we shall see in the next section, there may be other ways of approaching strategy processes that get closer to everyday strategic practices and the way in which they are reproduced and adapted and that take into account multiple perspectives.

218

ANN LANGLEY AND CHAHRAZAD ABDALLAH

TWO TURNS The two approaches to qualitative analysis of strategy process phenomena described above are not of course the only ones. However, we chose to present them because they are not only powerful and useful but also representative of the most common sets of epistemological assumptions, methodological toolkits, and rhetorical frames supporting qualitative research in this ﬁeld. In the second part of this chapter, we move toward some more recent and less traditional approaches to qualitative studies in strategy and management. These approaches are broader and less codiﬁed than the templates described above, so our mode of presentation will be somewhat different. However, they are currently generating a great deal of interest. Each has different epistemological assumptions, suggests different methodologies, and may involve different styles of writing. We begin by focusing on the ‘‘practice turn’’ and then move on to the ‘‘discursive turn’’ drawing on selected methodological texts and empirical exemplars in each case (for a summary of this discussion, see Table 2).

The Practice Turn: Studying Strategy as a Social Practice Epistemological Foundations and Empirical Exemplars The practice turn in strategy research, or the ‘‘strategy as practice’’ perspective (Whittington, 2006; Jarzabkowski, 2005; Johnson, Langley, Melin, & Whittington, 2007) has developed considerable momentum in recent years building on an interest in practice-based studies that has spread from philosophy and sociology (Schatzki, Knorr Cetina, & Von Savigny, 2001; Reckwitz, 2002; Giddens, 1984; Bourdieu, 1977) into various subﬁelds of organization theory and management including strategy (Feldman & Orlikowski, forthcoming; Miettinen, Samra-Fredericks, & Yanow, 2009; Corradi, Gherardi, & Verzelloni, 2010). Speciﬁcally, scholars of strategy as practice argue that rather than being seen as something that organizations have, strategy should be viewed as ‘‘something people do’’ (Whittington, 2006; Jarzabkowski, Balogun, & Seidl, 2007). Practice thinking thus begins with an empirical focus on activity, and in this case with the concrete micro-level activities that strategy practitioners, broadly deﬁned, engage in, and with the regularities constituted and reproduced by these activities. For some, practice thinking ends where it begins: the ‘‘doing of strategy’’ is an interesting empirical phenomenon that can be and indeed has been studied in a variety of different ways using methods that are often not all

Templates and Turns in Qualitative Studies of Strategy and Management

219

Table 2. Two ‘‘Turns’’ in Qualitative Research on Strategy and Management. ‘‘Strategy as Practice’’ Empirical focus

‘‘Strategy as Discourse’’

The ‘‘doing’’ of strategy: Activities of strategy practitioners and regularities emerging from or underlying them Foundational Whittington (2006, 2007), references Jarzabkowski (2004), Johnson et al. (2007), Rasche and Chia (2009), Feldman and Orlikowski (forthcoming) Epistemological Practices as constitutive of social foundations and world; diverse theoretical roots but key theoretical some key common elements: elements – Knowledge as embedded in practices – Socio-material nature of practice – Recursivity of practices

Language and strategy: How discourses are shaped and shape understandings of strategy and organizational direction Phillips et al. (2008), Vaara (2010), Phillips and Hardy (2002), Vaara and Tienari (2004)

Empirical exemplars Methodological and rhetorical elements

Heracleous and Barrett (2001), Vaara and Monin (2010) – Detailed analyses of content of texts (e.g., themes, structure, etc.) – Need for ethnographic or process data on context (writers, readers, intentions, events, practices surrounding text) – Longitudinal data to capture temporality – Writing including both detailed analysis of text and as well as data on how texts are used in context

Rouleau (2005), Kaplan (2011), Jarzabkowski (2008) – Ethnographic observation to detect elements of practice (e.g., implicit knowledge; sociomateriality) not usually consciously perceived – Need for in-depth longitudinal studies to capture recursivity of practices – Writing around detailed vignettes to reveal underlying dynamics – Use of temporal bracketing to structure recursive analysis

Social world created and maintained through discourse; Key elements: – Hermeneutic: focus on meaning – Critical: revealing politics and power – Interdiscursive: focus on interplay among discourses at multiple levels

that different from those we described earlier. Indeed, the studies of Eisenhardt (1989b) on fast decision making and Gioia and Chittipeddi (1991) on sensemaking and sensegiving in strategic change can be seen as studies of strategy as practice in that sense (Johnson et al., 2007). This empirically driven notion of practice has renewed interest in the human and

220

ANN LANGLEY AND CHAHRAZAD ABDALLAH

practical elements of strategy making, giving rise to some innovative and interesting studies [e.g., Johnson, Prashantham, Floyd, and Bourque’s (2010) multiple case studies of success and failure in strategy workshops drawing on ritual theory; Maitlis and Lawrence’s (2003) single case study of strategy failure; Balogun and Johnson’s (2004) interpretive study of the role of middle-manager sensemaking in strategic change using diaries and focus groups]. However, the notion of strategy as practice can become deeper and more distinctive if the notion of practice is taken to refer not just to an empirical interest in the doing of strategy but to include a commitment to theories of social practice, and eventually to a practice-based ontology in which ‘‘practices are understood to be the primary building blocks of social reality’’ (Feldman & Orlikowski, forthcoming, p. 3; Schatzki et al., 2001). This point has been argued in different ways by both proponents (Whittington, 2007; Rasche & Chia, 2009) and critics (Chia & MacKay, 2007; Carter, Clegg, & Kornberger, 2008; Corradi et al., 2010) of the strategy as practice perspective. However, what exactly this means is obscured by the fact that, as Miettinen et al. (2009, p. 1312) note, ‘‘social practice theory is not a uniﬁed theory, but rather a collection of authors and approaches interested in studying or theorizing practice, each of whom has his or her own distinctive vocabulary’’ (see also Corradi et al., 2010). Nevertheless, some common features of practice theorizing can be identiﬁed (Miettinen et al., 2009; Rasche & Chia, 2009; Feldman & Orlikowski, forthcoming) and we will draw on three of these to illustrate the implications for empirical research, using exemplars for each. First, practice theorizing emphasizes the way in which knowledge is embedded in and regenerated through practical activity (Cook & Brown, 1999; Gherardi, 2006). Thus when individuals engage in practices, they draw on unconscious tacit understandings of how to ‘‘go on’’ in speciﬁc situations that have been learned over time and that are enacted collectively (Rasche & Chia, 2009). From this perspective, the knowledge of how strategy or indeed any practical activity is accomplished may not be easily available only from asking questions in interviews, the dominant methodology in qualitative studies of strategy and management. Rather, it is implicit in what people do in speciﬁc situations. To appreciate and to a degree capture this form of knowledge requires close ethnographic observation, and sensitivity not just to surface activity but to the skills and competencies that underlie it (Rasche & Chia, 2009). Rouleau’s (2005) study of everyday sensemaking and sensegiving practices illustrates this focus. Speciﬁcally, through a ﬁnegrained analysis of incidents and conversations observed among middle

Templates and Turns in Qualitative Studies of Strategy and Management

221

managers and clients in a clothing ﬁrm, Rouleau (2005) shows how enacting a new strategy in the everyday involves adjusting stories to the people addressed (‘‘translating the new orientation’’), drawing on broad cultural repertoires associated with gender and ethnic origin (‘‘overcoding the strategy’’), mobilizing space, the body and displayed emotions to channel attention (‘‘disciplining the client’’) and framing legitimate reasons for strategic change (‘‘justifying the change’’). All these micro-practices and their embedded skills appear to be enacted subtly, smoothly, and naturally with little readily apparent conscious reﬂection. A second common tenet of practice theory is that material objects ranging from sophisticated technologies to the everyday tools of living are deeply intertwined in everyday practices, mediating how and what is accomplished (Latour, 2005). Practices are thus often qualiﬁed as ‘‘socio-material’’ to encompass the notion of the inseparability of human and nonhuman agency (Feldman & Orlikowski, forthcoming). This too has implications for research, again suggesting a need for ﬁne-grained attention to how material elements intervene within the context of practice. An interesting recent ethnographic study by Kaplan (2011) reveals how PowerPoint technology is deeply implicated in the ways in which strategic decisions are constructed. Through the ﬁne analysis of strategy making negotiations, Kaplan shows how the materiality, mutability, modularity and digitality of PowerPoint slides contributes to enabling both collaboration among people holding different perspectives (through information sharing and idea generation), but also to what she calls ‘‘cartography’’ – the political effort to pin down and ‘‘draw boundaries around the scope of the strategy’’ (Kaplan, 2011, p. 21) by selective inclusion of information and actors manifested materially in the slides themselves and in the way in which they are diffused and presented. Finally, a third important notion in practice theory is the idea that practices are recursive (Feldman & Orlikowski, forthcoming; Jarzabkowski, 2004). Ongoing activity leads to the stabilization and reiﬁcation of social orders or social structures that become resources for subsequent activity. For example, in Giddens’ (1984) theory of structuration, social structures constituted through practice include power dependencies (‘‘structures of domination’’), shared meanings or interpretive schemes (‘‘structures of signiﬁcation’’), and norms (‘‘structures of legitimation’’). Ongoing activities are constrained and enabled by these social structures, but they are simultaneously the means by which they are produced and reproduced over time. The mutually constitutive nature of structure and agency implicit in these theories of practice can be hard to pin down in empirical research and

222

ANN LANGLEY AND CHAHRAZAD ABDALLAH

detailed ethnographic observation again seems desirable. In addition however, the ability to capture the recursive nature of practices requires fairly long time frames. For example, in a seven-year study of university strategy making, Jarzabkowski (2008) used a structuration theory framework to examine how strategizing iteratively involved ad hoc decisions about speciﬁc strategies (interactive strategizing), the enactment of embedded routines and structures that generated decisions while reproducing those routines (procedural strategizing), and activity that creating new routines and structures that would serve to embed later decisions (integrative strategizing). Doing and Writing Research from the Practice Turn As we have suggested above, studying strategy from the perspective of the practice turn often requires deeper and closer contact with the doing of strategy than is often seen in other approaches. Thus, ethnography has been a favored research method because it enables researchers to capture what participants themselves are unable to articulate, at least not as well (Rouleau, 2005; Rasche & Chia, 2009) and to physically see how material objects, the body, space, and time are mobilized within practices (Rouleau, 2005; Kaplan, 2011). For example, strategy as practice scholars have begun to use video ethnography and photographs to capture systematically what is happening beyond the merely verbal component of strategic practices (Molloy & Whittington, 2005; Liu & Maitlis, 2010). In addition, longitudinal observations over long time periods are required to capture the recursive nature of practices as in Jarzabkowski’s (2008) seven-year study. Clearly however, such work generates immense databases of disparate kinds of information, and the researcher is faced with another complex task in communicating it in the context of journal articles. Without suggesting that these are the only ways of analyzing and communicating insight about practice, we observe two interesting ways in which authors reveal their ﬁndings that are somewhat different from those described earlier. The ﬁrst is particularly evident in the Rouleau (2005) and Kaplan (2011) articles and involves the detailed elaboration and unfurling of highly speciﬁc but powerfully illustrative vignettes. For example, Rouleau’s (2005) ethnographic study took place over six months with four days per week of presence on the site. However, she uses six small vignettes (three routines and three conversations) to build her in-depth analysis of the practices. She draws an interesting analogy between her own approach and that of the natural scientist when she says, ‘‘Just as using a microscope helps understanding of the whole through its tiny parts, routines and

Templates and Turns in Qualitative Studies of Strategy and Management

223

conversations offer an interesting insight to examine strategic change’’ (p. 1419). As each of the microscopic samples reveals similar underlying phenomena whose workings are ﬁnely traced out, cumulative understanding becomes increasingly layered and credible. Similarly, Kaplan undertook an 18-month ethnography. However, her analysis draws intensively on two sequences of PowerPoint-based negotiations with detailed illustrations and a complex table in which modiﬁcations over time are illustrated. The explicit showing of how the practices she is describing are manifested in every element of these concrete sequences adds to the credibility of her theoretical insights. A second analytical and rhetorical device that has been useful in practicebased studies draws on Barley’s (1986, p. 82) sawtooth representation of the recursive nature of actions and institutions (or structures) where the realm of action and institution are shown as horizontal parallel lines that interact (see also Barley & Tolbert, 1997). In this representation, institutions are shown as directly inﬂuencing the practices carried out in the action realm. Each iteration of a practice implies its recursive reproduction or adaptation. Over time, ad hoc adaptations progressively cumulate and eventually result in sharper shifts in the institutional frame itself. This classic sawtooth model is used by Jarzabkowski (2008) in her study of strategizing in universities, by Howard-Grenville (2007) in her study of shifts in issue-selling practices in a chip-making company, and by Rerup and Feldman (2011) in their study of evolution in interpretive schemes in a research unit. The framework provides a heuristic for breaking down analysis into successive temporal brackets (Langley, 1999, p. 703) to explicitly examine how iterative actions taken during one period lead over time to changes in the context that will affect action in subsequent periods. Assessing the Turn: Limitations and Variations The practice turn offers potential to understand the doing of strategy and management rather differently, throwing light on its implicit, sociomaterial and recursive nature, something that is largely absent in the two templates we presented earlier. The practice turn also has a natural afﬁnity for qualitative and ethnographic research methods because of its empirical focus on the situated and particular. As Feldman and Orlikowski (forthcoming) note, however, this does not mean that practice theorizing has no generality. Rather, strong practice-based studies like those mentioned above generate new concepts and understandings that have much broader relevance. In a striking example of this potential, Feldman’s ethnographic study of practices in a university housing department generated broadly applicable theories of

224

ANN LANGLEY AND CHAHRAZAD ABDALLAH

the performative and ostensive aspects of routines (Feldman, 2000) as well as the development of the notion of ‘‘resourcing’’ (Feldman, 2004). Both these ideas have many interesting applications far beyond the original context of their production, and more particularly in the area of strategy. The key limitation of the practice turn in strategy may be that as some critics have suggested (Chia & MacKay, 2007; Carter et al., 2008), it is not quite yet a ‘‘turn’’ in the epistemological sense. ‘‘Strategy as practice’’ is more in the nature of an ‘‘umbrella’’ concept (Corradi et al., 2010) that enables the grouping together of a community of people interested in similar empirical phenomena and drawing on a loose collection of theoretical lenses that have something to do with practice. So far, this seems to be leading to a renewal and enrichment of qualitative methodology in strategy and management, a positive trend it seems to us. As the perspective develops through its own empirical research practice, its theoretical reach will no doubt recursively shift and hopefully deepen. The emphasis on practice has also in many ways fed into the second turn we examine here.

The Discursive Turn: Studying Strategy as Discourse Epistemological Foundations and Empirical Exemplars As the result of a more general ‘‘linguistic turn’’ in organization studies (Alvesson & Karreman, 2000), and building on the progression of socioconstructivist epistemologies inspired by Berger and Luckmann (1967), discursive approaches have become increasingly prevalent in organization and management research (Phillips, Sewell, & Jaynes, 2008; Vaara, 2010). In particular, a wide variety of linguistic approaches to strategy have been proposed varying from critical discourse analysis (CDA) (Phillips et al., 2008) to narrative analysis (Barry & Elmes, 1997) to conversation analysis (SamraFredericks, 2003). In this section, we will focus more particularly on exemplars of discursive approaches used to study multi-level strategy processes over time. One of the most widely shared deﬁnitions of discourse was offered by Parker (1992) for whom discourse does not refer simply to text, but is a set of texts and of the practices related to their production, dissemination, and reception. Texts can take on different forms: written, spoken, images, symbols, and other artifacts (Grant, Keenoy, & Oswick, 1998). Discourse analysis involves examining how discourses shape understandings of social reality, and how they are in turn shaped through discursive practices

Templates and Turns in Qualitative Studies of Strategy and Management

225

including the production, distribution, transformation, movement, and interpretation of texts. It aims to understand how social phenomena are produced or constructed and maintained through time (Phillips & Hardy, 2002). Thus, there are clear links between this approach and the practicebased approach described in the previous section. Paralleling the traditional themes of strategy research, discourse studies in strategy ‘‘all share an interest in exploring how organizations, industries and their environment are created and maintained through discourse’’ (Phillips et al., 2008, p. 770). As in the case of practice studies, there is no strong coherence among discursive approaches, but three main concerns are featured in this type of research that we identify as hermeneutic, critical and interdiscursive. The hermeneutic dimension is related to the need to understand how certain meanings are discursively constructed and interpreted and how they evolve over time (Heracleous & Barrett, 2001). Discourse studies also share a critical concern that calls for a multidimensional or intertextual analysis of discourse to bridge micro, meso, and meta levels of analysis and to critically examine the shaping of various organizational processes (Phillips et al., 2008; Vaara, 2010). Finally, while some discourse analyses tend to be static focusing on speciﬁc documents or narratives, as noted by Vaara (2010), the greatest potential of discursive approaches for strategy comes from analyses of the interplay of discourses over time and across multiple levels, what he labels ‘‘interdiscursivity.’’ This could involve for example looking at how macro-level discourses about the nature of strategy are taken up in speciﬁc organizations (Mantere & Vaara, 2008), how multiple discourses interact and conﬂict (Heracleous & Barrett, 2001), or how dominant discourses come to emerge or are contested over time (Ezzamel & Willmott, 2008). Discursive approaches can therefore offer a new way of introducing complexity into the study of strategic processes by examining their nonlinearity, their linguistic nature, and the various forms of their internal dynamics (Vaara, 2010). The two exemplars of discourse studies we chose to present in this chapter represent two very different ways of studying strategy processes from a discursive standpoint. The ﬁrst by Heracleous and Barrett (2001) uses discourse analysis with a primarily hermeneutic concern. It examines organizational change from a discursive perspective through an exploration of the implementation process of a risk-placing support system in the London Insurance Market over a ﬁve-year period. The paper, one of the ﬁrst of its nature to be published in the Academy of Management Journal, makes a strong case for a structurationist conceptualization of discourse as made up of both deep meaning structures and surface communicative

226

ANN LANGLEY AND CHAHRAZAD ABDALLAH

actions and defends this conceptualization as a means of reconciling the social dualisms of structure and action (Giddens, 1984). Again the linkage with the previous perspective is clear, although the emphasis here is clearly on communicative actions and their underlying meaning, rather than on practices. The paper is a longitudinal (ﬁve-year) investigation of how a change process (the implementation of a new IT system) is shaped by the discourses of different stakeholders over time. It is both an inquiry into the nature of the discourse employed by various stakeholders and an inquiry into its role in shaping the change process. Interestingly, a combined discourse analysis method, termed ‘‘Rhetorical-Hermeneutic’’ by the authors, was used and constitutes an original way of bridging between multiple levels of analysis: the deep discursive structure level, the surface communicative action level and the contextual level through interpretive schemes that are used as modalities that mediate between the two discursive levels. This methodological bridging apparatus generated a systematic processual analysis that tracks shifts and transformations in the change process over time. The study shows how the deep structures of discourse act as stable patterns that shape action in various ways for different stakeholders through contextual elements of interpretation. Its approach is ‘‘interdiscursive’’ in that it examines the struggles among alternate meanings inherent in stakeholder’s communicative actions. The second article by Vaara and Monin (2010) is a study of a process of discursive legitimation in a post-merger situation using a multimethod critical approach. The paper also shows the recursivity of discourse and action in that the discursive legitimation process unfolds by simultaneously shaping and being shaped by organizational action. The interesting aspect of the process as described in the paper is how a key discursive ‘‘device’’ of justiﬁcation, termed ‘‘theranostics’’ (a combination of the two strategic resources of the merging entities, respectively ‘‘therapy’’ and ‘‘diagnostics’’) was taken up and echoed in media discourse, creating enthusiasm around this concept not only in the business press but by ricochet within the ﬁrm itself as its members came increasingly to believe it, and indeed attempted to enact it despite its origin as a useful ‘‘story’’ developed to legitimate a merger that had been promulgated for other reasons. The study illustrates the potentially performative nature of discourse (producing that of which it speaks) and its role in the merger outcome. It shows the process of transformation of theranostics from a discursive resource of legitimation into a source of unrealistic expectations, as the ideas underlying it ultimately proved to be illusory.

Templates and Turns in Qualitative Studies of Strategy and Management

227

In their paper, Vaara and Monin (2010) interestingly also echo themes like sensemaking, sensegiving, or sensehiding often examined by others through the ‘‘Gioia method,’’ but they analyse them using a discursive approach that is based on a multidimensional conception of discourse as made up of texts but also of a set of material actions that transform or are transformed by it. Doing and Writing Research from the Discursive Turn Aside from the two examples of published research from a discursive perspective described above, it is important to note that a large number of studies have been using this perspective in recent management research. In their recent review of and call for applying CDA to strategy research, Phillips et al. (2008) show the increase in the number of published papers including CDA since 1995, reaching around 140 in 2005. Although a wide range of methodologies can be found under the discourse analysis umbrella, three main elements structure the discursive approach methodologically in relation to studying processes: the multiple forms of text(s), the crucial role of context, and the temporality of discourse. First, the textual dimension of discourse analysis is of course fundamental since it is mainly through texts in their various forms that any discursive work can be done. The juxtaposition of written, spoken and other symbolic textual devices characterizes the aim of discursive approaches to accentuate in more depth the internal circumvolutions of process and its interdiscursive nature. The studies we describe each contain speciﬁc ways of systematically analyzing the content of texts, for example, looking at ‘‘ethymeme components’’ or rhetorical structures in the texts for Heracleous and Barrett (2001) and looking at legitimation strategies inherent in the texts for Vaara and Monin (2010). Other kinds of textual analysis methods such as conversation analysis or narrative analysis would be possible. However, it is not only the text as a micro analytical device that is of interest here but texts as multiple forms of discursive manifestations embodied in their practices of production, dissemination and consumption that are at the heart of this relatively new methodological approach. It is important to note here the differences in the way textual data (interviews, documents, and other materials) are treated in this perspective as compared with the approaches we presented in the ﬁrst half of this chapter. The Eisenhardt method involves analyzing such data to establish facts while the Gioia method would treat the same data as interpretations. In the discursive approach, texts are discourses that are analyzed not only for what they say but for what they do: for example, the meanings they

228

ANN LANGLEY AND CHAHRAZAD ABDALLAH

construct, reproduce, contest or maintain, the effects they have and the precise means by which these effects are achieved (Vaara, 2010). These effects may include the propagation of managerial concepts (e.g., ‘‘theranostics’’; strategy itself), the transformation of institutional ﬁelds (Suddaby & Greenwood, 2005) or the reproduction of power relations (Knights & Morgan, 1991; Ezzamel & Willmott, 2008), with critical researchers being particularly concerned with revealing the latter. Second, in almost all the studies that use a discursive approach to understand organizational processes, the notion of context is presented as the stepping stone upon which a strong analysis should be built. No ‘‘thick description’’ is possible without it and no sense of unfolding or of temporality can be conveyed if context is not addressed. For example, in Heracleous and Barrett’s study (2001), context is taken into account through the collection of ethnographic data that is used in conjunction with the textual data in the analysis of the change process. In a constant hermeneutic interplay between texts and discourses deﬁned as ‘‘constituted of the totality of single texts’’ (p. 762), the analysis illustrates the importance of their ‘‘textsin-context’’ approach (interviews, written texts, ethnographic data) to understand the temporal unfolding of the process. Similarly, in recent research by Vaara and his colleagues, (e.g., Vaara, Kleymann, & Seristo¨, 2004; Vaara & Monin, 2010; Mantere & Vaara, 2008), context is always given a preponderant role in explaining the dynamics of the processes under examination. Elements of context are drawn from data collected during lengthy contact with the studied organizations and are included in the narrative constructions around the unfolding of the examined processes. Generally, context gives the necessary depth and grounding to studies that move from the meso to the micro levels of analysis. Finally, temporality is one of the main issues in studying processes and it seems that recent discursive approaches with their multidimensional and multilevel methodological choices are tackling the temporality issue in an interestingly relevant manner. Echoing the methodological opening-up to multiple dimensions, the conceptualization of temporality is broader here than in the more traditional process research studies. The temporality revealed in these studies is not simply a linear progression through time but a dynamic interdiscursive process that evolves in sinuous, nonlinear ways. For example, in the Heracleous and Barrett (2001) study, temporality is crucial and is shown through the description of the evolution of both levels of discourse and their mutual structuring broken down into distinct phases of evolution. In their description of the legitimation process of a merger, Vaara and Monin (2010)’s conception of temporality is anchored within the

Templates and Turns in Qualitative Studies of Strategy and Management

229

particular interpretive context of individuals in the two merging organizations. Temporality becomes a relative notion that might have to be taken into account in a different way in different contexts and for different organizational actors. Assessing the Turn: Limitations and Variations Like its main proponents (Phillips et al., 2008; Vaara, 2010), we believe that the discursive turn offers potential to open up research on strategy processes, through a more performative conception of discourse, to a multidimensional examination of organizational processes. In its critical manifestation, the discursive turn also draws attention to the ways in which realities that favor certain groups over others are socially constructed but also to how those relations might be thought of differently (Ezzamel & Willmott, 2008; Mantere & Vaara, 2008). Nevertheless, we see several ways in which discursive studies might be developed and improved. First, some of the earlier difﬁculties associated with publishing discourse-based studies in major journals were perhaps associated with the relatively opaque nature of some of their analyses. Recent work including the studies by Heracleous and Barrett (2001) (see also Heracleous, 2006), by Vaara and colleagues (see also Vaara & Tienari, 2004; Mantere & Vaara, 2008) and by Phillips and Hardy (2002) have begun to render the methods more accessible, providing more methodological detail and worked examples to build conﬁdence in and understanding of ﬁndings that this type of analysis can generate. Second, greater emphasis could be placed on the pragmatic aspects of discourse studies in strategy research to enable them to reach a wider audience. An understanding of the way in which discursive practices contribute to deﬁning the realities organizations live with ought to have serious practical implications, but these have not necessarily been strongly emphasized. As with any academic enterprise, there is a risk of becoming too self-referential (Luhmann, 1995), and this arises particularly with approaches that build on their own specialized methodological language. Put differently, the knowledge generated by the more traditional templates has perhaps in the past been a little easier to consume.

CONCLUSIONS This chapter has considered four different ways in which qualitative research can contribute to developing valuable knowledge about strategy

230

ANN LANGLEY AND CHAHRAZAD ABDALLAH

processes. By describing two somewhat institutionalized approaches to conceptualizing qualitative research and of writing qualitative articles (the two ‘‘templates’’), we illustrate some ways in which positivist and interpretive conceptions of reality and knowledge development have been successfully mobilized to generate insight. We have also shown how these approaches achieve their persuasive effects by examining not only the logic behind the methods used, but also by revealing the related rhetorical moves underlying their presentation and argumentation. Second, we attempted to move beyond the positivist and interpretive frames reﬂected in the two more traditional templates to consider alternative ways in which qualitative data might be used to throw light on strategic management processes. Drawing on a number of illustrative exemplars, we showed the potential for the practice and discursive turns in strategy research to offer important and original ways of seeing these processes. From these perspectives, qualitative data is not simply something that can be valuable in the ‘‘early stages’’ of research as is often assumed in the positivist paradigm, but something that is inherent to the ability to uncover certain types of knowledge about organizational phenomena, for example, knowledge that is embedded in strategic practices or that is itself constructed through language. We hope that the ideas presented in this chapter will encourage researchers interested in using qualitative research methods to examine the approaches presented here for themselves, perhaps by delving into some of the exemplars we identiﬁed. We also hope that through their own reading and research, they might discover, articulate and/or invent others. There is, fortunately, still ample room for innovation and creativity in the area of qualitative research on strategy and management.

NOTES 1. We thank Dennis Gioia for an instructive telephone conversation about his approach to qualitative research. 2. Note that while Eisenhardt (1989a) indicated that the data do not have to perfectly ﬁt the proposed model, in most published papers, it is hard to observe any lack of ﬁt in the tabulated evidence that almost always exhibits perfect correlation.

REFERENCES Ahrens, T., & Dent, J. F. (1998). Accounting and organizations. Journal of Management Accounting Research, 10(1), 1–39.

Templates and Turns in Qualitative Studies of Strategy and Management

231

Alvesson, M., & Karreman, D. (2000). Varieties of discourse: On the study of organizations through discourse analysis. Human Relations, 53(9), 1125–1149. Ambos, T. C., & Birkinshaw, J. (2010). How do ventures evolve? An inductive study of archetype change in science-based ventures. Organization Science, 21(6), 1125–1140. Anand, N., Gardner, H., & Morris, T. (2007). Knowledge-based innovation: Emergence and embedding of new practice areas in management consulting ﬁrms. Academy of Management Journal, 50(2), 406–428. Balogun, J., & Johnson, G. (2004). Organizational restructuring and middle manager sensemaking. Academy of Management Journal, 47(4), 523–549. Barley, S. R. (1986). Technology as an occasion for structuring: Evidence from observations of CT scanners and the social order of radiology departments. Administrative Science Quarterly, 31, 78–108. Barley, S. R., & Tolbert, P. S. (1997). Institutionalization and structuration: Studying the links between action and institution. Organization Studies, 18(1), 93–117. Barry, D., & Elmes, M. (1997). Strategy retold: Toward a narrative view of strategic discourse. Academy of Management Review, 2, 429–452. Berger, P. L., & Luckmann, T. (1967). The social construction of reality. New York: Doubleday Anchor. Bingham, C. B. (2009). Oscillating improvisation: How entrepreneurial ﬁrms create success in foreign market entries over time. Strategic Entrepreneurship Journal, 3, 321–345. Bourdieu, P. (1977). Outline of a theory of practice. Cambridge: Cambridge University Press. Brown, S. L., & Eisenhardt, K. M. (1997). The art of continuous change: Linking complexity theory and time-paced evolution in relentlessly shifting organizations. Administrative Science Quarterly, 42(1), 1–34. Buchanan, D., & Dawson, P. (2007). Discourse and audience: Organizational change as a multistory process. Journal of Management Studies, 44(5), 669–686. Carter, C., Clegg, S. R., & Kornberger, M. (2008). Strategy as practice? Strategic Organization, 6(1), 83–99. Chia, R., & MacKay, B. (2007). Post-processual challenges for the emerging strategyas-practice perspective: Discovering strategy in the logic of practice. Human Relations, 60(1), 217–242. Clark, S. M., Gioia, D. A., Ketchen, D. J., & Thomas, J. B. (2010). Transitional identity as a facilitator of organizational identity change in a merger. Administrative Science Quarterly, 55, 397–438. Cook, S. D. N., & Brown, J. S. (1999). Bridging epistemologies: The generative dance between organizational knowledge and organizational knowing. Organization Science, 10(4), 381–400. Corley, K. G. (2004). Deﬁned by our strategy or our culture? Hierarchical differences in perceptions of organizational identity and change. Human Relations, 57(9), 1145–1177. Corley, K. G., & Gioia, D. A. (2004). Identity ambiguity and change in the wake of a corporate spin-off. Administrative Science Quarterly, 49(2), 173–208. Corradi, G., Gherardi, S., & Verzelloni, L. (2010). Through the practice lens: Where is the bandwagon of practice studies heading? Management Learning, 41(3), 265–283. Dyer, W. G., & Wilkins, A. L. (1991). Better stories, not better constructs, to generate better theory: A rejoinder to Eisenhardt. Academy of Management Review, 16(3), 613–619. Eisenhardt, K. M. (1989a). Building theories from case study research. Academy of Management Review, 14(4), 532–550.

232

ANN LANGLEY AND CHAHRAZAD ABDALLAH

Eisenhardt, K. M. (1989b). Making fast strategic decisions in high velocity environments. Academy of Management Journal, 32(4), 543–576. Eisenhardt, K. M., & Bourgeois, L. J. (1988). Politics of strategic decision making in high velocity environments: Toward a midrange theory. Academy of Management Journal, 31(4), 737–770. Eisenhardt, K. M., & Graebner, M. E. (2007). Theory building from cases: Opportunities and challenges. Academy of Management Journal, 50(1), 25–32. Ezzamel, M., & Willmott, H. (2008). Strategy as discourse in a global retailer: A supplement to rationalist and interpretive accounts. Organization Studies, 29, 191–217. Feldman, M. S. (2000). Organizational routines as a source of continuous change. Organization Science, 11(6), 611–629. Feldman, M. S. (2004). Resources in emerging structures and processes of change. Organization Science, 15(3), 295–309. Feldman, M. S., & Orlikowski, W. (forthcoming). Theorizing practice and practicing theory. Organization Science. Galunic, D. C. (2001). Architectural innovation and modular corporate forms. Academy of Management Journal, 44(6), 1229–1249. Galunic, D. C., & Eisenhardt, K. M. (1996). The evolution of intracorporate domains: Divisional charter losses in high-technology multidivisional corporations. Organization Science, 7(3), 255–282. Giddens, A. (1984). The constitution of society. Berkeley, CA: University of California Press. Gilbert, C. G. (2005). Unbundling the structure of inertia: Resources versus routine rigidity. Academy of Management Journal, 48(5), 741–763. Gioia, D. A. (2004). A renaissance self: Prompting personal and professional revitalization. In: R. E. Stablein & P. J. Frost (Eds), Renewing research practice (pp. 97–114). Stanford, CA: Stanford University Press. Gioia, D. A., & Chittipeddi, K. (1991). Sensemaking and sensegiving in strategic change initiation. Strategic Management Journal, 12, 433–448. Gioia, D. A., Price, K. N., Hamilton, A. L., & Thomas, J. B. (2010). Forging an identity: An insider-outsider study of processes involved in the formation of organizational identity. Administrative Science Quarterly, 55(1), 1–46. Gioia, D. A., & Thomas, J. B. (1996). Identity, image and issue interpretation: Sense making during strategic change in academia. Administrative Science Quarterly, 41(3), 370–403. Gioia, D. A., Thomas, J. B., Clark, S. M., & Chittipeddi, K. (1994). Symbolism and strategic change in academia: The dynamics of sensemaking and inﬂuence. Organization Science, 5(3), 363–383. Gherardi, S. (2006). Organizational knowledge: The texture of workplace learning. Oxford, UK: Blackwell Publications. Glaser, B. G., & Strauss, A. L. (1967). The discovery of grounded theory. Chicago, IL: Aldine. Golden-Biddle, K., & Locke, K. (1993). Appealing work: An investigation of how ethnographic texts convince. Organization Science, 4(4), 595–616. Golden-Biddle, K., & Locke, K. (2006). Composing qualitative research. Thousand Oaks, CA: Sage Publications. Graebner, M. E. (2004). Momentum and serendipity: How acquired leaders create value in the integration of technology ﬁrms. Strategic Management Journal, 25, 751–777. Graebner, M. E. (2009). Caveat venditor: Trust asymmetries in acquisitions of entrepreneurial ﬁrms. Academy of Management Journal, 52(3), 435–472.

Templates and Turns in Qualitative Studies of Strategy and Management

233

Graebner, M. E., & Eisenhardt, K. M. (2004). The seller’s side of the story: Acquisition as courtship and governance as syndicate in entrepreneurial ﬁrms. Administrative Science Quarterly, 49, 366–403. Grant, D., Keenoy, T., & Oswick, C. (1998). Discourse and organization. London: Sage Publications. Guba, E., & Lincoln, Y. S. (1994). Competing paradigms in qualitative research. In: N. K. Denzin & Y. S. Lincoln (Eds), Handbook of qualitative research (pp. 105–117). Thousand Oaks, CA: Sage Publications. Hallen, B. L., & Eisenhardt, K. M. (2009). Catalyzing strategies: How entrepreneurs accelerate inter-organizational relationship formation to secure professional investments, Social Science Research Network. Available at http://ssrn.com/abstract¼1372328. Retrieved on January 15, 2011. Heracleous, L. (2006). Discourse, interpretation, organization. Cambridge: Cambridge University Press. Heracleous, L., & Barrett, M. (2001). Organizational change as discourse: Communicative actions and deep structures in the context of information technology implementation. Academy of Management Journal, 44, 755–778. Howard-Grenville, J. A. (2007). Developing issue-selling effectiveness over time: Issue selling as resourcing. Organization Science, 16(6), 618–636. Jarzabkowski, P. (2004). Strategy as practice: Recursiveness, adaptation, and practices-in-use. Organization Studies, 35(4), 529–560. Jarzabkowski, P. (2005). Strategy as practice: An activity-based approach. London: Sage Publications. Jarzabkowski, P. (2008). Shaping strategy as a structuration process. Academy of Management Journal, 51(4), 621–650. Jarzabkowski, P., Balogun, J., & Seidl, D. (2007). Strategizing: The challenges of a practice perspective. Human Relations, 60(1), 5–27. Johnson, G., Langley, A., Melin, L., & Whittington, R. (2007). Strategy as practice. Cambridge, UK: Cambridge University Press. Johnson, G., Prashantham, S., Floyd, S. W., & Bourque, N. (2010). The ritualization of strategy workshops. Organization Studies, 31(12), 1589–1618. Kaplan, S. (2011). Strategy and PowerPoint: An inquiry into the epistemic culture and machinery of strategy making. Organization Science, 22(2), 320–346. Knights, D., & Morgan, G. (1991). Corporate strategy, organizations, and subjectivity: A critique. Organization Studies, 12, 251–273. Langley, A. (1999). Strategies for theorizing from process data. Academy of Management Review, 24(4), 691–710. Langley, A. (2007). Process thinking in strategic organization. Strategic Organization, 5(35), 271–282. Langley, A. (2009). Studying processes in and around organizations. In: D. Buchanan & A. Bryman (Eds), Sage handbook of organizational research methods (pp. 409–429). London: Sage Publications. Latour, B. (2005). Reassembling the social. Oxford: Oxford University Press. Lee, T. W., Mitchell, T. R., & Sablynski, C. J. (1999). Qualitative research in organizational and vocational psychology, 1979–1999. Journal of Vocational Behavior, 55, 161–187. Lincoln, Y. S., & Guba, E. (1985). Naturalistic enquiry. Beverly Hills, CA: Sage Publications.

234

ANN LANGLEY AND CHAHRAZAD ABDALLAH

Liu, F., & Maitlis, S. (2010). Changing plans: Emotional dynamics in the construction of a strategic plan. Workshop on talk, text, tools in the practice of strategy, Montreal, August 5, 2010. Luhmann, N. (1995). Social systems. Stanford, CA: Stanford University Press. Maguire, S., & Phillips, N. (2008). ‘Citibankers’ at Citigroup: A study of the loss of institutional trust after a merger. Journal of Management Studies, 45(2), 372–401. Maitlis, S. (2005). The social processes of organizational sensemaking. Academy of Management Journal, 48(1), 21–49. Maitlis, S., & Lawrence, T. (2003). Orchestral manoeuvres in the dark: Understanding failures in organizational strategizing. Journal of Management Studies, 40(1), 109–140. Mantere, S., & Vaara, E. (2008). On the problem of participation in strategy: A critical discursive perspective. Organization Science, 19, 341–358. Martin, J. A., & Eisenhardt, K. M. (2010). Rewiring: Cross-business-unit collaborations in multi-business organizations. Academy of Management Journal, 53(2), 265–301. Miettinen, R., Samra-Fredericks, D., & Yanow, D. (2009). Re-turn to practice: An introductory essay. Organization Studies, 30(12), 1309–1327. Miles, M. B., & Huberman, A. M. (1994). Qualitative data analysis. Thousand Oaks, CA: Sage Publications. Mohr, L. B. (1982). Explaining organizational behavior: The limits and possibilities of theory and research. San Francisco, CA: Jossey-Bass. Molloy, E., & Whittington, R. (2005). Practices of organising inside and outside the processes of change. Advances in Strategic Management, 22, 497–521. Nag, R., Corley, K. G., & Gioia, D. A. (2007). The intersection of organizational identity, knowledge and practice: Attempting strategic change by knowledge grafting. Academy of Management Journal, 50(4), 821–847. Ozcan, P., & Eisenhardt, K. M. (2009). Origin of alliance portfolios: Entrepreneurs, network strategies and performance. Academy of Management Journal, 52(2), 246–279. Parker, I. (1992). Discourse dynamics: Critical analysis for social and individual psychology. London, UK: Routledge. Pettigrew, A. M. (1990). Longitudinal ﬁeld research on change: Theory and practice. Organization Science, 1(3), 267–292. Phillips, N., & Hardy, C. (2002). Discourse analysis. Investigating processes of social construction. Thousand Oaks, CA: Sage Publications. Phillips, N., Sewell, G., & Jaynes, S. (2008). Applying critical discourse analysis in strategic management research. Organizational Research Methods, 11(4), 770–789. Pratt, M. G. (2009). For the lack of a boilerplate: Tips on writing up (and reviewing) qualitative research. Academy of Management Journal, 52(5), 856–862. Rasche, A., & Chia, R. (2009). Researching strategy practices: A genealogical social theory perspective. Organization Studies, 30(7), 713–734. Ravenswood, K. (forthcoming). Eisenhardt’s impact on theory in case study research. Journal of Business Research, doi:10.1016/j.jbusres.2010.08.014. Reckwitz, A. (2002). Toward a theory of social practices: A development in cultural theorizing. European Journal of Social Theory, 5(2), 243–263. Rerup, C., & Feldman, M. (2011). Routines as a source of change in organizational schema: The role of trial-and-error learning. Academy of Management Journal, 54(3), forthcoming. Rindova, V., Dalpiaz, E., & Ravasi, R. (2011). A cultural quest: A study of organizational use of new cultural resources in strategy formation. Organization Science, 22(2), 413–431.

Templates and Turns in Qualitative Studies of Strategy and Management

235

Rouleau, L. (2005). Micro-practices of strategic sensemaking and sensegiving: How middle managers interpret and sell change every day. Journal of Management Studies, 42(7), 1413–1441. Samra-Fredericks, D. (2003). Strategizing as lived experience and strategists’ everyday efforts to shape strategic direction. Journal of Management Studies, 40(1), 141–174. Santos, F. M., & Eisenhardt, K. M. (2009). Constructing markets and shaping boundaries: Entrepreneurial power in nascent ﬁelds. Academy of Management Journal, 52(4), 643–671. Schatzki, T. R., Knorr Cetina, K., & Von Savigny, E. (2001). The practice turn in social theory. London: Routledge. Strauss, A., & Corbin, J. (1990). Basics of qualitative research. Newbury Park, CA: Sage Publications. Suddaby, R. (2006). What grounded theory is not. Academy of Management Journal, 49(4), 633–642. Suddaby, R., & Greenwood, R. (2005). Rhetorical strategies of legitimacy. Administrative Science Quarterly, 50, 35–67. Vaara, E. (2010). Taking the linguistic turn seriously: Strategy as a multifaceted and interdiscursive phenomenon. Advances in Strategic Management, 27, 29–50. Vaara, E., Kleymann, B., & Seristo¨, H. (2004). Strategies as discursive constructions: The case of airline alliances. Journal of Management Studies, 41(1), 1–35. Vaara, E., & Monin, P. (2010). A recursive perspective on discursive legitimation and organizational action in mergers and acquisitions. Organization Science, 21(1), 3–22. Vaara, E., & Tienari, J. (2004). Critical discourse analysis as a methodology for international business studies. In: R. Piekkari (Ed.), Handbook of qualitative research methods for international business (pp. 342–362). Cheltenham, UK: Edward Elgar. Van de Ven, A. H. (1992). Suggestions for studying strategy process: A research note. Strategic Management Journal, 13(Summer special issue), 169–188. Whittington, R. (2006). Completing the practice turn in strategy research. Organization Studies, 28, 613–634. Whittington, R. (2007). Strategy practice and strategy process: Family differences and the sociological eye. Organization Studies, 28(10), 1575–1586. Yin, R. K. (2009 [1984]). Case study research. Thousand Oaks, CA: Sage Publications. Zott, C., & Huy, Q. N. (2007). How entrepreneurs use symbolic management to acquire resources. Administrative Science Quarterly, 52, 70–105.

THE REVOLUTION WITH A SOLUTION: ALL IS NOT QUIET ON THE STATISTICAL AND METHODOLOGICAL MYTHS AND URBAN LEGENDS FRONT Robert J. Vandenberg ABSTRACT Purpose – The purpose of this chapter is to present a subset of seven statistical and methodological myths and urban legends (SMMULs). When present, SMMULs degrade the overall research process and make manuscript evaluation problematic during the review process. SMMULs covered here included those pertaining to accepting the theoretical model, conventional cutoff values, exploratory factor analysis, common method bias, moderation analysis, Baron and Kenny’s four-step mediation test, and permitting correlated item residuals. Design/Methodology – Given that the details underlying the SMMULs have already been published, the present chapter was a summary of each. The summaries presented the urban legend and sources for it. Subsequently, the kernel of truth underlying the SMMUL was presented, and how this truth may have been lost and distorted. Each summary ends Building Methodological Bridges Research Methodology in Strategy and Management, Volume 6, 237–257 Copyright r 2011 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-8387/doi:10.1108/S1479-8387(2011)0000006009

237

238

ROBERT J. VANDENBERG

with the recommended ‘‘good’’ practices as presented by the original authors. Findings/Implications – The implication for researchers is to modify their current practices to strengthen their research and to make better inferences. And for editors and reviewers, the implication is to develop accurate decision rules to strengthen the review process. Originality/Value – The overall value of the chapter is to improve the research process in general. Keywords: statistical and methodological myths and urban legends; moderation; mediation; false inferences regarding model ﬁt; common method bias; correlated residuals For well over a decade at this point, my personal crusade in the organizational sciences has been to identify statistical and methodological myths and urban legends (SMMULs). A more thorough deﬁnition of a SMMUL is provided below. In brief, though, SMMULs result in frequent poor methodological choices during study design and execution, the misapplication of certain analyses and inferences from the results that are wholly or partially inappropriate or outright false (Lance & Vandenberg, 2009; Vandenberg, 2006). Researchers are not alone. In addition, editors and reviewers have developed decision scripts, heuristics, or criteria from the SMMULs against which to judge the worthiness of manuscripts submitted for publication. However, as is the case with researchers, those heuristics and criteria are frequently wholly or partially inappropriate. While this chapter’s focus is on SMMULs, I would ask the reader’s indulgence and permit me to digress and elaborate brieﬂy on the chapter’s title as it puts SMMULs in context. As is often the case, change begins with a revolution – a conscious decision not to accept the status quo when corrections are required. I am really oversimplifying in the following statements many serious and in-depth conversations and discussions among us within the qualitative and quantitative methodological house of our discipline over the years. The general theme underlying these exchanges was the sense of frustration among us with respect to the frequency of poor methodological and statistical practices, the disproportion of theory relative to what the statistical results can support, mislabels of key constructs, poor construct and measurement development, the SMMULs (my thing), etc. We recognized that these actions most certainly did not occur in all published articles or characterize all the decisions underlying the review process. However, they occur frequently enough to be of concern in our opinions.

The Revolution with a Solution

239

It was a conversation in 2008 between Professor Jeffrey Edwards (UNC – Chapel Hill) and me one evening after the daily events during the annual meeting of the Society of Organizational Behavior that really provided an operational context to all of these discussions. Speciﬁcally, after we lamented on some of the issues again, Professor Edwards suddenly stated, ‘‘You know what we need? It’s a ‘revolution with a solution!’’’ By ‘‘revolution’’ is meant for us to not just complain but to be ‘‘adult about it’’ and use our seniority and reputation in our discipline as catalyst for change. However, by ‘‘solution’’ is meant to drive change through education and not confrontation; that is, to use our knowledge and skills to show others what is perhaps a stronger practice than is currently the case. We considered the latter absolutely critical. Speciﬁcally, we recognized that the vast majority of researchers and those in the review process are not willfully engaging in these weak practices. Rather, they are simply following ‘‘scripts’’ that they were taught without really knowing that these scripts are not wholly or partially appropriate. In short, we felt strongly that this revolution should not be undertaken in a manner to makes people feel stupid, dumb, or otherwise inadequate for practicing what they know and understand to be good. As the conversation progressed, we soon realized that there were many exemplars either underway or that had been published that typiﬁed what ‘‘revolution with a solution’’ meant. The following is by no means exhaustive. However, there was Professor Edwards’ own work on ﬁt (e.g., Edwards, 1996, 2001). There was Becker’s article (2005) on the misuse of control variables. In most respects, the Bergh and Ketchen volumes on research practices in strategy fall under this general title. There are the numerous feature topics in Organizational Research Methods such as the one in the July 2008 issue on what is quality in qualitative research or the one in the July 2009 issue on common method variance (CMV). And most certainly there is my own passion with SMMULs. While we agreed that there are many ﬁne examples, they were not united efforts under the ‘‘revolution with a solution’’ banner. Thus, we have spent considerable time in recent years in recruiting members into this revolution. Uniﬁcation is important because like all revolutions, it is a campaign that has many fronts. SMMULs are just one of those fronts. Thus, what is a SMMUL? Vandenberg (2006, pp. 195–196) ﬁrst attempted to deﬁne it by characterizing it as follows: Doctoral students may be taught or told something to do within the research process as if it were an absolute truth when in reality it is not, and yet, being who they are, they accept that presumed fact as the ‘‘truth.’’ Similarly, authors may accept something from an editor or a reviewer who in turn was told that ‘‘this’’ is the way it must be as well. The unfortunate outcome is that the truism being perpetuated is anything but true. These are aspects of the research process that are, in reality, myths or urban legends. At one point,

240

ROBERT J. VANDENBERG

there may have been a kernel of truth to it, but that kernel has long been forgotten or altered in such a way as to be lost. Rather, unbeknownst to the student, author, reviewer, and editor applying the criterion, it is a criterion of the legendary kind (i.e., ‘‘my grandpappyy’’). There are all kinds of deleterious side effects to this, not the least of which may be the unfair evaluation of a manuscript against criteria that are mythical in nature or the application of the criteria in undertaking some aspect of the research process resulting in a ﬁnished study of questionable quality. The overall end result, however, is a degradation of the whole research process.

Spector (2006) also noted that ‘‘the term urban legend is appropriate in that it reﬂects something that is based on truth but has been distorted and exaggerated as it is passed from person to person over time. We have all heard it so often and from so many sources that it does not occur to many of us to question the extent to which it is true’’ (p. 222). This includes statements such as ‘‘our reliability of .81 exceeds the cutoff criterion of .70.’’ Or ‘‘the ﬁt indices support the ﬁt of our model to the data.’’ Or ‘‘we permitted the error terms to correlate to deal with the unmeasured variables problem.’’ For reasons stated in forthcoming pages, these are all urban legends in which there lies a kernel of truth but that truth has long been lost and distorted. My passion with SMMULS resonated quite strongly with many others as well as evidenced by a 2004 All Academy Symposium in which attendance was well over 300 people. Furthermore, the April 2006 issue of Organizational Research Methods (ORM) contained a SMMUL feature topic, and this issue remains historically one of the most cited issues. Given the latter’s success, another SMMUL feature topic appeared in the April 2011 issue of ORM. At the time of this chapter’s writing, we do not know yet about its success. Another indicant of the strong interest in SMMULs is the six symposia that Professor Charles Lance and I have organized at each conference of the Society of Industrial and Organizational Psychology since 2006. Each one is typically scheduled in a large room, and there is standing room only by the time the symposium starts. Finally, our (Lance & Vandenberg, 2009) edited book on SMMULs has sold nearly 2,000 copies since its release. I would like to think that the resonance of the topic is an indication that ‘‘all is not quiet on the SMMULs front’’ in this revolution with a solution. In what follows is a brief review of some but not all of the published SMMULs to date. Brief is the operative term because these publications represent hundreds of pages at this point. This in turn puts the onus squarely on the reader to thoroughly consult the primary references associated with each SMMUL. I also selected out those SMMULs that I believe are of less interest to the strategy researcher. This was purely a judgment call on my part and thus, I may be wrong. Furthermore, I apologize in advance to any of the

The Revolution with a Solution

241

SMMUL contributors who feel that I made a mistake in leaving them out of this chapter. Brevity necessitates a quick focus on the SMMUL, and as such, full treatment of it is not possible in this chapter. Please view each of the following as whetting your appetite but to satisfy it, you will need to consult the original source.

SEVEN DEADLY SMMULS SMMUL 1: The One True Model This SMMUL was ﬁrst introduced by Vandenberg (2006) and elaborated upon by Vandenberg and Grelle (2009). While the authors focused solely on structural equation modeling (SEM) in specifying the SMMUL, it is also equally relevant to regression. The myth is the belief among researchers that favorable model ﬁt indices (i.e., chi-square goodness of ﬁt, Tucker–Lewis index, normed ﬁt index, root mean square error of approximation, and the many others) permit them to accept their well articulated theoretically derived model. The truth, however, is that these ﬁt indices support the null hypothesis that the reproduced or model-implied variance–covariance matrix equals the observed variance–covariance matrix. Not rejecting the null hypothesis is not the same thing at all as accepting the model of theoretical interest – it only means we fail to reject it (Vandenberg, 2006). The legend, therefore, is a conﬁrmation bias based upon making a false inference about the theoretical model from the set of ﬁt indices. This bias and the potential consequences for engaging in it were succinctly summarized by McCoach, Black, and O’Connell (2007, p. 464) in the following paragraph: In SEM, it is impossible to conﬁrm a model. Although we may fail to conﬁrm a model, we can never actually establish its veracity (Cliff, 1983). Statistical tests and descriptive ﬁt indices can never prove that a model is correct (Tomarken & Waller, 2003). Rather, they suggest that the discrepancy between the observed variance covariance and the modelimplied variance covariance matrix is relatively small. Therefore, one can reasonably conclude that the model ‘‘provides an acceptable description of the data examined’’ (Biddle & Marlin, 1987, p. 9), in the sense that the covariance matrix implied by the speciﬁed model sufﬁciently reproduces the actual covariance matrix. Moreover, ‘‘when the data do not disconﬁrm a model, there are many other models that are not disconﬁrmed either’’ (Cliff, 1983, p. 117), given the number of untested models that are statistically equivalent to the speciﬁed model. Therefore, in the best case scenario, when we achieve good ﬁt, we can conclude our model ‘‘is one plausible representation of the underlying structure from a larger pool of plausible models’’ (Tomarken & Waller, 2003, p. 580).

242

ROBERT J. VANDENBERG

Vandenberg and Grelle (2009) noted that this legend has had two interrelated consequences over the past two decades. First is simply the continued practice of ‘‘accepting’’ the theoretical model of interest based on strong ﬁt when this is clearly inappropriate. In part, this is driven by the plethora of published articles in which the inference was ‘‘to accept’’ and thus, as noted by Spector (2006) no one questions the appropriateness of this inference. The second consequence is represented by the very last sentence in the above quote; that is, strong ﬁt means that the theoretical model is only one plausible model from a much larger pool of plausible models. Owing to the conﬁrmation bias of ‘‘accepting’’ the model, the latter does not typically enter into the researcher’s thinking. One may ask at this point how many plausible alternative models exist in the larger pool? MacCallum, Wegener, Uchino, and Fabrigar (1993) put this question in perspective by noting that there are exactly 33,925 mathematically equivalent models with a saturated block of six latent variables. There are also many hundreds if not thousands of other plausible models when one also considers potential nested and nonnested models as well (Vandenberg & Grelle, 2009). Thus, the larger pool of plausible models is quite large indeed. Vandenberg and Grelle (2009) noted that it is time for us in the organizational sciences to stop making the wrong inference about our theoretical models based on the ﬁt indices. Furthermore, once we do, we then have to accept that out of the pool of many thousands of plausible alternative models, there is going to be at least one (probably many more) representing a competing model that is as conceptually plausible as the model of interest. This is a good situation because one can engage in strong inference by pitting one conceptualization against at least one other to see which is best (Aguinis & Adams, 1998). It’s good because it puts our theories at risk as Popper (1959) encouraged us to do.

SMMUL 2: Those Crazy Cutoff Criteria Lance, Butts, and Michels (2006) noted that methodological lore over the decades has established a number of ‘‘conventionally’’ accepted cutoff values for criteria that if met or exceeded purportedly provide some index of study quality. They examined the veracity of four such ‘‘conventional’’ cutoff criteria by returning to the cited original source of the convention to address: (a) Did that source really say that; and (b) if not, what did the source really say? The following paragraphs summarize Lance et al.’s (2006) ﬁndings for two out of the four cutoff criteria: (a) the .90 value for

The Revolution with a Solution

243

goodness-of-ﬁt indices; and (b) the .70 value for reliability. The level of detail provided by Lance et al. (2006) to anchor their ﬁndings and conclusions is too much to enumerate here. Thus, once again, the reader is encouraged to read their article in full. Lance et al. (2006) noted that it is very common for authors to claim that the GFIs or goodness-of-ﬁt indices [e.g., Tucker–Lewis index, normed ﬁt index, comparative ﬁt index (CFI), and the many others standardized between 0 and 1] equal to or greater than .90 indicate well ﬁtting models. Furthermore, more often than not, these authors will claim this cutoff value was established by Bentler and Bonett (1980). The reality is that relatively little of their article was devoted to GFIs, and even then limited to just the Tucker–Lewis and normed ﬁt indices. Furthermore, what they actually stated about the latter was ‘‘experience will be required to establish values of the indices that are associated with various degrees of meaningfulness of results. In our experience, models with overall ﬁt indices of less than .9 can usually be improved substantially’’ (Bentler & Bonett 1980, p. 600). Thus, nothing was stated about .9 or greater indicating well ﬁtting models. Furthermore, Lance et al. (2006) noted that this urban legend morphed over time to include ﬁt indices such as rho, delta values and the CFI. The CFI is particularly interesting because it was ﬁrst reported 10 years after the Bentler and Bonett (1980) article. Without going into detail, Lance et al. (2006) noted a similar phenomenon occurring with the .95 values mentioned by Hu and Benter (1998, 1999) but which was later criticized because of some very restrictive assumptions regarding their simulation (Marsh, Hau, & Wen, 2004). The end result has been the same, however, and that is, despite the criticisms, researchers are now claiming in studies that ‘‘models of .95 or better indicate strong ﬁt as stated by Hu and Bentler.’’ As was the case with GFIs, Lance et al. (2006) noted that it is also very common for researchers to state ‘‘all reliabilities exceeded the .70 criterion’’ or some variant of the same statement (e.g., reliabilities were well above the recommended value .70). Furthermore, more often than not, the authors cite Nunnally (1978) as having established that cutoff value. As used in Lance et al. (2006), the following is a direct quotation from the Nunnally (1978) book: what a satisfactory level of reliability is depends on how a measure is being used. In the early stages of research y one saves time and energy by working with instruments that have only modest reliability, for which purpose reliabilities of .70 or higher will sufﬁcey In contrast to the standards in basic research, in many applied settings a reliability of .80 is not nearly high enough. In basic research, the concern is with the size of correlations and with the differences in means for different experimental treatments,

244

ROBERT J. VANDENBERG

for which purposes a reliability of .80 for the different measures is adequate. In many applied problems, a great deal hinges on the exact score made by a person on a testy . In such instances it is frightening to think that any measurement error is permitted. Even with a reliability of .90, the standard error of measurement is almost one-third as large as the standard deviation of the test scores. In those applied settings where important decisions are made with respect to speciﬁc test scores, a reliability of .90 is the minimum that should be tolerated, and a reliability of .95 should be considered the desirable standard. (pp. 245–246)

As noted by Lance et al. (2006), most researchers using a .70 cutoff and attributing it to Nunnally would agree probably that they are not in an early stage of research and thus can get by using measures of modest quality. Indeed, they note that most of these researchers should probably have reliabilities that do not fall below .80, and in cases where critical decisions will be made from the ﬁndings, the reliabilities should be near perfect. Finally, it is often the case that researchers in addition to attributing .70 to Nunnally (1978) will also claim that Nunnally was referring speciﬁcally to Chronbach’s alpha or some other means to operationalize the reliability value. As seen in the above quote, Nunnally does not reference any speciﬁc operationalization of reliability whatsoever. In the case of GFIs, Lance et al. (2006) astutely note that the question as to what constitutes an appropriate GFI cutoff value for strong model ﬁt is very much still in debate to this day. Thus, statements such ‘‘as .90 and above indicate strong model ﬁt’’ are unfounded and should be avoided. Similarly, in the case of reliability, researchers should take into account the maturity of their research and the measures being used and select reliability values appropriate for that stage and measures. SMMUL 3: Common Method What? Among all the SMMULs, CMV is one of the least understood by researchers, editors, and reviewers, and yet it is among the ﬁrst to be addressed or evoked when undertaking a research project or evaluating a manuscript. It is unfortunate, however, that when applied or evoked, it is wrongly done so in the majority of cases and its effects are grossly overstated. While the following summary is based on the Spector (2006) article, the reader is also encouraged to read Chan’s (2009) chapter on the self-report data SMMUL, and the Lance, Baranik, Lau, and Scharlau (2009) chapter on the multitraitmultimethod SMMUL. Chan (2009) notes the reason self-report data are wrongly rejected is due in part to the misperception that it is all biased by some form of CMV. Similarly, Lance et al. (2009) note that what some call

The Revolution with a Solution

245

measurement methods are indeed anything but measurement methods when evaluating method effects in multitrait-multimethod designs. Finally, the reader is also encouraged to read articles in the feature topic on CMV found in the July 2010 issue of Organizational Research Methods (guest edited by Paul Spector and Michael Brannick), and the article by Richardson, Simmering, and Sturman (2009). Spector (2006) succinctly captures the CMV urban legend as follows. Interestingly, the concern for CMV seems to be raised almost exclusively when crosssectional, self-report surveys are used. Monomethod studies (those using the same method for assessing all variables) using other approaches, such as reports about other people (e.g., assessment centers or job performance ratings), are less criticized for the same shortcoming, although some have noted that source bias can be a problem in these other domainsy. This automatic criticism of the cross-sectional self-report has become invoked so broadly and often so automatically that I argue it has achieved the status of a methodological urban legend. (p. 222)

Further in the article, Spector (2006, p. 223) states: ‘‘The legend part is the assumption that the method alone is sufﬁcient to produce biases, so that everything measured with the same method shares some of the same biasesy The reason this is legend is because there are few scientiﬁc data to unequivocally support this view and there are data to refute it (italics added). Spector (2006) then proceeds to review the so-called evidence for CMV (e.g., raw correlations, social desirability, etc.) noting in the end that there is so little evidence that ‘‘the time has come to retire the idea’’ (p. 228). Five years later nothing has occurred to change the validity of the last few sentences. Spector (2006) is careful to note that there is a kernel of truth to the CMV SMMUL. However, the truth is a complex one that cannot be simply distilled into blanket statements such as, ‘‘I am rejecting your manuscript because it is obvious that your reliance upon self-report data has produced CMV of such proportions as to invalidate your ﬁndings.’’ In summary, Spector (2006) notes the method of measurement may impact the observed data in two ways. First, depending upon the nature of the topic and how the questions are asked, questions may impact, for example, the attitude of interest, a person’s behavior, or other objective outcome. Second, characteristics of people, instruments, situations and the construct of interest can allow bias into the measurement process and thus, distort it. It is the latter way that is of most concern to researchers, editors, and reviewers. Spector (2006), though, notes regardless of how the observed data may be impacted by the method of measurement, there is a complex conceptual foundation underlying the impact if it exists at all. Rarely, however, is the conceptual foundation considered by researchers when addressing method effects in

246

ROBERT J. VANDENBERG

study design and execution or by reviewers and editors when evoking the CMV hammer during manuscript review. Spector’s (2006) major point in the end is, ‘‘More sophisticated thinking about method variance is needed than the often knee-jerk complaints of CMV or monomethod bias we hear from both authors and reviewers’’ (p. 231). Spector provides some useful guidance and advice as to that thinking from pages 228 to 231.

SMMUL 4: EFA Practices Bandalos and Boehm-Kaufman (2009, p. 61) commented that ‘‘in recent reviews of exploratory factor analysis (EFA) applications, researchers have described the state of the art as ‘routinely quite poor’ (Fabrigar, Wegener, MacCallum, & Strahan, 1999, p. 295), leading to ‘potentially misleading factor analytic results’ (Preacher & MacCallum, 2003, p. 14).’’ They note that the latter is due in large part to researchers clinging onto four common myths or misconceptions as they engage in EFA procedures. These are: The choice between component and common factor extraction procedures is inconsequential. Orthogonal rotation results in better simple structure than oblique rotation. The minimum sample size needed for factor analysis is y (insert your favorite guideline). The ‘‘Eigenvalues Greater than One’’ rule is the best way of choosing the number of factors. For brevity, the following is a summary of only the ﬁrst three myths listed above. Bandalos and Boehm-Kaufman (2009) reviewed the methodological, philosophical, and statistical differences between component and common factor procedures and concluded that the differences are anything but inconsequential. For example, one difference between the two extraction procedures is the underlying goal for each. Component analysis where all the variance among the variables is analyzed should only be used where the goal is data reduction; that is, the researcher wishes to ‘‘boil down’’ a larger set of variables into a smaller set of composite variables or components. The composite could then be used as predictors rather than the variables avoiding possible multicollinearity issues (Bandalos & Boehm-Kaufman, 2009). In contrast, only the shared variance among the variables is used in common factor analysis where the goal is to uncover a potential set of latent variables. The problem is that just over half of the studies reporting EFA

The Revolution with a Solution

247

results used component analysis and implied that the goal was to uncover the latent structure among their variables. Furthermore, it appears that the primary reason they did so was because others have done the same in past publications – not because the goal of their studies warranted the use of one extraction procedure over the other. Their recommendation obviously is to have researchers select the extraction method that is aligned with the goal of the factor analysis. With respect to the orthogonal (uncorrelated factors) vs. oblique (correlated factors) rotation choice, Bandalos and Boehm-Kaufman (2009) reported that orthogonal is selected most often. Interestingly, the majority of authors provided no rationale for the choice leading one to believe that they were unquestioningly following published practices. When a reason was provided, it was often a belief that orthogonal rotations provide simpler or cleaner solutions because the extracted factors are statistically independent from one another. However, for reasons provided on pages 73–74, Bandalos and Boehm-Kaufman (2009) note that the opposite is likely the case; that is, orthogonal rotation produces weaker simple structures relative to oblique rotation. Thus, their recommendation is to always undertake an oblique rotation and if the factors are truly statistically independent, the correlations between them will be near zero. Finally, with respect to the sample size myth, Bandalos and BoehmKaufman (2009) noted that this myth has emerged in the research literature in one of two variations: (a) researchers specifying a cutoff value for the minimum sample size; or (b) those specifying a sample size to number of variables ratio (N:p ratio). ‘‘However, recent studies of these guidelines by Velicer and Fava (1998), MacCallum, Widaman, Zhang, and Hong (1999), and Hogarty, Hines, Kromrey, Ferron, and Mumford (2005) have all reached the same conclusion, which is that there is no absolute minimum N or N:p ratio’’ (Bandalos & Boehm-Kaufman, 2009, p. 75). Bandalos and Boehm-Kaufman (2009) are careful to note that there is a kernel of truth to sample size as it relates to factor analysis. However, it’s an issue with respect to the number of factors and the commonalities among them, and not with the number of variables as is so often claimed.

SMMUL 5: The Myths of Moderation As noted by Edwards (2009), moderation analysis is an important tool in the organizational sciences, and one used in almost all of its underlying disciplines. He notes further, though, that there is still a great deal of

248

ROBERT J. VANDENBERG

confusion in conducting moderation tests. An unfortunate side effect to this confusion is the development of certain myths that have crept into published studies and are now being adopted by others as truth. Edwards (2009) identiﬁes seven deadly myths with regard to moderation tests. They are ‘‘deadly because they lead researchers to make unwise choices, waste time and effort, and draw conclusions that are misleading or incorrect’’ (Edwards, 2009, p. 143). The seven myths are: Myth 1: Product terms create multicollinearity problems Myth 2: Coefﬁcients on ﬁrst-order terms are meaningless Myth 3: Measurement error poses little concern when ﬁrst-order terms are reliable Myth 4: Product terms should be tested hierarchically Myth 5: Curvilinearity can be disregarded when testing moderation Myth 6: Product terms can be treated as causal variables Myth 7: Testing moderation in SEM is impractical As in the previous section, brevity requires that I only summarize a couple of the myths – once again, encouraging the reader to study the complete Edwards (2009) chapter. A common misunderstanding is the sources of multicollinearity in moderation analysis (see Equations 6.1 and 6.2 on pp. 144–145, and the discussions of them; Edwards, 2009). It is commonly believed that the high correlations between the two ﬁrst-order terms (X and Z) and the product term representing their interaction (XZ) is multicollinearity when in fact it is not. The high correlations are simply a natural by-product of including a product term in the equation with terms deﬁning that product, and they can be greatly reduced (often times to zero) through mean centering. As stated by Edwards (2009), ‘‘the only real source of multicollinearity in Equation 6.1 involves rX,Z, the correlation between X and Z, and this correlation is not affected by mean-centering’’ (p. 146). The major point is to put a stop to referring to the high correlations between the ﬁrstorder terms and their product term as multicollinearity. Multicollinearity only refers to the relationship between the two ﬁrst-order terms (there could be more than two) and impacts their associations with the outcome variable and will do so regardless of whether the ﬁrst-order terms are mean centered or not. With respect to Myth 4, product terms should be tested hierarchically, Edwards (2009) highlights two drawbacks to this approach. Directly quoting from Edwards (2009, pp. 150–151), ‘‘First, when a moderating effect is captured by a single product term, such as XZ in Equation 6.1, hierarchical analysis is unnecessary because the F-ratio in Equation 6.9 will give the same

The Revolution with a Solution

249

result as the t-test of the coefﬁcient on XZ (Cohen, 1978; Jaccard, Wan, & Turrisi, 1990; Kromrey & Foster-Johnson, 1998; McClelland & Judd, 1993).’’ Equation 6.9 in Edwards (2009) simply illustrated the F-test evaluating whether the inclusion of the interaction contributed in the prediction of the outcome above that already captured through the ﬁrst-order terms X and Z. ‘‘A second drawback of the hierarchical approach is that it can generate interpretations of the coefﬁcients on X and Z that are misleading’’ (Edwards, 2009, p. 151). This is because the X and Z terms are typically entered ﬁrst followed by an equation including their product term. ‘‘Hence, when the second step indicates that moderation exists, the coefﬁcients on X and Z in the ﬁrst step should be disregarded because, by deﬁnition, moderation means that the effects of X and Z on Y are not each represented by a single value, but by a range of values that vary across levels of the other variable. This variation is not captured by the coefﬁcients on X and Z from the ﬁrst step, and reporting these coefﬁcients invites their interpretation, which is unwarranted when the second step gives support for moderation’’ (Edwards, 2009, p. 151, italics added). The remaining ﬁve myths are as equally informative as the two that I selected to summarize. Thus, I encourage once more that anyone considering any form of moderation analysis should consult Edwards (2009) ﬁrst before even collecting data.

SMMUL 6: Legends Underlying the Four-Step Test of Mediation As noted by Lebreton, Wu, and Bing (2009) and others (e.g., Mathieu, DeShon, & Bergh, 2008), mediation, like moderation, is a primary analytical tool within the organizational sciences. Also like the topic of moderation, there remains a great deal of confusion surrounding its application. One result of this confusion has been an over reliance upon and use of the Baron and Kenny (1986) four-step approach for evaluating mediation hypotheses (LeBreton et al., 2009). Indeed, it continues to be used with very high frequency by researchers, but when a rationale is provided for its use, it is typically of the variety that the ‘‘four-step approach is an optimal and sufﬁcient test for mediation hypotheses’’ (LeBreton et al., 2009, p. 113). As noted by LeBreton et al. (2009), however, it is the latter rationale that has become an urban legend, and as with the other SMMULs, its veracity is seldom, if ever, questioned by those applying it. For the sake of brevity, I cannot specify how it became a legend here, but LeBreton et al. do so quite

250

ROBERT J. VANDENBERG

nicely on pages 113 through 115 and specify ﬁve reasons why this has occurred. LeBreton et al. (2009) viewed this urban legend as composed of three related parts. These are: Legend 1: A test of a mediation hypothesis should consist of the four steps articulated by Baron and Kenny (1986). Legend 2: Their four-step procedure is the optimal test of mediation hypotheses. Legend 3: Fulﬁlling the conditions articulated in their four-step test is sufﬁcient for drawing conclusions about mediated relationships. LeBreton et al. (2009) provide a very detailed and technical explanation from pages 116 through 120 as to fallacy for believing that the four steps are necessary to completing tests of mediation hypotheses. In summary, among the key issues raised by them are as follows. First, strict adherence to step or condition 1 (requiring a signiﬁcant association between X and Y) may prevent a test of mediation when full mediation is the true model, particularly when dealing with relatively small sample sizes. Furthermore, the ﬁrst step is even more problematic when the mediation hypotheses represent complex chain models with multiple mediators. An issue as it relates to step or condition 4 (that association between X and Y drops to zero when controlling the mediator) is that this step is in reality data mining and exploratory in nature. LeBreton et al. (2009) raise other issues as well for the other steps and conditions. In fairness, they also note that even Kenny and his colleagues (Kenny, Kashy, & Bolger, 1998) have recognized some of these issues and have themselves stated that steps 1 and 4 may not be necessary. ‘‘However, after nearly 10 years, this retraction has gone largely unnoticed by users of the Baron and Kenny four-step testyThus, the original article continues to be the predominant one used to deﬁne and justify mediation via the four-step test, even after one of the original authors revised and retracted portions of the four-step test’’ (LeBreton et al., 2009, p. 120, italics added). This alone should assuage any remaining doubts among this chapter’s readers as to why the use of the four-step strategy is included as a SMMUL. With respect to the belief among researchers that the four-step approach represents the ‘‘optimal’’ test of mediation hypotheses, LeBreton et al. (2009) dispel this one as well. The problem with this belief is it has distracted researchers from the myriad of alternative, yet stronger strategies for testing mediation hypotheses. LeBreton et al. (2009) largely summarize the MacKinnon, Lockwood, Hoffman, West, and Sheets (2002) article to underscore the fallacy in this belief. MacKinnon et al. (2002) organized the

The Revolution with a Solution

251

mediation tests into three categories – the causal step approach, difference of coefﬁcients approach, and the products of the effects approach. The fourstep strategy falls in the ﬁrst, causal step approach. MacKinnon et al. (2002) empirically compared the strategies along several criteria (i.e., power requirements, occurrence of type 1 errors, etc.). As stated by LeBreton et al. (2009, p. 121, italics added), ‘‘Although no universal, clear-cut ‘winner’ emerged among the fourteen tests of mediation, clear-cut losers did emerge – techniques relying on testing a set of causal steps, such as Baron and Kenny’s (1986) four-step testythe overwhelming conclusion of this article was that the Baron and Kenny (1986) four-step test was not the optimal strategy for detecting mediated relationships.’’ I italicized the material in the above paragraphs for one reason. Like the CMV SMMUL, there is an adherence to the four-step strategy that deﬁes logic and reason in my opinion given such overwhelming evidence to the contrary. Researchers need to look at this evidence and modify their analyses accordingly. Furthermore, I would strongly encourage looking into other strategies for testing mediation hypotheses. An excellent starting point is the April 2008 issue of Organizational Research Methods, which contains a feature topic on mediation with commentaries at the end by Lawrence James and David Kenny. SMMUL 7: Really!? You Allowed the Residuals to Correlate? I could not identify a better means to introduce the correlated residual SMMUL other than to use the introduction provided by Landis, Edwards, and Cortina (2009). It is as follows: Imagine a situation in which an innocent researcher wishes to explain variance in some critical criterion variable. For expedience, he uses only a single predictor in his validation study. After data collection, the researcher observes a rather unimpressive correlation (e.g., rxy ¼ .10). The researcher then explains in his Discussion section that his model is perfect because, if he had measured all relevant variables, his ‘model’ would have explained all the variance in the criterion. Absurd, you say! Ridiculous! J’accuse! How can one argue for the integrity of a model based on unmeasured variables and/or unexpected relationships? Despite the lunacy of the preceding example, a similar practice occurs with some frequency in applications of structural equation modeling (SEM). Speciﬁcally, the practice of allowing for correlated residuals among indicators in SEM is, in many cases, tantamount to capitalizing on ‘what could have been’ and serves as the focus of the current chapter. (p. 193)

In a review of the top organizational science journals over a ﬁve-year time period and focusing only on published articles in which SEM was used,

252

ROBERT J. VANDENBERG

Landis et al. (2009) found that up to 12% of those articles included correlated residuals. The good news is that it was only up to 12%, but the bad news is also that it was up to 12%. Examining the rationale underlying its use, Landis et al. (2009) concluded it was of the urban legend type where there is an apparent conviction among some researchers that it is a sensible practice to permit item residuals (IRs) to be correlated in SEM applications to obtain better model ﬁt. They trace the origins of this urban legend on pages 196 and 197 and why researchers hang onto this conviction (e.g., unmeasured variables). Landis et al. (2009) conclude, however, that in the vast majority of cases permitting correlated residuals should not be applied in the analyses at all. At the risk of greatly oversimplifying the authors detailed explanation as to why it is inappropriate (see pages 195–209), the primary problem is severe capitalization upon chance. Furthermore, the severity increases dramatically for each pair of residuals permitted to be correlated. The end result is estimating parameters in a model that cannot be generalized to the population as a whole. Given that most researchers wish to make general inferences about the population from their sample parameter estimates (i.e., results), permitting correlated residuals means that technically they cannot do so. Furthermore, interjecting my own understanding of the issue means that most published studies using correlated residuals to date are not all that meaningful in light of having engaged in this practice. Landis et al. (2009) state that there are very limited circumstances where correlated residuals may be permitted. One is when using longitudinal (repeated measures) data such as a latent growth model. Correlated residuals are permissible is this case to account for auto-regressiveness among like items across time (see Bentein, Vandenberg, Vandenberghe, & Stinglhamber, 2005). Landis et al. (2009) note further that correlated residuals may be permitted where there are shared components such as in testing a latent interaction term. Other than these very limited circumstances, there is no justiﬁcation for permitting this practice. If some speciﬁcation search (e.g., modiﬁcation index) during model test uncovers correlated residuals, it is typically due to a problem with the underlying measure and not because there is an unmeasured variable problem or the other justiﬁcations used by researchers to permit correlated residuals. Landis et al. (2009) highly recommend that the researcher consult Anderson and Gerbing (1988) and examine the unidimensionality of the measure. Perhaps the researcher can remove one of the items from the pair without jeopardizing the validity of the measure itself. If the researcher truly feels it is an unmeasured variables issue, then Landis et al. (2009) state that the researcher has no choice but to collect new data to test this hypothesis. Their point is that using correlated residuals

The Revolution with a Solution

253

on the belief that there is an unmeasured variable is pure conjecture unless one can demonstrate it. The odds are that the researcher is capitalizing on pure chance.

SUMMARY AND CONCLUSIONS The seven SMMULs summarized in the previous section are not the only ones. There are many more – some of which have appeared in press and even more that have not. The text above as well as the references below should guide readers to those that have been published. I’m quite positive that Charles Lance and I will produce another volume from those that have not yet appeared in print. However, the seven presented in this chapter are summarized in Table 1. Now that you have completed your reading of this chapter, you have become a full-ﬂedged member of the ‘‘revolution with a solution.’’ Welcome! Your part in this revolution now is to alter your own practices if you were prone to engaging in some of the SMMULs presented here or in the other sources. Your most important role in my opinion, however, is to educate your colleagues, students, editors, and reviewers and alert them to these SMMULs; that is, to spread the word. The whole research process will beneﬁt tremendously to the degree that researchers engage in stronger methodological and analytical practices than was the case before, and to the degree that decisions regarding manuscript worthiness are based on sound heuristics and scripts. Some topics such as CMV and the Baron and Kenny (1986) four-step strategy to testing mediation hypotheses are going to be more difﬁcult to ‘‘change’’ than will other topics. This is due simply to the plethora of publications in which these topics have appeared over the decades, and thus, the perception among researchers and reviewers that ‘‘so many publications couldn’t be wrong.’’ However, they are, and that is the point here as well as in the original sources. Just because some practice or belief has been mentioned over and over does not make it correct. It simply means that it continues to be accepted without question. In closing, I would like to recruit your help in one other manner. If there is an urban legend of which you are aware and that has not been mentioned, please contact me. You should have a strong sense at this juncture as to the types of legends we are seeking. However, please beware that if you contact me, I may ask you to take the lead on writing an article or chapter enumerating it. ¡Viva la Revolucio´n!

254

ROBERT J. VANDENBERG

Table 1.

Summary of This Chapter’s Severn Deadly SMMULs.

The SMMUL Strong ﬁt permits me to accept my theoretical model

GFIsZ.90 indicate well ﬁtting models; Conventional reliabilityZ.70 Assumption that the method alone is sufﬁcient to produce biases, so that everything measured with the same method shares some of the same biases EFA consists of principle components, orthogonal rotation produces the best simple solution and the recommended sample to item (N:p) ratio isy Moderation analyses are easy to interpret Baron & Kenny four-step test of mediation is the best strategy and all four steps must be completed to conclude there is mediation Correlated item residuals are okay to achieve strong measurement model ﬁt

The Relative Truth

Recommended Practice

Strong ﬁt means that my model is one possible representation of the underlying structure from a larger collection of plausible models Unsure yet what GFI indicates strong ﬁt; Reliability depends on measure and decisions to be made Historically, very little scientiﬁc evidence, and researchers not understanding the complex nature of the underlying issues Goals of component analysis are not what most researchers are seeking; orthogonal produces complex solutions; no N:p ratios or minimum sample sizes were ever speciﬁed Interpretations are often maligned due to seven unaccounted for myths There are better strategies than the four-step approach, and Kenny has long retracted that all steps are needed

Recognize what strong model ﬁt really means and specify competing models

Produces results that cannot be unambiguously interpreted due to severe capitalization of chance

Consult latest research on GFIs; Use reliability value ﬁtting of the study and its consequences Retire the term ‘‘common method variance,’’ and educate yourself as to the real state of the issues instead of making automatic assumptions Use common factor analysis; oblique rotation produces proper simple solutions; Research what sample size really means in factor analysis and don’t use arbitrary rules Attend to the myths – too much to put in a summary table Stop using the four-step approach and adopt strategies appropriate to the study in which mediation hypotheses appear Stop the practice as researchers and reviewers should seriously question manuscripts in which this occurs

Note: GFI ¼ refers to goodness-of-ﬁt indices standardized between 0 and 1 such as the Tucker– Lewis index; EFA ¼ exploratory factor analysis.

255

The Revolution with a Solution

REFERENCES Aguinis, H., & Adams, S. K. R. (1998). Social-role versus structural models of gender and inﬂuence use in organizations: A strong inference approach. Group & Organization Management, 23, 414–446. Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103, 411–423. Bandalos, D. L., & Boehm-Kaufman, M. R. (2009). Four common misconceptions in exploratory factor analysis. In: C. E. Lance & R. J. Vandenberg (Eds), Statistical and methodological myths and urban legends: Received doctrine, verity, and fable in the organizational and social sciences (pp. 61–87). New York: Routledge. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Becker, T. E. (2005). Potential problems in the statistical control of variables in organizational research: A qualitative analysis with recommendations. Organizational Research Methods, 8, 274–289. Bentein, K., Vandenberg, R., Vandenberghe, C., & Stinglhamber, F. (2005). The role of change in the relationship between commitment and turnover: A latent growth modeling approach. Journal of Applied Psychology, 90, 468–482. Bentler, P. M., & Bonett, D. G. (1980). Signiﬁcance tests and goodness of ﬁt in the analysis of covariance structures. Psychological Bulletin, 88, 588–606. Biddle, B. J., & Marlin, M. M. (1987). Causality, conﬁrmation, credulity, and structural equation modeling. Child Development, 58, 4–17. Chan, D. (2009). So why ask me? Are self-report data really that bad? In: C. E. Lance & R. J. Vandenberg (Eds), Statistical and methodological myths and urban legends: Received doctrine, verity, and fable in the organizational and social sciences (pp. 309–336). New York: Routledge. Cliff, N. (1983). Some cautions concerning the application of causal modeling methods. Multivariate Behavioral Research, 18, 115–126. Cohen, J. (1978). Partialed products are interactions: Partialed powers are curve components. Psychological Bulletin, 85, 858–866. Edwards, J. R. (1996). An examination of competing versions of the person-environment ﬁt approach to stress. Academy of Management Journal, 39, 292–339. Edwards, J. R. (2001). Ten difference score myths. Organizational Research Methods, 4, 264–286. Edwards, J. R. (2009). Seven deadly myths of testing moderation in organizational research. In: C. E. Lance & R. J. Vandenberg (Eds), Statistical and methodological myths and urban legends: Received doctrine, verity, and fable in the organizational and social sciences (pp. 143–164). New York: Routledge. Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating the use of exploratory factor analysis in psychological research. Psychological Methods, 4, 272–299. Hogarty, K. Y., Hines, C. V., Kromrey, J. D., Ferron, J. M., & Mumford, K. R. (2005). The quality of factor solutions in exploratory factor analysis: The inﬂuence of sample size, communality, and overdetermination. Educational and Psychological Measurement, 65, 202–226.

256

ROBERT J. VANDENBERG

Hu, L., & Benter, P. M. (1998). Fit indices in covariance structure modeling: Sensitivity to underparameterization model misspeciﬁcation. Psychological Methods, 3, 424–453. Hu, L., & Benter, P. M. (1999). Cutoff criteria for ﬁt indices in covariance structure analysis. Structural Equation Modeling, 6, 1–55. Jaccard, J., Wan, C. K., & Turrisi, R. (1990). The detection and interpretation of interaction effects between continuous variables in multiple regression. Multivariate Behavioral Research, 25, 467–478. Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social psychology. In: D. T. Gilbert, S. T. Fiske & G. Lindzey (Eds), The handbook of social psychology (4th ed., pp. 233–265). Burr Ridge, IL: McGraw-Hill. Kromrey, J. D., & Foster-Johnson, L. (1998). Mean centering in moderated multiple regression: Much ado about nothing. Educational and Psychological Measurement, 58, 42–68. Lance, C. E., Baranik, L. E., Lau, A. R., & Scharlau, E. A. (2009). If it ain’t trait it must be method: (Mis)application of the multitrait-multimethod design in organizational research. In: C. E. Lance & R. J. Vandenberg (Eds), Statistical and methodological myths and urban legends: Received doctrine, verity, and fable in the organizational and social sciences (pp. 337–360). New York: Routledge. Lance, C. E., Butts, M. M., & Michels, L. C. (2006). The sources of four commonly reported cutoff criteria: What did they really say? Organizational Research Methods, 9, 202–220. Lance, C. E., & Vandenberg, R. J. (2009). Statistical and methodological myths and urban legends: Received doctrine, verity, and fable in the organizational and social sciences. New York: Routledge. Landis, R. S., Edwards, B. D., & Cortina, J. M. (2009). On the practice of allowing correlated residuals. In: C. E. Lance & R. J. Vandenberg (Eds), Statistical and methodological myths and urban legends: Received doctrine, verity, and fable in the organizational and social sciences (pp. 193–215). New York: Routledge. LeBreton, J. M., Wu, J., & Bing, M. N. (2009). The truth(s) on testing for mediation. In: C. E. Lance & R. J. Vandenberg (Eds), Statistical and methodological myths and urban legends: Received doctrine, verity, and fable in the organizational and social sciences (pp. 107–141). New York: Routledge. MacCallum, R. C., Wegener, D. T., Uchino, B. N., & Fabrigar, L. R. (1993). The problem of equivalent models in applications of covariance structure analysis. Psychological Bulletin, 114, 185–199. MacCallum, R. C., Widaman, K. F., Zhang, S., & Hong, S. (1999). Sample size in factor analysis. Psychological Methods, 4, 84–99. MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002). A comparison of methods to test mediation and other intervening variable effects. Psychological Methods, 7, 83–104. Marsh, H. W., Hau, K., & Wen, Z. (2004). In search of golden rules: Comment on hypothesistesting approaches to setting cutoff values for ﬁt indexes and dangers in overgeneralizing Hu and Bentler’s (1999) ﬁndings. Structural Equation Modeling, 11, 320–341. Mathieu, J. E., DeShon, R. P., & Bergh, D. P. (2008). Mediational inferences in the organizational sciences: Then, now, and beyond. Organizational Research Methods, 11, 203–223. McClelland, G. H., & Judd, C. M. (1993). Statistical difﬁculties of detecting interactions and moderator effects. Psychological Bulletin, 114, 376–390.

The Revolution with a Solution

257

McCoach, D. B., Black, A. C., & O’Connell, A. A. (2007). Errors of inference in structural equation modeling. Psychology in the Schools, 44, 461–470. Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York: McGraw-Hill. Popper, K. R. (1959). The propensity interpretation of probability. The British Journal for the Philosophy of Science, 10, 25–42. Preacher, K. J., & MacCallum, R. C. (2003). Repairing Tom Swift’s electric factor analysis machine. Understanding Statistics, 2, 13–43. Richardson, H. A., Simmering, M. J., & Sturman, M. C. (2009). A tale of three perspectives: Examining post hoc statistical techniques for detection and correction of common method variance. Organizational Research Methods, 12, 762–800. Spector, P. E. (2006). Method variance in organizational research: Truth or urban legend? Organizational Research Methods, 9, 221–232. Tomarken, A. J., & Waller, N. G. (2003). Potential problems with ‘‘well ﬁtting’’ models. Journal of Abnormal Psychology, 112, 578–598. Vandenberg, R. J. (2006). Statistical and methodological myths and urban legends: Where pray tell did they get this idea? Organizational Research Methods, 9, 194–201. Vandenberg, R. J., & Grelle, D. (2009). Alternative model speciﬁcations in structural equation modeling: Facts, ﬁctions, and truth. In: C. E. Lance & R. J. Vandenberg (Eds), Statistical and methodological myths and urban legends: Received doctrine, verity, and fable in the organizational and social sciences (pp. 167–193). New York: Routledge. Velicer, W. F., & Fava, J. L. (1998). Effects of variable and subject sampling on factor pattern recovery. Psychological Methods, 2, 231–251.

QUALITATIVE COMPARATIVE ANALYSIS AND STRATEGIC MANAGEMENT RESEARCH: CURRENT STATE AND FUTURE PROSPECTS Thomas Greckhamer and Kevin W. Mossholder ABSTRACT Purpose – This chapter examines the potential of qualitative comparative analysis (QCA) for strategy research. Methodology/approach – We introduce the set-theoretic framework of QCA and provide an overview of recent methodological developments. Findings – We utilize a variety of examples relevant to strategy research to illustrate the action steps and key concepts involved in conducting a QCA study. Originality/value of paper – We develop examples from core research areas in strategic management to illustrate QCA’s potential for examining issues of causality and diversity in strategy research, and in settings involving medium-N samples. We conclude by emphasizing that QCA offers an alternative mode of inquiry to open and redirect important lines of strategy research. Building Methodological Bridges Research Methodology in Strategy and Management, Volume 6, 259–288 Copyright r 2011 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-8387/doi:10.1108/S1479-8387(2011)0000006010

259

260

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

Keywords: Qualitative comparative analysis; QCA; strategy; set theory; crisp sets; fuzzy sets The ﬁeld of strategic management has developed a strong focus on issues regarding the amount of diversity among ﬁrms, the differential success achieved by some ﬁrms, and managerial interventions that make certain ﬁrms more successful (e.g., Bowman, Singh, & Thomas, 2002; Bromiley, 2005; Carroll, 1993; Rumelt, Schendel, & Teece, 1994). Research agendas and streams spawned by the pursuit of these issues form the core of strategic management. Recently, strategic management researchers have been presented with a theoretical and methodological framework – qualitative comparative analysis (QCA) (Ragin, 1987, 2000) – that speciﬁcally focuses on exploring the diversity of ﬁrms as well as the interdependence of causal effects underlying outcomes like ﬁrm performance (Greckhamer, Misangyi, Elms, & Lacey, 2008; Kogut, MacDufﬁe, & Ragin, 2004). The purpose of this chapter is to provide an introductory overview of QCA and highlight its potential utility for strategy research. We introduce QCA basics, review current strategy-related applications, and connect recent methodological developments with strategy scholarship. Addressing both methodological issues and potential future applications, we demonstrate how strategy researchers might beneﬁt from this evolving analytical approach.

QUALITATIVE COMPARATIVE ANALYSIS As in many social sciences, research questions in strategic management can be explored in various ways. The methods employed are frequently organized into qualitative (i.e., focusing on in-depth studies examining small numbers of speciﬁc cases) and quantitative (i.e., focusing on relationships between variables across many cases) (e.g., Mahoney & Goertz, 2006; Newman & Benz, 1998; Ragin, 1994). Building on arguments that qualitative and quantitative research approaches possess complementary features, Ragin (1987) introduced QCA with the aim to close the gap between these research orientations. Starting from the premise that most empirical social science research involves comparisons of one kind or another, Ragin (1987) developed QCA to remedy two fundamental problems presented by empirical cross-case analyses. The ﬁrst problem is identifying the types of cases in a way that coherently integrates their key similarities and differences. The second problem is assessing causal complexity when an outcome can result from

Qualitative Comparative Analysis and Strategic Management Research

261

several different combinations of case attributes. To resolve these problems, QCA was proposed as a synthetic method featuring both case-oriented and variable-oriented analytical procedures. It offers a formal approach to map the diversity of cases as well as the variety of ways an outcome is reached, with the objective of constructing empirical evidence for causal necessity and causal sufﬁciency (Ragin, 2000). A condition is considered necessary if it must be present for a certain outcome to occur, and sufﬁcient if it can produce a certain outcome by itself. QCA is set theoretic, as opposed to correlational, in both its theoretical and methodological approach. Set theory pertains to a branch of mathematics dealing with collections of elements (sets) and relationships among these elements. It provides a framework that relies on Boolean algebra, a system of symbolic mathematical logic, to study relations among sets (for an introduction to Boolean algebra and set theory, see e.g., Mendelson, 1970; Smithson & Verkuilen, 2006; Whitesitt, 1995). In classical or crisp set theory, an element either belongs inside a set or it does not based upon certain rules. Thus, crisp sets are dichotomous and cases are classiﬁed as either ‘‘in’’ (1) or ‘‘out’’ (0) of the set, signifying a truth value. For example, a speciﬁc industry may belong or not belong to the set of muniﬁcent industries. If this industry has full membership in the set (membership ¼ 1), the statement that the industry belongs to the set of muniﬁcent industries is true, else it is false (membership ¼ 0). A more versatile approach in classifying cases involves the concept of fuzzy sets. In fuzzy set theory, a case may belong partially to a set with membership scores in the range from 0 to 1, permitting gradations of set membership and thereby blurring set boundaries (Smithson, 1987). To continue with the muniﬁcence example, in addition to be fully in (1) or out (0) of the set of muniﬁcent industries, a speciﬁc industry may be classiﬁed as having a partial degree of membership in the set (e.g., it may belong more in than out of the set of muniﬁcent industries and have set membership score of 0.75). Set membership and relations are expressed through Boolean algebra. The combinatorial logic of Boolean algebra allows researchers to link combinations of the presence or absence of causally relevant attributes to an outcome, as well as reduce expressions of causal complexity. The two basic Boolean operators – logical and and logical or – represent the primary means of designating set relations. The operator and represents the intersection of sets, and is used when conditions A and B combined may lead to an outcome. For example, ﬁrms may show superior performance if they are located in a muniﬁcent industry and have abundant slack resources. The operator or represents the union of sets, and is used when either one condition or another may lead to the same outcome. For example, ﬁrms may show

262

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

inferior performance if they lack slack resources or if they are located in a highly competitive industry. The Boolean negation not denotes the complement of a deﬁned set and contains cases not in the deﬁned set. For example, the set of not large ﬁrms encompasses all those cases that do not satisfy the criterion for membership in the set of large ﬁrms. It does not denote any other set such as the set of small ﬁrms, however, which would have to be deﬁned separately.

Advantages of QCA for Strategic Management Research As a theoretical framework and methodological approach, QCA has several desirable properties for studying strategic management issues. To begin with, it is compatible with theories of organization and strategy as conﬁgurations of design elements (e.g., Fiss, 2007; Ketchen et al., 1997; Miller, 1986; Miller & Friesen, 1984). Indeed, the unit of analysis in QCA is the conﬁguration of included case attributes. QCA’s basic premise that aspects of cases should be examined as packages and that a single difference between cases may constitute a difference in kind permits researchers to capture organizations’ diversity as conﬁgurations of resources, structures, and strategies. This paradigm differs greatly from linear statistical approaches often employed in strategy research (Fiss, 2007; Greckhamer et al., 2008). In QCA, causal explanations become multiple conjunctural (Becker, 1992). They are conjunctural in that causes operate in combination rather than independently, and multiple (i.e., equiﬁnal) because more than one combination may produce the same outcome. Implications of this perspective are that outcomes of interest rarely have a single cause, causes rarely operate in isolation, and speciﬁc causes may have opposite effects depending upon context (Ragin, 1987). Strategy scholars generally recognize that outcomes like ﬁrm performance or joint venture success are inﬂuenced by a multitude of organizational and environmental attributes. They also accept that these causes do not operate in isolation, a position held by conﬁgurational theories. Finally, scholars have noted that some strategies may lead to success in one competitive environment but lead to failure in different settings. Social science concepts describing ideal attributes of cases can be somewhat vague (Smithson, 1987; Verkuilen, 2005). In strategy, for example, dimensions of organizations’ task environments have been characterized as muniﬁcent-scarce, stable-dynamic, and simple-complex (Burns & Stalker, 1961; Pfeffer & Salancik, 1978; Scott, 1998), but many environments fall somewhere in between these prototypical delineations. Sets and set theory

Qualitative Comparative Analysis and Strategic Management Research

263

allow researchers to use categorization in a manner more in line with theories and scholarly communication processes about strategy and organization. It would, for instance, be especially advantageous to use QCA in studies involving organization environments because environmental dimensions are interrelated (Pfeffer & Salancik, 1978) and potentially equiﬁnal in their impact on ﬁrm performance (e.g., Greckhamer et al., 2008). Through fuzzy set classiﬁcation of environmental dimensions, an industry could be demarcated as fully in, fully out, or partially in the sets of muniﬁcent, dynamic, or complex environments. Finally, QCA can be applied with both large and medium-sized samples. It was originally developed for rigorous analyses of medium-N samples (i.e., 10–50 cases) that are too small for linear statistical methods, but too large for in-depth qualitative analyses. Many strategy-related phenomena involve medium sample sizes, including comparative industry analyses (i.e., studies using industries as cases) and intra-industry analyses of ﬁrm performance in many industries.

QCA Applications in Strategy Research We now offer a brief overview of existing QCA applications, highlighting how these studies contribute to key strategy research areas. Although an exhaustive review of QCA applications in ﬁelds other than strategy is beyond the scope of this chapter, we touch on selected QCA studies having indirect implications for strategic management. Kogut et al. (2004) utilized fuzzy set QCA (fsQCA) to identify combinations of complementary technological and organizational practices affecting manufacturing performance in the international auto industry. They started from the assumption that strategic capabilities are embedded in human– technology relationships that are complementary rather than additive in their combined effects. Fuzzy set analysis enabled these researchers to study how technological and organizational practices could be combined to achieve higher performance in terms of quality and productivity. Four conﬁgurations of automotive plant attributes were identiﬁed as usually sufﬁcient for achieving high performance. To illustrate, one conﬁguration concerned plants that organized work using a broad division of labor, implemented a minimal buffer lean production approach, produced relatively older model designs, were highly automated, and were not large in scale. Utilizing QCA, Kogut et al. (2004) were able to show that plants could achieve performance

264

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

advantages through alternative combinations of complementary interdependencies across key areas of production organization. Greckhamer et al. (2008) utilized QCA to examine the potentially interdependent effects of industry, corporate, and business-unit attributes in determining business-unit performance. They showed how QCA can be utilized to examine combinations of industry, corporate, and business-unit attributes that are linked to superior or inferior performance. Redirecting the debate from questions such as whether corporate strategy matters (e.g., Bowman & Helfat, 2001) toward questions such as under what conditions corporate strategy may matter, Greckhamer et al. (2008) also demonstrated how combinations of attributes usually sufﬁcient for superior and inferior performance differ across broad industry sectors. They further illustrated QCA’s potential to capture causal asymmetry by showing that the determinants of superior and inferior performance are not plain opposites. Fiss (2007, 2011) applied QCA to the literature on organizational conﬁgurations. He noted that even though this literature has stressed nonlinearity, synergies, and equiﬁnality, empirical research largely has drawn on methods that assume linearity, additive effects, and uniﬁnality. Fiss (2007) discussed how set-theoretic approaches constitute a viable alternative for overcoming this mismatch. He subsequently tested Miles and Snow’s (1978) generic typology of organizational conﬁgurations, and these conﬁgurations’ link to very high performance, high performance, and not high or low performance (Fiss, 2011). This line of research suggests that QCA could reinvigorate the organizational typologies literature by combining conﬁgurational theory with a suitable conﬁgurational approach. O¨z (2004) applied QCA to the systematic analysis of evidence of multiplecase studies and demonstrated how this enhances the possibility of deriving empirical generalizations from such data. Utilizing research based on Porter’s (1990) diamond model of key home-country attributes that determine international competitive advantage, she illustrated how QCA can be used to examine commonalities in competitive conditions implied by multiple-case studies of Turkish and Greek industries. Analysis of these case studies revealed that there are four different paths to international competitiveness. This study is also instructive for researchers considering the importance of multiple-case study research designs in strategy research (Ridder, Hoon, & McCandless, 2009). In a similar vein, Schneider, Schulze-Bentrop, and Paunescu (2010) utilized fsQCA to investigate the institutional sources of national competitive advantage in high-technology manufacturing. They examined conﬁgurations of institutional features representing characteristic forms of capitalism linked

Qualitative Comparative Analysis and Strategic Management Research

265

to export performance. Starting from the premise that the performance of ﬁrms is likely to be inﬂuenced by the complex interplay of institutions rather than characteristics of a single one of them, these researchers found no single institutional characteristic was sufﬁcient for explaining export performance in high-technology industries. To the contrary, several conﬁgurations were found to be sufﬁcient for leading to export performance outcomes, with the most important of these involving a combination of extensive university training and a large stock market. Based on the premise that understanding the preferences of multinational enterprises in making long-term direct investments is vital for developing effective global strategies, Pajunen (2008) utilized fsQCA to explore how institutional factors inﬂuence the relative attractiveness of countries for foreign direct investments (FDIs). His results showed institutional factors have diverse inﬂuences on a country’s attractiveness for multinationals’ FDIs. Similar kinds of institutions may be associated with different outcomes in different regions, and in developed versus nondeveloped countries. Pajunen (2008) also found that speciﬁc countries may not be attractive (or unattractive) because of a single institutional factor. Differential policy recommendations for groups of countries emerged regarding how to retain attractiveness as well as avoid unattractiveness for FDIs. Applications from areas indirectly related to strategic management issues suggest further potential QCA contributions. For example, Stokke (2007) utilized QCA to analyze case studies on successful and unsuccessful shaming – criticizing the target’s undesirable behavior before a third party whose opinion matters – as a way of improving the effectiveness of international ﬁsheries management regimes. This study illustrated the application of QCA to data gleaned from a relatively small number (N ¼ 10) of case studies. Three alternative combinations of attributes led to successful shaming. To provide an example, one of these combinations included explicit advice by the regime’s scientiﬁc body, commitment violations by the regime, and that behavioral changes that shamers were seeking to elicit not be inconvenient to the shamed party. Elsewhere, in a cross-cultural study, Greckhamer (2011) used a conﬁgurational approach to investigate set-theoretic combinations of cultural and macro-environmental attributes leading to crossnational differences in compensation level and compensation inequality. His approach to cross-cultural research using fsQCA has important implications for strategy research because various cultures are likely to respond to similar strategic issues in different ways (cf. Schneider & de Meyer, 1991). To give a ﬁnal example, focusing on 11 European and North American countries, Varone, Rothmayr, and Montpetit (2006) utilized QCA to investigate

266

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

variations in policy design processes and resulting policy outputs within the domain of assisted reproductive technology (ART). They illustrated that relevant actors’ policy preferences as well as key institutional rules need to be considered jointly to understand countries’ adoption of ART policies to regulate biomedicine. This study has implications for strategy in that it highlights how QCA can contribute to theory salient for formulating public policy to regulate industries. Lastly, we note that QCA has been applied in political science and sociology to study a range of topics such as the rise of social movements (Amenta, Carruthers, & Zylan, 1992; Cress & Snow, 1996), social movements’ success of impacting political party formation and associated political systems (Redding & Viterna, 1999; Veugelers & Magnan, 2005), evolution of pension and social spending programs in the United States (Amenta, Caren, & Olasky, 2005; Amenta & Halfmann, 2000; Amenta & Poulson, 1996), collective and individual resistance of workers including strikebreaking and solidarity in labor strikes (Brown & Boswell, 1995; Roscigno & Hodson, 2004), performance variation in national medical systems (Hollingsworth, Hanneman, Hage, & Ragin, 1996), and improvements in infrastructure systems (Lam & Ostrom, 2010). Such studies are indirectly relevant to strategy research in that they generally demonstrate the utility of QCA for exploring how conﬁgurations of macro level forces affect organizational and political systems.

QCA IN STRATEGIC MANAGEMENT: A CONCEPTUAL OVERVIEW The process of conducting a QCA study can be organized into ﬁve steps. The ﬁrst step is to select an outcome of interest and cases for which this outcome is relevant and potentially present. For example, a researcher interested in the determinants of superior ﬁrm performance in high-technology industries may select the ﬁrms operating in one or several selected such industries. The second step involves the calibration of crisp or fuzzy sets that assess the outcome as well as key attributes suspected of being causally linked to this outcome based on previous theory and substantive knowledge of the cases. The third step requires the construction of a truth table, comprising all possible combinations of the included causal attributes. The fourth step entails set-theoretic reduction of the truth table; Boolean logic is used to simplify truth table rows and identify attribute conﬁgurations consistently linked to an outcome. The ﬁfth and ﬁnal step involves the representation and

Qualitative Comparative Analysis and Strategic Management Research

267

interpretation of results. It is important to note that one or more of the above steps may be reﬁned and repeated as part of the QCA process. In this section, we illustrate these ﬁve action steps by using a variety of examples and highlighting concepts that are commonly addressed in conducting a study using QCA. More speciﬁcally, we discuss current approaches to the calibration of crisp and fuzzy sets, the construction and analysis of truth tables associated with sets, and the representation and interpretation of results. We illustrate these aspects with examples emphasizing case attributes typically studied in strategic management. It should be noted that this overview is drawn at a conceptual rather than operational level. We purposefully limit the discussion in this manner so as to introduce QCA rather than provide technical instruction. Readers interested in greater operational detail are referred to Ragin (2000, 2008).

Calibration of Sets Exploring the diversity of cases is at the core of QCA; crisp and fuzzy sets capture this diversity in different ways. Crisp sets evaluate set membership in terms of mutually exclusive membership or nonmembership states, thus identifying qualitative differences in kind. For example, membership in the set of ﬁrms with superior performance may delineate cases into those with superior performance (membership ¼ 1) and those lacking superior performance (membership ¼ 0). In contrast, allowing for partial set membership, fuzzy sets involve differentiation that is simultaneously qualitative and quantitative (Ragin, 2000). As noted above, many strategy theories examine potential causal associations by positing rather ill-deﬁned prototypical cases, such as ﬁrms with superior or inferior performance, industries with high or low entry barriers, or operations with high or low asset speciﬁcity. This makes precise categorization of cases difﬁcult. Additionally, many of these descriptive attributes involve both differences in kind and differences in degree. For example, entry barriers in many industries are not merely high or not high (differences in kind), but rather reﬂect gradations of impediments to industry entry (differences in degree). To create crisp or fuzzy sets that capture differences in kind and/or degree, a researcher must decide upon criteria determining set membership. This process is referred to as calibration (Ragin, 2000, 2008). Crisp and fuzzy set calibration infuses sets with distinguishable thresholds based upon empirical and theoretical knowledge. Crisp sets require merely setting thresholds of full membership, whereas fuzzy sets require setting thresholds

268

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

of full membership, full nonmembership, and degrees of membership. Fuzzy sets combine precision akin to quantitative measurement with substantive description that is emphasized by qualitative research. For example, in deciding what comprises set membership in a given study, researchers could be forced to proactively deﬁne degrees of membership in sets such as ﬁrms with superior performance or industries with high entry barriers. As a discovery mechanism, calibration could be informative because it requires understanding what constitutes full membership, full nonmembership, and partial membership in the sets studied. Calibration should be rigorous and transparent as to enable comprehension, evaluation, and replication. In set-theoretic analyses, unlike conventional statistical approaches, not all variation is equally relevant. Calibration demands researchers distinguish between theoretically relevant and irrelevant variance (Ragin, 2000). For example, performance variation among ﬁrms unambiguously excluded (i.e., full nonmembership) from the set of ﬁrms with superior performance is not relevant for studying the causes of superior ﬁrm performance; neither is performance variation among ﬁrms that are unambiguously included (i.e., full membership) in the set of ﬁrms with superior performance. Similarly, in linking ﬁrm size as causal attribute to superior performance, one would not be concerned with variation in size among ﬁrms that are fully in (or out) of the set of large ﬁrms. To express partial set membership, researchers can create a variety of sets, depending upon information available from the data. Table 1 provides an illustration of possible sets (adapted from Ragin, 2000, 2008). Some sets of interest to strategy researchers will be naturally crisp. For example, a researcher studying CEO succession may be concerned with whether a new CEO came from inside or outside of the ﬁrm. In set theory terms, the question becomes whether a ﬁrm belongs or does not belong to the set of ﬁrms with insider CEO succession. Another naturally crisp set could be exempliﬁed by the set of ﬁrms with CEO duality, that is, the same person holding the position of CEO and chairman of the board. The ﬁrst possible reﬁnement in moving from crisp to fuzzy sets is a threevalue fuzzy set. In addition to indicating full (non) membership, this set identiﬁes intermediary membership of cases. The crossover point of 0.5 is the membership score signifying maximum ambiguity or fuzziness of the set’s truth value. Cases with this membership score are neither in nor out of the set, and thus the truth value expressed by this set membership is neither true nor false. Fuzzy set scores greater than 0.5 would indicate relatively strong but partial membership in the set, whereas those smaller than 0.5 would indicate weak set membership cases that are more out than in.

Qualitative Comparative Analysis and Strategic Management Research

Table 1. Two-Value or Crisp Set

269

Examples of Types of Sets.

Three-Value Fuzzy Set

Four-Value set

Five-Value Fuzzy set

Continuous Fuzzy Set

0 ¼ Full 0 ¼ Fully out 0 ¼ Fully out 0 ¼ Fully out 0 ¼ Fully out nonmember- 0.5 ¼ Crossover point; 0.33 ¼ More 0.25 ¼ More out 1 ¼ Fully in ship neither fully in out than in than in nor fully out 1 ¼ Full 1 ¼ Fully in 0.67 ¼ More in 0.5 ¼ Neither Numerical scores membership than out fully in indicating partial nor fully out membership, with 0.5 ¼ crossover point 1 ¼ Fully in 0.75 ¼ More in than out 1 ¼ Fully in Note: Adapted from Ragin (2000, 2008).

Continuing the CEO succession example, in addition to determining whether a ﬁrm’s new CEO came from inside or outside, suppose a researcher was also interested in whether the new CEO was recruited from one of the ﬁrm’s close business contacts (e.g., a main consulting ﬁrm). In this instance, it might be useful to denote partial membership for CEOs with a history of consulting with the ﬁrm, based on the assumption that this work has led them to establish some connections within the ﬁrm. A more reﬁned fuzzy set could include ﬁve possible values, adding membership categories of more in than out and more out than in to the range of possible membership values; Table 1 further includes a variation with four values that includes these membership values but not a crossover point. Calibration schemes of this nature may be useful when researchers have available a substantial amount of information about cases, but the nature of the evidence is not identical or strictly comparable across cases. For example, researchers analyzing data from multiple-case studies of ﬁrms may be able to distinguish gradient membership in sets such as vertical integration or diversiﬁcation by using such and possibly even more ﬁne-grained multi-value fuzzy sets. Finally, a continuous fuzzy set allows set membership values anywhere in the interval from 0 to 1. The current best practice for calibrating conventional interval scale indicators of concepts into continuous fuzzy sets involves the direct method (Ragin, 2008). With this approach, the researcher sets three qualitative break points (setting criteria for full membership and full nonmembership as

270

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

well as the crossover point) based upon theoretical and empirical knowledge. The direct method ﬁrst estimates the log of the odds of full membership in a set as an intermediate step. Then, it calculates the degree of fuzzy set membership by applying a standard formula for converting log odds of membership to gradients between full membership (1) and full nonmembership (0). Again, 0.5 is considered the crossover point of being neither in nor out of a set. A detailed explanation of the direct method is beyond the scope of this chapter, but is available elsewhere (Ragin, 2008). Assume we want to calibrate set membership in the fuzzy set of largest U.S. companies, drawing on the Fortune rankings as external criteria for calibration. These indices are widely considered to discriminate among the largest for-proﬁt organizations in the United States, and feature established set thresholds like the Fortune 100, 500, and 1000. For our demonstration, we chose the size (in revenues) of the 100th largest company as qualitative break point for full membership in the set of largest U.S. ﬁrms, the size of the 500th largest company as crossover point, and the size of the smallest company on the Fortune 1000 ranking as break point for full nonmembership (Fortune, 2010). We calibrated this set applying the direct method. Table 2 presents a randomly selected listing from our calibration of Fortune 1000 ﬁrms, along with their gradient membership in the set of largest U.S. companies. This example demonstrates that a calibration scheme is tied to the deﬁnition of a set. It involved calibrating the set of the largest U.S. companies. If, however, we were interested in calibrating the set of large (rather than largest) U.S. companies, we might decide different qualitative break points were appropriate for demarcating degrees of set membership. Depending on the sets deﬁned, there might be no cases in the sample having full membership or full nonmembership. Thus, in deﬁning a set for a speciﬁc research purpose, a researcher must carefully decide what relevant and irrelevant variation is (Ragin, 2008).

Truth Table Construction and Analysis After calibrating outcome and causal attributes into sets, the data are analyzed using truth tables (Ragin, 2000, 2008). As such, a truth table is constructed to map the logically possible and empirically observed diversity of cases. In Boolean algebra, a truth table is deﬁned as a chart with 2k rows (k ¼ number of included sets) for displaying all logically possible combinations of sets representing truth values for logical statements (Caramani,

271

Qualitative Comparative Analysis and Strategic Management Research

Table 2.

Sample of Fortune 1000 Companies Calibrated into Set of Largest U.S. Firms.

Company

Lowe’s Allstate Staples Oracle Duke Energy BB&T Corp. Textron Quest Diagnostics Harris Group 1 Automotive Northern Trust Corp. Wyndham Worldwide Diversey Holdings LifePoint Hospitals Harman International Ind. LPL Investment Holdings HealthSpring PolyOne Triple-S Management MGIC Investment

Fortune 1000 Rank (2010)

Revenue (in million U.S. Dollars)

Membership Score

42 68 101 105 181 217 220 303 371 457 497 541 624 650 661 686 697 811 838 933

47220 32013 24276 23252 12731 10818 10548 7455 5600 4526 4193 3750 3111 2980 2891 2750 2666 2061 1989 1709

1 0.98 0.95 0.94 0.78 0.73 0.72 0.62 0.55 0.51 0.5 0.38 0.23 0.21 0.19 0.17 0.15 0.08 0.08 0.06

2009; Mendelson, 1970). Each row represents a logically possible conﬁguration of included attributes and constitutes a potential difference in kind among cases. A truth table also describes cases’ limited diversity. Limited diversity describes a situation where not all theoretically possible conﬁgurations exist in empirical reality due to case attributes’ tendency to emerge in coherent patterns (Meyer, Tsui, & Hinings, 1993). By listing all logically possible attribute conﬁgurations of causal attributes and the consistent outcomes associated with each attribute conﬁguration present, researchers make truth tables a tool for analyzing the diversity and causal complexity of crisp and fuzzy sets (Ragin, 1987, 2008). We ﬁrst discuss truth table construction and analysis using a simple crisp set because this entails the same logic as used in more complex analyses. Assume there are two key potential causal attributes for high ﬁrm performance in an industry, large size and high vertical integration, and assume that we have

272

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

already calibrated the involved sets. We use these sets and the outcome to construct the relevant truth table in Table 3. Each row of this table includes one of the four logically possible combinations of the two attributes. That is, ﬁrms can be not large and not highly vertically integrated, not large and highly vertically integrated, large and not highly vertically integrated, and large and highly vertically integrated. In our example, Table 3 shows that the high performance outcome is associated with the cases falling into two of the four conﬁgurations: high performance ﬁrms are either large and not highly vertically integrated or they are large and highly vertically integrated. Applying Boolean algebra, this statement can be reduced to the simple statement that high performance ﬁrms are large, applying the logic that: A and B or A and not B ¼ A. This ﬁnding supports the notion that large size is both necessary (i.e., it must exist for high performance to occur) and sufﬁcient (i.e., it produces high performance by itself). More complex truth table analyses follow the same logic, applying Boolean algebra and logic to reduce the conﬁgurations that are linked with the outcome to a minimal Boolean statement. Evidence expressed by this statement, in combination with theoretical rationales for expected causal effects, can be used to construct arguments regarding necessity and sufﬁciency of causes (Ragin, 2000). Truth table analysis has several desirable characteristics for strategic management research. First, it unpacks the diversity of cases studied into distinct conﬁgurations, and can even include ones not present in the data. Cases differing by just one attribute belong in a different conﬁguration in the truth table, reﬂecting QCA logic that cases can be conceptualized as conﬁgurations of causal conditions, and that a single difference between cases may constitute a difference in kind. Second, expressing limited diversity through truth table analysis is important because arguments of causation – in the absence of simplifying assumptions – should be restricted to conﬁgurations that are actually represented by cases with strong membership in the data. Third, truth

Table 3. Large Size 0 1 0 1

Truth Table Combining Two Conditions. Vertically Integrated

High Performance

0 0 1 1

0 1 0 1

0, Not membership in set; 1, membership in set.

273

Qualitative Comparative Analysis and Strategic Management Research

table analysis also allows for counterfactual evaluations, which consider conﬁgurations that are plausible and substantively relevant for the outcome but that may not be empirically observed in the existing truth table. This is vital for strategic management because counterfactual analysis demands positing what-if questions, a process that has been advocated as key to analysis of causation in strategy research (Durand & Vaara, 2009). However, to keep the following truth table analysis simple, in our example we purposely avoid incorporating counterfactual evaluation. A detailed description of truth table analysis employing counterfactual evaluation can be found in Ragin (2008). To illustrate truth table analyses with fuzzy sets, we present a limited hypothetical example focusing on theoretical drivers of high ﬁrm performance in a selected manufacturing industry. Assume that based upon knowledge of the ﬁrms and industry, we expect four ﬁrm attributes to be critical for understanding high performance. These are a ﬁrm’s: size, length of time operating in the industry (i.e., age), R&D intensity, and degree of vertical integration. After data were collected, fuzzy sets were calibrated and a truth table produced using fsQCA (Ragin, Drass, & Davey, 2006). The resultant truth table appears in Table 4. Row by row, it represents the 16 logically possible combinations of the 4 conditions (16 ¼ 24) comprising

Table 4. High Age

1 0 1 0 0 0 0 0 0 0 1 1 1 1 1 1

Truth Table.

Large Size

R&D Intensity

Vertical Integration

Number of Cases

High Performance

0 1 1 0 0 1 1 0 0 1 0 1 0 0 1 1

1 0 1 1 0 1 1 0 1 0 0 0 0 1 0 1

1 0 1 1 0 0 1 1 0 1 0 0 1 0 1 0

14 8 5 4 3 3 2 1 1 1 1 1 0 0 0 0

0 1 0 0 1 1 0 0 0 1 0 1 0 0 0 0

1, Membership in set; 0, not membership in set.

274

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

the fuzzy set vector space. They have been sorted by number of cases with strong membership in each conﬁguration. In fuzzy set analysis, each case has strong membership in only one conﬁguration (row) in a truth table, determined by assigning a 1 to set memberships W0.5 and a 0 to set memberships o0.5. Additionally, it has partial membership in other conﬁgurations. Table 4 demonstrates how a truth table arrays the sample of ﬁrms in the dataset into distinct conﬁgurations. For example, the conﬁguration most frequently represented by strong members in the data involves 14 ﬁrms that are old, not large in size, have high R&D intensity, and are vertically integrated. The table further shows that four logically possible conﬁgurations are not represented among the cases. To pick one example, the last conﬁguration in the truth table indicates the sample does not include any ﬁrms that are old, large, R&D intensive, and not vertically integrated. To code the outcome in the truth table, consistency scores (further explained below) are used to decide which attribute conﬁgurations are consistently associated with the outcome. This requires researchers to select a minimum number of cases comprising the conﬁguration and a consistency score minimum for accepting that a conﬁguration is associated with the presence of the outcome. A consistency score of 1 indicates that the cases with membership in the conﬁguration consistently show high performance, whereas one close to 0 indicates that these cases consistently do not show high performance. In our example, to be consistently associated with the presence of the outcome, conﬁgurations need to fulﬁll the conditions of (a) a recommended consistency level of X0.85 (Ragin, 2008) and (b) contain at least one case with strong membership (as explained in the previous paragraph). We complete the truth table by coding high performance for the six conﬁgurations fulﬁlling these criteria as a 1.

Result Representation and Interpretation Once truth table analysis has derived the logically reduced solution, one must represent and interpret these results. Representation involves exhibiting conﬁgurations that were found, using set-theoretic reduction of the truth table, to be consistently linked to the outcome in a table. Each column in such a table represents a conﬁguration found to be consistently linked to the outcome. In this fashion, the alternative conﬁgurations associated with the outcome are arrayed (Ragin & Fiss, 2008). For purposes of representing conﬁgurations in this table, full circles (K) indicate the presence of a condition and barred circles (~) indicate their absence.

Qualitative Comparative Analysis and Strategic Management Research

275

Measures of consistency and coverage are critical for evaluating and interpreting the reduced truth table (for details, see Ragin, 2008). Recall that consistency assesses the degree that cases sharing a given conﬁguration also display the outcome. The concept of consistency is easy to convey in connection with crisp sets. Here, consistency simply describes the proportion of cases belonging to a conﬁguration of attributes associated with the outcome. For example, if 18 of 20 cases sharing the conﬁguration large size and high vertical integration also show high performance (and by implication 2 do not), the consistency of this conﬁguration would equal 0.90 (i.e., 90 percent of cases in the conﬁguration share the outcome). If only 2 out of these 20 cases show high performance, consistency would equal 0.10, and the conﬁguration is consistently associated with the absence of the outcome. By convention, the minimum consistency below which no subset relationship with the outcome can be inferred is X0.75 (Ragin, 2008). Intermediate consistency scores (ranging from about 0.30 to 0.70) indicate contradictory conﬁgurations whose cases are divided with respect to the presence or absence of the outcome. An in-depth study of these cases may allow researchers to identify additional causal attributes that resolve these contradictions if added to the truth table. With fuzzy sets, cases with strong membership in a conﬁguration are ones most relevant for providing information about consistency. Generally, the calculation of consistency scores includes substantial penalties for large inconsistencies. Determining adequate consistency is a precondition for calculating its coverage (as described below) because without it one cannot infer that a set relationship between a conﬁguration and an outcome exists in the ﬁrst place. Coverage assesses the degree to which a causal conﬁguration accounts for instances of the outcome, thus gauging its empirical importance to achieving the outcome. The concept of coverage is easy to convey in connection to crisp sets. Coverage simply captures the proportion of cases showing the outcome that fall into a given conﬁguration. This proportion is calculated by dividing the number of cases showing a speciﬁc conﬁguration by the total number of instances of the outcome. For example, if 15 of 40 cases showing superior performance share the conﬁguration large size and high vertical integration (presuming that this conﬁguration was found to be consistently linked to the outcome), the (raw) coverage of this conﬁguration is 0.375. That is, 37.5 percent of the cases showing superior performance combine large size and high vertical integration. The overall coverage of a conﬁguration that may overlap with that of others is deﬁned as its raw coverage. Occurrences of an outcome explained only by a certain conﬁguration comprise that conﬁguration’s unique coverage. If multiple conﬁgurations are consistently linked to an

276

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

outcome, raw and unique coverage provide assessments of their relative empirical importance and extent of their overlap. The combined coverage of all conﬁgurations consistently linked to the outcome is considered the solution coverage. Continuing with our hypothetical example, in Table 5 we present truth table analysis results showing three conﬁgurations in this industry are consistently linked to high performance (see Solutions 1–3). Framed in terms of a Boolean equation, the elements of each of these conﬁgurations would be linked by and, and the three conﬁgurations would be connected with each other by or. Also reported in Table 5 are consistency and coverage measures for the solution overall as well as for the individual conﬁgurations. These three conﬁgurations are equiﬁnal paths to high performance in this industry. Solution 1 shows that ﬁrms combining a lack of high R&D intensity and of vertical integration consistently are members in the set of ﬁrms with high performance. The second and third conﬁgurations demonstrate further that ﬁrms that combine large ﬁrm size and not being old with either a lack of R&D intensity or a lack of vertical integration are also consistently linked to high performance. QCA results should only be interpreted as conﬁgurations. Drawing inferences regarding independent effects of individual attributes from these results would be inappropriate. For example, referring again to Table 5, the ﬁrst conﬁguration does not imply that a lack of vertical integration is positively associated with high performance. Rather, it implies that ﬁrms combining a lack of R&D intensity with a lack of vertical integration consistently are members in the set of ﬁrms with high performance. It should also be noted that conditions that are not part of a conﬁguration have no

Table 5.

Conﬁgurations Achieving High Performance.

Conditions High age Large size R&D intensity Vertical integration Consistency Raw coverage Unique coverage Overall solution consistency Overall solution coverage

Solution 1

~ ~ 0.9 0.66 0.20

Solution 2

Solution 3

~ K ~

~ K

0.92 0.5 0.03 0.88 0.73

K, Core causal condition present; ~, core causal condition absent.

~ 0.95 0.48 0.05

277

Qualitative Comparative Analysis and Strategic Management Research

impact on high performance of this conﬁguration regardless of their presence or absence. Taking again the ﬁrst conﬁguration in Table 5 as an example, ﬁrms lacking high R&D intensity and vertical integration are linked to high performance regardless of their size or age. At the same time, the third conﬁguration links a combination of large size, lack of vertical integration, and not being old to high performance irrespective of R&D intensity. Assessments of both raw and unique coverage provide the individual solutions’ relative empirical weight. In our example, the ﬁrst conﬁguration is the empirically dominant conﬁguration, embracing about two thirds of ﬁrms with high performance (raw coverage ¼ 0.66) and uniquely embracing 20 percent of them (unique coverage ¼ 0.20). The second and third conﬁgurations have relatively high raw coverage but relatively low unique coverage, implying that they have a lot of overlap with other conﬁgurations in accounting for cases with high performance. Finally, solution consistency and coverage for the overall solution (the Boolean combination of Solutions 1–3) are 0.88 and 0.73, respectively. Thus, this combination of conﬁgurations is consistently linked to high performance, and it represents 73 percent of the cases with high performance. As noted earlier, set-theoretic results can manifest asymmetry (Fiss, 2011; Greckhamer et al., 2008; Ragin, 2008). This implies that causes for the absence of an outcome can be quite different from the inverse causes of the presence of an outcome. Therefore, in QCA causes for the presence and absence of an outcome should be analyzed separately. To illustrate this idea, we brieﬂy present the results of analyzing the conditions associated with the negation of the set of high performance (which is not the same as the set of low performance). Table 6 shows that two conﬁgurations are consistently

Table 6.

Conﬁgurations Achieving Not High Performance.

Conditions High age Large size R&D intensity Vertical integration Consistency Raw coverage Unique coverage Overall solution consistency Overall solution coverage

Solution 1

Solution 2 ~ ~ K

K K K 0.92 0.68 0.5

K, Core causal condition present; ~, core causal condition absent.

0.87 0.31 0.13 0.89 0.81

278

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

linked to the absence of high ﬁrm performance in this industry. Neither of these is the simple inverse of a high performance conﬁguration. The ﬁrst conﬁguration combines vertical integration, high R&D intensity, and old age, whereas the second conﬁguration combines high R&D intensity with not being old or large in size. These results indicate that high R&D intensity is a possible necessary condition, since it is shared by both solutions. Even so, the solutions presented in Table 6 do not encompass all cases of ﬁrms lacking high ﬁrm performance. When researchers have a theoretical basis to expect a condition to be necessary, separate tests for necessity can be conducted.

POTENTIAL CONTRIBUTIONS OF QCA TO STRATEGY RESEARCH In this section, we ﬁrst highlight likely contributions of QCA for strategy research and then use examples from core strategy research areas to illustrate how QCA could complement or redirect research. As discussed above, strategic management has a strong focus on the differential success achieved by some ﬁrms as well as processes underlying managerial interventions leading to their success. However, establishing causation in strategic management research has been problematic (Durand & Vaara, 2009; Powell, 2001). One reason for this is that traditional approaches have sought to understand causation by studying average relationships across many kinds of cases, as is done with multiple regression analyses. In contrast, QCA focuses on uncovering causal conﬁgurations consistently associated with outcomes. This enables strategy researchers to take a more conditional tack in constructing alternative empirical evidence regarding causal relationships of necessity and sufﬁciency. A second area of contribution of QCA is its strengths in studying the diversity of cases. Strategy researchers have been concerned with the issue of organizational heterogeneity, which recognizes that modern economies are populated by a mixture of many different types of ﬁrms (Carroll & Hannan, 1999). A thorough understanding of diversity is vital for strategic management to explain how certain ﬁrms can be made more successful. Moreover, in the absence of counterfactual analysis, mapping limited diversity conﬁnes arguments of causation to conﬁgurations represented by cases with strong membership in the data (Ragin, 2008).

Qualitative Comparative Analysis and Strategic Management Research

279

A third area of contribution is QCA’s facility in analyzing smaller samples of cases. Although it has been shown to be applicable in contexts using large samples (e.g., Greckhamer et al., 2008; Ragin, 2008), QCA was initially developed for rigorous analyses with medium-N samples (i.e., 10–50 cases). Many potential research settings in strategy are limited to medium-N samples. We note, however, that it is important to balance the number of included causal attributes with the number of cases in the sample (see Marx, 2010).

Potential Substantive Applications in Strategic Management A traditional area of research that would beneﬁt from QCA pertains to industry, corporate, or business-unit factors that determine ﬁrm performance. Much past research has concentrated on the relative magnitude of industry, corporate, and business-unit effects on this criterion (e.g., McGahan & Porter, 1997, 2002; Rumelt, 1991; Schmalensee, 1985). The performance effects of speciﬁc strategic factors within each of these classes have also been explored recently (e.g., Misangyi, Elms, Greckhamer, & LePine, 2006). Rather than trying to understand the relative independent contributions of various industry, corporate, and business-unit level effects to performance, a QCA approach would examine combinations of industry, corporate, and businessunit attributes that are necessary and/or sufﬁcient for superior or inferior performance across multiple levels (Greckhamer et al., 2008). Further research identifying conﬁgurations of environmental and organizational attributes linked to high performance could provide insights closer to ﬁrms’ actual operating conditions as opposed to relying on average effects of ﬁrm attributes. Executives looking to improve organizational performance could beneﬁt from understanding where their organization ﬁts in the range of organizational diversity and learning from those organizations that perform particularly well in conditions similar to their own (McKelvey, 1975, p. 223; McKelvey & Aldrich, 1983). In concert with analyses of multiple-case studies, QCA could also facilitate a more rigorous examination of determinants of ﬁrm performance for smaller samples within industries. In the domain of corporate governance, a fundamental question concerns the separation of ownership from control of the ﬁrm. Foundational concepts for this line of research are found in agency theory and the theory of managerial capitalism (Eisenhardt, 1989; McEachern, 1975; Tosi, 2008), which both posit that different kinds of control lead to differences in ﬁrm behavior. For example, strategy scholars have debated whether relationships

280

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

exist between ﬁrm diversiﬁcation and managerial motives and control (Amihud & Lev, 1981, 1999; Lane, Cannella, & Lubatkin, 1999, 1998). This debate has been important for strategy because it examines whether agency theory (and by extension managerial capitalism theory) explains ﬁrms’ strategic behaviors. Amihud and Lev (1981) studied risk reduction as motive for conglomerate mergers. They found manager-controlled ﬁrms engaged in more conglomerate acquisitions than did owner-controlled ﬁrms, and that the operations of manager-controlled ﬁrms were more diversiﬁed than the operations of owner-controlled ﬁrms. However, Lane et al. (1998), using a different measure for diversiﬁcation, did not ﬁnd support for these ﬁndings. Utilizing QCA, researchers could redirect this debate by emphasizing research questions informed by a conﬁgurational approach. Speciﬁcally, researchers could ask not merely whether management control is associated with diversiﬁcation but also explore under what combinations of conditions different forms of control lead to diversiﬁcation. Researchers could also construct outcome sets such as the set of diversiﬁed ﬁrms with high performance (i.e., intersection of membership in the sets of diversiﬁed ﬁrms and of ﬁrms with high performance) and explore the combinations of conditions leading to successful diversiﬁcation. Other issues of dependency could also shape the degree of managerial discretion (Pfeffer & Salancik, 1978). Combining theoretical lenses, conﬁgurational approaches could explore combinations of conditions associated with expressions of managerial discretion such as risk reduction through diversiﬁcation. Two concepts associated with this line of research, ﬁrm control and relatedness of diversiﬁcation, are operationalized in ways that lend themselves particularly well to set-theoretic approaches. Researchers have generally resorted to categorizing ﬁrms into manager- versus owner-controlled ﬁrms, and may further distinguish them as owner-managed. Studies often use a ﬁve percent equity threshold held by a single equity holder to make these distinctions (e.g., Gomez-Mejia, Tosi, & Hinkin, 1987; Tosi & Gomez-Mejia, 1989). The manager–owner distinction can be represented using crisp sets indicating differences in kind: the sets of manager-controlled, ownercontrolled, and owner-managed ﬁrms. Fuzzy sets could be employed when needed to incorporate partial membership in these sets, thereby capturing gradual increases in control levels with increasing stockholder concentration. Indeed, Amihud and Lev (1981), based on the percentage of shares held by one owner, included a category ‘‘weak owner-control’’ that in fuzzy sets could constitute partial membership in the set of owner-controlled ﬁrms. Also, QCA could be applied to study whether different causal combinations may lead to

Qualitative Comparative Analysis and Strategic Management Research

281

diversiﬁcation in owner-controlled and manager-controlled ﬁrms as well as to the success or failure of these efforts. Additionally, QCA could open up alternative ways of capturing the relatedness and unrelatedness of diversiﬁcation efforts. Building on Rumelt’s (1974) classiﬁcation of various forms of related and unrelated diversiﬁcation, a commonly used measure of diversiﬁcation in strategy, researchers could calibrate fuzzy sets capturing the degree of relatedness and/or unrelatedness regarding instances of diversiﬁcation. Previous research has also used qualitative relatedness assessments, categorizing businesses as related or unrelated based on their similarity with a ﬁrm’s traditional domain (e.g., Haveman, 1992). Fuzzy sets could be used to enhance such qualitative assessments by enabling partial membership based on the number of domain dimensions that are evaluated as similar or dissimilar as well as on judgments regarding the relative importance of the individual domains. Relatedly, QCA could be applied to further investigations of diversiﬁcation through joint ventures. For example, previous research has found that in the context of international joint ventures (IJV), product relatedness of an IJV with both parents was positively associated with the joint venture’s performance. This relationship was found to be moderated by dynamics of the IJV’s organizational form and dynamics of the market environment (Luo, 2002). A fuzzy set approach to studying this issue could capture the relatedness of an IJV’s products to all parents through a set summarizing the individual relatedness of the IJV with all partners’ products. Researchers could then examine the conﬁgurations of relatedness, organizational characteristics, and environmental characteristics that lead to successful diversiﬁcation through IJVs. Research in the area of transaction cost economics (TCE) could also beneﬁt from applications of QCA. TCE has become a leading perspective in studying strategy and organizations (David & Han, 2004; Williamson, 1991). The unit of analysis in TCE is the transaction. Among principal attributes of transactions are locational, human, and physical asset speciﬁcity, transaction frequency, and uncertainty (Williamson, 1981, 1991). The outcome in the TCE model is governance form chosen, including market, hybrid, and hierarchy. Speciﬁc characteristics of TCE lend themselves to QCA’s set-theoretic approach. First, the theory speciﬁes prototypes of governance, market, or hierarchy forms representing differences in kind, with hybrid forms as intermediate governance arrangements between ‘‘make’’ and ‘‘buy’’ decisions. Second, TCE anchors transaction characteristics in dichotomous prototypes. For example, it describes human asset speciﬁcity as either high or low, and the evaluation of human asset productivity as easy or difﬁcult (Williamson, 1981).

282

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

Finally, characteristics of transactions are not independent but rather complementary in nature. For example, David and Han (2004) pointed out that the impact of uncertainty on transaction costs is conditional upon the degree of asset speciﬁcity. A QCA approach could capture attributes of transactions by deﬁning appropriate sets, such as sets of transactions with (1) high (low) uncertainty; (2) high (low) frequency; (3) high (low) site speciﬁcity; (4) high (low) physical asset speciﬁcity; and (5) high (low) human asset speciﬁcity. Similarly, the governance form chosen for transactions could be cast into sets. For example, the set of market-governed transactions would give full membership to transactions governed by classical contract law, full nonmembership to hierarchical or internally organized transactions, and partial membership to various hybrid governance forms, depending on their respective similarity to market or hierarchy forms. Delineating different combinations of transaction characteristics might clarify some of the mixed empirical support obtained over the years for TCE predictions (David & Han, 2004). A ﬁnal example of QCA’s utility lies in the area of supply chain management. Supply chain relationships enhance variables such as order fulﬁllment cycle, quality, or delivery dependability, that drive ﬁrm and supply chain performance (Hult, Ketchen, & Arrfelt, 2007; Hult, Ketchen, & Slater, 2004; Jin, Hopkins, & Wittmer, 2010). Benseau and Venkatraman (1995) advocated a conﬁgurational approach to identify the limited diversity of empirically occurring as well as equally effective conﬁgurations of interorganizational relationships. A conﬁgurational approach is important because key management and environmental attributes interact with one another to create effective interorganizational relationships in supply chains. With QCA, strategy researchers could examine performance outcomes of individual ﬁrms within a supply chain or those of entire supply chains. Researchers could identify sets of attributes expected to be causally related to salient outcomes, such as sets capturing dimensions of environmental uncertainty including market growth, turbulence, and product complexity and customization (Benseau & Venkatraman, 1995; Hult et al., 2007). Other supply chain attributes of potential interest would be the culture of competitiveness (Hult, Ketchen, & Nichols, 2002), level of mutual trust (Benseau & Venkatraman, 1995), or ﬁrms’ human capital at managerial and workers’ levels (Jin et al., 2010). Rather than trying to parse out the independent effects of these attributes, QCA would enable researchers to study the complex conﬁgurations in which these attributes combine to affect supply chain performance outcomes.

Qualitative Comparative Analysis and Strategic Management Research

283

CONCLUSION The purpose of our chapter was to discuss QCA as a relatively novel alternative approach to address research questions in strategy, as well as to take stock of the current state and future prospects for QCA in strategy research. We have highlighted QCA’s potential to inform and potentially redirect central areas of strategy research, thereby contributing to the development of the discipline, its theories, and empirical accomplishments. Strategic management is a relatively young ﬁeld within the administrative sciences (Hambrick & Chen, 2008). Research methods play a vital role in a ﬁeld’s advancement by providing ways to empirically support, question, or advance theories. Accordingly, research methods have contributed substantially to the development of the strategy ﬁeld’s domain and its theories (Ketchen, Boyd, & Bergh, 2008). QCA is a valuable addition to the methodological tool kit of strategic management researchers. Its unique features enhance access to the theoretical and methodological pluralism in strategy research and hold vital potential for contributing to theories in the ﬁeld. QCA is not inherently superior to conventional linear analyses, and as with any other method, it is not a panacea for all research questions. However, it may open up empirical contexts involving medium-N samples that are too small for conventional statistical analyses but too large for indepth qualitative studies to rigorous inquiry. It also offers strategic management researchers a valuable alternative in contexts where causal forces may operate in conﬁgurations, thus offering new insights on pressing issues in the ﬁeld. In such instances, QCA’s focus on multiple conjunctive causality and its deduction of necessary and sufﬁcient causes could uncover causal connections with outcomes heretofore unnoticed by strategy scholars. An important part of this focus is using truth table analysis to map the logically possible and empirically observed diversity in cases, in addition to identifying attribute conﬁgurations that are linked with focal outcomes. Our selective review of the literature suggested that QCA’s potential is starting to be tapped, as exempliﬁed by studies on organizational conﬁgurations (Fiss, 2011), the causes for the relative attractiveness of countries for FDIs (Pajunen, 2008), and the determinants of ﬁrm performance (Greckhamer et al., 2008). As we have highlighted in this chapter, a key for choosing QCA as research strategy is whether a researcher is interested in isolating independent effects or in studying combinations of effects. When the latter is the case, QCA may be preferred as it emphasizes interdependencies and causal complexity among causal factors. These emphases could uncover causal

284

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

relationship patterns among case phenomena that strategy scholars study, including ﬁrms, joint ventures, supply chains, or transactions. Although only scratching the surface in presenting this conceptual overview, we hope to have conveyed QCA’s potential value as an alternative mode of inquiry and provided an impetus for using it to pursue new insights in the strategic management domain.

ACKNOWLEDGMENTS We wish to thank Tim Chandler, Sebnem Cilesiz, and Brian Connelly for their insightful comments on an earlier draft of this chapter.

REFERENCES Amenta, E., Caren, N., & Olasky, S. J. (2005). Age for leisure? Political mediation and the impact of the pension movement on US old-age policy. American Sociological Review, 70, 516–538. Amenta, E., Carruthers, B. G., & Zylan, Y. (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, 308–339. Amenta, E., & Halfmann, D. (2000). Wage wars: Institutional politics, WPA wages, and the struggle for U.S. social policy. American Sociological Review, 67, 506–528. Amenta, E., & Poulson, J. (1996). Social politics in context: The institutional politics theory and social spending at the ending of the New Deal. Social Forces, 75, 33–60. Amihud, Y., & Lev, B. (1981). Risk reduction as a managerial motive for conglomerate mergers. The Bell Journal of Economics, 12, 605–617. Amihud, Y., & Lev, B. (1999). Does corporate ownership structure affect its strategy towards diversiﬁcation? Strategic Management Journal, 20, 1063–1069. Becker, H. (1992). Cases, causes, conjunctures, stories, and imagery. In: C. Ragin & H. Becker (Eds), What is a case? Exploring the foundations of social inquiry (pp. 205–216). Cambridge: Cambridge University Press. Benseau, M., & Venkatraman, N. (1995). Conﬁgurations of interorganizational relationships: A comparison between U.S. and Japanese automakers. Management Science, 41, 1471–1492. Bowman, E. H., & Helfat, C. E. (2001). Does corporate strategy matter? Strategic Management Journal, 22, 1–23. Bowman, E., Singh, H., & Thomas, H. (2002). The domain of strategic management: History and evolution. In: A. Pettigrew, H. Thomas, & R. Whittington (Eds), Handbook of strategy and management (pp. 31–51). London: Sage. Bromiley, P. (2005). The behavioral foundations of strategic management. Wiley-Blackwell.

Qualitative Comparative Analysis and Strategic Management Research

285

Brown, C., & Boswell, T. (1995). Strikebreaking or solidarity in the great steel strike of 1919 – A split labor-market, game-theoretic, and QCA analysis. American Journal of Sociology, 100, 1479–1519. Burns, T., & Stalker, G. M. (1961). The management of innovation. London: Tavistock Publications. Caramani, D. (2009). Introduction to the comparative method with Boolean algebra. Thousand Oaks, CA: Sage. Carroll, G. R. (1993). A sociological view on why ﬁrms differ. Strategic Management Journal, 14, 237–249. Carroll, G. R., & Hannan, M. T. (1999). The demography of corporations and industries. Princeton, NJ: Princeton University Press. Cress, D. M., & Snow, D. (1996). Mobilization at the margins: Resources, benefactors, and the viability of homeless social movement organizations. American Sociological Review, 61, 1089–1109. David, R. J., & Han, S.-K. (2004). A systematic assessment of the empirical support for transaction cost economics. Strategic Management Journal, 25, 39–58. Durand, R., & Vaara, E. (2009). Causation, counterfactuals, and competitive advantage. Strategic Management Journal, 30, 1245–1264. Eisenhardt, K. M. (1989). Agency theory: An assessment and review. Academy of Management Review, 14, 57–74. Fiss, P. (2007). Towards a set-theoretic approach for studying organizational conﬁgurations. Academy of Management Review, 32, 1180–1198. Fiss, P. (2011). Building better causal theories: A fuzzy set approach to typologies in organization research. Academy of Management Journal, 54. Fortune. (2010): Fortune 500: Our annual ranking of America’s largest corporations. Available at http://money.cnn.com/magazines/fortune/fortune500/2010/full_list/. Retrieved on October 2010. Gomez-Mejia, L. R., Tosi, H. L., & Hinkin, T. (1987). Managerial control, performance, and executive compensation. Academy of Management Journal, 30, 51–70. Greckhamer, T. (2011). Cross-cultural differences in compensation level and inequality across occupations: A set-theoretic analysis. Organization Studies, 32, 85–115. Greckhamer, T., Misangyi, V. F., Elms, H., & Lacey, R. (2008). Using qualitative comparative analysis in strategic management research: An examination of combinations of industry, corporate, and business-unit effects. Organizational Research Methods, 11, 695–726. Hambrick, D., & Chen, M.-J. (2008). New academic ﬁelds as admittance-seeking social movements: The case of strategic management. Academy of Management Review, 33, 32–54. Haveman, H. A. (1992). Between a rock and a hard place: Organizational change and performance under conditions of fundamental environmental transformation. Administrative Science Quarterly, 37, 48–75. Hollingsworth, R., Hanneman, R., Hage, J., & Ragin, C. (1996). The effect of human capital and state intervention on the performance of medical systems. Social Forces, 75, 459–484. Hult, T. M., Ketchen, D. J., & Arrfelt, M. (2007). Strategic supply chain management: Improving performance through a culture of competitiveness and knowledge development. Strategic Management Journal, 28, 1035–1052. Hult, T. M., Ketchen, D. J., & Nichols, E. L. (2002). An examination of cultural competitiveness and order fulﬁllment cycle within supply chains. Academy of Management Journal, 45, 577–586.

286

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

Hult, T. M., Ketchen, D. J., & Slater, S. F. (2004). Information processing, knowledge development, and strategic supply chain performance. Academy of Management Journal, 47, 241–253. Jin, Y., Hopkins, M. M., & Wittmer, J. L. S. (2010). Linking human capital to competitive advantages: Flexibility in a manufacturing ﬁrm’s supply chain. Human Resource Management, 49, 939–963. Ketchen, D. J., Boyd, B. K., & Bergh, D. D. (2008). Research methods in strategic management: Past accomplishments and future challenges. Organizational Research Methods, 11, 643–658. Ketchen, D. J., Combs, J. G., Russell, C. J., Shook, C., Dean, M., Runge, J., et al. (1997). Organizational conﬁgurations and performance: A meta-analysis. Academy of Management Journal, 40, 223–240. Kogut, B., MacDufﬁe, J. P., & Ragin, C. (2004). Prototypes and strategy: Assigning causal credit using fuzzy sets. European Management Review, 1, 114–131. Lam, W. F., & Ostrom, E. (2010). Analyzing the dynamic complexity of development interventions: Lessons from an irrigation experiment in Nepal. Policy Sciences, 43, 1–25. Lane, P. J., Cannella, A. A., & Lubatkin, M. H. (1998). Agency problems as antecedents to unrelated mergers and diversiﬁcation: Amuhud and Lev reconsidered. Strategic Management Journal, 19, 555–578. Lane, P. J., Cannella, A. A., Jr., & Lubatkin, M. H. (1999). Ownership structure and corporate strategy: One question viewed from two different worlds. Strategic Management Journal, 20, 1077–1086. Luo, Y. (2002). Product diversiﬁcation in international joint ventures: Performance implications in an emerging market. Strategic Management Journal, 23, 1–20. Mahoney, J., & Goertz, G. (2006). A tale of two cultures: Contrasting quantitative and qualitative research. Political Analysis, 14, 227–249. Marx, A. (2010). Crisp-set qualitative comparative analysis (csQCA) and model speciﬁcation: Benchmarks for future csQCA applications. International Journal of Multiple Research Approaches, 4, 138–158. McEachern, W. A. (1975). Managerial control and performance. Lexington, MA: Heath & Co. McGahan, A., & Porter, M. (1997). How much does industry matter, really? Strategic Management Journal, 18, 15–30. McGahan, A., & Porter, M. (2002). What do we know about variance in accounting proﬁtability? Management Science, 48, 834–851. McKelvey, B. (1975). Guidelines for the empirical classiﬁcation of organizations. Administrative Science Quarterly, 20, 509–525. McKelvey, B., & Aldrich, H. (1983). Population, natural selection, and applied organizational science. Administrative Science Quarterly, 28, 101–128. Mendelson, E. (1970). Schaum’s outline of theory and problems of Boolean algebra and switching circuits. New York: McGraw-Hill. Meyer, A. D., Tsui, A. S., & Hinings, C. R. (1993). Conﬁgurational approaches to organizational analysis. Academy of Management Journal, 36, 1175–1195. Miles, R. E., & Snow, C. C. (1978). Organizational strategy, structure, and process. In: collaboration with A. D. Meyer and with contributions by H. J. Coleman, Jr. New York: McGraw Hill.

Qualitative Comparative Analysis and Strategic Management Research

287

Miller, D. (1986). Conﬁgurations of strategy and structure: Towards a synthesis. Strategic Management Journal, 7, 233–249. Miller, D., & Friesen, P. H. (1984). Organizations: A quantum view. Englewood Cliffs, NJ: Prentice Hall. Misangyi, V. F., Elms, H., Greckhamer, T., & LePine, J. (2006). A new perspective on a fundamental debate: A multi-level approach to industry, corporate, and business-unit effects. Strategic Management Journal, 27, 571–590. Newman, I., & Benz, C. R. (1998). Qualitative-quantitative research methodology: Exploring the interactive continuum. Carbondale, IL: Southern Illinois University. O¨z, O¨. (2004). Using Boolean- and fuzzy-logic-based methods to analyze multiple case study evidence in management research. Journal of Management Inquiry, 13, 166–179. Pajunen, K. (2008). Institutions and inﬂows of foreign direct investment: A fuzzy-set analysis. Journal of International Business Studies, 39, 652–669. Pfeffer, J., & Salancik, G. R. (1978). The external control of organizations: A resource dependence perspective. New York: Harper & Row. Porter, M. E. (1990). The competitive advantage of nations. New York: Free Press. Powell, T. C. (2001). Competitive advantage: Logical and philosophical considerations. Strategic Management Journal, 22, 875–888. Ragin, C. (1987). The comparative method: Moving beyond qualitative and quantitative strategies. Berkeley, CA: University of California Press. Ragin, C. (1994). Constructing social research. Thousand Oaks, CA: Pine Forge Press. Ragin, C. (2000). Fuzzy-set social science. Chicago: University of Chicago Press. Ragin, C. (2008). Redesigning social inquiry: Fuzzy sets and beyond. Chicago: University of Chicago Press. Ragin, C., Drass, K. A., & Davey, S. (2006). Fuzzy-set/qualitative comparative analysis 2.0. Tucson, AZ: University of Arizona. Ragin, C., & Fiss, P. (2008). Net effects versus conﬁgurations: An empirical demonstration. In: C. Ragin (Ed.), Redesigning social inquiry: Fuzzy sets and beyond (pp. 190–212). Chicago: University of Chicago Press. Redding, K., & Viterna, J. S. (1999). Political demands, political opportunities: Explaining the differential success of left-libertarian parties. Social Forces, 78, 491–510. Ridder, H.-G., Hoon, C., & McCandless, A. (2009). The theoretical contribution of case study research to the ﬁeld of strategy and management. In: D. D. Bergh & D. J. Ketchen (Eds), Research methodology in strategy and management (Vol. 5, pp. 137–175). Bingley, UK: Emerald. Roscigno, V. J., & Hodson, R. (2004). The organizational and social foundations of worker resistance. American Sociological Review, 69, 14–39. Rumelt, R. P. (1974). Strategy, structure, and economic performance. Boston: Harvard University Press. Rumelt, R. P. (1991). How much does industry matter? Strategic Management Journal, 12, 167–185. Rumelt, R. P., Schendel, D. E., & Teece, D. J. (Eds). (1994). Fundamental issues in strategy. Boston: Harvard Business School Press. Schmalensee, R. (1985). Do markets differ much? The American Economic Review, 75, 341–351. Schneider, S., & de Meyer, A. (1991). Interpreting and responding to strategic issues: The impact of national culture. Strategic Management Journal, 12, 307–320.

288

THOMAS GRECKHAMER AND KEVIN W. MOSSHOLDER

Schneider, M. R., Schulze-Bentrop, C., & Paunescu, M. (2010). Mapping the institutional capital of high-tech ﬁrms: A fuzzy-set analysis of capitalist variety and export performance. Journal of International Business Studies, 41, 246–266. Scott, R. W. (1998). Organizations: Rational, natural, and open systems (4th ed.). Upper Saddle River, NJ: Prentice Hall. Smithson, M. (1987). Fuzzy set analysis for behavioral and social sciences. New York: Springer. Smithson, M., & Verkuilen, J. (2006). Fuzzy set theory: Applications in the social sciences. Thousand Oaks, CA: Sage. Stokke, O. S. (2007). Qualitative comparative analysis, shaming, and international regime effectiveness. Journal of Business Research, 60, 501–511. Tosi, H. L. (2008). Quo vadis? Suggestions for future corporate governance research. Journal of Managerial Governance, 12, 153–169. Tosi, H. L., & Gomez-Mejia, L. R. (1989). The decoupling of CEO pay and performance: An agency theory perspective. Administrative Science Quarterly, 34, 169–189. Varone, F., Rothmayr, C., & Montpetit, E. (2006). Regulating biomedicine in Europe and North America: A qualitative comprative analysis. European Journal of Political Research, 45, 317–343. Verkuilen, J. (2005). Assigning membership in fuzzy set analysis. Sociological Methods and Research, 33, 462–496. Veugelers, J., & Magnan, A. (2005). Conditions of far-right strength in contemporary Western Europe: An application of Kitschelt’s theory. European Journal of Political Research, 44, 837–860. Whitesitt, J. E. (1995). Boolean algebra and its applications. New York: Dover. Williamson, O. (1981). The economics of organization: The transaction cost approach. The American Journal of Sociology, 87, 548–577. Williamson, O. (1991). Strategizing, economizing, and economic organization. Strategic Management Journal, 12, 75–94.

WORTH A THOUSAND WORDS: PHOTOGRAPHS AS A NOVEL METHODOLOGICAL TOOL IN STRATEGIC MANAGEMENT Joshua L. Ray and Anne D. Smith ABSTRACT Purpose – The purpose of this chapter is to review and categorize how photographs have been used in management research and to provide strategic management researchers with suggestions about how to use photographs to enhance their qualitative research methodologies. Methodology/approach – We develop a typology of photographic uses in management research by reviewing several scholarly journals. Findings – We identify two dimensions that differentiate how photographs have been used in management journals. First, photographs can be used to illustrate scenes from a ﬁeld setting or they can be interpreted as data. Second, the role of ﬁeld participants can be one of active collaboration or no involvement in the photographic aspect of the qualitative research project. For instance, ﬁeld subjects can collaborate in research by aiding in the photo-documentation process and/or aiding in the photo-elicitation process. Choosing which of our four identiﬁed photographic approaches represents a

Building Methodological Bridges Research Methodology in Strategy and Management, Volume 6, 289–326 Copyright r 2011 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-8387/doi:10.1108/S1479-8387(2011)0000006013

289

290

JOSHUA L. RAY AND ANNE D. SMITH

critical decision for qualitative researchers interested in incorporating photographs in their research. Practical implications – We suggest ideas for strategic management researchers related to use of photographs in their research. Also, we describe how speciﬁc strategic management research projects can be approached with photography, which we argue can lead to enhanced theoretical contributions. Originality/value of paper – To date, little has been written in the strategic management ﬁeld about the use of photography. This chapter provides a succinct review of photographic methods in management research. Moreover, this chapter provides suggestions for how strategy researchers, study participants, and interested readers of management research could beneﬁt from incorporating photographs into research accounts. Keywords: general review; photographs; photo-elicitation; qualitative methods The strategic management discipline is relatively young as a ﬁeld of academic study (Nag, Hambrick, & Chen, 2007; Pfeffer, 1993), especially as compared to other social sciences such as economics, anthropology, psychology, and sociology. The strategy ﬁeld has incorporated theories and borrowed research methodologies from disciplines in the hard and social sciences (McKiernan, 1996), but traditionally, research in strategic management has been dominated by quantitative approaches (Ketchen, Boyd, & Bergh, 2008). More recently, however, a growing body of qualitative studies (Bartunek, Rynes, & Ireland, 2006) and growing interest in content analysis indicate that strategy researchers are becoming interested in other research methods (Short, Broberg, Cogliser, & Brigham, 2010). Interestingly, despite the cross-fertilization of methodological approaches, an inclusion of visuals – speciﬁcally photographs in qualitative research – has had minimal use. The paucity of research using visual imagery is surprising because other disciplines have leveraged photographs for some time. For example, the sociology discipline has at least one peer-reviewed academic journal – Visual Studies – in which photographs play a prominent role. Likewise, the anthropology discipline has a long tradition of using photographs in ﬁeld studies (Bateson & Mead, 1942; Stasz, 1979). To wit, we are surprised that photographs have yet to be incorporated as a salient dimension of qualitative research in strategic management.

Photographs as a Novel Methodological Tool in Strategic Management

291

In this chapter, we begin by reviewing key features of the use of photographs from sociology and anthropology. Next, we identify recent studies in management and organization journals that have utilized photographs – a process we compare to ﬁnding needles in a haystack. Based on this literature review, we identify two discriminating dimensions in the way that photographs have been used. We then develop four approaches for using photographs in strategic management research. We suggest ways of introducing photographic methods in management research and argue that incorporating this methodological tool offers researchers an array of beneﬁts. Finally, we provide examples of photographic use from one of our research projects.

PHOTOGRAPHIC ELEMENTS IN SOCIOLOGY AND ANTHROPOLOGY RESEARCH In the sociological and anthropological disciplines, photographic research methods involve two processes – photo-documentation and photo-elicitation (Buchanan, 2001; Parker, 2009; Warren, 2002). Photo-documentation refers to the collection of photographs as a means of recording aspects of the research setting (Banks, 2007; Collier & Collier, 1986; Harper, 1988). This aspect of photographic research is not limited to any particular element of the research setting and can involve multiple aspects of the research setting including individuals, physical setting, and changes in the physical setting over time. Within the photo-documentation process, we have identiﬁed two general means of collecting photographs: primary and secondary collection. Primary collection of photographs involves either the researcher(s) or study participants taking photographs at the time of the study in the research setting. Secondary collection involves scanning archival resources (e.g., organizational documents, popular press accounts, and historical documents) for photographs produced by individuals not associated with the research project. Each collection technique has associated costs and beneﬁts that should be considered with regard to the research questions to be answered. For example, primary collection provides the researcher with considerable control over which photographic data are obtained because the researcher can direct what is photographed. In addition, photographs are able to capture details that a fatigued ﬁeld researcher may miss, thus facilitating subsequent analysis. As a ﬁnal example, researchers engaged in primary collection have insight into why a particular photograph was framed as it was or what it was

292

JOSHUA L. RAY AND ANNE D. SMITH

intended to capture, which is not always the case when relying on secondary sources (Parker, 2009). However, data collection of this type can involve considerable expense, such as the ﬁnancial cost of cameras, developing the photographs, and software needed for viewing and editing the photographs. These economic costs have rapidly decreased but the method still requires a substantial investment of the researcher’s time to take photos and organize digital photographic ﬁles. Furthermore, special permissions may be required for using photo-documentation and particular care must be taken when collecting photographs of human subjects. Secondary collection avoids some of the expense associated with personally taking photographs, and this approach can facilitate longitudinal study if photographs are available. Researchers can scan press releases, annual reports, and prior research for photographs related to the themes of their study. The downside of secondary collection is that a researcher is dependent on the photographs that are available; the researcher loses control over what data are collected and how the data are generated. As well, the researcher may need to negotiate use or reproduction rights with the photograph’s owner. Finally, as mentioned above, any researcher using secondary sources has less insight into what, how, and why particular elements were photographed. As some photographs can be highly staged – or manipulated – care must be taken to ensure the legitimacy of the photographic content (Guthey & Jackson, 2005). The second process used by anthropologists and sociologists related to photography in research is photo-elicitation (Collier & Collier, 1986). This involves using photographs for their representative value as objects of interpretation (Banks, 2007; Buchanan, 2001; Harper, 1998; Wagner, 1979; Winston, 1998). Photo-elicitation is the process by which photographs are viewed by the researcher and/or study participant in order to prompt memories or interpretations including values, beliefs, and meanings (Parker, 2009). Hence, photo-elicitation is less interested in the physical elements captured in the photograph and more interested in how beliefs and intentions are manifested in the photograph and how the photographs can elicit meanings during subsequent discussions with research subjects, as well as why the photographers chose to take the picture. Photo-elicitation can take many forms and involves both primary and secondary sources (see Parker, 2009 for a review). As with any research process, photo-elicitation has strengths and weaknesses. The obvious strength of this approach is that it can provoke discussion of research subjects’ tacit knowledge. Another beneﬁt is the information gathered during interviews in which photographs are the focal objects of discussion. This approach relieves

Photographs as a Novel Methodological Tool in Strategic Management

293

some stress on research participants by relocating the focus from the individual to the photograph (Warren, 2002). Warren (2002) also points out a signiﬁcant limitation to this process in that ‘‘interviewing with images assumes either that the meaning is wholly contained within the image, with the respondent being required to extract it; or that the photograph is only a prompt, eliciting comment ‘contained within’ the respondent’’ (p. 239). She suggests that a more realistic understanding of photo-elicitation acknowledges that any meaning extracted is a ‘‘joint effort’’ of both the photograph’s meaning and the participant’s interpretation. Given the rich elaboration of photographic use within the sociology and anthropology traditions and the strategic management ﬁeld’s close connections to these disciplines, we expected to ﬁnd some use of photographs in strategic management research, but we were disappointed as we describe below.

PHOTOGRAPHS IN MANAGEMENT RESEARCH: FINDING NEEDLES IN A HAYSTACK In order to identify the use of photography in management research, and because we were particularly interested in photographs used in empirical research on business organizations, we looked for articles that contained photographs in the research process description or in the published paper. We chose the period of 1990–2010 to parallel the growth of qualitative methods (Aguinis, Pierce, Bosco, & Muslin (2009)) and affordable digital photography. We began by searching for photographs (i.e., the truncated search term ‘‘photo’’ in article text) in management journals that publish empirical research. We conducted broad searches on the term ‘‘photo’’ in the text of an article and a journal name with ‘‘business,’’ ‘‘manage,’’ ‘‘organiz,’’ ‘‘administrat,’’ ‘‘strateg,’’ ‘‘entrepren,’’ or ‘‘plan.’’ Using the EBSCO Business Source Premier database, we reviewed these search terms for our speciﬁed period in many management journals, including: Academy of Management Journal, Administrative Science Quarterly, Strategic Management Journal, Organization Science, Journal of Management Studies, Organization Studies, British Journal of Management, Journal of Management, Journal of Management Inquiry, Journal of Management Studies, Journal of International Business Studies, Entrepreneurship Theory and Practice, Journal of Business Venturing, Journal of Small Business Management, Journal of Business Research, Management Science, Journal of Organizational Behavior, Organizational Dynamics, Long Range Planning, Business Strategy Journal, Scandinavian Journal of Management, Organization, Strategy Organization, and Qualitative Research in Organizations and Management.

294

JOSHUA L. RAY AND ANNE D. SMITH

We also looked for photographs in the Advances in Strategic Management series by undertaking a page-by-page review of all 21 issues published since this series began. We initially identiﬁed over 50 articles from this effort. We then removed articles that picked up on articles on the photographic industry as well as any articles that were simple photograph displays. For instance, the Journal of Management Inquiry published a series of photographs between 2001 and 2002 that were taken by management authors (e.g., Adler, 2001). Finally, we removed other articles that included stock photographs of advertising (Munir & Phillips, 2005) and articles in which no methods were identiﬁed (e.g., Rippin, 2006). Several academic friends who were aware of our project suggested other sources, such as book chapters, that we may have missed (Dougherty & Kunda, 1990; Molloy & Whittington, 2005). What remained from our search was a handful of articles that described or contained photographs used in an empirical research project. Some of these articles focused speciﬁcally on the methodology of using photographs (i.e., Guthey & Jackson, 2005; Warren, 2008), which led us to original research in online organization journals (Warren, 2002). From these efforts, we identiﬁed articles that used photographs in the research or publication of empirical research. We refer to these articles as ‘‘needles in a haystack’’ because our ﬁnal list contains only 10 articles: Barry and Rerup (2006), Buchanan (2001), Dougherty and Kunda (1990), Guthey and Jackson (2005), Molloy and Whittington (2005), Smith (2002), Sood and Pattinson (2006), Venkatraman and Nelson (2008), Warren (2002), and Widener (2007). In addition, we reviewed four individual studies that used photographs and were summarized in a book chapter (Felstead, Jewson, & Walters, 2004). We recognize that we may have missed some articles, but our goal was to ﬁnd a representative group of articles to show the different uses of photographs in management and organizational research. After a close reading of the articles we had collected, we began to identify dimensions that differentiated among our group of articles. Both authors came up with a list of discriminating features, many of which overlapped. We honed in on two ways in which the photographs were used – as illustration for the reader or as a component of the data collection used for analysis. We also identiﬁed the degree of collaboration with ﬁeld participants – close or arms length, if at all. With these two dimensions – collaboration and use of photographs in research – we constructed a table in which we could ‘‘see’’ how photographs had been used. From this table, we identify how topics in strategic management might be studied or augmented using photographic approaches in empirical research.

Photographs as a Novel Methodological Tool in Strategic Management

295

TYPOLOGY OF PHOTOGRAPHIC USE IN MANAGEMENT RESEARCH Our identiﬁcation of the two discriminating dimensions allowed us to categorize all of the articles we found. From this categorization, we examined common elements across the research methodologies of the studies located within a particular cell. Below we describe the facets of our typology in more detail and identify four approaches to using photographs in management research. The description of these combinations is followed by descriptions of the empirical studies from which they were derived. Illustration vs. Interpretation Our literature review suggests that organizational researchers who employ photographs as part of their research methodology leverage the photographs in one of two ways. First, some researchers use photographs as a means of observation through which they hope to capture some aspect of organizational reality. This type of photo-documentation typically involves taking pictures of elements in the organizational environment with the intent to use pictures to illustrate organizational facets per se. In this way, photographs can be used to depict internal organizational elements such as work space conﬁgurations and environmental conditions as well as organizational functions such as tracking product movements along an assembly line or document ﬂows related to administrative activities. The essential element in this type of photo-documentation is the researcher’s intent to capture some characteristic of the organization in as objective a fashion as possible. These photographs are then leveraged to provide illustrations that can augment narrative accounts or used to bolster research materials (e.g., ﬁeld notes). The second way photographs are leveraged in organizational research is as a vehicle for capturing meaning. This type of photo-documentation is less concerned about the nominal elements of the photograph itself (e.g., the juxtaposition of desks in an ofﬁce space) and more concerned with the interpretations of the elements of a photograph that lead to insights about the organizational meanings attributed to the elements. In this way, photographs provide insights into organizational processes such as identity formation, sensemaking, and decision-making. This use of photographs can assess organizational phenomena such as climate, culture, and organizational identity. For this type of photo-documentation, the intent of the

296

JOSHUA L. RAY AND ANNE D. SMITH

researcher is to use the photographs to prompt the researcher and/ or research subjects to elaborate on organizational interpretations and beliefs.

No Collaboration vs. Collaboration Researchers differ in how they approach collaboration with study participants. Some researchers do not involve research participants (i.e., organizational members) at all. This type of research typically takes the form of archival analysis wherein the researcher can pull photographs from various media (e.g., popular press accounts and annual reports) and analyze them without any direct contact with organizational members. Alternatively, the researcher can also choose to take photographs without soliciting input from participants, thus negating the need for collaboration. In fact, in some instances, the researcher who plays the role of the photographer can provide a buffer between themselves and the study participants by emphasizing their role as an outside observer (e.g., Widener, 2007). Finally, for some researchers, collaboration is not an issue because they choose to leverage the use of photographs as a reﬂexive activity of their own work (e.g., Smith, 2002). In this situation, the researchers can use photographs to augment their ﬁeld notes, to prompt their recall, or to provide illustration for the research project. When researchers choose to collaborate with participants, they have several options. Some research designs involve participants taking photographs (i.e., photo-documentation) as a means of demonstrating what the participants ﬁnd particularly relevant in the research situation. This is generally accomplished in two ways. First, the researcher can allow the participants to take photographs of anything as long as it relates to the research project in some way. This allows the participant to dictate the content of the photographs and provides a means through which the researcher can identify important organizational aspects without overtly inﬂuencing the participant. Moreover, allowing participants free reign to photograph whatever they choose also taps into the photographer’s intentions, motives, beliefs, etc., in a way that interviewing alone cannot (Warren, 2002). An alternative means of collaboration involves providing guidelines for the participants to follow. The degree to which the researcher chooses to provide criteria for what should be photographed can focus the data collection and help guide the participant toward organizational elements of interest. In this

Photographs as a Novel Methodological Tool in Strategic Management

297

situation, the participant still has some ﬂexibility in regard to which speciﬁc elements they photograph and how they photograph them, but the researcher exerts a greater degree of control. The context of organizational life is vast; without direction, participants could provide a range of photographic material that would be more or less related to the research questions of interest. Although we discuss these approaches as two different options, they are probably best conceived as two ends of a continuum. On one end of the spectrum, the participant would have almost complete autonomy in what content they capture and how they choose to capture it, while on the other end the researcher could provide an extensive set of instructions that limit the content to be captured. Moving from one extreme to the other involves one party or the other losing a measure of control. Decisions about the amount of control that either party enjoys should be based on the research questions to be addressed. The other aspect of collaboration between researcher and participant involves methods of photo-elicitation (Buchanan, 2001; Parker, 2009). As mentioned, photo-elicitation is the process through which researchers and participants use photographs to elicit meaning about the elements represented in the photograph. As with photo-documentation, several avenues are available as far as research design is concerned. First, the researcher can choose to analyze photographs without the help of study participants or organizational members. This can be accomplished by examining archival photographs (e.g., annual reports and popular press) or by leveraging the aforementioned photo-documentation methods. Once the photographs are obtained, the researcher can analyze the photographs at his or her discretion. Perhaps a more popular approach, however, involves photo-elicitation, wherein the researcher presents photographs to study participants and then engages in a dialogue aimed at discerning the deeper meanings of the content. This form of photo-elicitation requires that photographs are introduced during an interview with the intent of spurring discussion with the study participant. The photographs can be photos taken by the study participant, photos taken by the researcher, or archival photos. This information can be valuable in its own right to the extent that it sheds light on the attitudes, beliefs, and interpretations of the participant. Furthermore, this process can be valuable in that it directs the interview in novel directions as the researcher and participant mutually discover surprising insights during their focus on the photograph. In other words, photo-elicitation techniques can add both content and process to the interview experience.

298

JOSHUA L. RAY AND ANNE D. SMITH

Typology In Table 1, we present a typology that incorporates researcher intentions (i.e., photographs as illustration vs. photographs as objects of interpretation) and collaboration with study participants (i.e., with help vs. without help). Below we describe the various combinations of researcher use of photographs and collaboration and provide examples of empirical work that employ photographic research methods in the speciﬁed manner. The studies in Cell 1 represent situations in which photographs are leveraged for their meaning (i.e., as objects of interpretation) and involve collaboration with study participants. Researchers in this mode are not concerned with capturing objective elements in the organizational environment; instead, they are looking for the meaning inherent in the photograph. Collaboration can take many forms in the studies in this cell in that the photo-documentation, and interpretation duties can be the responsibility of the researcher or the participants. Cell 2 includes studies in which the researcher intends to capture some element of the research setting at face value. Researchers in this mode are interested in photographs for their illustrative or recording value. Collaboration in this cell, as in Cell 1, can take many forms depending on how study participants are leveraged in the photodocumentation process. The studies in Cell 3 represent situations in which photographs are leveraged for their interpretive value, but the researchers did not collaborate with study participants. Researchers pursuing this strategy will typically be interested in symbolic value and meaning or change over time Table 1.

Typology of Photographic Use and Participant Collaboration. Interpretation of Photos

Photos as Illustrations

Collaboration with study participants

Cell 1 Venkatraman and Nelson (2008); Warren (2002); Felstead et al. (2004– Studies 3 and 4)

Cell 2 Buchanan (2001)

No collaboration

Cell 3 Dougherty and Kunda (1990); Guthey and Jackson (2005); Molloy and Whittington (2005); Felstead et al. (2004– Study 1)

Cell 4 Barry and Rerup (2006); Smith (2002); Sood and Pattinson (2006); Widener (2007); Felstead et al. (2004 – Study 2)

Photographs as a Novel Methodological Tool in Strategic Management

299

and will choose to elicit that meaning without consulting study participants. This methodology often takes the form of archival analysis using popular press accounts and publicly available organizational documents. Finally, Cell 4 includes studies that demonstrate photographs as leveraged for their illustrative value without collaboration with study participants. Generally, the researcher might be interested in using photographs to depict organizational reality to augment ﬁeld notes, jog researcher memory of the ﬁeld setting, track physical changes over time, or provide illustration for their audience.

Empirical Examples As mentioned, Cell 1 of our typology describes a research methodology in which photographs are used to elicit meaning, and researchers collaborate with study participants (i.e., organizational members). Our review of the literature provided two examples for this type of photographic research. First, Warren (2002) advocates for the use of photographic techniques in organizational research and provides an example of the methodology in her examination of organizational aesthetics. As part of her ethnographic exploration of a website design department of a global technology ﬁrm, she employs photographs as a tool that organizational members can use to describe to her how it feels to work in their ﬁrm. Warren argues that the aesthetic aspects of an organization are difﬁcult to capture and that a methodology beyond typical interviews is necessary to explore this aspect of organizational life. She describes how, at the outset of the study, many participants were eager to show her different elements and aspects of the organization. Instead of imposing her own views and biases by collecting photographs of these things herself, she realized that an alternative would be to provide the participants with a camera and capture their interpretations of what, how, and why they chose to photograph the different elements. To that end, she solicited the help of organizational members by providing them with a digital camera and asking them to take photographs representative of their work environment. Warren and the study participants then discussed these photographs. These discussions led to several insights concerning the participants’ aesthetic experience within the organization. Moreover, with respect to the photographic methodology employed, these discussions led to insights about the integration of photographs and interviews. Warren notes several unintentional effects of incorporating photography. First, she states that not only did she enjoy employing this methodology but the participants also

300

JOSHUA L. RAY AND ANNE D. SMITH

seemed to enjoy taking the photographs, which facilitated unexpected interest in the study. Second, she argues that the photographs add to the verbal data by providing affective content that would be difﬁcult to verbalize. Finally, Warren (2002) argues that neither the verbal component nor the photographs dominate the study; rather, their ‘‘symbiotic’’ effect provides the ultimate value because neither ‘‘words nor the images would be adequate alone’’ (p. 238). The photographic aspect of Warren’s work augments her contribution to aesthetic organizational research. A more recent example of collaborative photo-elicitation comes from Venkatraman and Nelson (2008), who examine how consumers create personally meaningful interactions with retailers. They argue that consumers transform the facade retailers provide in their shops to inﬂuence consumer behavior. The authors scheduled two meetings with the study participants. The initial meeting was used to collect demographic information, explain the study and the photographic portion of the study, and put study participants at ease. They instructed the participants to take photographs that illustrated their experience in the store and assured them that all their photographs would be useful photographs in the study. The participants were given a disposable camera and told that they could take any number of photographs they wished (up to the 24 exposures on the disposable camera). The authors solicited permission for the participants to take photographs at a local Starbucks store in Beijing. The second meeting was held within three days of the Starbucks visit and involved interviews employing photo-elicitation methods. During the interview, the authors presented the photographs one at a time to study participants in the order in which the subjects took the photos. The researchers asked questions such as, ‘‘Describe this photograph to me’’ and ‘‘What do you see here’’ so as to minimize biasing the participants’ answers. The interviews were recorded and became the main source of data analyzed in this research. Each author examined the transcripts for meaningful elements (e.g., themes, quotes, and interpretations). The authors then came together to discuss their individual ﬁndings and negotiate disagreements. Finally, two independent judges examined the ﬁnal categorization scheme, which provided material agreement calculations. The authors argue that photo-elicitation methods were essential to their study in that this procedure elicits rich descriptions of the physical layout and emotional reactions to the [retail space] by facilitating deep dives by the informants or helping them go below conscious, surfacelevel observations to connect to deeper, submerged feelings, symbols, myths, and metaphors. (Venkatraman & Nelson, 2008, p. 1013)

Photographs as a Novel Methodological Tool in Strategic Management

301

As beneﬁts, they cite the relatively cheap and efﬁcient data collection process and the productive interview protocol. The authors also mention that although the procedure was successful, requiring participants to take pictures might have altered their experience somewhat in so far as individuals are not accustomed to systematically taking photographs during trips to a retail shop. However, in general, the authors argue that similar photographic methods have extraordinary potential for providing insight into business and organizational processes. In Cell 2, our typology includes photographic approaches in which the author intends to capture some element of organizational or research setting reality. This can take many forms, such as photographs of physical change and organizational processes whether they are intended to bolster ﬁeld notes or provide illustration. Furthermore, in this cell we see a researcher who engages in collaboration with study participants. Buchanan (2001) provides an example of leveraging photographs as an ancillary tool in a multi-method exploration of organizational processes. Buchanan suggests that incorporating photographs as one element in a multi-method research agenda provides a common context and mitigates some issues concerning variability in interpretation. As an example of approaching organizational research with photographs, Buchanan (2001) examines a process reengineering project in a National Health Service hospital. He used photographs to document different aspects of the in-patient process in concert with archival document research, in-depth interviews, surveys, and on-site inspections. Speciﬁcally, he took photographs of each stage of the in-patient process. These photographs were then transformed into a slide show for hospital administrators and employees. Those individuals who attended the slide show presentations were asked to provide feedback about any accuracy or omissions in the process as well as suggestions for improving the process. This feedback was incorporated into written accounts and ultimately into the ﬁnal report presented to the hospital committee in charge of implementing reengineering initiatives. Beyond providing additional information vital to mapping the in-patient procedure process, Buchanan argues that the use of photographs broadened staff awareness and assuaged uncertainty and suspicion, thus encouraging participation and supporting accuracy. In Cell 3, we present examples of empirical research that employs photographs for their interpretive value. In these cases, however, the researchers do not enlist the aid of study participants. Rather, the researchers take their own photographs or use publicly available documents (e.g., annual reports and public press accounts) that contain photographs. Molloy and

302

JOSHUA L. RAY AND ANNE D. SMITH

Whittington (2005) provide an example of the ﬁrst approach in their study of reorganization processes at eight ﬁrms. They utilize the ability of photographs to capture aspects of organizational reality so that they can ‘‘freeze moments’’ during a planning event related to a signiﬁcant reorganization (Molloy & Whittington, 2005, p. 498). They found that photographs can be used to emphasize mundane elements of the organizational setting that might be overlooked and therefore difﬁcult to recall and discuss. They also point out that photographs, in part, offset some of the cost associated with the deep immersion typically required of ethnographic studies. These cost savings are a result of the time savings from facilitating collecting, analyzing, and writing purely narrative accounts across multiple organizations. For those researchers interested in using publicly available documents, the favored source of photographs is annual reports. Annual reports provide fertile ground for analysis in that they are available for a large number of organizations, are published yearly, and are subject to scrutiny by a large number of stakeholders (e.g., corporate shareholders, regulators, employees, and competitors). More importantly, beyond the ﬁnancial and strategic information provided in these documents, it is widely held that these documents are the conduit by which organizational decision makers attempt to convey the personality or philosophy of the organization (Anderson & Imperia, 1992). To wit, organizational researchers have looked to the photographs embedded in annual reports as a means of documenting a range of organizational phenomena. Dougherty and Kunda (1990) examine photographs from annual reports in order to study whether the theories of consumers that emerge in the photographs are the result of public relations style or unique cultural artifacts. Their results suggest unique theories of consumers among their sample organizations, as evidenced by consistent depictions of consumers in the photographs contained in their annual reports. In this case, the researchers provide evidence to suggest that the use of photographs in corporate annual reports not only demonstrate consistent and systematic patterns but also provide a means of differentiating between organizations by eliciting tacit organizational perspectives. Another example of noncollaborative interpretation of photographs is provided by Guthey and Jackson (2005), who examine the nature of CEO portraits with respect to the ‘‘authenticity paradox’’ (p. 1057). The authenticity paradox arises from the staged nature of CEO portraits in that, although the portrait is intended to demonstrate some aspect of organizational identity, the artiﬁcial nature of the portrait makes the legitimacy of the claims suspect. The authors conclude that the staged nature of such

Photographs as a Novel Methodological Tool in Strategic Management

303

photographs does require scrutiny; however, they emphasize that the use of photography is an underappreciated tool for examining organizations. In addition, they argue that certain aspects of organizational research (e.g., corporate image, organizational identity, and leadership) cannot afford to continue to overlook the importance of visual media in organizational communications. Finally, in Cell 4, we see an approach through which researchers use photographs intended to capture some aspect of reality, and the photographs are obtained without collaborating with study participants. Several empirical works have utilized photographs in this manner. For example, Smith (2002) includes photographs of notes, ﬁgures, and ofﬁce space to illustrate and bolster her narrative concerning the sensemaking processes associated with her qualitative data analysis. Smith’s use of photographs augments her descriptions of her sensemaking processes with the process data she collected as a graduate student. The combination of narrative and photographic records provides insight into the extensive work required to make sense of imposing amounts of raw, qualitative data. Likewise, Sood and Pattinson (2006) use photographs to facilitate longitudinal analysis of the development of ‘‘brainports’’ (p. 701), which they deﬁne as knowledge and commerce centers. They collected archival photographs of seaports in Sydney, Australia and Kuala Lumpur, Malaysia then combined them with photographs of the current landscape. These photographs provide illustration of the changes in the port cities over time and provide support for the descriptions of the brainport concept as it manifests across these two cities. Other examples of leveraging photographs for illustrative purposes include Barry and Rerup (2006), who examine aesthetic considerations of organizational design by examining the Learning Lab Denmark. The Learning Lab Denmark is a research institute that has been keenly interested in its own aesthetic design and provided a suitable focal organization to support comparison with constructivist designs. The authors use analogical reasoning to discern elements of organizational design necessary to be aesthetically appealing. Photographs are leveraged in the work by providing examples of the artistic elements that were used for the aesthetic appeal of the organization. Additionally, Widener (2007) explores the conﬂict between resistance groups and the oil industry in Ecuador through a photographic essay. Widener used photographs to augment ﬁeld notes taken over a four-year period. Beyond their support of the narrative, Widener’s photographs and the process of collecting them served other purposes. For example, she argues that possessing photographs might be used as currency to be exchanged with various groups for future access to the research setting. In addition, Widener

304

JOSHUA L. RAY AND ANNE D. SMITH

surmises that by assuming the role of photographer she was able to distance herself from the other groups involved in the conﬂict. The camera acted as a buffer and provided her with evidence as a third party observer; hence, not only did she not collaborate with study participants but the process of photodocumentation was instrumental in articulating the boundary between researcher and participant. These studies arrayed in the cells in Table 1 provide examples of researchers’ use of photography. Next we identify how a research team’s use of photography evolved over time through more experience with the process and outcomes of photographic efforts. Examination of this group of studies, accompanied by the authors’ insights, provides a unique opportunity to examine the evolution of photographic techniques for the research team. As described below, the emergent nature of the use of photographs for this group was based on unforeseen results that ultimately led to novel insights and augmented theory exploration and development.

EVOLUTION OF PHOTOGRAPH USE IN A RESEARCH TEAM Felstead et al. (2004) recount their use of photographs across four studies that examine professionals who work from home and mobile workers who routinely conduct work as they travel. The use of photographs in their research shows a trajectory over time through our table, beginning in Cell 4 and moving through Cell 3 to Cell 1. They note an ‘‘unfolding awareness of y the roles played by images, interviews and interpretations in each study y as time went on, visual methods increased in salience’’ (Felstead et al., 2004, p. 106). We brieﬂy present the evolutionary path taken in their four studies. Felstead et al.’s (2004) initial study of professional work at home was based on 72 interviews with managerial and professional employees who work from home. After each interview was completed, the researcher asked permission to take photographs of the home worker’s ofﬁce. Photographs were not presented as an intrinsic part of their data analysis at this stage; rather, the photographs were used for illustration of existing theoretical categories of home ofﬁces. The photographs were printed before the interview texts arrived because of the lag in transcription time. As these researchers inspected the 254 photographs of 72 workplaces, they saw that instead of the home ofﬁces illustrating dimensions of an existing classiﬁcation, the photographs provided

Photographs as a Novel Methodological Tool in Strategic Management

305

evidence and details of a more reﬁned classiﬁcation scheme. The researchers began to interpret ‘‘anomalous images,’’ which led to a ‘‘reﬁnement of the categories’’ from existing theoretical models (Felstead et al., 2004, p. 107). The intention in this study was to use photographs as a record of all the research sites. Their project began in Cell 4, where photographs were intended to be illustrative without collaboration from ﬁeld participants beyond having them consent to having their home ofﬁce photographed. The interviewees did not identify what should be photographed; instead the researcher decided the shots of each ofﬁce. Additionally, the interviewees were not asked to interpret or comment on the photographs. The researchers found that ‘‘the photographs contained more information about socio-spatial boundaries in the home than ever could have been captured by interview y careful scrutiny of photographs revealed aspects of domestic interiors that respondents were likely to take for granted or even seek to hide’’ (Felstead et al., 2004, p. 109). Their photographs were interpreted for new theoretical insights. Their project was intended to be in Cell 4 – photographs as illustrative – but the researchers realized the value of interpreting the photographs, and the methodology that emerged was more in line with the Cell 3 – photographs for interpretation – approach. In another project on mobile workers (referred to as Study 2), Felstead et al. (2004) used photographs similarly to the above study on home ofﬁce, and the intended use of photographs was in keeping with Cell 4. The researchers took photographs in an ad hoc fashion in order to have a reliable and accessible picture of the various places that these workers traveled to each day. As well, they noted that their photographs ‘‘allowed the research team to see the something of the texture of the working lives’’ of these mobile workers (Felstead et al., 2004, p. 110). The interpretation rested on the interviews with these workers and photographs ‘‘functioned primarily as illustrations to respondents’ words’’ (Felstead et al., 2004, p. 110). Hence, this study used photographs in a manner consistent with Cell 4 (illustrative/ noncollaborative); the photographs provided a common understanding among the researchers based on their illustrative value. In the ﬁrst two studies, the intention was to supplement interviews with photographs (Cell 4), but in the case of the home ofﬁces, the photographs took on a primary role, allowing for interpretations and theoretical contributions (Cell 3). As these two research projects advanced, the researchers’ use of photographs changed in that photographs were incorporated into the research design. In Study 3, the researchers took a more collaborative approach with the mobile workers. They gave the workers cameras and speciﬁc directions about the types of photographs to take during their daily life. After reviewing

306

JOSHUA L. RAY AND ANNE D. SMITH

the photographs and creating questions to ask about the photographs, the researchers conducted extensive interviews with the workers. Their intent was to ‘‘bring the words and images of respondents into a productive dialogue with one another’’ (Felstead et al., 2004, p. 111). This photo-elicitation approach added signiﬁcant new understanding about mobile workers to the study, especially in terms of the tacit knowledge that allowed them to function in spite of their nomadic work life. The photographs also led to stories, myths, histories, and anecdotes that the researchers believed would not have been evoked in a traditional interview. They also noted several issues related to their data collection such as the incidence of respondents forgetting to take photographs, the difﬁculty of discerning elapsed time between photographs, the notable absence of the research subject in the photographs, and the staged nature of some of the photographs. The researchers stated that ‘‘images, interviews and interpretation all played a major role in the research process, informing each other and driving the research process forward through interaction’’ (Felstead et al., 2004, p. 113), which is consistent with Cell 1 of our typology. In Study 4, the researchers focused on workers without permanent ofﬁces. Felstead et al. (2004) adopted the same methodology as the prior study in that the subjects were given cameras and instructions on which photographs to take. When the researchers developed the ﬁlm, they found that the subjects had taken what they believed to be trivial photographs. During subsequent interviews, the researchers asked probing questions about each photo, but this approach was found to be less revealing than placing all photographs on the table and asking broadly about their working lives. At this point in the interviews, Felstead et al. (2004) noted that ‘‘the images [from the group of photographs] triggered off rich verbal reﬂections on working in collective ofﬁces’’ (p. 114). The photographs were incorporated into the verbal transcripts to create a photo-script, which led to theoretical insights. This study is in keeping with Cell 1 but provides an interesting variant on photo-elicitation in that the researchers joined images and text into one transcript that they labeled a photo-script; other scholars have called this an image/text (Warren, 2002). This photo-script ‘‘created a new synthesis y [that] proved very useful in disseminating [the] research results’’ (Felstead et al., 2004, p. 114). These studies by the Felstead research team show a variety of approaches to photographs in research, beginning with photographs that were designed to illustrate what the researcher was seeing (the intent of Study 1 and the realized use of photographs in Study 2), but for Study 1, the photographs became the main data source that was interpreted to lead to theoretical

Photographs as a Novel Methodological Tool in Strategic Management

307

contributions. In Studies 3 and 4, the researchers used a photo-elicitation approach in keeping with Cell 1, but provided a variant in terms of the creation of a photograph script in Study 4. Over time, the researchers’ use of photographs became more central to the research design as well as the data collection and analysis. This progression across our table of cells may reﬂect a learning process for strategic management researchers who are new to photographic research. These studies provide no evidence of Cell 2, but later in this chapter we identify how strategic management researchers can consider a time line interviewing approach with photographs. This may be a stepping stone or intermediate project toward movement to Cell 1, but limited data exists beyond Felstead et al. (2004) about researcher learning paths with photographs. We now turn our attention to ideas for use of photographs by strategic management researchers.

USING PHOTOGRAPHS IN STRATEGIC MANAGEMENT RESEARCH: SOME SUGGESTIONS In this section, we provide ideas for researchers about how to approach the inclusion of photographs in future research. We use our typology to organize advice to strategy researchers, and our advice is summarized in Table 2. In each cell, we consider topics in the strategic management discipline that might be enhanced by inclusion of photographs in future research. We also provide tangible advice and pros and cons related to the use of photographs for each cell. Similar to Felstead et al.’s (2004) use of photographs over time, we start in Cell 4, then discuss Cell 3 followed by Cell 2, and end with a discussion of strategy research using photographs in Cell 1. Cell 4: Noncollaborative, Illustrative Photographs in Strategic Management Research In Cell 4, photographs are primarily used to illustrate the ﬁndings from a research project. In this cell, photographs are not analyzed or interpreted for their intrinsic meaning. Rather, they are used to bring to life models created from ﬁeldwork or existing theories. Several areas in strategic management research may beneﬁt from this use of photographs including strategic change processes, organizational identity, and culture. Strategy researchers who are approaching a strategic change process should consider using photographs as a part of their method. Historical

308

JOSHUA L. RAY AND ANNE D. SMITH

Table 2.

Suggested Strategic Management Research Using Photographs. Interpretation of Photos

Photos as Illustrations

Collaboration with study participants

Cell 1 A photo-elicitation study about how strategy is understood in an organization. Photos are made by a wide cross section of members of an organization and then interviewed to understand the meaning of the photos. These photos and interviews could be compiled into a photo script, analyzed, and compared to the stated strategy of top management. Study participants, such as top managers, could be interviewed about stock photos of team decisionmaking processes and how these photos relate to their team processes.

Cell 2 Time line with photos to use during ﬁeld interviews about a change process (to aid study participant’s recollection of events). Researcher can take photos of organizational conﬁgurations that may exemplify the power relationships and processes uncovered from interviews. Study participants take photos (or suggest photos to researchers) of activities related to the research focus; these photos could be used to illustrate ﬁndings from ﬁeld interviews.

No collaboration

Cell 3 Analyze photos in annual reports for changes in leadership, commitment to status quo, strategic emphasis, or identity over time. Photos that capture events in an organization are later analyzed for details of the events. For instance, a researcher could capture board meetings or executive planning sessions; these photos could be analyzed later in conjunction with a study on TMT–board interactions or strategic planning processes.

Cell 4 Photos from annual reports (or other historic or stock photos) are incorporated in the manuscript to reﬂect ﬁndings from qualitative ﬁeld study. Documented changes to workplace such as historic photos of headquarters space (e.g., portraits on walls) or work settings over time illustrating cultural change.

Photographs as a Novel Methodological Tool in Strategic Management

309

Table 2. (Continued ) Interpretation of Photos Analyze documented changes over time to a workplace such as historic photos of headquarters space (e.g., portraits on walls) or work settings over time to identify cultural change.

Photos as Illustrations Researcher photos of objects mentioned in interviews about culture or an organization’s identity.

photographs can provide a window into seeing change more vividly. For instance, Pitcher and Smith (2001) identiﬁed how the change in CEO (and resulting different personality of the new CEO) affected ﬁrm strategy, top management team dynamics, and ﬁrm performance. The artistic CEO was replaced with a technocratic CEO who drove out goals of growth with retrenchment and diversiﬁcation with refocusing. These changes over three time periods could have been further illustrated by photographs from the company’s annual report; for instance, the colors used in the annual reports, proﬁle of the CEO (e.g., clothes, posture, and demeanor), inclusion or exclusion of photographs of other employees, and photographs of subsidiaries could have documented the changes noted from interview and archival data. Photographs could have provided a visible mapping of the three eras of top management. These photographs could have helped readers of the published article to see, perhaps more clearly, the key points about each era. We also can imagine well-known studies of strategic change over time, such as Chiles, Meyer, and Hench’s (2004) study of the growth of Branson, MO, and Dutton & Dukerich’s (1991) study of the New York Port Authority’s image change, could have been enhanced by photograph illustrations of their ﬁndings. Chiles et al. (2004) mention in their article that artifacts such as photographs and many documents were reviewed from the Branson museum, but do not elaborate on the degree to which the photographs were utilized in their analysis. We can imagine how changes and forces that they identiﬁed from their rich case study could be traced through the physical changes in this area captured in historic photographs. Another area of strategic management research that may beneﬁt from using photographs to illustrate ﬁndings is research on culture and organizational identity. Stiles (2004) argued that organizational identity research is an appropriate application of photographs, but he could not ﬁnd one instance of

310

JOSHUA L. RAY AND ANNE D. SMITH

photographs in this research stream. Stiles (2004, p. 127) stated that ‘‘given the qualitative power that such images convey, why are academics y so reluctant to embrace the pictorial forms as a means of understanding their worlds’’? Organizational identity seems to be a ripe area for exploration with photos. We can imagine a photo-elicitation study in which subjects across many levels in a company are asked to photograph images that best represent their organization. These images could be discussed with the research team to tease out what are commonly understood aspects of the organization’s identity. Photographs documenting physical changes to artifacts in the workplace over time may reﬂect the ﬁndings from research in the area of culture. For instance, Biggart’s (1977) study on the U.S. Post Ofﬁce identiﬁed how symbols can be changed to reﬂect the reorganization process and new identity. During the study period, the Post Ofﬁce’s image changed dramatically from a colonial rider to the eagle, and colors were updated from drab green to red white and blue. Inclusion of photographs of the old and new symbols could have provided the reader with further connection to Biggart’s analysis of the situation. We can imagine that photographs of the physical space of organizations – taken by the researchers or in archival records such as annual reports or historic archives – can illustrate models and ﬁndings from intensive ﬁeld studies. Another example of where visuals might have enhanced or illustrated a study’s ﬁndings is Edmondson’s (2004) work on psychological safety in different nurse settings. The differences in nursing stations were vividly described in her article, but having photographs to analyze and illustrate the dirty scrubs, messy desks, and informal meetings with nurses during the day may have reinforced Edmondson’s key ﬁndings (derived primarily from observations and interviews) about differences in unit characteristics associated with the nurses’ perceived ability to speak up. These are just two studies that provide a vivid description of an organization or organizational unit that might have been further imagined by the reader if photographs were included. These studies provide obvious examples of how photographs could add value but researchers must take into account certain considerations in using photographs in the aforementioned ways. Some organizations allow researcher access only if the organization or organizational unit is disguised (e.g., Edmondson, 2004; Pitcher & Smith, 2001). Reproduction of photographs by researchers in these settings would be difﬁcult because of concerns over identiﬁcation of the setting; however, these photographs could still be used by the researcher or research team as a vivid reminder of key takeaways while analyzing interview data or writing up results. Another issue in the use of

Photographs as a Novel Methodological Tool in Strategic Management

311

photographs in this research is obtaining permission to reproduce archival photographs. Obtaining these permissions can be complicated; for instance, locating the person who is holding the rights to the photographs can be difﬁcult. It may even be costly if the owner of the photograph demands payment for reproduction. These costs must be weighed against the beneﬁts of creating photographs to illustrate ﬁndings, one of which is the ability to create a vivid and memorable image in the mind of the reader of the academic article or book. Qualitative studies have been characterized as dense; having photographs to illustrate and guide readers might allow for easier retrieval and memorable study ﬁndings. Inclusion of photographs in a published paper may lead the reader to see different elements that might have been overlooked in a solely textual presentation of a study. Probably the most important concern related to using photographs is the need to identify the context of a photo, especially in historic photographs. This is a concern expressed in both anthropology and sociology visual literatures (Banks, 2007; Becker, 1995; Prosser & Schwartz, 1998). Researchers should document their ﬁeld research setting and photographic process; they should keep a detailed journal of what and why they took photographs, and perhaps their emotional reactions to the photographic setting. This context may be reported in abbreviated form in an article’s methodology. Ignoring this critical information could lead to confusion about why the photograph is being integrated into the text. Furthermore, it could lead to difﬁculties for individuals attempting to leverage the research in subsequent projects. These difﬁculties, however, can be minimized by including detailed descriptions of context either in the narrative, footnotes, or appendices. Paying attention to the inclusion of this material will go a long way in maximizing the value of adding photographs.

Cell 3: Photographs as Interpretation Without Collaboration Photographs in Cell 3 are open to interpretation alongside text from interviews or archival records and can inform the ﬁndings and theoretical contributions. Areas of strategy that could beneﬁt from this use of photographs are numerous, but researchers using content analysis of texts such as CEO letters to shareholders and researchers using a strategy-as-practice lens might particularly beneﬁt from inclusion of photographs as data. Content analysis is a growing methodology used in strategic management (Short & Palmer, 2008). Many studies have analyzed the text in CEO letters to shareholders to further understanding of entrepreneurial orientation

312

JOSHUA L. RAY AND ANNE D. SMITH

(Short et al., 2010) and commitment to status quo (McClelland, Liang, & Barker, 2010). Instead of using text alone, we suggest analyzing photographs included in annual reports in conjunction with the text in the report. In the management discipline, only a few researchers have analyzed annual report photographs (e.g., Dougherty & Kunda, 1990), but this has been a rich and growing area in the accounting discipline (Preston, Wright, & Young, 1996; Warren, 2005), where annual report photographs are analyzed to discern the global orientation of the ﬁrm (Preston & Young, 2000), branding (Davison, 2007), stakeholder emphasis (Chwastiak & Young, 2003), and managerial ideology (Prasad & Mir, 2002). The emphasis in many of these accounting studies is the photography as opposed to the text in an annual report. We suggest that strategic management researchers include photograph analysis in conjunction with text analysis from annual reports. For instance, in the McClelland et al.’s (2010) study cited above, we can imagine that features of photographs in the report over time could also signal commitment to the status quo; the use of texts and photographs could provide a source of convergence and promote conﬁdence in ﬁndings. In these ways, including photographs has the potential to add new insights into a text-only approach. Another area that could incorporate photographs is the emerging strategy-as-practice orientation. Strategy-as-practice researchers focus on actors and their actions related to forming strategy in organizations. Molloy and Whittington (2005) leveraged photographs to analyze the minutiae of situations that they might otherwise have missed. We can imagine that these researchers who have access to organizations and their strategy-making processes could take photographs in their role as participant observer. These photographs could be taken to illustrate an emerging model (in keeping with Cell 4), but the researcher should consider whether the details and images in the photographs might contain further insights or elaboration of an emerging model from ﬁeld notes. Again the issue with the above uses of photographs is the need to understand the context of the photographs; making the context speciﬁc is critical during interpretation of the photographs. Interpreting the photograph can be aided by qualitative software (e.g., QDA Miner), which can handle coding of articles in a systematic fashion through existing categorization from theory or through inductive development of dimensions in the photographs. A particular challenge with using photographs in annual reports is if the annual report is only available in microform or microﬁlm format. The black and white photographs may be very grainy and dark, thus making analysis of photographs in the report difﬁcult if not impossible. Another stumbling block

Photographs as a Novel Methodological Tool in Strategic Management

313

for analysis of photographs in strategic management research may be reviewers’ skepticism of the systematic nature of photograph analysis. These concerns can be overcome in time by use of qualitative software to analyze photographs, meticulous documentation of the process, and rich description of the photographic analysis process in the methodology section of papers – as well as by increasing the number of publications with photographs that are associated with theoretical contributions.

Cell 2: Photographs as Illustrative and Part of Field Collaborative Strategic management researchers can use photographs for illustration while leveraging participant collaboration. This might be helpful to researchers who study detailed organizational processes over time such as market expansion or new product innovation. Most studies of change processes beneﬁt from detailed tracking of a process through a ﬁrm, which is elicited through either retrospective interviews or ethnographic research. Many times a time line is constructed in interviews to remind an interviewee of the series of events surrounding a strategic process. For instance, in research on the internationalization process, Smith and Zeithaml (1999) used a time line of events to anchor comments about the process and key decisions and turning points. We can imagine that this time line could be augmented by photographs, such that pictures of a key acquisition, change in CEO, or international partner appear as a photo on a time line to trigger memories of the process. For product innovation, if a researcher has access to an organization, photographs can be used to keep a visual record of what the researcher is seeing unfold. For instance, in innovation research, cameras could be given to key organizational members important to the process; Dougherty and Bowman (1995) have identiﬁed that there are critical middle managers who play important roles in innovative processes. The researcher can ask these linchpin managers to take pictures during a speciﬁed time frame of meetings outside his or her department. Likewise, the researcher could trail these managers and take photographs of important activities related to bringing new products to market. Either way – whether the middle manager or the researcher takes the photographs – these visual images might augment emerging ﬁndings from interview data. In this research approach, the photographs would not be interpreted per se, but would form a visual portrayal of the processes that would be helpful to convey ﬁndings.

314

JOSHUA L. RAY AND ANNE D. SMITH

Thus, this approach to the use of photographs requires that researchers provide directions for what to photograph and how to photograph it – at certain points in each day of what they ﬁnd interesting, particular elements, number of photographs, etc. Consideration must be given to how broad or narrow to make the instructions, when to give out the cameras, and when to give photographs back at end of a project. Another shortcoming of participant photo-documentation is that it would be likely that the participant would not be captured in the photograph. If the researcher takes photographs, a different set of considerations emerge, an example of which is how to develop the appropriate level of trust and communication with the subject while balancing the needs of the study. Furthermore, consideration of what is not included in the photo and which photographs are not allowed can be as important as what is included.

Cell 1: Photographs Are Interpreted and Part of Field Collaborative Probably the most advanced use of photographs among the four cells is in Cell 1, where many studies following a photo-elicitation approach would fall. Although many areas of strategic management could be enhanced by photoelicitation approaches, we discuss three particular research areas where the use of photographs could further theory development: sensemaking about the strategic direction of the ﬁrm, top management decision-making processes, and ambidexterity in practice. We propose that photo-elicitation might be used to explore these three strategic management areas. Strategy exists only in the minds of interested parties (Mintzberg, 1987), so shaping the understanding of the strategic direction of an organization requires attention, top management effort, and sustained communication, or what Gioia and Chittipeddi (1991) called a sensegiving process. Strategic management researchers discriminate between intended and realized strategy, which parallels sensegiving by top managers and sensemaking by lower level employees; however, research on the connection or disconnection between what is communicated or intended by management and the resulting understanding by organizational participants is an underexplored research area. Photo-elicitation could provide a way to assess organizational members’ understanding of the strategic direction; for instance, a cross section of employees from different departments and levels are given a disposable camera and speciﬁc instructions to capture a set number of photos to reﬂect how they understand the strategic thrust of their organization. Then, after the photos are developed, the researcher interviews each organizational member

Photographs as a Novel Methodological Tool in Strategic Management

315

about his or her understanding of their organization’s strategy. A photo-script could be created in which key photos and comments are placed in juxtaposition in a document. This document can be further analyzed by a researcher using a qualitative software program that can analyze photos (e.g., QDA Miner). The analysis of the photo-scripts can be both deductive and inductive: deductive in comparing previous stipulated intended strategic elements gleaned from interviews with the CEO and top managers and inductive in identifying new features outside the stated strategy. This photoelicitation approach will require the support of management, but the return to management for this intrusion can be a report on the alignment of strategy and perhaps emerging new areas within the organization. It is a timeconsuming process, but one that promises rich contributions to our understanding of sensemaking of strategy within an organization. Another active area of strategic management research involves top management team strategic decision-making processes. Unfortunately, this area of research is hampered by problems of access to the upper echelon and the need for deep researcher immersion in the ﬁeld. This area of strategy research would beneﬁt from a photo-elicitation approach, though used in a different way than the above example. Because top executives are unlikely to have the time or inclination to pursue a photo protocol, the researcher may need to use stock photos of decision-making processes and query top managers to determine which photos resonate with their experiences. These photos could be staged by the researcher and related to existing theoretical insights from the literature or from the researcher’s ﬁeld experiences. These photos could include meetings of executives before the main decisionmaking meeting; heated red-faced, ﬁst-pounding meetings; meetings with ﬂip charts, handouts, and PowerPoint presentations; a meeting of the CEO with one close conﬁdant; use of consultants, etc. The photos would exclude members of the organization to reduce the potential for personal loyalties that could bias a top manager’s comments during a photo-elicitation interview. This interview about top management decision-making could lead to a photoscript, which can be analyzed in a qualitative software program. The photos in this case are used to elicit information to understand an organization’s top management team decision-making approach. This is a less invasive approach to photo-elicitation, but we believe an effective one to use with top managers who probably do not have the time or inclination to take photos for a research study. The ﬁndings from the research should be of interest to the top executives as well as researchers who can provide new theoretical insights. Finally, a photo-elicitation approach may provide new insights and tangible evidence of ambidexterity in organizations, which is deﬁned in the

316

JOSHUA L. RAY AND ANNE D. SMITH

strategic management literature as the ongoing daily tension between efﬁciency concerns (exploitation) or building new capabilities (exploration) (Gibson & Birkinshaw, 2004). This area of research in strategic management has been investigated through surveys and some qualitative observation but remains an area in need of ﬁeld-based insights (Raisch & Birkinshaw, 2008). We can imagine a research project in which a similar manufacturing unit across three or more cases is compared by the degree of ambidexterity in these units. A photo protocol could be set up to require a manufacturing manager to take photos at intervals when he or she is notiﬁed by the researcher, perhaps via text messaging. A photo could be taken by the manager with a few words added and transmitted by cell phone back to the researcher. After about a week, the researcher could sit down with the manager to talk about the photos and text. The intent of the conversations would be to categorize the photos as to whether the manager engaged in activities of exploitation or exploration. This would allow a more ﬁne-grained understanding of these concepts and how a manager acts ambidextrously on a daily basis. Of course, this photoelicitation approach would require company access and cooperation, but the feedback on different managerial approaches to daily activities may provide direction for managerial training and insights to corporate management. Overall, the photo-elicitation process as described above in Cell 1 can take many forms to tackle areas of strategic management in need of fresh theoretical insights. This cell requires the most researcher skill to obtain access to organizational members, conduct photo-elicitation interviews, and craft a photo-script for analysis. Organizations may allow researchers to conduct this type of research if the resulting insights are solid and convincing; having photos to share with upper management could make for a compelling presentation of results and perhaps facilitate further access to the ﬁrm.

USE OF PHOTOGRAPHS IN A RESEARCH PROJECT Our interest in the use of photographs emerges from a ﬁeld research project in which we explore the everyday sensegiving of values in the workplace. Similar to Schwartz (2001) in which espoused values were linked to manifestations of ethical behavior, we focus on how stated values are connected to daily activities as well as incorporating objects in our investigation. We chose a small and medium enterprise (SME) because of some unique qualities of this type of organization and the fact that most research on values has taken place in large organizations (Hammann, Habisch, & Pechlaner, 2009; Thompson & Smith, 1991). An SME offers fewer hierarchical layers between the owners

Photographs as a Novel Methodological Tool in Strategic Management

317

and lowest level workers providing a more direct link between top management initiatives and implementation. In addition, study of these organizations promotes a look at most, if not all, organizational functions due to the small size of these ﬁrms (less than 500 employees). The organization selected for our study is a U.S. apparel ﬁrm, the Marena Group in Lawrenceville, GA. This ﬁrm manufactures and sells postoperative compression garments and everyday shapewear. Marena produces domestically in the United States and has a growing global sales strategy. Our goal for this research was to cultivate an understanding of how values are practiced in everyday organizational life. The methodology chosen for this study was a case study (Eisenhardt, 1989; Yin, 1984), which included observation, interviews (mostly taped), archival data, and photographs of activities and objects. The researchers did not begin the research with any preconceived ideas about how values might be manifested in the ﬁrm or what theoretical lens might apply to the ﬁndings. Researchers visited the facility four times over a year period during 2009 and 2010. On-site visitation comprised over 30 h of observation and about 12 h of audio-taped interviews. Although the research included observation, interviews, and photographs, we focus on one element of the methodology – the role of photographs in this project. Photographs we obtained during our visits provide examples of our typology. Examples of the photos for each cell of our typology are shown in Table 3 and are explained below. We began the research by attending a daylong series of presentations by the Marena managers that provided an overview of their operations, operating philosophy, and history of the ﬁrm. During this daylong meeting, we took extensive pictures of the facility as we toured the building. We also took photographs of organizational elements that we identiﬁed as unusual from our experiences at other manufacturing plants. Many of our photographs from this initial immersion are related to the types of photographs in Cell 4. For instance, we took several photographs of a large number of employees dancing. Each day, at 2:30, all work stops for 15–20 min. During this time, many workers go to a room carved out of the manufacturing ﬂoor to dance. We did not analyze these photographs, but rather used these photographs to show to the rest of the research team to demonstrate the idiosyncrasy of the ﬁrm. (Later in the research, we tried to use the dancing photograph in a photoelicitation approach in keeping with the approach in Cell 1, but, as we will discuss shortly, this photo-elicitation approach using a photograph of the employees dancing failed.) In some photographs that were taken in the ﬁrst visit, we decided to categorize the objects in the photographs so that we would not forget the details of what we saw in the hopes of subsequent interpretation. This

318

JOSHUA L. RAY AND ANNE D. SMITH

Table 3.

Examples of Photos from Our Research Project. Interpretation of Photos

Photos as Illustrations

Collaboration with study participants

Cell 1

Cell 2

No collaboration

Cell 3

Cell 4

approach ﬁts in Cell 3 of our typology. For instance, we took a photograph of the building foyer. This foyer contained details that, we believed, were important features (perhaps linked to) the conveyance of culture in this ﬁrm. We used QDA Miner qualitative software to capture details of this photography through an initial coding of objects in the foyer. In Fig. 1, objects in this photograph are tagged using the codes from the ‘‘physical objects’’ code list – such as books, plants, light, and art. We also added comments about each object (a yellow square rectangular shape on the code) – to indicate that the plant was an orchid, the books were free and pertained to spiritual matters, and the foyer was connected to both the sewing area and manager ofﬁces. However, this analysis of this photograph was preliminary and very basic. After our second visit in which we conducted in-depth interviews with the founders and managers, the foyer took on new meaning, with a closer connection to the culture of the organization. For instance, the foyer was ﬁlled with several orchids, but their signiﬁcance was more than being ‘‘a plant’’ as we categorized them. Rather, the co-owner recounted stories of these orchids being delivered to the foyer in a near death state and

Photographs as a Novel Methodological Tool in Strategic Management

Fig. 1.

319

Screen Shot of Photo Analysis (Cell 3): Foyer at Marena.

that they had all been extraordinarily revived. Another feature, which we noted in the initial coding, was that there was no company sign outside the building. During a subsequent interview, we discovered that the founders had designed the building to be light and airy, but they never got around to getting a sign out front. ‘‘It is really about what happens within these walls that matter,’’ stated a manager in the ﬁrm. We added new codes to our original coding of objects that were closer to the espoused values of the founder/owners and seen in daily activities. These new codes, seen under the ‘‘values’’ code list, were added after in-depth interviews took place in which several elements in the foyer were discussed. The photograph of the foyer was not shown in any interviews, but the comments about elements of the foyer during our interviews and walking around the building were incorporated into the photograph analysis. The coding of the foyer includes codes for objects and values, seen in Fig. 1. During our second visit to the ﬁeld site, there were many in-depth interviews primarily with managers. During these interviews, study participants

320

JOSHUA L. RAY AND ANNE D. SMITH

would reference objects, and we would take photographs of these objects. This approach to photographs is in keeping with Cell 2 of our typology in that we did not analyze many of these photographs that were supplemental to the interviews, but the study participant had prompted us to the importance of the object. For instance, seen in Cell 2 of Table 3, we have a picture of a box that was in the shipping area. The lead designer had brought this box back to the shipping area while the researcher was interviewing the shipping clerk. The shipping clerk then explained that the box contained garments that had been worn in a Marena fashion show the previous week and were being donated to hospitals in Africa for burn or postoperative surgery patients. The box was photographed, and the interview continued to discuss other ways that the organization, speciﬁcally the founder/owners, gives back to the community. There were many examples of taking photographs during interviews in order to capture vividly what was being discussed. In Cell 2, the interviews were the main source of data, but the photos of objects, pointed out by study participants, provided vivid reminders of an important story or aspect of organizational life. For our third ﬁeld visit to the Marena ﬁeld site, we planned to do a photoelicitation with several, what we believed at the time to be, compelling photographs that were taken during the previous two visits. Selecting three photographs, we decided to ask what these photographs meant to the workers – if they could explain their importance to daily life at Marena. Instead, when the study participants were shown the photos of their dancing or other daily activities, they laughed and did not know what to say. It was their daily life, and they had a difﬁcult time articulating what was the meaning of the activity. Despite several questions used to elicit photointerpretation, none of the questions were successful in gaining much insight from the study participants. Due to the surprising results of the photo-elicitation process, we decided to take a different approach and ask study participants about their photographs framed and hung in the lunchroom (see the Cell 1 photograph in Table 3). Every employee created a frame in which Marena management asked them to identify their country of origin and what was important in their lives; these frames are arrayed by the employee’s birth month. When we were in the lunchroom (where most of the interviews with nonmanagers took place), we asked the employees to tell us about the frames, beginning with their particular frame. The discussion of these photographs led to a rich description of their life, in general, and their employment at Marena, in particular, and why they stayed with the organization. They explained how they put their frame together and that country differences were important to

Photographs as a Novel Methodological Tool in Strategic Management

321

the organization. Several mentioned that it reﬂected the diversity of the workforce. We used their comments about the photographs and analyzed the wall photographs as data. Their comments are reﬂected in the analyzed photographs that reﬂect organizational values, shown in blue in Fig. 2. These examples also demonstrate the dynamic nature of photographic research. In some instances, we learned that our understanding of the photographs was incomplete (e.g., the initial foyer photograph), and subsequent interviews added to our understanding of a photography. The insights provided from the interpretation of the photographs led to new ﬁndings concerning the daily sensegiving of values. Photographs capturing aspects of the organizational environment were probably the most memorable – tied to stories or activities that were unusual. Yet, the dancing activity was not unusual for the study participants; they were perplexed about what to say about a photograph of this activity. Hence, the photographs’ value was providing context for the research team and not necessarily for interpretation.

Fig. 2.

Screen Shot of Photo Analysis (Cell 1): Photos on Wall of Lunchroom.

322

JOSHUA L. RAY AND ANNE D. SMITH

Taken together, our research experience provides many examples of the utilization of photographs across the four cells of our typology (see Felstead et al., 2004 above). All in all, the photographs taken at Marena could be accurately classiﬁed by our typology although their use changed over time. Hence, our own experience suggests that the use of photographs is somewhat malleable and may deviate from the original intention of the researcher. This should be seen as an advantage in that the methodology can facilitate data collection and interpretation as unforeseen elements arise during qualitative ﬁeld investigations

CONCLUSIONS Our chapter is predicated on the assumptions that photos can be a meaningful part of a research project, can provide new theoretical insights, and can be viewed as having intrinsic value to the published paper. We provide several examples of how photos have been and can be incorporated into research. Organizational research publications, in general, and strategic management research publications, in particular, have not been accustomed to facilitating photographs in print. However, that does not deny the fact that photographs can provide tangible beneﬁts for the research team and the audience of these publications. Additionally, we have provided evidence that researchers are now at the cusp of incorporating photographs. As the number of researchers employing these techniques grows, we should see a resulting increase in their inclusion in journals, perhaps an initial growth of publications with photographs in online journals. As technology improves that facilitates the publication of photographic images, it is likely this trend will hasten. Ultimately, we believe this underused methodology has tremendous potential for the ﬁeld of strategic management. Most of this chapter has assumed that use of photographs is linked closely to ﬁeld research, but photographs also allow researchers the ability to reﬂect on their research process, which can apply to both quantitative and qualitative researchers. Documenting the processes, key decisions, and even research team meetings might provide a lens into our research experience, pulling back the veil to reveal heretofore hidden nuances of strategic management research. For those researchers curious about this methodology, if as yet unconvinced, we suggest that they take a camera into the ﬁeld and begin to experiment. We have cited examples of how photographs have been used in the literature, and we make suggestions for future uses. However, we do not

Photographs as a Novel Methodological Tool in Strategic Management

323

claim that the information here is exhaustive or static. As with any practical research methodology, we fully expect this approach to grow with use and understanding. We are part of a visually oriented species (Guthey & Jackson, 2005; Parker, 2009). The use of photographs in strategic management research need no longer be overlooked.

ACKNOWLEDGMENTS We appreciate the data collection efforts of Gowthami Chinnaswamy and the paper-shaping efforts of Laura Madden for this chapter.

REFERENCES Adler, N. J. (2001). Time is the author of authors. Journal of Management Inquiry, 10(2), 134. Aguinis, H., Pierce, C. A., Bosco, F. A., & Muslin, I. S. (2009). First decade of organizational research methods: Trends in design, measurement, and data-analysis topics. Organizational Research Methods, 12(1), 69–112. Anderson, C. J., & Imperia, G. (1992). The corporate annual report: A photo analysis of male and female portrayals. The Journal of Business Communication, 29(2), 113–128. Banks, M. (2007). Using visual data in qualitative research. London: Sage. Barry, D., & Rerup, C. (2006). Going mobile: Aesthetic design considerations from Calder and the constructivists. Organization Science, 17(2), 262–276. Bartunek, J., Rynes, S., & Ireland, D. (2006). What makes management research interesting and why does it matter. Academy of Management Journal, 49(1), 9–15. Bateson, G., & Mead, M. (1942). Balinese character, a photographic essay. New York: New York Academy of Science. Becker, H. S. (1995). Visual sociology, documentary photography, and photojournalism: It’s (almost) all a matter of context. Visual Sociology, 10, 5–14. Biggart, N. W. (1977). Creative destructive process of organizational change. Administrative Science Quarterly, 22(3), 410–426. Buchanan, D. A. (2001). The role of photography in organization research: A reengineering case illustration. Journal of Management Inquiry, 10(2), 151–164. Chiles, T., Meyer, A., & Hench, T. (2004). Organizational emergence: The origin and transformation of Branson, Missouri’s musical theatre. Organization Science, 15(5), 499–519. Chwastiak, J., & Young, J. (2003). Silence in annual reports. Critical Perspectives on Accounting, 14(5), 533–552. Collier, J., & Collier, M. (1986). Visual anthropology: Photography as research method. Albuquerque, NM: University of New Mexico Press. Davison, J. (2007). Photographs and accountability: Cracking the codes of an NGO. Accounting, Auditing and Accountability Journal, 20(1), 133–158. Dougherty, D., & Bowman, E. H. (1995). The effects of organizational downsizing on product innovation. California Management Review, 37(4), 28–44.

324

JOSHUA L. RAY AND ANNE D. SMITH

Dougherty, D., & Kunda, G. (1990). Photograph analysis: A method to capture organizational belief systems. In: P. Gagliardi (Ed.), Symbols and artifacts: Views of the corporate landscape (pp. 185–206). New York: Aldine de Gruyter. Dutton, J. E., & Dukerich, J. M. (1991). Keeping an eye on the mirror – Image and identity in organizational adaptation. Academy of Management Journal, 34(3), 517–554. Edmondson, A. C. (2004). Learning from mistakes is easier said than done: Group and organizational inﬂuences on the detection and correction of human error. Journal of Applied Behavioral Science, 40(1), 66–90. Eisenhardt, K. (1989). Building theories from case research. Academy of Management Review, 14(4), 532–550. Felstead, A., Jewson, N., & Walters, S. (2004). Images, interviews and interpretations: Making connections in visual research. In: C. Pole (Ed.), Seeing is believing? Approaches to visual research (pp. 105–122). Oxford, UK: Elsevier Science. Gibson, C. B., & Birkinshaw, J. (2004). The antecedents, consequences, and mediating role of organizational ambidexterity. Academy of Management Journal, 47(2), 209–226. Gioia, D. A., & Chittipeddi, K. (1991). Sensemaking and sensegiving in strategic change initiation. Strategic Management Journal, 12(6), 433–448. Guthey, E., & Jackson, B. (2005). CEO portraits and the authenticity paradox. Journal of Management Studies, 42(5), 1057–1082. Hammann, E.-M., Habisch, A., & Pechlaner, H. (2009). Values that create value: Socially responsible business practices in SMEs – Empirical evidence from German companies. Business Ethics: A European Review, 18(1), 37–51. Harper, D. (1988). Visual sociology: Expanding sociological vision. The American Sociologist, 19(1), 54–70. Harper, D. (1998). An argument for visual sociology. In: J. Prosser (Ed.), Image-based research (pp. 24–41). New York: RoutledgeFalmer. Ketchen, D. J., Boyd, B. K., & Bergh, D. D. (2008). Research methodology in strategic management: Past accomplishments and future challenges. Organizational Research Methods, 11(4), 643–658. McClelland, P. L., Liang, X., & Barker, V. L. (2010). CEO commitment to the status quo: Replication and extension using content analysis. Journal of Management, 36(5), 1251–1277. McKiernan, P. (1996). Historical evolution of strategic management. Dartmouth, UK: Aldershot Press. Mintzberg, H. (1987). The strategy concept: 5 Ps for strategy. California Management Review, 30(1), 11–24. Molloy, E., & Whittington, R. (2005). Practices of organizing: Inside and outside the processes of change. Advances in Strategic Management, 22, 491–515. Munir, K. A., & Phillips, N. (2005). The birth of the ‘Kodak moment’: Institutional entrepreneurship and the adoption of new technologies. Organization Studies, 26(11), 1665–1687. Nag, R., Hambrick, D., & Chen, M. (2007). What is strategic management, really? Inductive derivation of a consensus deﬁnition of the ﬁeld. Strategic Management Journal, 28(9), 935–955. Parker, L. D. (2009). Photo-elicitation: An ethno-historical accounting and management research prospect. Accounting, Auditing & Accountability Journal, 22(7), 1111–1129.

Photographs as a Novel Methodological Tool in Strategic Management

325

Pfeffer, J. (1993). Barrier to the advance of organizational science: Paradigm development as a dependent variable. Academy of Management Review, 18(4), 599–620. Pitcher, P., & Smith, A. D. (2001). Top management team heterogeneity: Personality, power, and proxies. Organization Science, 12(1), 1–18. Prasad, A., & Mir, R. (2002). Digging deep for meaning: A critical hermeneutic analysis of CEO letters to shareholders in the oil industry. Journal of Business Communication, 39(1), 92–116. Preston, A. M., Wright, C., & Young, J. J. (1996). Imagining annual reports. Accounting Organizations and Society, 21(1), 113–137. Preston, A. M., & Young, J. J. (2000). Constructing the global corporation and corporate constructions of the global a picture essay. Accounting Organizations and Society, 25(4–5), 427–449. Prosser, J., & Schwartz, D. (1998). Photographs in the sociological research process. In: J. Prosser (Ed.), Image-based research (pp. 115–130). New York: RoutledgeFalmer. Raisch, S., & Birkinshaw, J. (2008). Organizational ambidexterity: Antecedents, outcomes, and moderators. Journal of Management, 34(3), 375–409. Rippin, A. (2006). Refusing the therapeutic: Marion Milner and me. Culture and Organization, 12(1), 25–36. Schwartz, M. (2001). The nature of the relationship between corporate codes of ethics and behavior. Journal of Business Ethics, 32(3), 247–262. Short, J. C., Broberg, J. C., Cogliser, C. C., & Brigham, K. H. (2010). Construct validation using computer-aided text analysis (CATA): An illustration using entrepreneurial orientation. Organizational Research Methods, 13(2), 320–347. Short, J. C., & Palmer, T. B. (2008). The application of DICTION to content analysis research in strategic management. Organizational Research Methods, 11(4), 727–752. Smith, A. D. (2002). From process data to publication: A personal sensemaking. Journal of Management Inquiry, 11(4), 383–406. Smith, A. D., & Zeithaml, C. (1999). The intervening hand: Contemporary international expansion processes of the regional bell operating companies. Journal of Management Inquiry, 8(1), 34–64. Sood, S. C., & Pattinson, H. M. (2006). Urban renewal in Asia-Paciﬁc: A comparative analysis of brainports for Sydney and Kuala Lumpur. Journal of Business Research, 59(6), 701–708. Stasz, C. (1979). The early history of visual sociology. In: J. Wagner (Ed.), Images of information (pp. 119–136). Beverly Hills, CA: Sage. Stiles, G. (2004). Pictorial representation. In: C. Cassell & G. Symon (Eds), Essential guide to qualitative methods in organizational research (pp. 127–139). London: Sage. Thompson, J. K., & Smith, H. L. (1991). Social responsibility and small business: Suggestions for research. Journal of Small Business Management, 29(1), 30–44. Venkatraman, M., & Nelson, T. (2008). From servicescape to consumptionscape: A photoelicitation study of Starbucks in the New China. Journal of International Business Studies, 39(6), 1010–1026. Wagner, J. (1979). Introduction: Information in and about photographs. In: J. Wagner (Ed.), Images of information (pp. 11–22). Beverly Hill, CA: Sage. Warren, S. (2002). Show me how it feels to work here: Using photography to research organizational aesthetics. Ephemera, 2(3), 224–245.

326

JOSHUA L. RAY AND ANNE D. SMITH

Warren, S. (2005). Photography and voice in critical qualitative management research. Accounting, Auditing, and Accountability Journal, 18(6), 861–882. Warren, S. (2008). Empirical challenges in organizational aesthetics research: Towards a sensual methodology. Organization Studies, 29, 559–580. Widener, P. (2007). Oil conﬂict in Ecuador: A photographic essay. Organization and Environment, 20(1), 84–105. Winston, B. (1998). The camera never lies: The partiality of photographic evidence. In: J. Prosser (Ed.), Image-based research (pp. 60–68). New York: RoutledgeFalmer. Yin, R. (1984). Case study research: Design and methods (1st ed.). Beverly Hills, CA: Sage Publishing.

Building Methodological Bridges

Crossing Boundaries, Building Bridges

Crossing Boundaries, Building Bridges

Building bridges: Between mathematics and computer science

Bridges

Building bridges between mathematics and computer science

Bridges

Methodological Individualism

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

Burning Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

Integral Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

Building Methodological Bridges

Crossing Boundaries, Building Bridges

Crossing Boundaries, Building Bridges

Building bridges: Between mathematics and computer science

Bridges

Building bridges between mathematics and computer science

Bridges

Methodological Individualism

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

Burning Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

Integral Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

London Bridges

Recommend Documents