International Review of Research in Mental Retardation

International Review of RESEARCH IN MENTAL RETARDATION VOLUME 16

Consulting Editors Ann M. Clarke THE UNIVERSITY OF HULL

J. P. Das THE UNIVERSITY OF ALBERTA

H. Carl Haywood VANDERBILT UNIVERSITY

Ted Nettelbeck THE UNIVERSITY OF ADELAIDE

International Review of RESEARCH IN MENTAL RETARDATION

EDITED BY

NORMAN W. BRAY ClVlTAN INTERNATIONAL RESEARCH CENTER A N D DEPARTMENT OF PSYCHOLOGY T H E UNIVERSITY OF ALABAMA A T BIRMINGHAM BIRMINGHAM. ALABAMA

VOLUME 16

ACADEMIC PRESS, INC. Harcourt Brace Jovanovich, Publishers

San Diego N e w York Boston London Sydney Tokyo Toronto

This book is printed on acid-free paper.

@

Copyright 0 1990 By Academic Press, Inc. All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.

Academic Press, Inc. San Diego. California 92101 United Kingdom Edition published by Academic Press Limited 24-28 Oval Road, London NWI 7DX

Library of Congress Catalog Card Number:

ISBN 0-12-366216-8

65-28627

(alk. paper)

Printed in the United States of America 9 0 9 1 9 2 9 3 9 8 7 6 5 4 3 2 1

Contents

.........................................................,........................... P r e j k e ...................................................,.................................................. Con/ribrtlors ..........

ix xi

Methodological Issues in Specifying Neurotoxic Risk Factors for Developmental Delay: Lead and Cadmium as Prototypes Stephen R. Schroeder I. Introduction ............................ ........................................................... 11. Brief Summary of Risk Literature in Child Development ........................... 111. Risk Models ....................................................................................... IV. Assessing Effects of Combinations of Risk Factors with Statistical Models ........................................................... .............. V. Specifying Risk Factors for Developmental Disabilities Due to Lead and Cadmium ......................................................................... VI. Recurring Issues in Neurotoxicology of Lead and Cadmium ............. VII. Summary ........................................................................................... References ...

i 6 7 II 14

22 28 28

The Role of Methylmercury Toxicity in Mental Retardation Gary J . Myers and David 0. Marsh I. Introduction ....................................................................................... 11. Mercury .......................... 111. Mercury Exposure ............

IV. Measurement of Mercury in Human

......................................... ......................................... Levels.. ............................. Fetal Sensitivity ........................................................... Studies of Metal Methylmercury Exposure at Low Levels ......................... ................ .......... References.. .......................................................................................

VI. Outbreaks of Human Exposure

VIII. IX.

............

V

33 34 35 36 37 39 43 44 45 47 48

vi

CONTENTS

Attentional Resource Allocation and Mental Retardation Edward C. Merrill 1. 11. 111. IV. V.

Introduction ........ ................................................................... Development of th onal Resource Hypothesis Attentional Resource Allocation ............................................................ Preliminary Research .......................... ................. ...... ............... ........... Discussion and Conclusions ............. ..................... ... References........................................ .................................................

51 54 60 65 XI 84

Individual Differences in Cognitive and Social Problem-Solving Skills as a Function of Intelligence Elizabeth J . Short and Steven W. Evans

I. 11. 111. IV. V. VI. VII. VI11. IX.

Introduction ....................................................... ............. .............__.. .. Models of Cognitive Problem Solving ...................... Individual Differences in Cognitive Problem Solving .............. ................... Models of Social Problem Solving ...................... Individual Differences in Social Problem Solving ...... ...... . ...... ......... . _ _ . . _... . Methods of Assessment .............................................. ................ Methods of Fostering Cognitive and Social Proble Limitations of Current Research ....................... Future Directions and Conclusions ......................... ............................ . ... ........................ References............................................ ..........

X9

92 95 104 105 109 112 I I5 116 I I8

Social Intelligence, Social Competence, and Interpersonal Competence Jane L. Mathias

I. Introduction ... ............. ............. ........................................................

..

125 126

I I I . Greenspan’s Models .. ......................................,...............................,... IV. Investigations of Social and Adaptive Intelligence in Mentally Retarded Adolescents ..................................................................... References ..........................

141

11. Investigations of Socially Based Competencie

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

147 151

Conceptual Relationships between Family Research and Mental Retardation Zolinda Stoneman

I. Introduction ... ............................................................................. ...,...

161

11. Issues Surrounding the Conceptualization of Mental Retardation. ....., .......... 111. Classification within Mental Retardation .......................................... ...,...

I 64 177

vii

CONTENTS

IV . Personality. Motivation. and Other Individual Differences .......................... V . Conclusions and Implications ................................................................ References .........................................................................................

.. ........................................................................................................

Itidi r

Co~trerrt.sof Previous Vo/rtrnus

.......................................................................

185 190 194

203

213

This Page Intentionally Left Blank

Contributors

Numbers in parentheses indicate the pages on which the authors’ contributions begin.

Steven W. Evans (89), Western Psychiatric Institute and Clinic, Pittsburgh, Pennsylvania 15213 David 0. Marsh (33), Department of Neurology, University of Rochester School of Medicine, Rochester, New York 14642 Jane L. Mathias (125), Dcpurtment of Psychology, University of Adelaide, Adelaide, South Austruliu 5001, Australia Edward C . Merrill (5 I ), Depurlmenl of Psychology, The University of Aluhumu. Tusculoosu, Aluhumu 35487 Gary J . Myers (33), Departments of Pediatrics and Neurology, University of Aluhatna (it Birminghum, Birmingham, Alabama 35294 Stephen R . Schroeder ( I ) , Burcuu of Child Research, University of KunS N S , Luwrence, Kansus 66045 Elizabeth J . Short (89). Depurtment of Psychology, Case Western Reserve University, Cleveland, Ohio 44106 Zolinda Stoneman (1611, University Affiliated Program for Persons with Developmentul Disuhilities, und Department of Child und Family Development, The University of Georgia, Athens, Georgia 30602

ix

This Page Intentionally Left Blank

Preface

The contributors to this volume address three important areas of current research in mental retardation. The first two articles focus on behavioral toxicology research and its implications for the cause and prevention of some types of mental retardation. The next two address aspects of cognitive processes in mentally retarded individuals and the significance of this issue for the further development of remediation techniques. The last two articles are concerned with social adaptation and family functioning and the importance of these topics for the study of mental retardation. In the first contribution, Stephen Schroeder presents a conceptual and methodological critique of research on the effects of chronic exposure to toxic substances. He discusses various risk models and applies these to studies on the contribution of lead and cadmium exposure to the risk of developmental disabilities. His disturbing conclusion is that long-term exposure to very low levels of neurotoxins such as lead and cadmium may have detrimental effects on cognitive abilities analogous to the gradual deterioration of the nervous system with age. Next. Gary Myers and David Marsh review research on the effects of methylmercury on human development. Awareness of the devastating effects of high levels of exposure to this toxin followed the tragic outbreak of methylmercury poisoning in Minimata Bay in Japan in 1953. Myers and Marsh discuss problems of chronic exposure to low levels of methylmercury in mothers and the effects of prenatal exposure on development. The research on lead and cadmium reviewed by Schroeder and on methylmercury reviewed by Myers and Marsh underlines the importance of environmental regulation of toxins in the prevention of mental retardation and developmental disabilities. In his article on attentional resource allocation, Edward Merrill reviews recent research, much of it from his own laboratory, indicating that mentally retarded individuals are less flexible in their allocation of attention than are nonretarded individuals. He suggests that it is not yet clear whether this difference reflects a fixed, structural feature of the cognitive system or if it represents a metacognitive deficiency. In either case, Merxi

xii

PREFACE

rill's review indicates that intervention techniques must take into account resource allocation in mentally retarded individuals. Elizabeth Short and Steven Evans review models and research on cognitive and social problem solving in mentally retarded individuals. There are important parallels between the component processes of these two general domains of problem solving. The major difficulty in both appears to be strategic flexibility. Like the chapter by Merrill on attention allocation, the research on problem solving suggests that intervention focusing on aspects of metacognition appears to be appropriate, a trend seen in several other areas of current research and consistent with the Vygotskian approach. In her contribution on social intelligence and social and interpersonal competence, Jane Mathias provides a historical review of the psychometric roots of the social intelligence concept and the human development roots of the social competence and interpersonal competence concepts. She presents Greenspan's model of personal competence as an integration of these overlapping but historically separate topics. She reviews her own research showing support for a distinction between social and nonsocial competence and suggests that this differentiation might modify our current conceptualization of adaptive behavior as a dimension of the definition of mental retardation. In the final article, Zolinda Stoneman presents a case for more interaction between research on mental retardation and family adjustment. She notes that family research cannot provide an adequate account of how a mentally retarded child impacts the family without a consideration of the nature of mental retardation, both in its cognitive aspects, as discussed by Merrill and Short and Evans, and in its social-adaptational aspects, as discussed by Mathias. Also, because mentally retarded individuals are embedded in an ecological context that may account for more of their behavior than the degree of mental retardation, research must incorporate the ecological context of the family as an important aspect of mental retardation. As illustrated by these contributions, this series will continue to publish integrative reviews on a wide range of current topics. These will include the psychological and social nature of mental retardation, the biological and neurological basis of behavioral and psychological problems associated with mental retardation, and the nature of problems of adaptation encountered by mentally retarded individuals. Reviews of both basic and applied research will be included. The majority of the articles are written by invitation, but unsolicited manuscripts will be considered. NORMAN W. BRAY

Methodological Issues in Specifying Neurotoxic Risk Factors for Developmental Delay: Lead and Cadmium as Prototypes STEPHEN R. SCHROEDER BUREAU OF CHILD RESEARCH UNIVERSITY OF KANSAS LAWRENCE. KANSAS 6h045

1.

INTRODUCTION

The passage of Public Law 99-457 (amending the 1975 “Bill of Rights for Handicapped Children”) in 1986, which requires identification and early intervention for infants and toddlers, 0-3 years of age, has renewed the interest of professionals in child development in the concept of risk for developmental delays. The publication of Air Quality Crireriu for Lead (U.S. Environmental Protection Agency [USEPA], 1986), together with the many controversies over setting standards for permissable levels of lead in the atmosphere, has drawn great interest by toxicologists and regulatory agencies in the longitudinal study of the neurotoxic effects of environmental pollutants on this very same 0-3 year population. Yet these two literatures have shared little overlap thus far. It would appear that child developmentalistshave viewed environmental neuorotoxic substances as one of many biological risk factors that may play only a minor role in developmental delays. Toxicologists, on the other hand, tend to think of critical organ damage or neurochemical pathology of the nervous system, often not even acknowledging behavior as an end point of neuroI INTERNATIONAL KEVIEW 01 KESEAKCH IN MENTAL KEI‘AKDATION. Vol. I6

Copyright 0 1990 by Academic Press. Inc. All rights of reproduction in any form reserved.

toxicity. Child development is considered a “soft science,” if any at all. Behavior is usually subsumed under the field of epidemiology, a practice which has often led to stressful controversies during the process of setting regulatory policy (Schroeder. in press). The purpose of this articlc is to merge these two literatures, in order to develop consistent definitions, theoretical frameworks, and a mutual awareness of the power and importance of these two disciplines to the study of mental retardation and developmental disabilities. Risk for disabilities due to environmental toxic substances looms large on the horizon of the entire planet, because pollution has been growing at an exponential rate since the Industrial Revolution began. Neurotoxins are being implicated in an increasingly wider array of cognitive and adaptive behavior problems. Needleman (1989).for instance, claimed the incidence of attcntion deficits, conduct disorders, juvenile delinquency, and adult criminality is significantly due to environmental lead exposure in the United States. The merger of child development behavioral science with the toxicology and epidemiology of environmental pollutants gives rise to a number of problems. The natural skepticism of behavioral scientists in dcaling with their independent and dependent variables is often puzzling for biological scientists, who prefer to deal with the presence or absence of disease entities. This difference in orientation often leads to disagreements about the definitions and the underlying dimensions of bchavioral end points studied. To illustrate these problems, 1 have chosen to contrast two toxic heavy metals, lead and cadmium; a great deal is known about the neurotoxicity of lead, but very little is known about cadmium. A.

Definition and Types of Risk

The term uf risk has been used to denote an individual or a population having an increased probability of having or acquiring a handicapping condition. Tjossem (1976) identified three categories of infants at risk as follows.

I . Established Risk. Infants whose early aberrant development is related to diagnosed medical disorders such as Down syndrome. 2. Biological Risk. Infants with a history of prenatal, perinatal, neonatal, and early developmental events suggestive of biological insult(s) to the developing central nervous system (CNS). Prematurity is an example. 3. Environmentd Risk. Infants who are biologically sound but for whom early life experiences are sufficiently limiting that they impart a high probability for delayed development; for example, by having parents

METHODOLOGY OF NEUROTOXIC RISK FACTORS

3

who have a disability or who lack knowledge of ways to stimulate development.

These categories of risk are not mutually exclusive. When more than one factor occurs, the risk combines to make the child significantly more at risk for developmental delay. A recent national survey by Graham and Scott (1988) of early intervention programs under Public Law 99-457 shows that this is the most widely used classification scheme. Neurotoxic insults would be classified as biological risk factors. The definition of risk in the neurotoxicology literature is based on a more narrow epidemiological concept. According to Kleinbaum, Kupper, and Morgenstern (1982), it is “any variable that the investigator determines to be ‘causally related’ (though not necessarily a ‘direct’ cause) and antecedent to illness outcome status on the basis of substantive knowledge or theory and/or on research findings” (p. 255). Operationally, the most common way to estimate risk is to divide the number of newly detected cases that developed during follow-up by the number of diseasefree subjects at the start of follow-up. This proportion is called crrmulutive incidence, and it assumes a fixed cohort observed over a set period of time. However, very often populations are dynamic in that some subjects may drop out or may be followed for different periods of time, or new members are added. These changes are likely to be related to the length of the period of follow-up or to whether a person developed a disease. In such cases an alternative measure is incidence rule, defined as the instantaneous potential for change in disease status per unit time. It is calculated by dividing the number of new cases by the amount of population-time of follow-up (e.g., person-years) of the disease-free population. Often this amounts to the total disease-free time during a follow-up period. lf these data are not available on all subjects, then they must be estimated. This ratio is called the average incidence density and is the subject of considerable epidemiological theory (Kleinbaum et ul., 1982). Epidemiological definitions of risk are often not considered in the child development risk literature, but they should be, especially when the inferential statistical analyses used are based on assumptions (e.g., sampling distributions, estimates of ascertainment bias, assessment of confounding) taken from epidemiological theory. B.

Dimensions of Neurotoxic Risk

A number of issues in the assesment of neurotoxic risk dimensions are indigenous to the field of behavior toxicology (Cory-Slechta, in press). The first issue is what behaviors should be measured. If one measures

4

Anosmia Appetite Loss Convulsions Depression Disorientation

I I

Headache

Jitteriness, Irritability Mental Retardation

Psychiatric Signs Somnolence Tremor Visual Disturbances Weakness

I

0

I 0

FIG. I . Matrix of toxic symptoms ascribed to metals. Adapted from Weiss (197x1.

the neurotoxic symptoms resulting from biological effects of metals, one finds that they tend to be diffused and not specific to a specific metal or other toxicant. Weiss (1978) gives a matrix of toxic symptoms ascribed to metals in Fig. I . This figure illustrates a number of points. Some vagueness is inescapable when symptoms represent only the early, incipient manifestations (thresholds) along continua of toxicity. Metals often do not attack a single critical organ or substrate, but whole organisms which vary in behavioral history, genetic susceptibility, dietary status, and a host of other variables. Figure I raises two inseparable issues. What if a response is not likely to occur? Many toxic effects on behavior occur at dose levels that do not produce symptoms. Behavior toxicology technology is largely devoted to making such sensitive measurements.

METHODOLOGY OF NEUROTOXIC RISK FACTORS

5

A second issue is that the wide individual differences in susceptibility to neurotoxic insults from metals, especially at very low exposure levels, limit the generality of the epidemiological approach. Misclassification of index cases increases when higher precision of measurement is required to achieve a statistically reliable effect. In such cases, the individual differences approach of behavior toxicology is especially important. A third important issue is that the proficiency and replicability of biological indices of exposure (e.g., blood, teeth, bone, hair) within and between laboratories is very important in assessing dose effects. For instance, hair levels are often considered a poor external measure of internal body burden. To understand better this view of exposure indices, it would be helpful to define the various terms which will be encountered. Dose is the amount or concentration of a substance which is presented over time to the specific, intracellular site where the effect is imparted. Since it is not usually possible to assess directly the quantity of that substance in the living organism at the affected site (ejfect site), the external and internal doses are considered as indices mirroring the effect-site concentration. External dose is the amount of the toxic substance in the external environment (air, water, food, etc.) to which an organism is exposed. Internal dose is the absorbed portion of the substance and is an integrated reflection of all contributing external exposures. Eflect is a biological change resulting from exposure to a toxic substance. Dose-eflect relationship is a quantitative relationship between the dose and a specific effect (i.e., it reflects changes in the intensity of an effect as a function of variations in dose). Dose-effect relationships vary among members of a population, and the frequency at which this occurs is expressed as the dose-response relationship. Response specifically is defined by Nordberg ( 1976) as that proportion of percentage of a population that exhibits a specific effect at a given internal dose level. Nordberg ( 1976) has defined the concept of critical organ, critical concentration in the critical organ, and critical effect. Critical organ is defined as that organ which first attains the critical concentration of a metal under defined situations and for a given population. The critical concentrarion is defined as the mean concentration of the toxic substance in the critical organ at which adverse functional changes appear. The critical eflect is the first adverse effect that occurs along the continuum of doseeffect relationships. A fourth assessment issue is the need to extrapolate from animal to human data. For ethical reasons, neurotoxicology research on humans needs to be observational or quasi-experimental. Only animals are used for truly experimental research which constitutes a crucial part of the

database. For instance. it would be very difficult to study experimentally the dietary interactions of lead, zinc, calcium, manganese, or iron and their effects on cadmium absorption. Such data are needed to estimate dose effects and neurotoxic symptoms, as well as the underlying behavioral and biological mechanisms of action of metals. With respect to lead neurotoxicity, the research findings from animal and human research have been remarkably consistent in finding a fairly smooth dose-response relationship between lead and CNS function (USEPA. 1986). The overwhelming proportion of neurotoxicity research in humans has been restricted to lead. There is not a comparable human database for other trace metals such as cadmium, for which data are almost exclusively from animal research. The reasons for this imbalance are likely to be the result of the ubiquity of the use of lead in so many industries and the political and economic ramifications of regulating its release into the atmosphere. Cadmium is much more toxic and has a longer and less reversible half-life (38 years).

II.

A.

BRIEF SUMMARY OF RISK LITERATURE IN CHILD DEVELOPMENT

Background

In the past 20 years, several researchers in child development were interested in developing lists of risk factors for developmental delay. subtyping risk factors, and developing tracking systems with periodic assessments which would permit them to predict and modify outcomes in later childhood. We have reviewed this literature (King, Logsdon, & Schroeder, 1990). This review indicated that the timing of the periodic risk assessments was related to the reliability of predicted outcomes in later life. Risk factors fluctuate over the life span in their power to predict developmental outcomes. In general, the earlier the risk factor was measured and the later the developmental outcome was measured, the poorer was the correlation between them. Furthermore, the literature is inconclusive as to which measures are most appropriate at which ages. B.

Developmental Framework for Concept of Risk

The recent literature has focused on a developmental approach. The basic elements are (a) the nature of the child’s behavioral repertoire, (b) the presence of individual difference, (c) the functioning of environments on a continuum of developmental facilitation, and (d) the possible exis-

METHODOLOGY OF NEUROTOXIC RISK FACTORS

7

tence of nodal developmental points that involve a change in the equation determining developmental outcome (Horowitz, 1988). Risk variables may reside in the organism or in the environment or both with different degrees of mutual influence at different developmental periods (Horowitz, 1988), and they may increase or decrease over time. For these reasons,developmental outcomes are difficult to predict with much certainty by using available statistical methods. It seems that single biological variables such as neurotoxic agents have been of limited value in predicting developmental outcomes, unless established risk factors such as encephalopathy are evident. It is possible that very severe biological variables (e.g., Apgar scores < 4 or very low gestational age or birth weight) may have some value in prediction; however, this has not been definitely established (Horowitz, 1988). Some environmental variables do have some value in predicting developmental outcome. The variables most often cited are socioeconomic status (SES), care-giving practices. maternal education, and maternal age. The best prediction comes from combining biological and environmental predictors. The best studies, however, have explained only about 30-50% of the variability in outcome (Horowitz, 1988). Neurotoxic factors usually account for less than 10% of the variance in such studies (Schroeder. in press). The earlier the risk evaluation, the less dependable are factors relative to neonatal events and capabilities of the child taken singly. It is also evident that some developmental problems appear early and disappear, while others will not appear until later in age. There is justification for following some families with children who are not exhibiting developmental delay but who do have biological and environmental risk factors. 111.

RISK MODELS

Traditionally, we have thought of two kinds of risk factors for developmental disabilities: nature and nurture. Fragile-X genes would be an example of the former, while poverty might be thought of as an example of the latter. However, rarely does delayed performance reduce to such a simplistic risk model. This model does not account adequately for the many complex interactions between heredity and environment which impact on retarded behavior. Modern-day epidemiological approaches contain multivariate risk factor models which incorporate biological and psychosocial variables, their interactions, and an assessment of their possible confounding effects. Combinations of risk factors can combine additively or multiplicatively to increase the probability of developmental delays.

Poverty not only sets the stage for psychosocial problems but also leads to medical problems (e.g.. very low birth weight). The effects are compounded multiplicatively. Modern epidemiological research attempts to model and weigh such effects appropriately. Most studies have assumed a linear relationship between risk factors and developmental outcome. This model assumes that if all risk factors were known, mathematical weightings could be applied, and the factors added together could estimate with a degree of precision developmental outcome or cumulative risk. The fact that researchers have been unable to predict developmental outcome with precision is explained by the assumption that we have not identified some essential risk factors. More and more thinkers in the area of risk are feeling that it is not the knowledge of risk factors which is the problem. It is more likely that known risk factors do not have a linear relationship to developmental outcome. There are three models which are competing with the cumulative risk model. These are as follows: the Transactional Model, as discussed by Sameroff and Chandler (1975); the Biosocial Systems Model, as discussed by Ramey, Yeates, and MacPhee (1984); and the New Morbidity, as discussed by Baumeister and Kupstas (1987) and by Baumeister. Dokecki, and Kupstas (1988). These models are discussed in detail and their meaning for the study of risk is described below.

A.

Transactional Model

The Transactional Model was identified by Sameroff and Chandler (1975) as the continuum of caretaking casualty. In this approach, there is increased focus on child-environment interactions and how these reciprocal relationships affect developmental outcomes. Biological factors arc seen as being important during the first several months of life and the environment more important later in development. Horowitz (1988) adds to these transactions the concepts of universal and nonunivcrsal behaviors. Universal components of behavior have a high probability of occurring regardless of environment, given minimal conditions for learning. An example would be walking. They are subject to compromise due to biological risk factors. The universal behavioral repertoire provides the basis for the acquisition of nonuniversal behavior. Nonuniversal behaviors can be interfered with by organic problems, the environment, or both. An example would be language communication. Horowitz (1988) also discusses periods during development when the child might be particularly sensitive to environmental deficits. This model assumes a degree of plasticity in both the child and the environment.

METHODOLOGY OF NEUROTOXIC RISK FACTORS

B.

9

Biosocial Systems Model

Ramey et CJI. (1984) developed an enhancement model of risk that gives equal weight to transactions and the environment or ecology within which those transactions take place. They highlight the fact that typical models of development have been inadequate in accounting for individual variation in development. They argue that one should abandon the use of variables related to developmental problems and highlight those that predict optimal outcomes. Biological and environmental factors are seen as stressors. Risk is the disequilibrium resulting from change and the inability to control that change. Intervention is based on developing strategies for adapting to the risk or teaching individuals ways of meeting the demands of the environment. Strengths in the individual or the environment make one more or less vulnerable to the effects of the risk condition. Development is viewed from a systems perspective in this model. The child’s development is seen as a product of a system of interacting units. Interaction between the child and the environment occurs at multiple levels and with a variety of degrees of complexity. Interactions are conceptualized as occurring at the following levels: child, care-giver, household, school, neighborhood, and society. Variables in a system vary in their stability. Any variable which forces a system beyond its range of stability produces stress, necessitating self-regulatory or coping mechanisms. When self-regulatory mechanisms are unable to reestablish equilibrium, disorder or maladaptive behavior occurs. The child is seen as plastic and an active participant in interactions with his or her environment. The developmental outcome results from a stream of cause-effect events measured within the context of the child’s ecology. C.

New Morbidity

Baumeister and Kupstas (1987) used the term new morbidity to describe poor health and developmental consequences of adverse social and environmental factors. This model is useful for identifying strategies which a community could use in dealing with prevention of developmental disabilities. The model emphasizes a combination of conditions and identifies predisposing variables (such as demographic, behavioral, and genetic factors), catalytic variables (such as poverty), and resource variables (such as limited personal/social resources) which lead to neonatal problems including poor pregnancy outcomes. Poor pregnancy outcomes lead to a variety of poor developmental outcomes. Intervention efforts are targeted to those families particularly susceptible to predispos-

ing variables. Poverty is seen as a predisposing variable in the new morbidity. D.

Establishing Probable Causes with Risk Models

Why do we need models such as those above to guide us‘? The main reason is that there are so many possible causes, so many possible outcomes, and so few variables which we can manipulate to prevent developmental disabilities that we must take a statistical modeling approach to maximize the chances of success. The fact that we must rely on epidemiological observational studies severely limits our ability to test how social variables mask or potentiate a risk factor’s effect on a developmental outcome. Ethical constraints limit our ability to do any experimental manipulations with the study population from which we can infer that factor A causes factor B. However. Cowan and Leviton ( 1980) have suggested several indirect criteria for evaluating probable causality from an epidemiological point of view that appear to be very reasonable. The first criterion is the strength of the association. That is, the higher the relative risk, the more likely the etiological factor is to be important in the development of the outcome. The second factor is the specificity of the association. This refers to the requirement that the suspected causal factor be associated with only one or a few related outcomes. Third. the consistency of association should be considered. The probability of a causal relationship between a factor and an outcome increases when similar results are repeatedly found in studies conducted by different investigators, under varying circumstances, at different times and locations. with different population groups. A fourth consideration is ternportrl seyrrence (i.e., the factor should be shown to precede the outcome in time). Fifth, if the.freqrrency or severity of outcome is found to increase with increasing levels of exposure of a risk factor, a dose-response relationship is said to exist and the association of interest is more likely to bc a direct (causal) one than an indirect one. Biologicd plarrsihility of the association is a sixth factor to consider. A seventh factor is coherence of the evidence, that is. whether a causal interpretation of an association is consistent with what is known about the history and biology of the disease or outcome. A final consideration is experimental or semiexperimental evidence. Thus, although deliberate, systematic exposure of children to a risk factor may not be ethical. it is possible to observe the effects of “accidental” exposure or of the reduction or elmination of exposure. Cowan and Leviton (1980) conclude that, although these criteria do not

METHODOLOGY OF NEUKOTOXIC RISK FACTORS

II

provide proof of a causal relationship, they can be used in determining a qualitative estimate of the probability that the relationship is causal. Questions of probable causality related to risk factors take on added complexity in longitudinal developmental studies where the relationship of cofactors to risk outcomes can change over different points in time. Establishing probable cause for developmental outcome using databased literature is difficult, and many problems become evident. Associations between developmental outcomes and predictor variables are weak. With the exception of certain established risk factors, the specificity of the association to certain causal factors is poor. Some consistency is being obtained in independent longitudinal studies between predictor variables and developmental outcome. Biological and environmental risk factors taken independently do not seem to increase prediction of developmental outcome. Taken together, however, the greatest amount of variance is explained differently if overlapping measures of biological and environmental measures are used to explain the developmental outcome. Temporal sequencing makes research straightforward but difficult in that it takes time before the impact of some risk variables is expressed in developmental outcome. With environmental variables it is often difficult to establish which came first. Severity of outcome is difficult to establish in young children with developmental delays. This is due both to the biological plasticity of the child and problems with accurate measurement. Problems with quantitative measurement of risk factors also complicate evaluating studies in terms of these criteria. Biological plausibility can be, and often is. argued with many groups establishing risk lists. Coherence with known history and biology as evidenced by research is an essential criterion. Accidental exposure is difficult to evaluate qualitatively due to difficulty in measurement and the plasticity of the nervous system in the child. IV.

ASSESSING EFFECTS OF COMBINATIONS OF RISK FACTORS WITH STATISTICAL MODELS

We are prohibited ethically from manipulating important variables experimentally in such a way as to demonstrate their deleterious effects in order to assess the effects of multiple risk factors on developmental outcomes. Instead, we must use observationalkorrelationanalyses and try to model the phenomenon statistically. A statistical model is simply a formalized presentation of ideas by specifying the mathematical relationship among variables observed to have a bearing on the phenomenon.

A.

Bivariate Models

In bivariate statistics, one may be interested in testing the strength of an association (e.g., by a correlation coefficient) in order to discern whether such associations can occur by chance. They afford a helpful descriptive statistic. but do not permit complex assessments among “predictor” (independent) variables and “outcome” (dependent) variables. When multiple predictors are involved, one must also be concerned about the et‘fects of interactions among variables and the effects of cofactors which may affect the accuracy and precision of a prediction. Interaction refers to the situation where the relationship among variables is different at different values among them. Cofactors are systematic sources of error when estimating relationships. If a cofactor is related to the predictor and the outcome variable, it is called a confounder. For instance, a child’s cognitive development might be at risk if his mother were an alcoholic because alcohol is directly toxic to the fetus or because alcohol may also affect the mother’s ability to care for the child pre- and postnatally. If we want to assess the toxic effects due to fetal exposure, we must control for confounding due to the mother’s diminished care-giving capacity. This requires a more sophisticated statistical modeling strategy. None of the studies reviewed used a correlation model because multiple predictor variables needed to be entered into the statistical model. B.

Multiple Regression Models

The multiple regression model is an extension of the bivariate model in which several predictor variables are used to predict a single outcome variable, while controlling for interactions and confounders statistically (Kleinbaum et ul., 1982). In our example above, the effects of fetal alcohol on the child’s IQ might be tested while controlling for mother’s caregiving ability. Here, mother’s 1Q and care-giving ability are considered nuisance variables. This appraoch has been debated because there is an issue about when a person has “undercontrolled” or “overcontrolled” for confounders. Some statisticians argue that selection of confounders should be based on developmental theory and previous research, while others feel that selection should be based on the empirical results in the data being examined. This becomes a very difficult problem when several independent variables are interrelated (e.g., SES, maternal IQ, and caregiving practices), where one variable might potentiate or suppress the effect of another. In such a case, one may create “latent” variables or “factors” by combining such variables statistically and treating them as a single factor, as. for instance, is done in factor analysis. Multiple regression

METHODOLOGY OF NEUROTOXIC RISK FACTORS

13

was used in many of the studies reviewed (Bradley et ul., 1989; Ramey & Gown, 1984; Sameroff & Seifer, 1987; Schraeder, Rappaport, & CourtWright, 1987). C.

Structural Equation Models

There are several extensions of the multiple regression model that have specific applications (e.g., multivariate regressional models and canonical correlation models) which we will not discuss here. But one that is often used in longitudinal studies of developmental risk and outcomes is the structural equation model(Bentler,1980;Seifer& SameroffJ982). The multiple regression model is good for assessing a cross section of developmental outcomes at a fixed point in time. However, we are often interested in longitudinal patterns of outcomes or changes over time and we want to know if the effects of covariates are getting stronger or weaker. Structural equations are a powerful generalization of the multiple regression model, where a set of “dependent” or outcome variables either (a) are hypothesized to be related to different sets of “independent” or predictor variables or (b) some of the dependent measures are specified to have influences on certain of the other “dependent” variables in the proposed statistical model. Variables may not all be the same (dependent/ independent) in each model. Statistical models may jointly specify the presumed relationships for the “dependent” variables. In this latter case, certain “dependent” variables actually serve as antecedents of other outcomes. Using our example from fetal alcohol syndrome, the structural equation model might attempt to specify whether the effects of alcohol on the mother’s cognitive ability preceded her becoming pregnant and/or engaging in poor prenatal care, which might impact directly or indirectly on developmental outcomes of the child in addition to a toxic insult directly to the fetus. One could then describe a path of several contributing variables, each appropriately weighted, to explain the child’s compromised cognitive development. Such data are often presented in a “path diagram” showing the various relationships among predictor and outcome variables. D.

Random Effects Models

These models are complementary to the previous models discussed, but they are designed for categorical data (occurrence vs. nonoccurrence) instead of data that vary on a continuous dimension (Laird & Ware, 1982; Swamy. 1971).They are also useful when serial measurements are nonlinear. For instance, the underlying relationship between predictors and out-

14

Stephen

R . Scliroeder

comes may vary substantially from one child to the next. One may respond positively to intervention, while another may respond negatively. The random effects model is designed to estimate parameters and standard errors of measurement for each individual, so as to allow inference to the population from which the sample has been drawn. We will not discuss the complex details of this model, because it is not currently in wide use in developmental risk-outcome studies, although it is likely to become popular in the future.

V.

SPECIFYING RISK FACTORS FOR DEVELOPMENTAL DISABILITIES DUE TO LEAD AND CADMIUM

Specifying neurotoxic risk factors involves (a) identifying exposure sources, (b) relating them to internal body burden, (c) identifying possible mechanisms of neurotoxicity, (d) description of various health effects and their relation to exposure levels, and (e) identifying the most sensitive population groups. A.

Lead

Lead is a good example of an environmental pollutant which, because of its ubiquity, is difficult to regulate. It is almost everywhere. It is in the earth’s crust, water, air, the food chain, and hundreds of commercial items to which we are exposed daily. Each of u s who lives in an industrialized country inhales, ingests, and excretes a substantial amount of lead daily. A certain amount remains in storage compartments in the body (i.e., the blood, soft tissue, bone, and teeth). The half-life in blood is on the order of days; in soft tissue, weeks; in bone, years; in teeth, it is present permanently (USEPA, 1986). Thus, there is concern about chronic cumulative effects of low levels of exposures, especially over the years of early childhood, as well as a single acute toxic dose (e.g.. as a result of eating leaded paint chips). 1. HEALTH EFFECTS OF LEAD

Health effects of lead are wide ranging. At its lowest detectable levels (
METHODOLOGY OF NEUROTOXIC RISK FACTORS

15

higher levels (ca. 40 pg/dl). Chronic high lead levels have also been associated with cancer. By far the most serious effects of concern are neurotoxic effects on the nervous system, that is, perceptual-motor, cognitive, and attentional deficits, slowed peripheral nerve conduction velocity, and changes in cortical evoked response activity at low levels (ca. 10 pg/dl). Overt symptoms of encephalopathy, coma, and death occur more often as the lead level increases above 50 kg/dl (USEPA, 1986). 2. NEUROTOXIC EFFECTS Neurotoxic effects have long been recognized as being among the more severe consequences of human lead exposure. Since the early 1900s, extensive research has focused on the elucidation of lead exposure levels associated with the induction of various types of neurotoxic effects and related issues, such as critical exposure periods for their induction and their persistence or reversibility. Such research, spanning more than 50 years, has provided increasing evidence indicating that progressively lower lead exposure levels, previously accepted as “safe,” are actually sufficient to cause notable neurotoxic effects. The neurotoxic effects of extremely high exposures, resulting in blood lead levels in excess of 80-100 pg/dl have been well documented, especially in regard to increased risk for fulminant lead encephalopathy (a well-known clinical syndrome characterized by overt symptoms such as gross ataxia, persistent vomiting, lethargy, stupor, convulsions, and coma). While the use of chelation therapy to mobilize the excretion of lead can be lifesaving, it does not prevent permanent sequelae among survivors after onset of symptoms. The persistence of neurological sequelae in cases of nonfatal lead encephalopathy has also been well established. The neurotoxic effects of subencephalopathic lead exposures in both human adults and children, however, continue to represent a major area of interest and controversy. Reflecting this, much research during the past 10-15 years has focused on the delineation of exposure-effect relationships. One focus has been on the occurrence of overt signs and symptoms of neurotoxicity in relation to other indicators of subencephalopathic overt lead intoxication. Another focus has been on the manifestation of more subtle, often difficult-to-detect indications of altered neurological functions in apparently asymptomatic (Le., not overtly lead poisoned) individuals.

3. DEFINING EXPOSURE-EFFECT OR DOSE-RESPONSE RELATlONSHIPS Defining exposure-effect or dose-response relationships between lead and particular neurotoxic responses in humans involves two basic steps.

16

Stephen R . Schrrwdcr

First, there must be an assessment of the internal lead burden resulting from external doses of lead received via various routes of exposure (such as air, water, food, occupational hazards, and house dust). Internal lead burdens may be indexed by lead concentrations in blood, teeth, or other tissue, or by other biological indicators. The second step involves an assessment of the relationship of internal exposure indices to behavioral or other types of neurophysiological responses. The difficulty of this task is reflected by current controversies over existing data (Rutter & Jones, 1983). Studies vary greatly in the quality of design, precision of assessment instruments, care in data collection, and appropriateness of statistical analyses employed. Many of these methodological problems are broadly common to research on toxic agents in general and not just to lead alone. Although epidemiological studies of lead’s effects have immediate environmental relevance at the human level, difficult problems are often associated with the interpretation of the findings, as we have noted in a recent extensive review (USEPA, 1986). The main problems are (a) inadequate markers of exposure to lead (Davis & Svendsgaard, 1987; Smith, Grant, & Sors, 1989),(b) insensitive measures of performance, (c) bias in selection of subjects, (d)inadequate handling of confounding covariates, (e) inappropriate statistical analyses, (f)inappropriate generalization and interpretation of results, and (g) the need for “blind” evaluations by experimenters and technicians. Each of these problems is briefly discussed below. Each major exposure route-food, water, air, dust, and soil-contributes to a person’s total daily intake of lead. The relative contribution of each exposure route, however, is difficult to ascertain; neurotoxic endpoint measurements, therefore, are most typically evaluated in relation to one or another indicator of overall internal lead boy burden. Subjects in epidemiological studies may be misclassified as to exposure level unless careful choices of exposure indices are made based on the hypotheses to be tested, the accuracy and precision of the biological media assays, and the collection and assay procedures employed. The most commonly used measure of internal dose is blood lead concentration, which varies as a function of age, sex, race, geographic location, and exposure. The blood lead level is a useful marker of current exposure but generally does not reflect cumulative body lead burdens as well as lead levels in teeth. Hair lead levels, measured in some human studies, are not viewed as reliable indicators of internal body burdens at this time (USEPA, 1986). Future research may identify a more standard exposure index, but it appears that a risk classification similar to that of the U.S. Centers for Disease Control (1985) in terms of blood lead and free erythrocyte protoporphyrin (FEP) levels will continue in the foreseeable future to be the standard approach

METHODOLOGY OF NEUROTOXIC RISK FACTORS

17

most often used for lead exposure screening and evaluation. Much of the controversy is, therefore, focused on defining dose-effect relationships for human neurotoxic effects in terms of blood lead levels versus pertinent tooth lead levels or hair lead levels. The frequency and timing of sampling for internal lead burdens represent another important factor in evaluating studies of lead effects on neurological and behavioral functions. For example, epidemiological studies often rely on blood lead and/or erythrocyte protoporphyrin (EP) levels determined at a single point in time to retrospectively estimate or characterize internal exposure histories of study populations that may have been exposed in the past to higher levels of lead than those indicated by a single current blood sample. Relatively few prospective studies exist that provide highly reliable estimates of critical lead exposure levels associated with observed neurotoxic effects in human adults or children, especially in regard to the effects of subencephalopathic lead exposures. Some prospective longitudinal studies on the effects of lead on early development of infants and young children are currently in progress, but results of these studies are only beginning to become available (Davis & Svendsgaard. 1987; Smith e? al., 1989).The present assessment of the neurotoxic effects of lead in humans must, therefore, rely most heavily on published epidemiological studies which typically provide exposure history information of only limited value in defining exposure-effect relationships and less than optimum cross-sectional study designs. Key variables that have emerged in determiningeffectsoflead on the nervous system including the duration and intensity of exposure and age at exposure. Much evidence suggests that young organisms with developing nervous systems are more vulnerable than adults with fully matured nervous systems. Particular attention is, therefore, called to children as a special group at risk to neurotoxic effects of lead. Since lead readily crosses the placental barrier, fetal exposure is also a matter of great concern. 4. PRECISION OF MEASUREMENT

Precision of measurement is a critical methodological issue, especially when research on neurotoxicity leaves the laboratory setting. Neurotoxicity is often measured indirectly with psychometric or neurometric techniques in epidemiological studies. The accuracy with which these tests reflect what they purport to measure (validity) and the degree to which they are reproducible (reliability)are issues central to the science of measurement theory. Many cross-sectional population studies make use of instruments that only briefly sample behavior thought to be representative of some relatively constant underlying traits such as intelligence.

Standardization of tests is the subject of much research in psychometrics. The quality and precision of specific test batteries have been particularly controversial issues in evaluating possible levels of lead exposure that could cause neurotoxic effects in children. Assessments should place most weight on results obtained with age-normed, standardized psychometric test instruments and well-controlled, standardized nerve conduction velocity (NCV) tests. Other measures, such as reaction time, fingertapping. and certain electrophysiological measures (e.g., cortical-evoked and slow-wave potentials), are potentially more sensitive indices, but are still experimental measures whose clinical utility and psychometric properties with respect to the neurobehavioral toxicity of lead need to be explored more fully.

5. SELECTION BIAS Selection bias is a critical issue in epidemiological studies in which attempts are made to generalize from a small sample to a large population. Volunteering to participate in a study and attendance at special clinics or schools are common forms of selection bias that often limit how far the results of such studies can be generalized. These factors may need to be balanced in lead neurotoxicity research since reference groups are often difficult to find because of the pervasiveness of lead in the environment and the many nonlead covariates that also affect performance. Selection bias and the effects of confounding can be reduced by choosing a more homogeneous stratitied sample, but the generalizability of the results of such cohort studies is thereby limited. 6. CONTROLLING FOR CONFOUNDING COFACTORS

Perhaps the greatest methodological concern in epidemiological studies is controlling for confounding cofactors, so that residual effects can be more confidently attributed to lead. Among adults, the most important covariates are age, sex. race, educational level, exposure history, alcohol intake, total food intake, dietary calcium and iron intake, and urban versus rural styles of living (USEPA, 1986). Among children, a number of developmental covariates are additionally important: parental SES; maternal IQ; pica; quality of the care-giving environment; dietary iron and calcium intake, vitamin D levels, body fat, and nutrition; and age at exposure. Preschool children between the ages of 1 and 5 years appear to be particularly vulnerable, in that the rate of accumulation of even a low body-lead burden is higher for them than for adults (National Academy of Sciences, Committee on Lead in the Human Environment, 1980). Potential confounding effects of cofactors become particularly important when trying to interpret threshold effects of lead and cadmium exposure.

METHODOLOGY OF NEUROTOXIC RISK FACTORS

19

Each covariate alone may not be significant but, when combined, may interact to pose a cumulative risk which could result in under- or overestimation of a small effect of lead (Ernhart, in press; Kleinbaum, Kupper, & Muller, 1988). Dealing with confounders is a complex problem with no single solution. For instance, Bellinger, Leviton, and Waternaux ( 1989) showed through computer simulation that, since social class is usually related to both lead level and IQ in lead studies, the relationship between lead and IQ can be over- or underestimated, depending on how social class is handled in the statistical regression analysis. This is probably due to the fact that social class does not lie on an interval scale and that the lead-1Q relationship is not homogeneous across all social strata.

7. STATISTICAL CONSIDERATIONS Statistical considerations important not only to lead and cadmium but to all epidemiological studies include adequate sample size, the use of multiple comparisons, and the use of multivariate analyses. Regarding sample size, false negative conclusions are at times drawn from small studies with low statistical power. It is often difficult and expensive to use large sample sizes in complex research such as that on lead neurotoxicity. This fact makes it all the more important to use sensitive assessment instruments which have a high level of discriminating power and can be combined into factors for multivariate analysis. Multiple statistical comparisons can then be made while reducing the likelihood of finding a certain number of significant differences by chance alone. This is a serious problem, because near-threshold effects are often small and variable. 8. BLIND CONDITIONS A final crucial issue in this and other neurotoxicology research revolves around the care taken to ensure that investigators are isolated from information that might identify subjects in terms of their lead exposure levels at the time of assessment and data recording. Unconscious biases, nonrandom errors, and arbitrary data correction and exclusion can be ruled out only if a study is performed under blind conditions or, preferably, double-blind conditions. 9. SUMMARY Since experimentation with lead exposure in humans is not possible, all such research is perforce observational in nature, whether it is retrospective or prospective. Experimental control is rarely possible. Therefore, there is no study without flaw. With the current state of the art, the above are considered minimum criteria for future studies in the area of

20

Stephen R . Schrordrr

neurobehavioral toxicology of lead exposure in humans. The need for behavior methodologists to be involved in lead research is apparent. B.

Cadmium

I . EXPOSURE SOURCES In terms of exposure sources, cadmium is a natural constituent of zinc ores, so it is a product of zinc refining (Chisolm, 1980). Though rarer than lead, it is nevertheless ubiquitous in the ecosystem. It is distributed widely in the earth’s crust, surface soils, underground water tables, and in surface waterways (USEPA. 1979). It is released into the atmosphere via mining and smelting, industrial uses (e.g., batteries), fossil fuel, diesel oil and gasoline use, and phosphate fertilizer and sewage-sludge disposal. It thereby makes its way into the food chain. It is generally not recycled, so it accumulates in urban centers and industrialized areas where these activities occur. Municipal incineration of waste materials is the single largest airborne cadmium exposure source (ca. 130 million tons annually), affecting the breathing of over an estimated 50,000,000 people in the United States. Widespread exposure to the toxicity of cadmium was not recognized until 1946, when Dr. Hagino, a Japanese physician who practiced downstream from a zinc- and lead-mining and smelting area, was visited by a number of patients with a painful bone disease which he called ltailtai By0 (Ouch! Ouch! Disease). This disease was traced to the high cadmium content of the rice grown near the smelters in the valley (Chisolm, 1980). Since then, the major food groups in which cadmium has been found are land plants, land animals and animal products, freshwater fish, marine fish and free-moving crustaceans, and shell fish. Another significant source of cadmium intake is active and passive cigarette smoking. Smoking can double the normal amount of cadmium inhalation. 2. HEALTH EFFECTS OF CADMIUM Health effects of cadmium are mainly on the lungs and kidneys. Renal tubular dysfunction is the chronic critical effect for the general population. It leads to tubular proteinuria. Other effects observed are reproductive system problems, that is, low birth weight (Huel, Boudene, & Ibrahim, 1981), immunosuppression effects, carcinogenic effects, and cardiovascular effects leading to chronic hypertension.

3. DOSE-EFFECT RELATIONSHIPS In terms of dose-effect relationships in humans, cadmium is different from lead in that it has a very long half-life (ca. 38 years), is more lethal

METHODOLOGY OF NEUROTOXIC RISK FACTORS

21

for its size and weight, and accumulates in soft tissue, chiefly the kidney cortex, rather than in bones or teeth where it might become inert. Unlike lead, there is no treatment like chelation therapy for cadmium to reverse body burden. Thus, even low-level accumulations over many years are a matter of considerable concern. Blood cadmium values have a toxicokinetic profile which prevent blood cadmium levels from being very useful in reflecting critical-organ exposure. Hair cadmium levels, like hair lead levels, have not been shown to reflect body burden accurately. Urinary cadmium levels seem the most appropriate measure of chronic exposure at present. Populations at risk in the United States are (a) those living near a point source (e.g., zinc, lead, iron, or steel smelters, trash incinerators), (b) workers in the cadmium industry exposed to cadmium oxide fumes, (c) heavy cigarette smokers, (d) and children in urban areas with a high content of cadmium in the atmosphere. Chronic cumulative exposure in children is of special concern. The neurotoxic effects of cadmium have been of much less concern because of its low uptake in the brain and neural tissue and the fact that cadmium does not cross the placental barrier to a great degree. Only a few human studies have been performed. Four studies using hair cadmium levels found a relationship between learning disabilities and cadmium (Ely, Mostardi, Woebkenberg, & Worstell, 1981; Pihl & Parkes, 1977; Stewart-Pinkham, 1989; Thatcher, Lester, McAlaster, & Horst, 1982). None of these studies controlled adequately for confounding, their sample sizes were small, they did not guard against ascertainment bias, and they did not control for exposure history or exposure to other heavy metals, as has been done in most recent lead studies. Nevertheless, they suggest that the effects of lead and cadmium are highly correlated and that lead and cadmium show the same gradient of effect on IQ performance in school-aged children, the same sensitivity to confounding variables, and the same increased susceptibility in the presence of zinc, iron, and vitamin D insufficiency. Several trace metals in the diet interact with cadmium by preventing its toxic accumulation in the body. Zinc especially shows this protective effect against cadmium (Buell, 1975) and has been claimed as an important dietary supplement to enhance brain functioning and cognitive performance (Thatcher & Lester, 1985). A number of other metals also show complex interactions with cadmium (e.g., selenium, calcium, iron, copper, and lead) which might indirectly mitigate its neurotoxic effects (USEPA, 1979). None of the required dietary studies to prove this have been done in humans. It may well be, therefore, that the neurotoxic effects of cadmium have been underinvestigated and underestimated be-

22

Srrphen

R . Schroeder

cause of the heavy emphasis on its covariable, lead. This could be an important area for future research.

VI.

A.

RECURRING ISSUES IN NEUROTOXICOLOGY OF LEAD AND CADMIUM

Lead

An assessmcnt of the impact of lead on human and animal neurobehavioral function raises a number of issues. Among the key points addressed here are (a) the internal exposure levels, as indexed by blood levels, at which various adverse neurobehavioral effects occur; (b) the reversibility of such deleterious effects; (c) the populations that appear to be most susceptible to neural damage; and, finally, (d) the question as to the utility of using animal studies to draw parallels to the human condition. 1. INTERNAL EXPOSURE LEVELS AT WHICH ADVERSE

NEUROBEHAVIORAL EFFECTS OCCUR Markedly elevated blood lead levels are associated with neurotoxic effects (including severe, irreversible brain damage as indexed by the occurrence of acute and/or chronic encephalopathic symptoms) in both humans and animals. For most adult humans, such damage typically does not occur until blood lead levels exceed 120 pg/dl. Evidence does exist, however, for acute encephalopathy and death occurring in some human adults at blood lead levels below 120 p,g/dl, down to about 100 pg/dl. In children, effective blood lead levels for producing encephalopathy or death are somewhat lower, encephalopathy signs and symptoms having been reported for some children at blood levels as low as 80-100 pg/dl. It should be emphasized that, once encephalopathy occurs, death is not an improbable outcome, regardless of the quality of medical treatment available at the time of acute crisis. In fact, certain diagnostic or treatment procedures themselves tend to exacerbate matters and push the outcome toward fatality if the nature and severity of the problem are not fully recognized or properly diagnosed. It is also crucial to note the rapidity with which acute encephalopathic symptoms can develop or death can occur in apparently asymptomatic individuals or in those apparently only mildly affected by elevated body burdens of lead. It is not unusual for rapid deterioration to occur, with convulsions or coma suddenly appearing and with progression to death within 48 hours. This strongly suggests that, even in apparently asymptomatic individuals, rather severe neural damage probably exists at high blood lead levels, although such darnage

METHODOLOGY OF NEUROTOXIC RISK FACTORS

23

is not yet overtly manifested in obvious encephalopathic symptoms. This conclusion is further supported by numerous studies showing that children with high blood levels (over 80-100 pg/dl), but not observed to manifest acute encephalopathic symptoms, are permanently cognitively impaired, as are most children who survive acute episodes of frank lead encephalopathy . There is an increasing amount of evidence that indicates that subencephalopathic lead intoxication in adults causes various overt neurological signs and symptoms at blood lead levels as low (30-50 pg/dl) as those at which other overt manifestations (e.g., gastrointestinal symptoms) of lead intoxication have been detected. In addition, among apparently asymptomatic. nonovertly lead-intoxicated adults, often more subtle (but important) central and peripheral nervous system effects (e.g., slowed nerve conduction velocities) have been observed at blood lead levels as low as 20 pg/dl. Other evidence confirms that various types of neural dysfunction exist in apparently asymptomatic children, sometimes at even lower blood lead levels. The body of studies on low- or moderate-level lead effects on neurobehavioral functions (USEPA, 1986) presents a rather impressive array of pointing toward average 5-point IQ decrements occurring in asymptomatic children at average blood levels of 50-70 pg/dl. Other evidence is indicative of average 1Q decrements of up to 4 points being associated with blood levels in a 30-50 pg/dl range. Below 30 pg/dl, the evidence for IQ decrements is quite mixed, with most studies showing no significant associations with lead once other confounding factors are controlled. Still, the I - to 2-point differences in 1Q seen with blood lead levels mainly in the 15-30 pg/dl range in some studies are suggestive of possible, very small lead effects that are typically dwarfed by other sociohereditary factors. Given the apparent nonspecific nature of some of the behavioral or neural effects probable at low levels of lead exposure, one would not expect to find striking differences in every instance. The lowest blood lead levels clearly associated with altered behavioral performance, both in apparently asymptomatic children and in developing rats and monkeys, generally appear to be in the range of 30-50 pg/dl. However, certain behavioral (e.g., reaction-time and reaction-behavior deficits) and electrophysiological (e.g., altered electroencephalogram [EEG] patterns, evoked potentials, and peripheral nerve conduction velocities) effects indicative of CNS and peripheral nerve functional pertubations have been reported at lower blood levels, supporting a likely continuous dose-response relationship between lead and neurotoxicity down to exposure levels as low as 15-30 pg/dl or, perhaps, somewhat lower. Timing, type, and duration of exposure are important factors in both

24

Stephen R . Schroeder

animal and human studies. It is often uncertain whether observed blood lead levels represent the levels that were responsible for observed behavioral deficits. Monitoring of lead exposures in pediatric subjects in all cases has been highly intermittent or nonexistent during the period of life preceding neurobehavioral assessment. In most studies of children, only one or two blood lead values are provided per subject. Tooth lead may be an important cumulative exposure index, but its modest, highly variable correlation to blood lead or FEP and to external exposure levels makes findings from various studies difficult to compare quantitatively. The complexity of the many important covariates and their interaction with dependent measures of modest validity (e.g., IQ tests) may also account for many of the discrepancies among the different studies. The precise medical or heatlh significance of the neuropsychological and electrophysiological effects associated with low-level lead exposure as reported in the above studies is difficult to state with confidence at this time. Observed 1Q deficits and other behavioral changes, although statistically significant in some studies, tend to be relatively small in most studies reported by investigators, but nevertheless may still affect the intellectual development, school performance, and social development of the affected children sufficiently to be regarded as adverse. This would be especially true if such impaired intellectual development or school performance and disrupted social development were reflective of persisting, long-term effects of low-level lead exposure in early childhood. The issue of persistence of such lead effects, however, remains to be more clearly resolved. Still, some study results reviewed above suggest that significant low-level lead-induced neurobehavioral and electrophysiological effects may, in fact, persist at least into later childhood, and a number of mammalian animal studies demonstrate long-term persistence into adulthood of neurological dysfunctions induced by relatively moderate- or low-level lead exposures early in postnatal development. 2. QUESTION OF IRREVERSIBILITY

Little research on humans is available on persistence of effects. Some work suggests the possibility of reversing mild forms of peripheral neuropathy in lead workers, but little is known regarding the reversibility of lead effects on CNS function in humans. A recent 2-year follow-up study by Otto et al. (1982) of 28 children of battery factory workers found a persistent relation between blood and lead and altered slow-wave voltage of cortical slow-wave potentials. Current human psychometric studies, however, will have to be supplemented by prospective longitudinal studies of the effects of lead on development in order to better elucidate per-

METHODOLOGY OF NEUROTOXIC RISK FACTORS

25

sistence or reversibility of neurotoxic effects of blood exposure early in infancy or childhood. Various animal studies provide evidence that alterations in neurobehavioral function may be long-lived, with such alterations being evident long after blood lead levels have returned to control levels. These persistent effects have been demonstrated in monkeys as well as rats under a variety of learning performance test paradigms. Such results are also consistent with morphological,electrophysiological,and biochemical studies on animals that suggest lasting changes in synaptogenesis, dendritic development, myelin and fiber tract formation, ionic mechanisms of neurotransmission, and energy metabolism. 3. EARLY DEVELOPMENT AND SUSCEPTIBILITY TO NEURAL DAMAGE On the question of early childhood vulnerability, the neurobehavioral data are consistent with morphological and biochemical studies of the susceptibility of the heme biosynthetic pathway to perturbation by lead. Various lines of evidence suggest that the order of susceptibility to neurotoxic effects of lead is young > adult and female > male. Animal studies also have pointed to the perinatal period of ontogeny as a particularly critical time for a variety of reasons: (a) it is a period of rapid development of the nervous system, (b) it is a period where good nutrition is particularly critical, and (c) it is a period where the care-giving environment is vital to normal development. However, the precise boundaries of a critical period for lead exposure are not yet clear and may vary depending on the species and function or end point that is being assessed. Nevertheless, there is general agreement that human infants and toddlers below the age of 3 years are at special risk because of in utero exposure, increased opportunity for exposure because of normal mouthing behavior of lead-containing objects, and increased rates of lead absorption due to various factors (e.g., iron and deficiencies). 4. UTILITY OF ANIMAL STUDIES IN DRAWING

PARALLELS TO THE HUMAN CONDITION Animal models are used to shed light on questions where it would be impractical or ethically unacceptable to use human subjects. This is particularly true in the case of exposure to environmental toxins such as lead. In the case of lead, it has been most effective and convenient to expose developing animals via their mothers’ milk or by gastric gavage, at least until weaning. Very often, the exposure is continued in the water

26

Slrphen R . Schroeder

or food for some time beyond weaning. This approach does succeed in simulating at least two features commonly found in human exposure: oral intake and exposure during early development. The preweaning postnatal period in rats and mice is of particular relevance in terms of parallels with the first 2 years or so of human brain development. However, important questions exist concerning the comparability of animal models to humans. Given differences between humans, rats. and monkeys in heme chemistry, metabolism, and other aspects of physiology and anatomy, it is difficult to state what constitutes an equivalent internal exposure level (much less an equivalent external exposure level). For example, is a blood lead level of 30 kg/dl in a suckling rat equivalent to 30 Fddl in a 3-year-old child? Until an answer is available to this question (i.e., until the function describing the relationship of exposure indices in different species is available), the utility of animal models for deriving dose-response functions relevant to humans will be limited. Questions also exist regarding the comparability of neurobehavioral effects in animals with human behavior and cognitive function. One difficulty in comparing behavioral end points such as locomotor activity is the lack of a consistent operational definition. In addition to the lack of standardized methodologies, behavior is notoriously difficult to “equate” or compare meaningfully across species because behavioral analogies do not demonstrate behavioral homologies. Thus, it is improper to assume, without knowing more about the responsible underlying neurological structures and processes, that a rat’s performance on an operant conditioning schedule or a monkey’s performance on a stimulus discrimination task necessarily corresponds directly to a child’s performance on a cognitive function test. Nevertheless, deficits in performance by mammalian animals on such tasks are probably indicative of altered CNS functions, and are likely to parallel some type of altered CNS function in humans as well. In terms of morphological findings, there are reports of hippocampal lesions in both lead-exposed rats and humans that are consistent with a number of independent behavioral findings suggesting an impaired ability to respond appropriately to altered contingencies for rewards. That is, subjects with hippocampal damage tend to persist in certain patterns of behavior even when changed conditions make the behavior inappropriate; the same sort of tendency seems to be common to a number of leadinduced behavioral effects, including deficits in passive avoidance, operant extinction, visual discrimination, an various other discrimination reversal tasks. Other morphological findings in animals, such as demyelination and glial cell decline, are comparable to human neuropathological observations only at relatively high exposure levels. Another neurobehavioral end point of interest in comparing human and

METHODOLOGY OF NEUROTOXIC RISK FACTORS

27

animal neurotoxicity of lead is electrophysiological function. Alterations of electroencephalographic patterns and cortical slow-wave voltage have been reported for lead-exposed children, and various electrophysiological alterations both in vivo (e.g., in rat visual-evoked response) and in vitro (e.g..in frog miniature end-plate potentials) have also been noted in laboratory animals. Thus far, however, these lines of work have not converged sufficiently to allow for much in the way of definitive conclusions regarding electrophysiological aspects of lead neurotoxicity. Biochemical approaches to the experimental study of lead effects on the nervous system have been basically limited to laboratory animal subjects. Although their linkage to human neurobehavioral function is at this point somewhat speculative, such studies do provide insight on possible neurochemical intermediaries of lead neurotoxicity. N o single neurotransmitter system has been shown to be particularly sensitive to the effects of lead exposure, but lead-induced alterations have been demonstrated in various neurotransmitters, including dopamine, norepinephrine, serotonin, and y-aminobutyric acid. In addition, lead has been shown to have subcellular effects in the CNS at the level of mitochondria1 function and protein synthesis. In particular, some work has indicated that delays seen in cortical synaptogenesis and metabolic maturation following prenatal lead exposure may well underlie the delayed development of exploratory and locomotor function seen in other studies of the neurobehavioral effects of lead. Further studies on the correlation between blood lead values in humans and lead-induced disruptions of tetrahydrobiopterin metabolism indicate that subsequent interference with neurotransmitter formation may be linked to small reductions in IQ scores. Given the difficulties in formulating a comparative basis for internal exposure levels among different species, the primary value of many animal studies, particularly in vitro studies, may be in the information they can provide on basic mechanisms involved in lead neurotoxicity. A number of key in vitro studies are summarized in the lead criterion document from USEPA (1986). These studies show that significant, potentially deleterious, effects on nervous system function occur at in situ lead concentrations of 5 pg/dl and possibly lower. This suggests that, at least intracellularly or at a molecular level, there may exist essentially no threshold for certain neurochemical effects of lead. The relationship between blood lead levels and lead concentrations of extra- or intracellular sites of action, however, remains to be determined. Despite the problems in generalizing from animals to humans, both the animal and the human studies show considerable consistency in that they both support a continuous dose-response functional relationship between lead and neurotoxic biochemical, morphological, electrophysiological, and behavioral effects.

28

Stephen

R . SchroeJer

6. Cadmium

What little evidence there is for the neurotoxic effects of cadmium comes mostly from in vitro animal studies. It has been implicated in impaired olfactory function (Friberg, 1950), in presynaptic suppression of acetylcholine release (Chen, 1975) from motor nerve terminals, and in decreases in regional brain serotonin (Hrdina, Peters, & Singhal, 1976). The lack of more research on the neurotoxic effects of cadmium is surprising, given its long half-life, irreversibility of effects, and its lack of any apparent physiological benefit for the organism.

VII.

SUMMARY

Implications of Chronic Neurotoxic Low-Level Exposure The risks due to high-level neurotoxic exposures (i.e., encephalopathy, coma, and death) are not controversial. Determination of the relative plasticity of the CNS and the reversibility of such gross neural insults is a matter of empirical research on an established risk factor. Much more disputed are the risks due to low-level neurotoxic exposures and their interactions with biological and social risk factors to development over the life-span. Some researchers minimize the effects of neurotoxins such as lead relative to the many other risks and threats to child development. Yet, as Weiss (1980) notes, the culmination of early chronic exposure to neurotoxins such as lead and cadmium may result in the behavioral expression of diminished functional capacity, for all persons affected. A useful analogy is to consider the slow, but exponential, deteriorating of functional CNS capacity in aging. The compounding of irreversible neurotoxic insults even with a normal trajectory of aging is likely to have an exponential, not just an additive, effect. Furthermore, it tends to affect the culturally disadvantaged and other more vulnerable groups in society the most. These are man-made risks that are preventable through environmental regulation. They are unique opportunities for prevention of developmental disabilities which should not be lost. REFERENCES Saumeister, A. A.. Dokecki. P. R., & Kupstas, F. D. (1988). Prevenfing /he newmorhidi/y: A guide for state planning for the prevention of mentul returdution und reluted disubilities ussociuted with socioeconomic conditions. Washington, DC: President's Committee on Mental Retardation. Baumeister, A. A , , & Kupstas. F. (1987, April). The new morbidity: Imp/icu/ion.s.forprevenrion find umelioration. Paper presented to the Royal Society of Medicine, London.

METHODOLOGY OF NEUROTOXIC RISK FACTORS

29

Bellinger, D.. Leviton. A.. & Waternaux, C. (1989). Lead, IQ. and social class. Internationul Journul of Epidemiology, 18, 180-185.

Bentler. P. M. ( 1980). Multivariate analysis with latent variables: Causal modeling. Annual Review c?f Psychology. 31, 419-456.

Bradley, R. H.. Caldwell, B. M., Rock. S. L., Barnard, K. E., Gray, C., Hammond, M. A.. Mitchell, S., Siegel, L., Ramey, C. T.. Gottfried, A. W.,& Johnson, D. L. (1989). Home environment and cognitive development in the first years of life: A collaborative study involving six sites and three ethnic groups in North America. Developmental P.sycholo~y,25, 217-235.

Buell. G. (1975). Some biological aspects of cadmium toxicity. Journal of Occupulional Medicine. 17, 193.

Chen. S. S. (1975). Effects of divalent cations and of catecholamines on the late response of the superior cervical ganglion. Journal of Physiology (London). 253, 443-457. Chisolm, J. J. (1980). Poisoning from heavy metals (mercury, lead, and cadmium). Pediurric Annuls, 9, 28-42.

Cory-Slechta. D. A. (in press). Developmental behavioral toxicology of metals: Recurring issues and emerging answers. In G. Melton, T.Sonderegger. & S. R. Schroeder (Eds.), Eehuviorul toxicology of childhood (Vol. I ). Lincoln: University of Nebraska Press. Cowan, L. D.. & Leviton, A. (1980). Epidemiologic considerations in the study of the sequelae of low level lead exposure. In H. L. Needleman (Ed.), Low level lead exposure: The clinicul implicutions of current research. New York: Raven. Davis, J. M.. & Svendsgaard. D. J. (1987). Lead and child development. Nature (London). 329, 297-300.

Ely. D. L., Mostardi, R. A., Woebkenberg, N., & Worstell. D. (1981). Aerometric and hair trace metal content in learning-disabled children. Environmenral Research, 25, 225239.

Emhart. C. B. (in press). Cofactors in observational research: Issues and examples from the lead effects literature. In G. Melton, T. Sonderegger, & S. R. Schroeder (Eds.), Eehuviorul toxicology of childhood. Lincoln: University of Nebraska Press. Friberg, L. (1950). Health hazards in the manufacture of alkaline accumulators, with special reference to chronic cadmium poisoning. Actu Medica Scandinavica, Supplementitm, 40, 1-124. Graham, M. A.. & Scott, K. G. (1988). The impact of definitiosn of high risk on services to infants and toddlers. Topirs in Eurly Childhood Special Education, w3). 23-38. Horowitz, F. D. (1988. April). The concept at risk: A re-evuluation. Paper presented at the Society for Research in Child Development, Kansas City. Hrdina, P. D., Peters. D. A., & Singhal, R. L. (1976). Effects of chronic exposure to cadmium. lead and mercury on brain biogenic amines in the rat. Research in Community Chemistry. Pathology, and Phurmucology. 15(3). 483493.

Huel, G.. Boudene, C., & Ibrahim, M. A. (1981). Cadmium and lead content of maternal and newborn hair: Relationship to parity, birth weight. and hypertension. Archives of Environmen/ul Health, 36, 22 1-227.

King. E. H.. Logsdon, D. A., & Schroeder, S. R. (1990). Risk factors for developmental deluy among infants and toddlers. Manuscript submitted for publication. Kleinbaum. D. G., Kupper, L.L.,& Morgenstern, H. (1982). Epidemiologic reseurch. London: Wadsworth. Kleinbaum, D. C . , Kupper, L. L., & Muller, K. E. (1988). Applied regression analysis and other mulrivariute methods (2nd ed.). Boston: PWS-Kent. Laird, N. M., & Ware, J. H. (1982). "Random effects models for longitudinal data. Eiomerrics, 38, %3-974.

30

Stephen

R . Schroeder

National Academy of Sciences. Committee on Lead in the Human Environment (1980). Lead in the humun environment. Washington. DC: National Academy of Sciences. Needleman, H. L. (1989). The persistent threat of lead: A singular opportunity. Americun Journal of Public Health, 79, 643-645.

Nordberg. G. F. (Ed.). (1976). Effects and dose-response relationships qf heavy metols. Amsterdam: Elsevier. Otto, D.. Benignus, V., Muller, K., Barton, C., Seiple, K.. Prah, J., & Schroeder, S. ( 1982). Effects of low t o moderate lead exposure on slow cortical potentials in young children: Two year follow-up study. Neurobehavioral Toxicology and Terutology, 4, 733-737. Pihl, R. 0.. & Parkes. M. (1977). Hair element content in learning disabled children. Srie w e , 198, 204-206.

Ramey, C. T.. & Gown, J. W. (1984). A general systems approach to modifying risk for retarded development. Early Child Development and Cure, 16, 9-26. Ramey. C. T., Yeates, K. 0.. & MacPhee, D. (1984). Risk for retarded development among disadvantaged families: A systems theory approach to preventive intervention. Advances in Special Education. 4, 249-272.

Rutter, M.. & Jones, R. R. (Eds.). (1983). Lead versus heulth. New York: Wiley. Sameroff, A. J., & Chandler, M. J. (1975). Reproductive risk and the continuum of caretaking casualty. In F. D. Horowitz, M. Hetherington, S. Scarr-Salapatek, & G. Siege1 (Eds.), Review of child development reseurch (Vol. 4, pp. 187-242). Chicago: Chicago University Press. Sameroff. A. J., & Seifer, R. (1983). Familial risk and child competence. Child Devrlopmen!, 54, 1254-1268.

Schraeder. 9. D., Rappaport, J.. & Courtwright, L. (1987). Preschool development of very low birthweight infants. Image: Journal qf Nursing Scholarship. 19, 174-178. Schroeder, S. R. (in press). Psychological issues related to human neurotoxicity due to lead exposure: A behavioral scientist’s view of the regulatory process. In G. Melton, T. Sonderegger, & S. R. Schroeder (Eds.), Behavioral toxicology of childhood. Lincoln: University of Nebraska Press. Seifer, R., & Sameroff, A. J. (1982). A structural equation model analysis of competence in children at risk for mental disorder. In H. A. Moss, R. Hess, & C. Swift (Eds.). Eur1.v intervention programs for infants (pp. 85-97). New York: Haworth. Smith, M. A., Grant, L. D., & Sors, A. E. (1989). Lead exposure and child development. London: Kluwer Academic Publishers. Stewart-Pinkham, S. M. (1989). The effect of ambient cadmium air pollution on the hair mineral content of children. Science ofthe Total Environment. 78, 289-2%. Swamy. P. A. V. 9. (1971). Statisticol inference in random coeflicient regression model. Berlin: Springer-Verlag. Thatcher, R. W.,& Lester. M. L. (1985). Nutrition, environmental toxins and computerized EEG: A mini-max approach to learning disabilities. Juurnul of Learning Disabilities, 18, 287-297.

Thatcher, R. W., Lester, M. L., McAlaster, R., & Horst. R. (1982). Effects of low levels of cadmium and lead on cognitive functioning in children. Archives of Environmentul Health, 37, 159-166.

Tjossem, T. D. (Ed.). (1976). Intervention strategies for high risk infants and young children (pp. 3-33). Baltimore, MD: University Park Press. U.S. Centers for Disease Control (1985). Preventing leud poisoning in young children: A statement by the Centers for Diseuse Control. Januury 1985 (pp. 99-2230). Atlanta, GA: U.S. Department of Health and human Services. U .S. Environmental Protection Agency ( 1979). Health assessment documentfor cudmiurn. EPA-600/8-79-003. Research Triangle Park, North Carolina.

METHODOLOGY OF NEUROTOXIC RISK FACTORS

31

U.S. Environmental Protection Agency (1986). Air quality criteria for lead (Vol. IV). EPA-

600/8-83-028A. Washington, DC. Weiss. 9. (1978). Behavior as a sentry for metal toxicity. In J. Rose (Ed.), Trace elements in heulth (pp. 185-1961. London: Butterworths. Weiss. B. (1980). Conceptual issues in the assessment of lead toxicity. In H . Needleman (Ed.). Low level lead exposure: The clinicul implications of current research. New York: Raven Press.

This Page Intentionally Left Blank

The Role of Methylmercury Toxicity in Mental Retardation GARY J. MYERS DEPARTMENTS OF PEDIATRICS AND NEUROLOGY UNIVERSITY OF ALABAMA AT BIRMINGHAM BIRMINGHAM. ALABAMA 35294

DAVID 0. MARSH DEPARTMENT OF NEUROLOGY UNIVERSITY OF ROCHESTER SCHOOL OF MEDICINE ROCHESTER. NEW YORK 14642

1.

INTRODUCTION

Development of the central nervous system begins within days of conception, and continues throughout gestation. During pregnancy, the fetus grows and develops in an environment protected by the maternal body with provision of nourishment and the removal of waste products handled exclusively by the mother’s circulatory system. The development of the fetus could be adversely affected if toxic substances are introduced into this environment during pregnancy. The embryopathic effects of toxic substances such as alcohol were suspected even in biblical times. In the Old Testament (Judges 13:7), it is commanded, “Behold thou shalt conceive and bear a son; and now drink no wine or strong drink.” Contamination of the fetal environment by another toxic substance, lead, was recognized as early as the beginning of the 20th century (Oliver, 191 I). Factory inspectors in Britain noted that exposure to lead increased stillbirths and abortions, and was associated with macrocephaly of the newborn. Both alcohol and lead readily pass from the maternal circulation through the placental barrier and enter the fetus. A host of other chemical 33 INTERNATIONAL REVIEW OF RESEARCH IN MENTAL. RETARDATION. Vol. 16

Copyright Q 1990 by Academic Press. Inc. All rights of reproduction in any form reserved.

34

Gary J . Mvers and h v i d 0.Mmdi

substances also pass into the fetus. These substances may be teratological, or neurotoxic, or have other effects. Most metals (e.g., arsenic. mercury, cadmium, chromium, lithium, thallium, aluminum, nickel, manganese, and tin) are known to affect the nervous system (Clarkson. 1987). Many of these are also known to enter the nervous system of experimental animals prenatally, and some have been associated with embryopathic or teratological effects in humans (Clarkson, Nordberg, & Sager, 1985). The list of chemical substances and elements that can have an adverse effect on the fetal nervous system is growing steadily. Severe effects, such as men?al retardation, hydrocephalus, and cerebral palsy, have been clearly related to alcohol (Jones. Smith, Ulleland, & Streissguth, 1973), lead, and other toxins affecting the fetus. Recently, evidence has shown that even the exposure of pregnant mothers to low levels of lead may have adverse effects in humans (Bellinger, Leviton, Waternaux, Needleman, & Robinowitz, 1987; Needleman, Robinowitz, Leviton, Linn, & Schoenbaum, 1984; Schroeder, this volume). Additionally, more subtle deficits produced by toxic substances such as methylmercury have been recognized in experimental animals (Spyker, 1972; Weiss, 1975). Environmental mercury is both of natural origin and from contamination caused by human activities (Clegg, 1971). Accidental experiments have indicated that organic mercury, specifically methylmercury, can cause a spectrum of human neurological dysfunction. The dose-response curve for fetal effects of methylmercury ranges from severe mental retardation and cerebral palsy at high exposures to milder cognitive impairments at lower doses. This fetal sensitivity and the fact that small concentrations of methylmercury are present in all fish has led to concern that a high intake of methylmercury in fish may be a major public health hazard. If this is true, it has potentially serious implications for large segments of the world’s population who depend on fish as an important protein source in their diet. In this article, the evidence supporting that concern and its relationship to mental retardation will be examined. II.

MERCURY

Mercury exists in a variety of chemical and physical forms. However, for evaluating its effects on humans, it can be divided into inorganic and organic forms. Inorganic mercury can exist in the metallic form, as vapor, or as various salts. Those compounds in which mercury is directly linked to a carbon atom by a covalent bond are considered organic mercury compounds. Among the organic mercury compounds, the short-chain al-

METHYLMERCURY TOXICITY AND MENTAL RETARDATION

35

kylmercurials are the most important toxic compounds for the human species, and, among the inorganic forms, mercury vapor is the most toxic. Methylmercury is a short-chain alkylmercurial. Mercury compounds, especially organic ones, readily enter the human body and pass into the nervous system. They are only slowly broken down and excreted. This may account for much of their toxicity (World Health Organization, 1976). Mercury is widely distributed in nature, but is generally present in only minute concentrations. The natural degassing of the Earth's crust is thought to put 25,000 tons of mercury a year into the environment. Industrial production of mercury was estimated to add an additional 10,000 tons in 1973 and the oceans are believed to contained over 70 million tons of mercury (Magos, 1975). Natural mercury deposits are found on most continents and local areas may be more heavily contaminated by accident. For example, Oak Ridge National Laboratory was unable to account for 2.4 million pounds of mercury which presumably was lost to the environment between 1953 and 1983 (Marshall, 1983). Large losses of mercury have also occurred in the past at most paper pulp factories. Such factories are widely distributed throughout the North American continent and elsewhere. 111.

MERCURY EXPOSURE

Humans can be exposed to mercury in many ways. Exposure to various compounds, such as methyl-, ethyl-, methoxyethyl-, or phenylmercury, in high concentrations may occur in a variety of industrial situations. The toxicity of certain organic mercurial compounds was first recognized in factories making these chemical compounds (Hunter, Bomford, & Russell, 1940). The toxicity of mercury vapor has been recognized for centuries. The mercury mines in Almaden, Spain, have been open since 700 BC, and the work schedule of miners is organized to reflect knowledge of mercury toxicity. Miners works only a certain number of hours per day and days per week, thus allowing their bodies to excrete mercury between exposures and reducing their chances of being poisoned (Goldwater, 1972). In recent years, scientists have recognized that methylation of inorganic mercury in the environment through the action of naturally occurring bacteria can produce the very toxic methylmercury. Consequently, methylmercury is found widely in nature. Mercury that enters the food chain concentrates in fish and nearly all of it is in the form of methylmercury. Methylmercury is efficiently absorbed through biological mem-

36

Guty J . M v u s iind Dwid 0.Mursh

branes. Biomagnification results in higher concentrations in fish such as shark and swordfish at the upper end of the food chain. Unfortunately, methylmercury is only slowly degraded and excreted from the living organism. The biological half-life for excretion of methylmercury is about 70 days in humans and primates; however, it is up to 1000 days for some species of fish. Since fish vary so widely in their diet and metabolism, it is difficult to predict the concentration of methylmercury in a given species. However, in general, carnivorous fish at the end of the food chain have the highest levels. While most fish have mercury levels below 200 pg/kg (0.2 ppm), species such as shark may reach levels of 5000 pg/kg ( 5 ppm) wet weight or even higher. Fish found in waters contaminated with mercury may have levels several times higher than those found naturally. The final stage of biomagnification occurs when humans consume fish that have high mercury levels. One source of methylmercury in terrestrial food chains is the use of mercury fungicides to preserve seed grain. If the seeds are planted and sprout, then the coated hull remains in the ground and the grain which grows is not contaminated. However, the coated seeds may be eaten by birds, rodents, and other small animals, and these may be eaten by carnivores such as hawks. Some outbreaks of human methylmercury poisoning (e.g., in Iraq) have been directly related to uninformed people eating seed grain treated with mercury compounds. After mercury enters the human body, it is widely distributed to the red blood cells, liver, kidney, brain, and other organs. The distribution is partly dependent on the specific type of mercury compound. The time between ingestion and toxicity may be from several weeks to months. Methylmercury is the most stable in the living organism. IV.

MEASUREMENT OF MERCURY IN HUMANS

Mercury entering the human body is rapidly taken up by the red blood cells. Blood taken near the time of mercury ingestion provides a good measure of the acute toxicity. If the intake of mercury is steady and does not vary over time, then random sampling of blood levels provides a fair measure of exposure. However, if the dose is pulsed, the blood level is only a useful index near the time of the exposure. Long-term exposure to mercury is reflected best by mercury deposits in hair. Mercury is deposited in growing hair in direct proportion to the blood concentration during the active hair growth phase. The deposits average about 250 times the amount present in the blood, with a range from below 200 to above 300

METHYLMERCURY TOXICITY AND MENTAL RETARDATION

37

(Inskip & Piotrowski, 1985). Once incorporated into hair, mercury is stable and not available for redistribution in the body. Thus, if the amount of mercury in the body rises acutely, the amount of mercury found in hair segments grown during that time period will be correspondingly high. However, as the blood levels decrease, the amount of mercury incorporated into the growing hair will decrease proportionally. Thus, when hair is analyzed segmentally, one can estimate the blood level of mercury during the growth of any specific segment of hair (Giovanoli-Jakubczak. Greenwood, Smith, & Clarkson, 1974). By segmental analysis, the maximum mercury level can be obtained and it can be determined whether it was a single exposure, multiple acute exposures, or a continuous high- or low-level exposure over a prolonged time period. When segmental measurement of hair mercury levels is combined with information about hair growth, it is possible to recapitulate the individual's exposure to mercury. If the individual has long hair, as is true of many women in developing countries, the mercury exposure and body burden over periods of months or even years can be retrospectively determined. This technique has been used extensively in field studies. The accuracy with which mercury exposures can be determined contrasts markedly with the exposure indices presently available for use with lead, alcohol, and many other toxic substances. Blood levels of lead and alcohol, for example, reflect the body burden for only a brief time span. Repeated blood levels over time would be needed to determine exposure accurately during pregnancy, but usually assumptions are made that one or two levels are representative. Teeth have been used in some studies to recapitulate lead exposure, but the correlation with blood or body levels or time of exposure are not accurately known. Such assumptions are not necessary with mercury. Hair analysis can provide an accurate recapitulation of the individual's exposure to mercury over time. The exposure of the fetus to mercury can thus be delineated. V.

CLINICAL SYNDROMES OF MERCURY POISONING

Inorganic mercury toxicity can follow a variety of industrial exposures. Neurologically, it usually presents as a tremor of the hands and neurasthenia (vague functional fatigue). There is also renal involvement and gingivitis. The mad hatter from Alice in Wonderland was a caricature of mercury vapor exposure. Mercury was used in the production of felt for hats, and hatters were constantly exposed to the vapor. When heat is applied to felt or mercury compounds, such as in shaping the hats, there is increased vaporization and consequently increased exposure of the hatter.

38

Gtiry

J . M y e r s and David 0. Mtirsh

In children, a syndrome referred to as acrodynia appears to be a hypersensitivity to inorganic mercury. The name derives from the erythema or redness of the hands and feet. These children also are irritable and have photophobia. This syndrome was relatively common when mercurous chloride was used in teething powders, but now it is rarely seen. Removing the children from continued exposure and giving chelating agents to increase the body excretion of mercury were the major treatments (Hirschman, Feingold, & Boylen, 1963). The toxicity of organic mercurials in humans differs from that of inorganic mercurials. Between exposure and symptomatology , there may be a lag phase of weeks to months followed by fairly rapid neurological changes. The clinical symptoms usually start with paresthesias of the limbs or visual field constriction and are rapidly followed by ataxia, deafness, spasticity, dysarthria, tremor, and other findings. If a large enough dose is taken, the patients progress to stupor, coma, and eventually death. Should the dose be less than lethal and the individual survive, spontaneous improvement may occur, but severe cases may be left with generalized central nervous system (CNS) deficits. In children, there may be arrest of further head growth and severe cerebral palsy and mental retardation. Fetal exposure to methylmercury occurs during pregnancy if there is an elevated level of methylmercury in the mother because transfer across the placenta into the fetus occurs readily (Koos & Longo, 1976). Indeed, levels of methylmercury in the fetus may be significantly greater than those in the mother, ranging up to 1.5 times the maternal concentration (Amin-Zaki et ul., 1974 a, 1974~;Bakir et ul., 1973). In the fetus, the methylmercury also concentrates in the brain, where it can damage tissues. Also, the growth and differentiation of the brain may be slowed or halted during this period of exposure to methylmercury. The result is that infants may be born with microcephaly, mental retardation, motor handicaps (cerebral palsy), and developmental delays (Snyder. 1971). One disturbing feature of this toxicity is that the mothers themselves may have minimal or no signs of symptoms, but they deliver infants who are seriously affected (Harada, 1968). Experimental studies of methylmercury have shown that in animals, the neurological effects are dose related and can be subtle. Spiker found that low doses of methylmercury in mice can produce neurological changes that are only recognized during behavioral testing (Spyker, 1972). Similar effects were found in rats treated prenatally with very low doses of methylmercury (Schalock, 1981). In the crab-eating macaque, fetal exposure to methylmercury at levels subclinical for the mother resulted in abnormal visual recognition in the infant (Gunderson, Grant,

METHYLMERCURY TOXICITY AND MENTAL RETARDATION

39

Burbacher, Fagan, & Mottet, 1986).These studies used the visual novelty preference paradigm developed by Fantz (1964) and found that macaques exposed in utero to methylmercury directed significantly less visual attention to novel stimuli than did controls. Snyder (1971) reported an infant born to a mother who had ingested methylmercury-contaminated pork from the third to the sixth month of her pregnancy. Several family members had clinical methylmercury poisoning, but the mother was asymptomatic. From birth, the infant had signs and symptoms of a diffuse central nervous system abnormality. The infant had an elevated urinary mercury level at birth, but this was normal by 6 weeks of age. The infant’s electroencephalograms (EEGs) were normal from 3 days to 2 months of age. Subsequently, however, all the EEGs were markedly abnormal. During a 9-year follow-up, the infant showed a severe developmental delay with myoclonic jerks and a marked spastic quadriplegia (Brenner & Snyder, 1980). I n this patient, there was no breast-feeding or other known postnatal exposure to methylmercury. The effects were thought to be entirely from in utero exposure. The infant’s urinary mercury level was elevated, but this is not a reliable measure of exposure and does not correlate with the body load. This incident suggested again the special sensitivity of the fetus to methylmercury. Korns (1972) described a single, but disputed, case of adult methylmercury poisoning from consumption of large quantities of swordfish. N o example of human fetal methylmercury toxicity, caused by maternal fish consumption, has been reported in the U.S. population.

VI. A.

OUTBREAKS OF HUMAN EXPOSURE

Japan (Minimata)

Although an isolated episode of in utero mercury poisoning had been reported earlier (Engleson & Herner. 19521, the first concern about methylmercury as an environmental toxin arose in the late 1950s. Starting in 1953, a strange illness was noted near Minimata Bay in Japan that caused unusual symptoms in fish, cats, and birds. The animals became ataxic and died (McAlpine & Araki, 1958). The disorder was a mystery for several years, but eventually its etiology was traced to intoxication with organic mercury. The Chisso Company factory near Minimata produced vinyl chloride and acetaldehyde, two intermediates in manufacturing plastics. The waste effluent from the factory contained inorganic mercury and methylmercury. The effluent was discharged into Minimata Bay and

40

Gary J . Myers and David

0.Mnrsh

heavily contaminated fish were consumed locally, predominantly by fishermen and their families. Fishing was banned in the Bay in 1956, even before the etiology of the disorder was identified. Early medical studies of human neurological problems associated with Minimata disease were reported by McAlpine and Araki (1958, 1959; Kurland, Faro, & Siedler, 1960). The fetal or congenital form of Minimata disease was recognized somewhat later by Harada (1968). His initial report was of 22 children born between 1956 and 1959 who had severe mental retardation and motor disturbances. In 17 of these families, the father was a fisherman. All were thought to have been exposed in iltero to excessive levels of methylmercury. None had other explanations for their cerebral dysfunction. Both maternal and infant hair levels of mercury were generally elevated. However, these were not measured until after 1960, some 2-6 years following the children’s births. Consequently, the level of fetal exposure could not be determined and no correlation with signs and symptoms could be made. About two-thirds of the infants had family members with acute Minimata disease. Thirteen of the 22 children had severe mental retardation, while 6 had moderate and 3 mild. The motor disturbance was severe in 7 children, moderate in 10, and mild in 5. A total of 40 Japanese congenital cases were eventually identified (Harada, 1976). These children were not known to have eaten contaminated fish. However, nearly all had been breast-fed by mothers who had elevated body levels of methylmercury. Maternal milk contains between 5 and 8% of the mercury level found in a simultaneous blood sample (Amin-Zaki et ul., 1976), so the infants could have received small amounts of mercury postnatally from breast milk. One disturbing feature of all the cases was the relatively asymptomatic state of their mothers. The only maternal abnormality described during pregnancy was a numb feeling in five mothers and hyperemesis in one. In addition, those cases diagnosed as congenital because of their early onset were generally more severely affected than those thought to have a postnatal onset of the disease. The possibility that the fetal brain was more sensitive to the toxin than the maternal brain or the brain at later ages was raised by these cases. However, the degree of exposure to methylmercury was unknown, and it was not possible to confirm this clinical impression. 6.

Iraq

In 1971, almost 100,OOOmetric tons of seed grain treated with a methylmercury fungicide was distributed widely in Iraq (Bakir et u l . , 1973). Some of it was used by farmers to make bread, and a large epidemic of mercury poisoning occurred. Some 6530 patients with poisoning were

METHYLMERCURY TOXICITY AND MENTAL RETARDATION

41

hospitalized, 459 of these died, and an unknown number of other people had poisoning. Early recognition of the course of this disorder and stringent warnings by the authorities probably limited the extent of the disaster. While Japanese families had consumed contaminated fish and shellfish for years, the toxic exposure in Iraq was different. In Iraq, the dose was pulsed and exposure occurred over a time period of less than 3 or 4 months. During the Iraq epidemic, attention was especially focused on fetal and postnatal consequences of methylmercury ingestion. Amin-Zaki et al. (1974a, 1974~)reported a series of 15 infants exposed in urero to methylmercury. They were all breast-fed, so the exposure may have continued postnatally. Six of the mothers had clinical manifestations of methylmercury poisoning, as did six of the infants. However, only one infant was affected while the mother remained asymptomatic, a finding that differed from the Japanese experience. A follow-up to this study expanded the series and reported 32 infants with prenatal methylmercury exposure followed over a 5-year period (Amin-Zaki er al., 1979). Fifteen of the 27 mothers who agreed to be examined had clinical manifestations of methylmercury poisoning. Fourteen of the 32 infants had clinical findings when first examined, but 18 infants were initially considered to be clinically normal. One of the 18, however, was microcephalic on initial examination. Those infants abnormal on the initial examination all had hyperreflexia, and some had extensor Babinski signs, microcephaly, and cerebral palsy. Five of those infants died before the age of 5 years. All of the remaining nine had delayed mental and motor development, usually severe. Of special interest were those felt to be normal on initial examination. Four of these children died before the age of 5 years. There were 9 who had speech delay, 3 with delayed mental development, 7 with motor delays, and 12 with hyperreflexia. This study showed that methylmercury can be responsible for a milder neurological syndrome than previously described. No fetal cases with minimal development of neurological findings were described in the Japanese studies. A separate group of 29 mothers who were pregnant during the epidemic were identified and their infants were followed to age 44-5 years (Marsh et al., 1979). This series differed from that of Amin-Zaki and colleagues (1979) in that the maternal peak hair mercury concentrations were generally lower. In the Amin-Zaki et al. (1979) series, the maternal peak hair concentrations ranged from 32 to 592 ppm with only 3 below 100 ppm. In the Marsh et al. (1979) series, the peak hair mercury of mofhers ranged from 2 to 384 ppm with only 9 mothers having a peak above 100 ppm. Among the 29 infants, there were only 5 who appeared normal and the mothers of these children all had peak hair mercury concentrations below

42

G c q J . Mvers cind David 0.Morsh

25 ppm. There were 2 children with severe neurological abnormalities and their mothers' peak hair mercury levels were 165 and 209 ppm. Fourteen of the mothers in this series were symptomatic. Thirteen of them had paresthesias, but 10 of these had no other symptoms of toxicity. Four mothers experienced weakness, impaired vision, ataxia, or difficulty walking during the pregnancy. It was clear that those having a maternal peak mercury concentration of 99 ppm or above had significantly more neurological abnormalities than those with lower exposure. This series agreed with the Japanese experience in that the mothers were relatively asymptomatic while the infants were sometimes seriously affected. Breast-feeding by the infant's mother occurred in all cases, so exposure of methylmercury was not totally prenatal. Methylmercury is excreted in breast milk, and the continued intake of methylmercury in breast milk slows the infant's excretion of mercury (Amin-Zaki et a / . , 1974b). This prolongs the period of time the infant is exposed to an elevated mercury level and may contribute to the toxic effect. Peak level and duration of exposure may both be important. In this series, however, peak levels above 99 ppm were closely correlated with delayed mental and motor development. Maternal alcohol consumption was not a variable in this Moslem population, and no other specific confounding factors could be identified. Marsh et 01. (1987) also described a group of 81 mother-infant pairs. For this report, the maternal peak concentration of hair mercury was measured in single strands by a new more accurate technique, X-ray fluorescence spectrometry, which provided better definition of peak concentration and duration of exposure. Using these new measurements, a significant correlation between peak levels and signs and symptoms was found in the infants. An increase in frequency of abnormalities was found at peak levels well below 99 ppm. Indeed, peak levels between 10 and 67.6 ppm correlated with abnormal findings in the children. These levels are considerably below those previously suspected of causing neurological abnormalities, and include the range achieved by individuals with a moderate intake of certain fish. lnvestigators began to suspect that methylmercury could be considerably more toxic than previously thought. However, they cautiously pointed out two things: First, a subacute exposure with a peak level may differ significantly from a longitudinal exposure resulting from regularly eating fish where there may be a relatively constant level of mercury in the mother's body throughout pregnancy; second, exposure in Iraq was to methylmercury as a fungicide. C.

Canada (Cree Indians)

In the early 197Os, elevated mercury levels were found in fish in Canadian lakes and rivers, especially those in northwestern Ontario and Que-

METHYLMERCURY TOXICITY AND MENTAL RETARDATION

43

bec (Wheatley, Barbeau, Clarkson, & Lapham, 1979). Subsequently, over 34,000 blood and hair samples were tested for mercury, and nearly 31% showed elevated mercury levels (over 20 ppb in blood). Blood levels were as high as 55 1 ppb, a level where clinical signs and symptoms might occur. However, clinical examinations of adults were unable to confirm toxicity, although there were conflicting opinions (Harada, 1976; Wheatley et < I / . , 1979). The wives of fishing guides, especially, were exposed to methylmercury from eating fish, and one newborn infant had a cord blood mercury level of 100 ppb (Clarkson, 1975). A study of prenatal exposure was undertaken and 234 Cree children were elevated without examiners’ knowledge of their mercury exposure (McKeown-Eyssen,Reudy, & Neims, 1983). Ninety-five percent of the children aged 12-30 months who were eligible for the study were examined. Four boys (3.5%) and 5 girls (4.1%) were found to have definite neurological abnormalities. An additional 9 boys and 9 girls had minor changes such as sluggish reflexes which were of unknown significance. When the neurological changes were compared to the maternal methylmercury exposure during pregnancy, a positive association was found only in boys. No confounding variables seemed to explain this relationship, although nearly two-thirds of the mothers were known to consume alcohol. In this study, a peak maternal hair mercury concentration between 13 and 24 ppm was associated with possible neurological abnormalities in Cree Indian boys exposed in urero to this toxin. This study suggested that methylmercury may cause neurological changes at levels considerably lower than previously thought toxic. VII.

RELATIONSHIP OF CLINICAL DISORDERS TO EXPOSURELEVELS

The first available dose-response data on congenital methylmercury poisoning came from studies on the Iraq epidemic. Previous fetal cases either did not have the indices of exposure measured adequately (Snyder, 1971; Engleson & Herner, 1952), or they were measured so late that it was not possible to recapitulate the exposure (Harada, 1968). However, in Iraq the cause of the epidemic was identified early, and thus both blood and hair mercury concentrations were obtained with clinical symptomatology. Iraqi data indicated that adults began having paresthesias, the earliest symptom, at blood mercury levels of 200-500 ppb (Bakir et al., 1973). This corresponds to hair mercury levels of 50-125 ppm. In adults, neurological dysfunction and irreversible damage do not occur until much higher levels of exposure. However, Amin-Zaki et al. (1974~)found that, in prenatal exposure, signs of neurological damage occurred at a blood

44

G ~ r yJ . Myers rind

David 0.Mtirsh

mercury concentration as low as 564 ppb. The findings in adults at this level of exposure were generally mild and reversible, but those in infants were more severe and fixed. Thirteen of the infants in the Amin-Zaki et al. (1979) study had mercury measured in whole blood within 60 days of birth, and 19 had mothers whose hair was long enough that mercury measurements recapitulated in ufero exposure. The maximum mercury concentration in mothers' hair for those 24 infants who were abnormal was from 32 to 532 ppm. In that series, the infant blood samples had mercury levels that ranged from 42 to 4220 ppb, but it is not clear at what level symptoms or signs first appeared. Peak maternal hair mercury concentrations were found in all 29 of the prenatally exposed infants reported by Marsh et al. in 1979. Levels above 99 ppm were significantly correlated with abnormal neurological findings. The series was later expanded to 84 and analyzed in groups using the frequency of signs of neurological damage (Marsh et al., 1981). When this was done, there was an increase of neurological signs and symptoms at a maternal peak hair concentration of 18 to 68 ppm. This level is considerably lower than previously suspected. A later reanalysis of these data was done measuring hair mercury concentrations by X-ray fluorescence analysis of single strands of hair rather than flameless atomic absorption, which requires clusters of hair and underestimates peak hair mercury concentrations (Cox et al., 1989. It was concluded that the lowest level for adverse effect was a maternal hair concentration of 10-15 ppm, a level very close to that found in the Canadian study. VIII.

FETAL SENSITIVITY

In vitro experimental studies undertaken to delineate the effects of methylmercury on neurons have shown effects at relatively low concentrations. Methylmercury inhibits DNA synthesis in fetal astrocytes, stops neuronal migration, impairs growth cones of neurites, and damages neurotubules (Choi, Cho, & Lapham, 1980. 1981). These are all processes which are essential in the developing brain to ensure that neurons migrate properly, align themselves in an organized manner, and form proper connections. Sager, Clarkson. and Nordberg ( 1986) reviewed experimental studies and concluded that methylmercury's known effect on microtubules in neurons and astrocytes may be responsible for these findings because microtubules are essential for the division of cells, cell migration, and extension of neural processes. In adult primates, methylmercury produces different neuropathological lesions depending on the dosage and duration of ingestion (Shaw, Mottet.

METHYLMERCURY TOXICITY AND MENTAL RETARDATION

45

Body, & Luschei, 1975). Moderate doses given daily for 17 days produced neuronal degeneration in deep nuclear groups of the thalamus, pons, and cerebellum, whereas smaller doses given daily for many months caused a pseudolaminar necrosis of the cerebral cortex. Some animals that received the low-dose chronic exposure had no neuropathological lesions apparent by light microscopy. These experiments do not address the question of developmental effects, but they do raise the issue of thresholds. Adult animal experiments suggest a higher threshold level of toxicity, but also that the length of time the animal is exposed is important. During the Japanese epidemic of mercury poisoning, brains of two infants with fetal Minimata disease were studied pathologically by Takeuchi (1968). He found important pathological changes in these brains which were not present in nonfetal cases. There was evidence of underdevelopment and malformation of the brain along with the atrophic changes characteristically seen at later ages. The focal findings usually present in the calcarine and paracentral cortex and cerebellum in affected adults and children were present in the fetal cases. However, there was much more widespread cortical involvement in the fetal cases so that the focal findings, even though present, tended to be obscured. Takeuchi felt the fetal pathological changes were similar to those at older ages, but more extensive. Following the mercury poisoning in Iraq, the brains of another two infants with in utero exposure to methylmercury were studied (Choi, Lapham, Amin-Zaki, & Saleem, 1978). The findings in these brains were those of abnormal neuronal organization, and were felt to be secondary to disturbances in neuronal migration. In addition, there were astrocytes present throughout the white matter which were considered evidence of a reaction to some destructive progress. Choi et al. postulated that the major effect of in utero methylmercury toxicity is a disruption of neuronal migration and cortical organization. Although they felt neuronal destruction had occurred, they saw the developmental interference as more significant. IX.

STUDIES OF METAL METHYLMERCURY EXPOSURE AT LOW LEVELS

As clinical evidence accumulated that the fetus was very sensitive to methylmercury exposure and experimental studies provided support and a rationale for this effect, interest in the threshold level for neurological abnormalities increased. Studies in Iraq suggested an effect level as low

as 10-15 ppm in hair, and studies in Canada were in the same range. Although the types of exposures were different, a pulsed dose in Iraq as opposed to a more uniform elevation from regularly eating fish in Canada, both studies concurred in their conclusions concerning the lowest effect level. Levels of this magnitude can be attained by regular fish consumption, especially if species of fish with higher levels of methylmercury like tuna, swordfish, and shark are consumed (Inskip & Piotrowski. 1985). Although the average daily consumption of fish varies widely, there are many populations throughout the world for whom fish is an important food source. I t is not necessary to consume fish contaminated with excessive levels of methylmercury to achieve hair levels of 10 to 15 ppm. In New Zealand, a survey of trace elements in the diet found that many women with elevated mercury levels regularly consumed fish (Dick, Hughes, Mitchell, & Davidson. 1978). Subsequently, Kjellstrom. Kennedy, Wallis, and Mantell (1986) undertook a study to determine if steady-state. low-level exposures of the human fetus to methylmercury have clinically detectable effects. They surveyed maternity hospitals and asked mothers to complete a dietary questionnaire and provide a hair sample. There were 10,930 mothers who completed the questionnaire and gave a hair sample, 20% said they ate fish once a week or more and I .S% ate fish daily. Subsequently, hair analysis identified 73 mothers with a hair mercury concentration of 6 ppm or greater. A case-reference study was then undertaken. Thirty-one of the children whose mothers had a mean hair mercury level of 6 ppm or greater were located when they were approximately 4 years of age and matched with a reference case. All of the children were given the D e n w r Devc~lopmrntalScreening Tcst and three sensory tests. Two of the sensory tests used light touch for finger identification and point localization and the third differentiated hot and cold. The authors found delays on the Denver Test and abnormalities on sensory testing more frequently in the exposed than in the reference population. Only 17% of the reference group had delays while 52% of the exposed group had a delay. Most of the exposed group had mothers with a mean hair mercury level of 6 to 12 ppm during pregnancy. This is equivalent to a peak maternal hair mercury level of about 9 to 18 ppm. In the New Zealand study, there were three ethnic groups and in several instances the reference case was mismatched. There were also problems in that the timing of the examinations of exposed and reference cases varied. Some of the reference cases were older and may have performed better on the Denver Test because of that. However, the authors felt that fetal effects of methylmercury exposure were detected at levels in the same range as identified in the Iraqi and Canadian studies. The public health significance of such a low level of methylmercury

METHYLMERCURY TOXICITY AND MENTAL RETARDATION

47

causing developmental delays is considerable. Consequently, the next step in this area of research would seem to be a well-planned prospective study of fetal effects conducted in a fish-eating population. This would require a large enough study population that a reliable estimate of risk can be made. The population should be cooperative so that a high proportion of all pregnancies can be included to avoid selection bias. Data on all pertinent confounding factors which might influence early child development should be collected. This would include information on parental education, socioeconomic status, intelligence, alcohol intake, and all maternal and perinatal factors, as well as illnesses after birth. Measures of outcome should include neurological and psychological tests of adequate sensitivity and discrimination to allow comparisons of outcome measures in relation to exposure as determined by maternal mercury levels during pregnancy. X.

SUMMARY

Mercury and many of its chemical forms are toxic substances. The organic form of methylmercury is especially toxic to the nervous system and is known to cause severe mental retardation and motor delays. Experimentally, methylmercury can cause more subtle abnormalities in the nervous system, but the milder forms of methylmercury toxicity have rarely been recognized in humans. The wide distribution of methylmercury in freshwater and oceanic fish and its rapid and total absorption through biological membranes can lead to elevated levels of methylmercury in people who eat fish. The developing fetal nervous system is particularly sensitive to methylmercury, which readily passes from the maternal to the fetal circulation. Consequently, the fetus may be seriously damaged while the mother remains asymptomatic. Recent studies suggest that the fetal brain may be adversely affected at levels much lower than previously suspected. This effect may be at levels easily achieved by eating fish. If this is confirmed, it would have serious implications for large populations who depend on fish as a primary source of protein. It would also be a potentially preventable cause of mental retardation. Identifying the mildest effect of an environmental toxin such as methylmercury presents a multiplicity of challenges. There are as yet no recognizable definitive clinical or biochemical markers specific for methylmercury. Many factors, such as alcohol, malnutrition. infectious diseases, and others, have similar effects. The appearance of symptoms may be delayed, only becoming apparent when the child is older. Small increments of neuronal loss or minor degrees of cerebral disorganization would

48

Gary J . M y e r s a n d D a v i d 0. M u r s h

be difficult to detect, even on neurohistological examination. There is even a suggestion that selenium in fish may have a protective effect from the toxicity of methylmercury. The challenge of sorting out the public health effects is enormous, but the serious implications for both society and individuals if methylmercury can cause toxicity at such low levels of exposure make the effort imperative. Simple public health measures would prevent the blunting of development if methylmercury indeed causes abnormalities at the low levels which fish eaters can readily achieve. REFERENCES Amin-Zaki. L.. Elhassani. S.. Majeed, M. A.. Clarkson. T. W., Doherty, R. A.. &Greenwood. M. ( 1974a). Intra-uterine methylmercury poisoning in Iraq. Pediutrics. 54,587595. Amin-Zaki. L.. Elhassani, S . . Majeed, M. A., Clarkson, T. W., Doherty, R. A., & Greenwood. B. S . (1974b). Studies of infants postnatally exposed to methylmercury. Joumul of Pediutrics. 85, 81-84. Amin-Zaki. L., Elhassani, S . , Majeed, M. A., Clarkson, T. W.. Doherty, R. A., & Greenwood, M. R. (1974~).Prenutul methylmercury poisoning in Iraq. Abstract of paper presented at the Congress0 International del Mercurio, Barcelona. Amin-Zaki. L., Elhassani. S., Majeed, M. A.. Clarkson, T. W., Doherty. R. A.. Greenwood, M. R., & Giovanoli-Jakubczak, T. (1976). Perinatal methylmercury poisoning in Iraq. Americun Journul of Diseuses of Children. 130, 1070-1076. Amin-Zaki. L.. Majeed. M. A., Elhassani. S. B.. Clarkson. T. W..Greenwood. M. R.. & Doherty, R. A. (1979). Prenatal methylmercury poisoning: Clinical observations over five years. Americun Journal of Diseuses of Children, 133, 172-177. Bakir. F., Damluji. S. F., Amin-Zaki, L., Murtadha. M., Khaladi. A., Al-Rawi. N. Y., Tikriti. S . , Shahir, H. 1.. Clarkson. T. W..Smith, J. C., & Doherty, R. A. (1973). Methylmercury poisoning in Iraq. Science. 181, 230-241. Bellinger. D., Leviton, A.. Waternaux, C.. Needleman, H., & Rabinowitz, M. (1987). Longitudinal analyses of prenatal and postnatal lead exposure and early cognitive development. New England Journul of Medicine. 316, 1037-1043. Brenner, R. P., & Snyder, R. D. (1980). Late EEG findings and clinical status after organic mercury poisoning. Archives of Neurology. 37, 282-284. Choi, B. H..Cho,K. H.,&Lapham,L. W.(1980). EffectsofmethylmercuryonDNAsynthesis of human fetal astrocytes: A radioautographic study. Bruin Reseurch, 202,238-242. Choi. B. H., Cho, K. H., & Lapham, L. W.(1981). Effects of methylmercury on human fetal neurons and astrocytes in vitro: A time-lapse cinematographic phase and electron microscopic study. Environmentul Research. 24,61-74. Choi. B. H.. Lapham. L. W.,Amin-Zaki, L., & Saleem, T. (1978). Abnormal neuronal migration, deranged cerebral cortical organization, and diffuse white matter astrocytes of human fetal brain: A major effect of methylmercury poisoning in utero. Journul of Neuroputhology und Experimentul Neurology, 37, 7 19-733. Clarkson, T. W. (1975). Exposure to methylmercury in grussy narrows und white dog reserves. Interim report to the Research Institute, Hospital for Sick Children, Toronto. Clarkson, T. (1987). Metal toxicity in the central nervous system. Environmentul Heulrh Perspectives. 75, 59-64.

METHYLMERCURY TOXICITY AND MENTAL RETARDATION

49

Clarkson, T., Nordberg, G. F., & Sager, P. R. (1985). Reproductive and developmental toxicity of metals. Scundinuviun Journul of Work, Environmental Health, 11, 145-154. Clegg. D. J. (1971). Mercury in muds environment (pp. 141-148). Ottawa: Royal Society of Canada. Cox, C.. Clarkson. T. W., Marsh, D., Amin-Zaki. L., Tikriti, S., & Myers, G. J. (1989). Dose-response analysis of infants prenatally exposed to methylmecury: An application of a single compartment model to single-strand hair analysis. Environmental Reseurch. 49, 318-332.

Dick. G. L.. Hughes, J. T., Mitchell, J. W..& Davidson, F. (1978). Survey of trace elements and pesticide residues in the New Zealand diet. I: Trace element content. New Zeulund Journul of Science. 21, 57-69.

Engleson. G., & Herner, T. (1952). Alkylmercury poisoning. Acta Puediutriu, 41,289-294. Fantz. R. (1964). Visual experience in infants: Decreased attention to familiar patterns relative to novel ones. Science, 146, 668-670. Giovanoli-Jakubczak, T., Greenwood. M. R., Smith, J. C., & Clarkson, T. W. (1974). Determination of total and inorganic mercury in hair by flameless atomic absorption and of methylmercury by gas chromatography. Clinical Chemistry, 20, 222-229. Goldwater, L. J. (1972). Mercury: A History of quicksilver. Baltimore: York Press. Gunderson, V. M., Grant, K. S.. Burbacher. T. M., Fagan, J. F., Ill, & Mottet. N. K. ( 1986). The effect of low-level prenatal methylmercury exposure on visual recognition memory in infant crab-eating macaques. Child Development. 57, 1076-1083. Harada, M. ( 1976). Intrauterine poisoning. Clinical and epidemiological studies and significance of the protein. Bulletin of the Issi4es of Constitutionul Medicine (Kumumoro University), 25, 1-59. Harada, Y. (1968). Congenital (or fetal) Minamata disease. In Study Group of Minamata Disease Minumutu Diseuse (pp. 93-1 17). Japan: Kumamoto University. Hirschman. S. Z.. Feingold, M., & Boylen. G. (1963). Mercury in house paint as a cause of acrodynia. New Englund Journul of Medicine, 269, 889-893. Hunter. D.. Bomford. R. R., & Russell, D. S. (1940). Poisoning by methyl mercury compounds. Quurt. J. Med.. 9, 193-213. Inskip. M. J., & Piotrowski, J. K. (1985). Review of the health effects of methylmercury. Journul of Applied Toxicology. 5, 113-133. Jones, K. I., Smith, D. W.. Ulleland. C. N.. & Streissguth A. (1973). Pattern of malformation in offspring of chronic alcoholic mothers. Lancet. 1, 1267-1274. Kjellstrom, T., Kennedy, P.. Wallis, S., & Mantell, C. (1986). Physicul and mentul development of children with prenutul exposure to mercury from fish (Report No. 3080, pp. 7-95). Solna: National Swedish Environmental Protection Board. Koos. B. J.. & Longo, L. D. (1976). Mercury toxicity in pregnant woman, fetus and newborn infant. Americun Joiirnul of Obstetrics und Gynecology. 126, 390-409. Korns, R. F. (1972). The frustrations of Bettye Russow. Nutrition Today, NovembedDecember, pp. 21-23. Kurland, L. T.. Faro. S. N.. & Siedler. H. (1960). Minimata disease. World Neurology, 1, 370-395.

Magos, L. (1975). Mercury and mercurials. British Medical Bulletin, 31, 241-245. Marsh, D. O., Clarkson, T. W., Cox, C. C., Myers, G. M.. Amin-Zaki. L.. & Al-Tikriti, S. ( 1987). Fetal methylmercury poisoning. Relationship between concentration in single strands of maternal hair and child effects. Archives of Neurology, 44, 10171022.

Marsh, D. 0.. Myers. G. M., Clarkson, T. W.,Amin-Zaki, L.. Tikriti, S., & Majeed, M. A. (1979). Fetal methylmercury poisoning: Clinical and toxicological data on 29 cases. Annuls of Neurology, 7 , 348-353.

50

Gary J . Myers and David 0 . Marsh

Marsh. D. 0..Myers. G. M.. Clarkson, T. W., Amin-Zaki, L., Tikriti. S . , Majeed, M. A., & Dabbagh, A. R. (1981). Dose-response relationship for human fetal exposure to methylmercury. Clinicul Toxicology, 18, 131 1-1318. Marshall, E. (1983). The “lost” mercury at Oak ridge. Science, 221, 130-132. McAlpine, D., & Araki, S. (1958). Minimata disease: an unusual neurological disorder caused by contaminated fish. Lancet. ii, 629. McAlpine, D., & Araki, S. (1959). Minimata disease. Archives of Neurology, 1, 522-530. McKeown-Eyssen, G., Reudy, J. E., & Neims. A. (1983). Methylmercury exposure in northern Quebec. 11. Neurologic findings in children. American Journul of Epidemiology, 118,470-479. Needleman, H. L., Rabinowitz, M., Leviton, L., Linn, S., & Schoenbaum, S. (1984). The relationship between prenatal exposure to lead and congenital anomalies. Journal of the Americun Medicul Associution, 251, 2956-2959. Oliver, T. (191I). A lecture on lead poisoning and the race. British Medicul Journal. i. 10961098. Sager, P. R., Clarkson. T. N.. & Nordberg. G. F. (1986). Reproductive and developmental toxicity of metals. In Hundbuok on the toxicology of’metals (2nd ed.). pp. 391-433. New York: Plenum. Schalock, R. L., Brown, W. J., Kark, R. A. & Menon, N. K. (1981). Perinatal methylmercury intoxication: Behavioral effects in rats. Dev. Psyrhobiol., 14,(3) 213-219. Shaw, C. M., Mottet, N. K.. Body, R. L.. & Luschei, E. S. (1975). Variability of neuropathologic lesions in experimental methylmercury encephalopathy in primates. Americcin Journal of Pathology. 80, 451-469. Snyder, R. D. (1971). Congenital mercury poisoning. New Englund Joitrnul qf Medicine. 284, 1014-1016. Spyker, M. J. & Smithberg. M. (1972). Effects of methylmercury on prenatal development in mice. Teratology. 5, 181-186. Spyker, M. J.. Sparber, S. B.& Goldberg, A. M. (1972).Subtle consequences of methylmercury exposure: Behavioral deviations in offspring from treated mothers. Science. 177, 621-630. Takeuchi. T. (1968). Pathology of Minimata disease. In Minimuta diseuse by Study Group on Minimata Disease (pp. 141-228). Kumamoto, Japan: Kumamoto University. Weiss. B. & Doherty, R. A. (1975). Methylmercury poisoning. Teratology. 12,(3) 311-314. Wheatley, B., Barbeau, A., Clarkson. T., & Lapham, L. W. (1979). Methylmercury poisoning in Canadian Indians-The elusive diagnosis. Cunadiun Journal of Neurologicid Sciences, 6, 417-422. World Health Organization ( 1976). Environmental health criteriu I : Mercury. Geneva.

Attentional Resource Allocation and Mental Retardation EDWARD C. MERRILL DEPARTMENT OF PSYCHOLOGY THE UNIVERSITY OF ALABAMA TUSCALOOSA. ALABAMA 35487

1.

INTRODUCTION

Various aspects of attention have long been assumed to play an important role in intellectual functioning. As is obvious from the amount of research and theory available that focuses on the comparison of the attentional processes of mentally retarded and nonretarded individuals (e.g., Bryant, Deckner, Soraci, Baumeister, & Blanton, 1988; Fisher & Zeaman, 1973; Heal & Johnson, 1970; Krupski, 1977; Nettelbeck & Brewer, 1981; Soraci, Barlean. Haenlein, & Baumeister, 1986; Zeaman & House, 1963, 1979), the nature of this potential relationship has also been evident to researchers gathering data about cognitive aspects of mental retardation. The fundamental reasons for this belief are clear. Because mentally retarded individuals exhibit performance deficits relative to nonretarded individuals across a wide range of cognitive tasks, researchers are compelled to consider the possibility that these deficits are at least partially the result of deficiencies in some fundamental cognitive abilities that are involved in the performance of many different activities. Attentional processes are presumed to mediate virtually all cognitive activities and are therefore logical candidates for the source of retarded-nowetarded differences in cognitive performance. When considering the possibility of attentional deficits in mentally re51 I N I ' I ~ R N A ' I I O N A I ,K I V I E W OF RESEARCH IN M E N T A L KIi1AKI)ATION. Vol. I6

Copyright 0 1990 by Academic Press. Inc.

All rights of reproduction in any form reserved.

52

Ednwrd C . Merrill

tarded individuals, it is important to recognize that over the years the concept of attention has referred to a variety of processes (see Parasuraman & Davies, 1984). These processes can be divided into two major categories. According to one view, attention can be described as a process of selection (Pick, Frankel, & Hess, 1975). In this view, attention operates to determine which stimuli from the internal and external environments are perceived and receive additional analysis and which stimuli are ignored. This conception of attention implies that individuals have both the ability to determine relevant from irrelevant aspects of the environment and some mechanism for filtering irrelevant aspects of the stimulus array, with performance deficits resulting from either a failure to attend to the relevant dimensions of the stimulus array or the inefficient execution of the filtering system. In a very different conceptualization, attention is described as a “capacity” rather than a process (e.g., Kahneman, 1973; Navon & Gopher, 1979; Norman & Bobrow, 1975). In this view, attention is described as a limited supply of processing resources that can be allocated to cognitive activities in a flexible manner. Cognitive processes can be activated simultaneously but, if they are, they must share the cognitive resources that are available for processing. In this case, performance deficits would result when the requirements of particular processes demand resources to an extent that obstructs the performance of other ongoing activities. The majority of empirical and theoretical work on the attentional capabilities of mentally retarded individuals has focused o n selective attention. In particular, Zeaman and colleagues (Fisher & Zeaman, 1973; Zeaman & House, 1963, 1979) have amassed considerable data detailing differences between the selective attention abilities of mentally retarded and nonretarded individuals on discrimination learning tasks. In discrimination learning, the basic procedure is to present a series of two or three stimuli that differ from each other along one or more stimulus dimensions (e.g., color and shape). The subject’s task is to determine which of the stimuli is “correct” on the basis of feedback about the correctness of his or her choices on previous trials. It is common for mentally retarded individuals to take longer than nonretarded individuals to learn which stimulus is correct when confronted with this kind of task. Zeaman and House ( 1963) concluded that this difference between retarded and nonretarded individuals is primarily a function of differences in selective attention; that is, mentally retarded individuals are much slower than nonretarded individuals at selecting the stimulus dimension targeted by the experimenter that is a prerequisite to determining the correct stimulus. Once the correct dimension has been specified, the learning of mentally retarded and nonretarded individuals appears to proceed in a very similar

A P E N T I O N A L RESOURCE ALLOCATION

53

fashion (Zeaman & House, 1963, 1979). Additional work has suggested the possibility that “breadth of attention,” or the number of stimulus dimensions that can be considered on each trial, may also covary with intelligence (see Zeaman, 1978).However, the details of this relationship have not been as well documented as those associated with the direction of attention. There is also some evidence to suggest that the efficiency of selection mechanisms is different for mentally retarded and nonretarded individuals (Hagen & Huntsman, 1971). For example, subjects may be shown a series of item pairs with instructions to attend to and try to remember the “animals” of the object pairs (the central information) without instructions to learn the other object, for example, a piece of furniture, of each pair (the incidental information).Current evidence suggests that older and more intelligent individuals exhibit a greater difference between the amount of central relative to incidental information that they are able to remember in comparison to younger and less intelligent individuals (Hagen & Huntsman, 1971; Hagen, Meacham, & Mesibov, 1970; Maccoby & Hagen. 1965). Apparently, the ability to focus on relevant information, even after that information has been singled out for further processing, is sensitive to differences in both developmental and intellectual level. In as much as a significant amount of information processing occurs in the absence of competing stimuli, it would appear that mechanisms of selectivity are not likely to be the only central component of the information-processing system associated with retarded-nonretarded differences in general information processing. Another possibility focuses on attention viewed as a capacity for processing (e.g., Kahneman, 1973; Navon & Gopher, 1979; Norman & Bobrow, 1975). However, in contrast to selective attention, there are relatively few studies that address retardednonretarded differences associated with attention when attention refers to a capacity for processing rather than a mechanism for the selection of stimuli. The major exceptions are studies that have focused on potential differences between retarded and nonretarded individuals associated with some aspect of the orienting response (see Luria, 19631, a relatively complex set of central and autonomic nervous system responses to novel and/ or meaningful stimuli that are associated with an increased ability to perceive and take in sensory information. For example, Krupski (1975) assessed the relationship between changes in heart rate and the performance of mentally retarded and nonretarded adults in a fixed reaction time procedure. Heart rate deceleration is generally assumed to reflect some aspect of attentional processing associated with preparing for the presentation of a stimulus or the generation of a response. Krupski found that, overall, mentally retarded subjects were slower to exhibit heart rate

54

Edward C . Merrill

deceleration during the reaction time task than were nonretarded subjects, and the retarded subjects exhibited a lower magnitude of heart rate deceleration than did the nonretarded subjects. The implication is that mentally retarded individuals are less efficient in making attentional resources available for the performance of reaction time tasks (see also Runcie & O’Bannon. 1975). There are many other aspects of attentional resource allocation that warrant research attention. The major premise of this article is that a significant portion of retarded-nonretarded differences in semantic processing speed may be the result of the inefficient execution of one or more components of the attentional resource allocation system. In the sections that follow the rationale underlying the development of the attentional resource hypotheses is reviewed, a conceptual framework in which to consider these hypotheses is presented, and the results of several preliminary studies designed to address the relationship between attentional resource allocation and retarded-nonretarded differences in cognitive performance are discussed.

II.

DEVELOPMENT OF THE AlTENTIONAL RESOURCE HYPOTHESES

It is common to observe a consistent relationship between measured intelligence and the speed of information transmission through the human information-processing system (see Jensen, 1982; Nettelbeck & Brewer, 1981). with individuals who score lower on standardized tests of intelligence commonly performing information-processing tasks more slowly than individuals who score higher on these tests. When comparing the performance of mentally retarded and nonretarded individuals this relationship is particularly evident (see Baumeister & Kellas, 1968; Sperber & McCauley, 1984). Some of the specific cognitive processes that have received a great deal of research attention include searching short-term memory (e.g., Dugas & Kellas, 1974; Harris & Fleer, 1974; Maisto & Jerome, 1977; McCauley, Kellas, Dugas, & DeVillas, 1976). retrieving information from long-term memory (e.g., Hunt, 1978; Hunt, Frost, & Lunneborg, 1975; Keating & Bobbitt, 1978) and making decisions about semantic category membership (e.g., Davies, Sperber, & McCauley, 1981; Merrill, 1985; Sperber, Davies, Merrill, & McCauley, 1982). The generality of differences in semantic processing speed between mentally retarded and nonretarded individuals has been particularly conspicuous. In fact, one is hard pressed to find conditions in which mentally retarded and nonretarded individuals do not differ in processing speed.

AITENTIONAL RESOURCE ALLOCATION

55

More remarkable, however, is the obvious similarity in the relative magnitude of retarded-nonretarded differences in processing speed across the different stimuli and cognitive processes tested in the various experiments. This has been most noticeable in the results of studies in which researchers have attempted to isolate individual components of the processing system and have tested mentally retarded subjects who have measured IQ test scores roughly between 50 and 70. Several investigators have used the Sternberg memory search procedure to examine differences between mentally retarded and nonretarded individuals in the rate at which information can be located in short-term memory (e.g., Dugas Ki Kellas, 1974; Harris & Fleer, 1974; Maisto & Jerome, 1977). The majority of these studies have reported reliable differences between mentally retarded and nonretarded subjects in memory search rate (cf. Silverman, 1974). In addition, the magnitude of this retarded-nonretarded difference has been very similar across studies. Dugas and Kellas (1974) reported that nonretarded subjects were able to search short-term memory twice as fast as were mentally retarded subjects (45 msechtem for the nonretarded and 90 msechtem for the retarded subjects). The relative magnitude of the group difference observed by Harris and Fleer (1974) was virtually identical to that of Dugas and Kellas (42 msec for nonretarded subjects and 88 msec for retarded subjects). Of course, this is as expected because these researchers used both similar stimuli (alphanumeric symbols) and similar subjects. However, Maisto and Jerome (1977) employed anomalous shapes as stimuli rather than alphanumeric symbols. These were irregularly shaped line drawings that did not have common names. Despite the fact that these new stimuli decreased memory search rate for all subjects, and the absolute value of the retarded-nonretarded difference was increased, the relative magnitude of this difference was essentially unchanged (65 vs. 126 msec for nonretarded and retarded subjects, respectively), with nonretarded subjects again performing at approximately twice the speed of mentally retarded subjects. Davies et al. (1981) examined differences in the ability of mentally retarded and nonretarded individuals to determine the category membership of common objects. Subjects were asked to determine whether or not pictures of common objects matched verbal labels provided by the experimenter, where the relationship between the label and the pictured object varied across several stimulus dimensions. For example, the labels were either the basic level names (e.g., dog, table) or the superordinate category names (e.g., animal, furniture) of the objects (see Rosch, Mervis, Gray, Johnson, & Boyes-Braem, 1976), or the pictured object was either a typical or atypical exemplar of the superordinate category

56

Edwwd C . Mrrrill

(Rosch, 19751, that is, a dog is usually considered to be a better exemplar of the superordinate category mammal than is the exemplar walrus, even though both are mammals by definition. Further, deciding that a less typical exemplar belongs to a particular category has proved to be more difficult than deciding that a more typical exemplar is a member of the category (see Smith, Shoben, & Rips, 1974). In the Davies er al. experiment, nonretarded subjects were faster than retarded subjects at making decisions about category membership in every condition, with t h e absolute difference in category decision time between the retarded and nonretarded subjects increasing with increases in decision difficulty. Nevertheless, the ratio of the difference between groups was again very similar across the various conditions, with the performance of the mentally retarded subjects ranging from 1.7 to I .9 times slower than that of the nonretarded subjects throughout the experiment. In addition, the relative magnitude of the retarded-nonretarded differences is very close to that observed for short-term memory search, as discussed above. The general similarity of group differences in semantic processing speed across stimulus materials and these two domains of semantic processing provide some measure of support for the suggestion that retarded-nonretarded differences in semantic processing speed are the result of the inefficient execution of a component of the information-processing system that is common to different cognitive activities. The consistency of retarded-nonretarded differences in semantic processing across the domains of short-term and long-term memory was subsequently assessed in a single study using a within-subjects design (Merrill, 1985). In this study, a modified Sternberg memory search procedure was used (Sternberg, 1969). In one condition, subjects were presented with memory sets that were composed of basic level object names and probes that were also basic level object names (e.g., dog, table). The subjects’ task in this condition was simply to determine if the probe was a member of the memory set. This condition is analogous to the standard Sternberg memory search procedure and assessed the subjects’ rate of searching short-term memory. In a second condition, subjects were presented with memory sets that were composed of superordinate category labels (e.g., animal, furniture) and probes that were basic level object names (e.g., dog, table). The subjects’ task in this condition was to decide whether or not the probe stimulus was a member of one of the categories in the memory set. This condition assessed the subjects’ ability to retrieve category information from long-term memory and make decisions about category membership. It was therefore possible to compare directly the relative difference in processing speed between mentally retarded and

AITENTIONAL RESOURCE ALLOCATION

57

nonretarded subjects in immediate memory versus long-term memory processing. As expected, across the two conditions mentally retarded subjects performed more slowly than did the nonretarded subjects, and the absolute difference in response times obtained for category decision times (longterm memory processing) was greater than that obtained for short-term memory processing (35 msec for immediate and 100 msec for long-term memory processing). Of primary importance to the argument presented here, the same kind of consistency in the relative magnitude of the retarded-nonretarded difference was observed across the two domains of processing, with the mentally retarded subjects being I .9 times slower than the nonretarded subjects in short-term memory processing and 2.2 times slower in long-term memory processing. In conjunction with a significant correlation obtained between the measures of immediate and long-term memory processing for the mentally retarded subjects ( r = .75), these data lend additional support for the suggestion that retardednonretarded differences in semantic processing speed across different tasks may be mediated by a common component of the information-processing system. The results of several additional studies converged on the possibility that a likely candidate for a difference in central processing between mentally retarded and nonretarded individuals may be some component of the attentional resource allocation system. Over the last 10-15 years, there have been numerous studies that suggest that differences between retarded and nonretarded individuals in semantic processing are observed primarily when subjects must actively use the semantic information and not under conditions that involve automatic or noneffortful processing (e.g., Cody & Borkowski, 1977; Meador & Ellis, 1987; Sperber, Ragain, & McCauley, 1976). For example, Meador and Ellis (1987) used a stimulus priming procedure developed by Posner and Snyder (1975) to examine differences in automatic versus effortful processing in mentally retarded and nonretarded individuals. Subjects were presented with pairs of letters and had to determine whether or not the two letters were the same. While the letter pairs matched half the time and did not match half the time, the most relevant data were obtained on the matching trials. Matching trials were preceded by the presentation of a letter that was identical to the letters in the matching pair, a letter that was different from the letters in the matching pair, or a plus sign. In general, the presentation of the same letter would be expected to facilitate matching times and the presentation of a different letter would be expected to interfere with matching times. In the Meador/Ellis procedure, the interval between the presentation of

58

Edwnrd C . Mtv-rill

the letter or plus sign (the prime) varied from 100 to 2000 msec. According to current views of automatic and effortful processing, the facilitating effect of a same letter prime on matching times at short intervals is assumed to reflect the operation of an automatic process and does not rely on an individual’s available processing resources. In contrast, the presence of facilitation from a same letter prime and/or the buildup of inhibition from a different letter prime on matching times at the long intervals is assumed to reflect the operation of an effortful or resource-demanding process. Meador and Ellis report that facilitation resulting from automatic processing was very similar for mentally retarded and nonretarded individuals. However, facilitation and inhibition due to effortful processing were slower to develop in mentally retarded than in nonretarded subjects. These findings seem to implicate the attentional resource allocation system as an important source of retarded-nonretarded differences in semantic processing speed. Ellis and colleagues (Ellis & Allison, 1988; Ellis, Katz, & Williams, 1987; Ellis & Meador, 1985; Ellis, Meador, & Bodfish, 1985; Ellis, Palmer, & Reeves, 1988; Woodley-Zanthos & Ellis, 1989) have, in fact, conducted an extensive series of investigations examining potential intellectual and developmental level differences in automatic aspects of memory. In their early work, the researchers sought to examine the memory processes of mentally retarded and nonretarded individuals under conditions that minimized the use of cognitive strategies (Ellis & Meador, 1985; Ellis et al., 1985). For example, Ellis and Meador (1985) used a delayed matching-to-sample procedure using squares of eight differing sizes as stimuli. The subjects were presented pairs of stimuli that were separated by retention intervals ranging from 0 to 20 seconds. Subjects simply had to report whether or not the squares were the same size. While some strategy use, possibly in the form of verbal labeling, was reported for the extreme sizes (e.g., “smallest” and “largest”), the authors convincingly argued that strategy use was much more difficult for the intermediate sizes. Hence, it was possible to examine the role of strategy use in memory for size by comparing differences between the relative abilities of mentally retarded and nonretarded individuals to remember easy versus hard to discriminate size differences. The results of this study indicated that, even under conditions of minimal strategy use, mentally retarded individuals performed more poorly than did nonretarded individuals on the matching-to-sample task. The authors therefore concluded that memory differences between mentally retarded and nonretarded individuals were not entirely the result of inefficient strategy use on the part of the mentally retarded individuals. However, while the authors were able to

ATTENTIONAL RESOURCE ALLOCATION

59

rule out the potential influence of cognitive strategies, it was not possible to conclude that the observed retarded-nonretarded differences in memory were definitely the result of differences in the execution of automatic aspects of memory processes. It is quite possible to exert cognitive effort in the performance of a task without translating that effort into a welldefined strategy. For example, one group of subjects simply may have concentrated more than the other groups. If the exertion of effort influenced memory performance, then the memory processes were not operating automatically. In subsequent research, Ellis and collaborators (Ellis & Allison, 1988; Ellis et d.,1988; Ellis, Woodley-Zanthos, & Dulaney, 1989; WoodleyZanthos & Ellis, 1989) focused on intelligence-related differences in memory for incidentally learned attributes of stimuli such as frequency of occurrence and spatial location. The theoretical importance of these attributes is that Hasher and Zacks (1979) have shown that they both may be successfully encoded without a subject's awareness or intention (i.e., automatically). Therefore, a comparison of mentally retarded and nonretarded individuals' relative abilities to remember spatial location and frequency information under incidental learning instructions would represent a comparison of automatic processing in mentally retarded and nonretarded persons. In general, Ellis found that mildly retarded individuals estimate frequency of occurrence as well as do nonretarded individuals and process spatial information as well as college students. Therefore, the available evidence supports the view that cognitive performance differences between mentally retarded and nonretarded individuals are obtained under task conditions that require effortful (not necessarily strategic) processing but not under conditions of automatic processing. This is as expected if retarded and nonretarded individuals differ in their abilities to use their attentional resource allocation system. There is one additional piece of evidence that deserves mention. It has become increasingly clear that retarded and nonretarded individuals exhibit larger differences in semantic processing when faced with more difficult activities. This has been reported in Davies et a / . (1981) and Merrill (1985) as discussed above, as well as by a number of other investigators (see, e.g., Merrill ez a / . , 1987; Mulhern & Baumeister, 1971; Sperber et a / . , 1982). If we can assume that one of the primary differences between easy decisions and difficult decisions is the degree of effort or attentional processing capacity that is required to complete the cognitive task, then the implication is again that some aspect of the attentional resource allocation system may mediate retarded-nonretarded differences in semantic processing speed (Merrill, 1982; Sperber & McCauley, 1984). The evi-

60

Edward C . Merrill

dence appears sufficient to offer this possibility as a general hypothesis that may help to explain differences in semantic processing speed between mentally retarded and nonretarded individuals.

111.

ATTENTIONAL RESOURCE ALLOCATION

A. General Framework

One general approach for conceptualizing potential difference between mentally retarded and nonretarded individuals in the allocation of attentional resources was briefly outlined in Sperber and McCauley (1984). What is presented here is an extension of that work. It is based on the principles of many current “capacity theories” of attention (e.g., Kahneman, 1973; Navon & Gopher, 1979; Norman & Bobrow, 1975). In these capacity theories, attention is viewed as one or more reservoirs of cognitive processing resources that can be allocated to particular tasks and their components in a flexible and continuous fashion (Wickens, 1984). An important feature of these theories is that the amount of cognitive resources that can be made available for the transmission and manipulation of information is limited in some way; that is, there is not an endless supply of these resources. If cognitive processes are activated at the same time, as is common in most complex cognitive activities such as reading (e.g., LaBerge & Samuels, 1974), then the activated processes must share the available resources. Limits on performance are observed when the amount of cognitive processing resources required to execute successfully one or more ongoing process fails to leave enough resources for other processes to be performed. So we find, for example, that less skilled readers who must devote a substantial portion of their available resources to decoding individual words relative to the requirements of more skilled readers would, as a result, exhibit severe problems with the comprehension of text (e.g., Vipond, 1980). Individual differences in the amount of processing resources needed to execute component processes of cognitive tasks may actually indicate differences in the acquisition of “automatic processing” (e.g., Schneider & Shiffrin, 1977; Shiffrin & Schneider, 1977). It is generally accepted that, with practice. processes that initially require a large portion of an individual’s available resources in order to be completed can come to require fewer and fewer of these resources without a corresponding decrease in level of performance. After quite extensive practice, a process may be successfully executed without the need for any cognitive resources (cf. Schneider, Shiffrin, & Dumais, 1984). When this occurs, pro-

ATTENTIONAL RESOURCE ALLOCATION

61

cessing is described as automatic, where automatic processing is characterized by fast, parallel, relatively effortless processing that is not limited by memory capacity and is not under an individual’s direct control (Hasher & Zacks, 1979; 1984; Posner & Snyder, 1975). Because of these characteristics, automatic processes are assumed to be extremely effective when used in the performance of well-developed skilled behaviors (Ackerman, 1987; Logan, 1985; Schneider, 1985). Most skilled behaviors require that a variety of simple and complex cognitive operations be executed simultaneously or in close succession. The efficiency of these behaviors may depend on the resource requirements of the component processes (see Ackerman, 1987). Since automatic processing is often the result of learning and has been found to vary with developmental level (Manis, Keating, & Morrison, 1980), it is quite possible that the rate at which processes become automated is slower for mentally retarded relative to nonretarded individuals. The result would be that fairly basic processes of information transmission would require more time and cognitive resources for their completion by these individuals. As a consequence, the amount of resources remaining to perform higher order cognitive activities like problem solving would be correspondingly small, thereby resulting in the less successful completion of the higher order activity (Sperber & McCauley, 1984). Hence, according to this argument, some of the cognitive deficits in the execution of complex tasks exhibited by mentally retarded individuals can be traced to the relatively high resource requirements of fundamental processes like stimulus encoding. In this sense, the deficit is stimulus driven because the cognitive processing requirements of particular stimuli limit the performance of mentally retarded individuals in more complex cognitive behaviors. Alternatively, because it is assumed that the allocation of cognitive processing resources is under some degree of cognitive control (Kahneman, 1973; Wickens, 1984), it is possible to conceptualize retarded-nonretarded differences in resource allocation as person variables as well. For example, we know that young children relative to older children exhibit less control and less flexibility in the manner in which they allocate cognitive resources under dual-task conditions (Lipps Birch, 1976; 1978). This may also be characteristic of mentally retarded relative to nonretarded individuals. So, rather than simple stimuli and basic processes requiring more of the attentional resources available to mentally retarded individuals, it may be that the mentally retarded actually allocate fewer of their available resources to the processing of simple stimuli (Sperber & McCauley. 1984). The result would be very similar to that described previously in that basic processes would be executed less efficiently as a

62

Edward C . Mcrrill

result of the allocation of insufficient resources, thereby interfering with the performance of more complex cognitive tasks. Further, in light of the well-known metacognitive deficits associated with mental retardation (e.g., Borkowski & Cavanaugh, 1979; Bray, 1979; Campione & Brown, 1978; Ellis, 1978), it is also important to recognize that metacognitive skills are likely to play a role in attentional resource allocation. In order to allocate the appropriate amount of cognitive resources to the various component processes of any task, an individual must have a great deal of knowledge about his or her own limitations and abilities. For example, it is necessary to be able to assess the resource demands of those processes and monitor task performance to determine when sufficient resources have been allocated. To the extent that mentally retarded individuals are unable to do this accurately, allocating too large or too small a portion of their available resources to the basic processes of information transmission, performance deficits in the tasks that rely on the efficient execution of those processes are again the likely result. There is at least one other way that cognitive processing capacity can be related to mental retardation. It may be that mentally retarded individuals simply have a smaller pool of processing resources at their disposal than do nonretarded individuals. If this is the case, then performance deficits would still be observed, even though the particular processes involved may not require different amounts of the available resources of retarded and nonretarded individuals and even if mentally retarded individuals are just as effective as nonretarded individuals at allocating their resources across processes. The processing resources of mentally retarded individuals would simply be depleted at lower levels of processing complexity than would the resources of nonretarded individuals, and the same pattern of performance deficits would be observed for the retarded individuals. Distinguishing between these alternative views will not be a straightforward assignment. Each alternative appears to predict the same general pattern of Performance differences between mentally retarded and nonretarded individuals. It is also important that these different alternatives not be viewed as mutually exclusive. It is quite possible that any deficiencies that may exist in attentional resource allocation will exist in combination. We may therefore find that individuals allocate resources differently because executing basic processes actually requires a different amount of their available resources, or because they are inefficient at allocating their resources to basic processing operations, or as the result of some unique combination of both differences. In addition, we cannot rule out the possibility that all observed deficiencies are ultimately tied to metacognitive

AITENTIONAL RESOURCE ALLOCATION

63

differences without first exhausting the available methods of modifying the resource allocation skills of mentally retarded individuals. B.

Methodological Considerations

The most common methods that are used to address issues concerning the allocation of cognitive resources are based on some form of the dualtask method (see, e.g., Kerr, 1973; Posner, 1978). In general, this approach requires that subjects perform two tasks simultaneously. In one version of this method, the cognitive task of interest is designated as the primary task and the other task as the secondary task (see Posner & Boies, 1971). The subject is instructed to allocate the necessary cognitive resources to the primary task to maintain maximum performance levels, and then to use whatever resources are left to perform the secondary task. The measure of interest is the degree to which performance on the primary task interferes with performance on the secondary task (relative to performing the secondary task alone). It is assumed that as the processing demands of the primary task increases, the amount of spare capacity, or processing resources, left to allocate to the secondary task decreases. Hence, performance on the secondary task decreases, and the magnitude of this decrease is assumed to reflect the relative amount of processing resources required to perform the primary task under two or more conditions. A second version of the dual-task method is exemplified by the concurrent memory load task (Logan, 1979). In this procedure, the cognitive task of interest is actually given secondary status. The subject is given a list of digits to commit to immediate memory just prior to performing the cognitive task. It is assumed that immediate memory operations make use of processing resources from the same source as other cognitive processes, and thus will interfere with the performance of the concurrent task (the task of primary interest) to the extent that this task also requires attentional resources. In this case, the degree of interference associated with performing the task of interest while maintaining a list of digits in immediate memory, relative to performing the task by itself, is taken as an index of the amount of resources required to perform the task of interest. Both versions of the dual-task method have been used effectively in the study of cognitive processing limitations in single population studies (cf. Duncan, 1980). However, important methodological difficulties arise when adapting these procedures for studying group differences in resource allocation ability. These difficulties come about because we are only able to assess the allocation of these resources indirectly; that is, we

64

Ed\curd C . Miwill

can only measure the amount of spare capacity left over from the processing of some cognitive task. One problem with group comparisons is that groups may differ in the amount of cognitive resources that they initially have at their disposal. Therefore, the amount of spare capacity that an individual has left over from performing a particular cognitive task may reflect either a difference in the amount of resources required to perform the task or a difference in the amount of resources initially available. In most respects, these two alternatives are functionally equivalent. Individuals who exhibit the fewest resources left over from the execution of basic information-processing operations will exhibit the greatest processing limitations when the processes are combined into more complex activities, regardless o f the source of the differences. Ultimately. however, the choice of techniques for remediation will depend on which of these alternatives most accurately describes any observed retarded-nonretarded differences in attentional resource allocation. A second problem that must be mentioned also exists in single group studies, but is aggravated by group comparisons. It is essentially a measurement scale problem and has no obvious resolution. The issue is that there is no assurance that equivalent changes in the magnitude of interference exhibited by individuals in these studies reflect equivalent changes in resource allocation either within the same individual or across groups of individuals. A reasonable analogy may be to consider the relationship between the speed of an automobile and the amount of fuel being consumed. The automobile necessarily increases fuel consumption as speed increases 10 mph from 20 to 30 mph. However, the amount of this change in fuel consumption is probably not the same if we increase speed 10 mph from 30 to 40 mph. Hence, using speed changes to infer fuel consumption changes potentially converts interval scale data to ordinal scale data. This problem will be further aggravated if we try to make inferences about fuel consumption on the basis of speed measures across different makes of automobiles that start out with different rates of fuel consumption. Automobile manufacturers solve the problem by measuring fuel consumption directly. This would also be the preferred way to measure the allocation of attentional resources. Unfortunately, we are not in a position to do this and must continue to infer attentional resource allocation from performance indices. There are no known methodologies that will eliminate these difficulties. It is obviously important that researchers in this domain remain sensitive to the interpretive difficulties that are likely to arise. To minimize these problems it will be essential to obtain data using a variety of procedures that converge on a single issue. For example, it will be important to use both versions of the dual-task approach discussed earlier, as well as a

AITENTIONAL RESOURCE ALLOCATION

65

variety of different secondary tasks (e.g., concurrent memory load, counting backwards, signal detection) in order to ensure that the results are not stimulus and/or task specific. This will at least provide a stronger basis for inferences to be drawn concerning attentional resource differences between mentally retarded and nonretarded individuals.

IV.

PRELIMINARY RESEARCH

This section includes the initial results of four studies that begin to assess the nature of the relationship between cognitive resource allocation and cognitive performance differences between mentally retarded and nonretarded individuals. All of the studies involve a comparison of the performance of mentally retarded and nonretarded adolescents or young adults matched on chronological age (CA). Across the various studies, the mentally retarded subjects had an average IQ of 60.9 (SD = 6.7) and an average CA of 19.6 years (SD = 3.3). The mentally retarded subjects were recruited from public high schools and local rehabilitation centers. The nonretarded subjects were recruited from public high schools and universities. A.

Mobilization of Cognitive Processing Resources

In order for cognitive processing resources to be effectively used in information-processing activities, an individual must be able to make them available for processing. One important aspect of cognitive resource allocation involves preparing for the presentation of upcoming stimuli (Posner and Boies, 1971). Attentional resources must be made available and directed toward the processing of particular stimuli, whether the stimuli are selected by an experimenter or the individual. Processing is facilitated by presenting a warning signal and allowing subjects time to prepare before a stimulus is presented. As discussed earlier, there is some evidence based on heart rate measures that mentally retarded and nonretarded individuals differ in their efficiency of making attentional resources available for stimulus processing in reaction time tasks (Krupski, 1975: Runcie & O'Bannon, 1975). Merrill and McCauley (1988) focused on the extent to which potential retarded-nonretarded differences in this ability help to account for previously reported group differences in stimulus encoding speed (Merrill et al., 1987). The general procedure used in Merrill and McCauley (1988) was adapted from a technique first used by Posner and colleagues (Posner & Mitchell, 1967) to examine parameters of stimulus encoding in a letter-

66

Edward C . Merrill

matching task. In the Merrill/McCauley version of the procedure, mentally retarded and nonretarded subjects were presented pairs of photographic slides of black-and-white line drawings. The stimulus pairs were presented sequentially, with the presentation of the two slides being separated by a variable stimulus onset asynchrony (SOA). The time between the slides varied from 0 to 1000 msec at 100-msec intervals. The subjects' task was to determine as rapidly as possible whether or not the two stimuli matched under either physical identity instructions or name identity instructions. Our ability to measure encoding times in this procedure is based on the SOA manipulation. When presented with short SOAs. subjects do not have enough time to encode the first stimulus of the pair completely prior to the presentation of the second stimulus. Hence, they must continue to process the first stimulus in order to make an accurate match-nonmatch decision. Therefore, response times, measured from the onset of the second stimulus of the pair, will include the time needed to complete the encoding for the first stimulus and will be relatively long. The performance of the subjects reaches maximum levels when the length of the SOA is just long enough to allow the first stimulus to be encoded prior to the appearance of the second. There is no further decrease in response times associated with the SOA manipulation after the first stimulus is encoded. Any additional processing requires the presentation of the second stimulus of the pair, and this coincides with the onset of the timing interval and is identical across all of the remaining SOAs. Thus, the shortest SOA at which subjects exhibit maximum performance levels corresponds to the time required to encode the first member of the stimulus pair. Using this procedure, Merrill et d.(1987) found that mentally retarded subjects are generally slower than nonretarded subjects matched with the retarded subjects on either CA or MA at both physical identity encoding (410 vs. 300 msec for the retarded and nonretarded, respectively) and name identity encoding (480 vs. 385 msec). To examine the extent to which these retarded-nonretarded differences in encoding specd could be attributed to corresponding differences in mobilizing cognitive processing resources, Merrill and McCauley (1988) assessed the encoding speed of mentally retarded and equal-CA nonretarded subjects when the length of time given between the presentation of a warning signal and the first stimulus of the stimulus pair was systematically varied. Four different alrrtirig intervals were tested: 0, 250, 500, and 1000 msec. It was assumed that if differences in the ability to alert for incoming stimuli were responsible for differences in encoding speed, then the magnitude of the encoding speed difference between the retarded and nonretarded subjects would vary as a function of the length of the alerting interval. More specifically. it was

67

ATTENTIONAL RESOURCE ALLOCATION

TABLE I MEAN RESPONSETIMES(IN DEVIATION SCORES) AT THE 0-MSEC STIMULUS ONSETASYNCHRONY Alerting interval Grouplencoding

0

250

500

lo00

Mentally retarded Physical identity Name identity

27 I 34 I

20 1 279

I98 222

202 215

Nonretarded Physical identity Name identity

I62 308

I I4

I24 I32

I I3

135

131

expected that the difference between retarded and nonretarded subjects would be largest at the short alerting intervals and smallest at the long alerting intervals. First, it was necessary to determine the effectiveness of the manipulation of cognitive processing resources; that is, did varying the length of the alerting interval influence any aspect of the subjects’ performance? To examine this issue, we examined the effect of alerting interval on response times at the 0-msec SOA. This SOA interval was chosen for comparison because it is the only interval for which the length of the alerting interval is the only variable that determines the amount of processing resources being deployed by the subjects. At all longer intervals, the presentation of the first stimulus of the pair also serves to alert the subject for an upcoming response. Thus, the effectiveness of the alerting interval manipulation is clearest at the 0-msec SOA (see Merrill and McCauley, 1988). In order to eliminate any unrelated differences between subjects in overall response times from influencing the results of the analyses, the raw response times were converted to difference scores (see Merrill et al., 1987). These scores were obtained by simply subtracting each subject’s response times at their fastest SOA from those of the 0-msec SOA. The obtained data are presented in Table I. Several aspects of these data are worth noting. First, under physical identity instructions, the length of the alerting interval had essentially the same influence on the performance of mentally retarded and nonretarded subjects. Both mentally retarded and nonretarded subjects were able to activate sufficient resources within 250 msec to perform at maximum levels. However, there was a difference between the two groups in the length of the optimal alerting interval under name identity instructions. The non-

68

Edward C . Merrill

retarded subjects were again able to mobilize sufficient resources to perform at maximum levels within 250 msec. In contrast, the mentally retarded subjects did not reach maximum performance levels until the length of the alerting was 500 msec. We therefore concluded that the alerting interval manipulation did influence performance in the matching task and, at least for name identity decision making, mentally retarded individuals required a longer time to mobilize their processing resources in preparation for the processing and manipulation of information. This conclusion is consistent with the results of Krupski (1975). In addition, because group differences were restricted to the name identity condition, it appears that accessing basic level object names and/or matching on the basis of name identity may be more susceptible to the influence of attentional resources than is physical identity encoding and matching on the basis of physical identity. One additional difference in the performance of mentally retarded and nonretarded subjects was observed. The retarded subjects exhibited less improvement in response time performance as the alerting interval increased relative to the nonretarded subjects ( 1 19 vs. 173 msec. respectively). This difference was only marginally significant (p < .lo), but is consistent with the possibility that some aspect of name identity matching in mentally retarded individuals is limited by the availability of cognitive resources. Either the operation requires more resources for mentally retarded relative to nonretarded individuals to execute or mentally retarded individuals allocate fewer resources to the name identity matching task. To obtain encoding times, the shortest SOA at which maximum performance levels were achieved was determined for each subject for both physical identity and name identity matches. This was done by comparing the subject’s response times at each SOA to those of the next three longer SOAs using t tests. An individual’s encoding time was defined as the shortest SOA for which the mean response time was not significantly greater than each of the next three SOAs. The mean encoding times obtained for subjects in each condition are presented in Table 11. As can be seen from these data, the length of the alerting interval had no effect on encoding times. Indeed, the overall differences between mentally retarded and nonretarded individuals were essentially identical to those reported in Merrill ef d.(1987). The mentally retarded subjects were approximately 100 msec slower than the nonretarded subjects at physical identity encoding (372 vs. 281 msec, respectively) and name identity encoding (481 vs. 375 msec). Apparently, the resource requirements of encoding are substantially lower than those of some other aspects of the matching task. This does not necessarily mean that physical and name identity encoding does not require any cognitive resources to be exe-

69

ATTENTIONAL RESOURCE ALLOCATION

TABLE I1 AVERAGEENCODING TIMES FOR SUBJECTS AS A FUNCTION OF ALERTINGINTERVAL Alerting interval Group/encoding

0

250

500

100

Mentally retarded Physical identity Name identity

388 500

362 415

362 475

375 475

Nonretarded Physical identity Name identity

2JS 388

28X

275 375

288 388

350

cuted. It may simply be that a sufficient amount of attentional resources was continuously being allocated to the experimental task to permit encoding to take place without interference. Issues concerning the resource requirements of basic encoding operations are addressed more directly in the next experiment. Nevertheless, on the basis of this experiment, it was tentatively concluded that retarded-nonretarded differences in encoding speed are not directly related to differences in the ability to prepare for the presentation of an upcoming stimulus event. B.

Cognitive Resource Requirements of Basic Processes

As mentioned earlier, differences between mentally retarded and nonretarded individuals in the amount of cognitive resources allocated to the basic processes of information transmission are likely to impact upon the performance of more complex tasks that rely on the efficient execution of these basic processes. It is therefore important to examine group differences in the resource requirements associated with the execution of these basic processes. The purpose of Merrill (1990) was to assess differences between mentally retarded and nonretarded individuals in the amount of cognitive resources needed to execute basic encoding and decision-making processes. These were studied in a task requiring matching stimuli on the basis of physical identity and name identity while maintaining a concurrent-memory-load (Logan, 1979). Mentally retarded and nonretarded subjects were required to memorize a series of numbers, and, while maintaining these numbers in memory, they performed the stimulus matching task described in the previous experiment (see Merrill & McCauley, 1988). An important feature of this version of matching task is that it is possible to separate the encoding from the decision portions of

70

Edward C . Merrill

the matching task. The time needed to execute encoding operations is dependent on the subjects’ performance on the first stimulus of the pair and is therefore reflected in the SOA at which maximum response time performance is first observed. In contrast, the time needed to decide whether or not the two stimuli match is dependent on the subjects’ performance on the second stimulus of the pair, which is identical across all SOAs and thus is reflected in the overall level of the response time function. It was therefore possible to separate the influence of the concurrentmemory-load task on encoding processes from its influence on decisionmaking processes. In this experiment, subjects were initially given a test of digit span. On each trial, the subjects were required to memorize either a full memory load, defined as the subjects’ digit span minus one, or half that amount prior to performing the matching task. The cognitive resource requirements of the stimulus matching task were assumed to be reflected in the degree of interference associated with performing the task while retaining a full memory load relative to a half memory load. To the extent that encoding operations require cognitive resources, the SOAs at which maximum performance levels are first reached will be longer when individuals are maintaining a full memory load relative to a half memory load. If decision processes require cognitive resources, then overall response times (independent of SOA) will be greater when individuals are maintaining a full memory load. The data from the experiment are presented in Fig. 1 . For purposes of data analysis, encoding times were again computed separately for each subject (see Merrill et al., 1987). They are reflected in the inflection point of the response time functions relating response time to SOA in the graphs. Encoding times were not substantially changed for the mentally retarded subjects under physical identity matching instructions (408 vs. 392 msec for full and half memory loads, respectively), whereas encoding times under name identity instructions were slower for mentally retarded individuals maintaining a full memory load relative to those maintaining a half load (575 vs. 467 msec, respectively). In contrast, the encoding times for the nonretarded subjects suffered as a function of memory load under both sets of matching instructions. Physical identity encoding slowed from 298 to 342 msec, and name identity encoding slowed from 408 to 450 msec. These results are in contrast to the commonly held belief that highly familiar stimuli are encoded automatically into memory (e.g., Keele, 1973; LaBerge & Samuels, 1974; Posner & Rogers, 1978). The results are more consistent with recent data suggesting that even the encoding of stimuli as familiar as letters of the alphabet (e.g., Ogden, Martin, & Paap,

Mentally retarded 1000

-\

900

800

700

600

I

0

I

I

I

I

I

I

I

I

I

I

100 200 300 400 500 600 700 800 900 1000

SOA (msec)

Nonretarded 800

700 al

-

\

2

v

'=E

600

a, C

P a:

500

400

1

0

1

1

1

1

1

1

1

1

1

1

100 200 300 400 500 600 700 800 900 1000

SOA (rnsec)

FIG. I. Encoding functions relating response time to SOA as a function of type of encoding and memory load. Key: 0 , W, physical identity encoding while maintaining a full memory load or a half memory load, respectively; 0.0, name identity encoding while maintaining a full memory load or a half memory load, respectively.

72

Edward C. Merrill

1980) and the identification of common objects (Kahneman, Treisman, & Burkell, 1983) may involve attentional resources for nonretarded adults. Perhaps the type of process assessed in this experiment can be viewed as what has come to be referred to as “partially automatic” (see Kahneman & Treisman, 1984). A partially automatic process is one that can be executed while attentional resources are focused elsewhere, but is facilitated when these resources are focused on completing that process (LaBerge. 1973, 1975). It may be that cognitive resources are not required for physical identity and name identity encoding in this experiment. However, subjects who did allocate some of their resources to encoding performed better than those who did not. In light of this possibility, the patterns of group differences in encoding times observed as a function of memory load were quite interesting. The nonretarded subjects exhibited essentially the same degree of interference under both physical identity and name identity matching instructions (44 vs. 42 ms, respectively). In contrast, the magnitude of interference exhibited across encoding conditions by the mentally retarded subjects was quite different. These subjects exhibited an interference effect of only 16 msec for physical identity encoding, but an effect of I08 msec for name identity encoding. It is therefore tempting to speculate that the locus of the interference effects was different for the two groups. For the nonretarded subjects, the resource limitations appear to be associated with some general aspect of encoding. Perhaps it is common for these subjects to generate effortfully some expectation for the second picture of the stimulus pair. When cognitive resources were limited by requiring subjects to maintain a full load in memory, they were unable to do this and performance suffered under both instruction conditions. In contrast, the resource limitation exhibited by the mentally retarded subjects appeared to be stimulus specific. Resource limitations were only observed under name identity matching instructions. Therefore, it would seem that mentally retarded individuals do not routinely make attentional resources available for stimulus encoding. Rather, they allocate resources only if it is essential, or at least obviously important, for the completion of the encoding process. For the retarded individuals it must have been necessary to allocate resources to name identity and not physical identity encoding. The one obvious difference between name identity encoding and physical identity encoding as examined in this experiment was that name identity encoding required access to semantic memory whereas physical identity encoding did not. It therefore seems reasonable to suggest that the differenceexhibited by the mentally retarded subjects reflects a difference in the resource requirements associated with accessing semantic memory and identifying pictures of objects at the basic level; that is, there may be a difference in the degree to which identifying objects at the basic

ATTENTIONAL RESOURCE ALLOCATION

73

level is automatic for mentally retarded relative to nonretarded individuals. However, since automatic processing is most often associated with particular stimulus-response relationships, it is important to recognize that this conclusion may turn out to be different for different sets of stimuli. Nevertheless, it is interesting to observe group differences in the resource requirements of such very basic information-processing operations, The next step will be to examine whether or not these differences are maintained across different basic level categories and if they truly impact upon the performance of more complex cognitive tasks. By comparing the absolute level of the response time function, it is also possible to examine the effects of memory load on the decision and response components of the stimulus matching task. As can be seen in Figs. I and 2, the pattern of results across groups was quite different from that observed for encoding times. In this case, a significant effect of memory load was observed for physical identity and name identity encoding for both the mentally retarded and nonretarded subjects. However, for both types of decision the effect of memory load was clearly greater for the nonretarded subjects than for the retarded subjects. Deciding whether or not two objects match apparently requires attentional resources for both groups of subjects, but the nonretarded subjects appear to allocate more of their available resources to this decision component than do the retarded subjects. Hence, when attentional resources are diverted elsewhere the nonretarded subjects exhibit greater interference than do the retarded subjects. It is not possible to ascertain the locus of this difference. It may be that the mentally retarded subjects did not efficiently allocate resources that were available to them. However, it may also be that the mentally retarded subjects were left with fewer resources available after encoding was completed. Distinguishing between these two possibilities will require a somewhat different experimental met hod.

C.

Individual Differences in the Development of Automaticity

The notion of automaticity is central to models of skilled performance across many cognitive and noncognitive domains (e.g., Fisk & Schneider, 1984; Neuman, 1984; Pew, 1974; West & Stanovich, 1978). Understanding how basic processes become relatively automatic and the extent to which the development of automatic processing differs for mentally retarded and nonretarded individuals is therefore an important topic for investigation. The assumption here is that if mentally retarded individuals differ from nonretarded individuals in the manner or rate at which automaticity develops, then these groups will also differ in the rate at which

74

E d n w d C. Merrill

complex skills develop, if they develop at all. The comparison of encoding times in the previous experiment led to the suggestion that mentally retarded and nonretarded individuals may differ in the degree to which basic information-processingoperations can be automatically executed, To the extent that this is true, it is reasonable to hypothesize that automatic processing develops at a slower rate for mentally retarded relative to nonretarded individuals. Merrill, Goodwyn, and Gooding (19%) examined this possibility. The general method used in the experiment involved a visual search task in which subjects determined the category membership of common objects (see Fisk and Schneider, 1983). Single slides contained either two, three, or four pictures of common objects from different natural language categories. Eight categories were used in the experiment (clothing. fourlegged animals, fruit, furniture, musical instruments, tools, toys, and vehicles). All of the pictures were highly typical exemplars (Rosch, 1975) from these categories. Four categories were designated as target categories and were learned by the subjects. The remaining categories were designated as nontarget categories. The subject’s task was to determine whether or not one of the objects pictured in the presented slide stimulus was a member of one of the target categories. The data of interest in the experiment were the slope values of the regression lines relating category decision times to the number of objects pictured in the visual search set. Under conditions in which stimuli are responded to in a consistent manner, the slope values obtained for nonretarded adults gradually decrease as a function of extended practice and become nonlinear, with subjects exhibiting no effect of set size (Fisk & Schneider, 1983). This is one of the characteristics of automatic processing. Automaticity allows for parallel processing (see Schneider & Shiffrin, 1977; Shiffrin & Schneider, 1977). Hence, it was used in this experiment as an index of the development of automaticity. Evidence for retardednonretarded differences in the rate at which automaticity develops would be obtained if the mentally retarded and nonretarded individuals exhibit a different rate of decline in slope values as a function of amount of practice. There is one important issue that had to be dealt with in this experiment. Category decision making was chosen as the learning task because it was felt that all of the subjects would be reasonably familiar with the process and, as a result, automaticity would develop fairly rapidly. In fact, under the right conditions, automaticity appears to develop within hours (see Ackerman & Schneider, 1985). However, this choice was also problematic in that mentally retarded and nonretarded individuals were likely to differ in the degree of knowledge (automaticity) that they would

ATTENTIONAL RESOURCE ALLOCATION

75

exhibit at the start of the experiment. At the very least, we know that category decision times of mentally retarded and nonretarded individuals are likely to differ (e.g.. Davies et d.,1981; Merrill, 1985; Sperber ef d., 1982). To alleviate this problem, it was necessary to create conditions under which the retarded and nonretarded subjects exhibited equivalent performance levels. The development of automaticity could then be traced from this point. One method of equating groups was built into the basic procedure. Since both groups of subjects received extensive practice, it was possible to assess the development of automaticity from both the actual starting point of the experiment and from the point at which mentally retarded subjects performed at levels equivalent to those of the nonretarded subjects at the start of the experiment. Therefore, two sets of analyses were performed. The data from the experiment are presented in Figs. 2 and 3. Figure 2 presents slope values of the mentally retarded and nonretarded subjects as a function of trial blocks from the start of the experiment, and Fig. 3 superimposes the performance of the mentally retarded subjects on that of the nonretarded subjects from their point of initial equivalence. Each trial block included 72 trials (36 target and 36 nontarget trials). Only the data from the target trials are included in the figures. An initial comparison was made between the average performance of the mentally retarded and nonretarded individuals in the first trial block. This was done to determine the degree to which the current procedure yielded results similar to those obtained in previous studies of category decision making. As expected. the average slope value for the retarded subjects was approximately twice that obtained for the nonretarded subjects (163 vs. 80 msec, respectively). The slopes of this visual search task were necessarily steeper than those of the memory search task of Merrill (1985) because in the current experiment subjects actually had to perform two searches at the same time: a memory search of the target category labels and a visual search of the slide stimulus. However, the ratio of retarded to nonretarded performance was quite similar to previous group comparisons of category decision making (Davies et a / . , 1981; Merrill, 1985). so it was assumed that the current experiment had not created an entirely new experimental situation. As can be seen in Fig. 2, the nonretarded subjects achieved a nonsignificant slope value and, hence, exhibited signs of automatic processing much more rapidly than did the mentally retarded subjects. For the nonretarded subjects, the effect of set size was no longer significant after five trial blocks. This was not evident for the mentally retarded subjects until the 15th trial block. We therefore reached the tentative conclusion that nonretarded individuals acquire automatic processing at a much faster

76

Ednwrd C . Mcv-rill

200

150

V

-E 2

100

3 m

B

v)

50

0

1111111111111111 1

2

3

4

5

6

7

8

9

10

11

121314

15

16

Stimulus block

FIG. 2. Slope values of mentally retarded sions.

( 0 )and

nonretarded (0) sub.jecrs over all ses-

rate than do mentally retarded individuals. As mentioned earlier, however, a portion of this effect may be associated with the initial differcnces between the two groups. To compensate for this, the data were reexarnined after equating groups on initial performance levels. This was done in a relatively arbitrary manner for this experiment. The trial block selected as the initial block for the mentally retarded subjects was simply the earliest one in which these subjects exhibited an average slope value that was smaller than that obtained in the first trial block by the nonretarded subjects. This was Trial Block 5, in which the retarded subjects achieved an average slope value of 70 rnsec. As can be seen in Fig. 3,

77

ATTENTIONAL RESOURCE ALLOCATION

11111111111 1

2

3

4

5

6

7

8

9 1 0 1 1

Stimulus blmk

FIG. 3 . Slope value5 of mentally retarded initial equivalence.

( 0 )and

nonretarded (0) subjects from point of

even if we consider Trial Block 5 to be the starting point for the mentally retarded subjects we still find that they exhibit signs of automatic processing much later than do the nonretarded subjects (Trial Block I I vs. Trial Block 5 , respectively). Thus, on the basis of both sets of analyses we were forced to conclude that mentally retarded individuals require considerably more practice at a given task to achieve automaticity than do nonretarded individuals. This conclusion clearly needs to be confirmed across a variety of task and stimulus manipulations before it can be generally accepted. In the case of this experiment, there was a problem in trying to equate the initial

78

Edwurd C. Merrill

performance levels of the two groups. It is possible that a recently acquired familiarity with stimuli is quite different from a long-standing familiarity, and that our effort to equate group performance in this experiment was inadequate. In fact, the difference in the shapes of the response time functions after equating group performance levels suggests that the two groups may have differed in some important ways. There are at least two other ways to equate initial performance levels that may prove more suitable. For example, it may be reasonable to examine the acquisition of automatic processing using artificial categories that are equally unfamiliar to all subjects. Another possibility may be to equate initial performance by varying exemplar typicality systematically across groups. Presenting high typical exemplars to the mentally retarded subjects and low typical exemplars to the nonretarded subjects may provide the two groups with equally familiar category/exemplar relationships. Nevertheless, the results of this initial experiment support the general hypothesis that mentally retarded individuals acquire automaticity at a slower rate than do nonretarded individuals. D.

Resource Allocation and the Development of Automaticity

If mentally retarded and nonretarded individuals actually differ in the rate at which processes become automatic, then it will be important to consider the possible causes of this difference. Some of the factors that influence the acquisition of automatic processing by nonretarded adults in search and detection tasks have been studied extensively by Schneider and colleagues (e.g., Fisk & Schneider, 1983; Schneider & Fisk. 1982a, 1982b; Shiffrin. Dumais, & Schneider, 1981). The most important factors appear to be the extent of the practice, the consistency with which the to-be-automated process is executed, the nature of the response required, and previous learning history. One factor that seems particularly relevant to individual differences in the development of automaticity has not been the focus of single group investigations. It is reasonable to suggest that the transition from effortful to automatic processing depends, in part, on how much effort is devoted to the cognitive task during the period of extended practice. When describing individual differences we must consider the possibility that some individuals devote less effort, or fewer processing resources, to practice than do other individuals. In the case of mentally retarded individuals, it may be that they allocate their resources poorly or have fewer resources available at the start of practice. In either case, this may result in mentally retarded individuals acquiring automaticity at a slower rate than do non-

ATTENTIONAL RESOURCE ALLOCATION

79

retarded individuals. Cha and Merrill (1990) conducted an experiment to determine whether or not there is a relationship between the amount of cognitive resources devoted to practice and the rate at which automatic processing develops. Only nonretarded subjects were used in this experiment. The general procedure combined a category decision task with the concurrent memory load procedure (Logan, 1979). Subjects were shown a series of slides picturing one common object per slide and were required to determine whether or not the object belonged to one of four natural language categories designated as target categories. At the same time, they were required to maintain a list of numbers in immediate memory. The number lists contained either two or eight numbers. The amount of cognitive resources that were available to perform the category decision task throughout the experiment was manipulated by varying the percentage of the time subjects were required to learn eight- versus two-number lists. One group of subjects received eight-number lists 50% of the time and two-number lists 50% of the time, a second group received eight-number lists 75% of the time and the two-number lists 25% of the time, and a third group received eight-number lists 100% of the time (excluding 24 test trials at two digits in each experimental session). The difference in category decision times when subjects were performing the category decision task with eight numbers relative to two numbers in memory was used as the measure of automatic processing in this experiment. Since automatic processes can be executed without attentional resources, diverting resources from the category decision task should not influence performance. Therefore, there should be no difference between the eight- versus two-number memory load conditions after automatic processing has developed. Further, if the rate of development for automatic processing depends on the amount of resources that can be devoted to practice, than subjects who received eight-number lists only 50% of the time would be expected to exhibit evidence of automaticity sooner than would the other two groups of subjects. All subjects took part in the experiment on four consecutive days. The first day was devoted entirely to practice in coordinating the two tasks. Therefore, subjects received a different category set every 10 trials during the first experimental session. Automaticity does not develop unless the same set is used consistently over trials, so this procedure afforded practice in the two tasks without changing the effort required to perform just the category decision task. Subjects received the same category set on each of the next three days of the experiment. Midway through and at the end of each experimental session subjects were tested for the develop-

80

Edwiird C'. Merrill

500

-8

-2 .-E

400

300

%

C

0 v) P

0

200

100

1

2

3

4

5

6

Test block

FIG. 4. Response time differences between two-number and eight-nunher memory load conditions a s a function of trial blocks. Key: 0.0 , 8. received eight-number memory load 50, 75, or 100% of the time, respectively.

ment of automaticity. During this phase, they received 24 category decision trials, 12 while retaining a two-digit memory set and 12 while retaining an eight-digit memory set. Figure 4 presents the difference in performance of each group in the eight-number versus two-number memory load conditions across these six testing blocks. As can be seen in this figure, the difference between these two conditions decreases rapidly at first and then gradually stabilizes over blocks. While these differences never reached zero. they did continue to approach zero throughout the sessions. Of primary interest was the finding that this difference was no longer significant for the 50% group in the fifth and sixth sessions. In contrast, the difference in response times between the eight- and two-number memory load conditions remained significant throughout this experiment for the 75 and 100% groups. It will be necessary to conduct longer duration experiments to determine when, if at all. automaticity develops under conditions in

A’ITENTIONAL RESOURCE ALLOCATION

81

which attentional resources are diverted from the primary task. Still, the results of this experiment suggest that the amount of resources available during practice does influence the rate at which automatic processing develops. Understanding the nature of this relationship may provide some insight into individual differences in the development of automatic processing. This study only examined the relationship between effort devoted to practice and the rate at which automaticity develops for nonretarded subjects. It will be important to ascertain whether or not this relationship also holds for mentally retarded individuals, and whether or not the relationship takes the same form for mentally retarded and nonretarded individuals. As these investigations continue, there are some important issues that must be considered. The relationship between the allocation of attentional resources to practice and the rate at which automatic processing develops will not be a simple one to evaluate. Again, the desire to assess this relationship across different groups of subjects magnifies the difficulties. For example, we should expect that the nature of this relationship may change as a function of the resource requirements of the process that is undergoing practice. In the experiment reported here, a category decision task was selected in which subjects were required to perform an already well-learned activity. Under these task conditions, relatively small but consistent differences were found between groups. It is reasonable to suggest that these differences would be greater if we had selected some process that required a significantly greater amount of the subjects’ available resources to be successfully completed. Similarly, starting with different groups of subjects with potentially different resource demands associated with the processing operation under investigation may make it difficult to distinguish the degree to which the relationship between resource allocation to practice and the development of automaticity is stimulus dependent or person dependent. Therefore, it will be important to proceed with caution when evaluating differences between individuals who may have different amounts of processing resources at the start of practice, but may also differ in the degree to which the process of interest demands access to those resources. V.

DISCUSSION AND CONCLUSIONS

The research described in this article illustrates some of the ways in which the cognitive performance differences of mentally retarded and nonretarded individuals may be related to their relative abilities to allocate attentional resources effectively to various semantic processing op-

a2

Edward C . Merrill

erations and their components. The effective allocation of these attentional resources is central to the successful completion of a wide variety of simple and complex cognitive activities (e.g., Ackerman, 1987; Logan, 1985). It is therefore essential that we understand how individuals may differ in this ability. It is clear from the studies reported in this article that this relationship may, and probably will, take many forms, The results of the studies reported here are quite promising at this early stage of the research program. Nevertheless, the conclusions that can be drawn from them must remain tentative. Still, it appears likely that mentally retarded and nonretarded individuals differ in many aspects of resource allocation. There were two major differences between mentally retarded and nonretarded individuals that were observed in the series of experiments reported here. First, mentally retarded and nonretarded individuals differed in the degree to which attentional resources were needed to complete some of the very basic information-processing operations associated with stimulus encoding and decision making. Second, mentally retarded and nonretarded individuals appeared to differ in the rate at which information-processing operations become automatic. These two related group differences may have profound effects on the ability of mentally retarded individuals to learn and efficiently execute complex cognitive tasks. As mentioned earlier, it is important that relatively basic processes be executed without depleting the individual’s available resources to ensure that a sufficient amount of resources is left to perform other aspects of the cognitive activity. To the extent that mentally retarded and nonretarded individuals differ in the resource requirements of these relatively basic processes, we should expect group differences in the performance of more complex tasks that rely on the efficient execution of these basic processes. However, this logical relationship has not been demonstrated empirically, and it is important that this be accomplished in the near future. One of the important implications of this “cumulative deficit” hypothesis is that the remediation of relatively complex skills in mentally retarded individuals will require some focus on improving cognitive activities that are already seemingly well learned. Individuals can perform tasks at 100% accuracy, and yet not be able to combine them in the performance of more complicated activities because each individual activity requires high levels of attentional resources to be completed. Trying to coordinate two or more component processes at this stage would only lead to failure. The coordination of component activities will be successful only when these activities can be executed with sufficiently reduced amounts of cognitive processing resources. Despite the general tone of this argument, it is important to recognize

ATTENTIONAL RESOURCE ALLOCATION

83

that it is unlikely that the cognitive difficulties of mentally retarded individuals will be overcome by simply giving sufficient training on the component processes of complex activities. In fact, automatic processing carries a cost as well as a benefit. When a process is executed automatically, it is no longer under the individual’s voluntary control and can actually interfere with the ability to perform other required activities. An early demonstration of this can be found in the work of Stroop (1935). As a function of the automatic nature of word reading, it is generally extremely difficult to ignore the meaning of a printed word and respond to some other aspect of the word stimulus that is inconsistent with the meaning. Hence, there is a great deal of interference associated with trying to name the color in which a word is printed when the word represents a different color name (e.g., saying blue to the word red printed in blue ink). Some recent work (Ellis, Woodley-Zanthos, Dulaney, & Palmer, 1989) suggests the possibility that the ability to override an automatic process may be much more difficult for mentally retarded individuals than it is for nonretarded individuals. So, an increase in the efficiency of specific activities associated with an increase in the relative automaticity of some basic operations may also decrease the relative flexibility of processing for mentally retarded individuals in particular: cognitive flexibility is often considered one of the hallmarks of intelligent behavior (see Sternberg, 1984). It is also important to acknowledge one serious deficiency in the work completed thus far. None of the research described in this article directly addressed the possibility that differences between mentally retarded and nonretarded individuals in resource allocation ability may reflect differences associated with metacognitive skills. This is an important consideration and needs to be addressed in the near future. It is quite possible that all of the retarded-nonretarded differences in performance described here may result from a failure to evaluate accurately the resource demands of information processing and then to allocate resources accordingly. To the extent that it is possible to teach mentally retarded individuals to use their attentional resource system effectively, many deficiencies in general cognitive performance may be alleviated. A full understanding of the causes of retarded-nonretarded differences in semantic processing speed will have to include an analysis of how differences in metacognitive skills influence the allocation of attentional resources to semantic processing. ACKNOWLEDGMENT Research reported in this article was supported in part by Grant #HD23325 from the National Institute of Child Health and Human Developnment.

84

Edward C. M e r r i l l REFERENCES

Ackerman. P. L. (1987). Individual differences in skill learning: An integration of psychometric and information processing perspectives. Psychological Bulletin. 102, 3-27. Ackerman. P. L., & Schneider, W. (1985). Individual differences in automatic and controlled information processing. In R. R. Dillon (Ed.), Individual dijjferences in cognition (Vol. 2, pp. 35-66). New York: Academic Press. Baumeister, A. A., & Kellas, G. (1968). Reaction time and mental retardation. In N. R. Ellis (Ed.). International review of research in mental returdation (Vol. 3, pp. 163193). New York: Academic Press. Borkowski, J. G . . & Cavanaugh. J. C. (1979). Maintenance and generalization of skills and strategies by the retarded. In N. R. Ellis (Ed.). Handbook of mental deficiency (2nd ed.. pp. 569-617). Hillsdale, NJ: Erlbaum. Bray. N. W. (1979). Strategy production in the retarded. In N. R. Ellis (Ed.), Handbook of mental deficiency (2nd ed., pp. 699-726). Hillsdale, NJ: Erlbaum. Bryant, J. T., Deckner. C. W.. Soraci. S. A.. Baumeister, A. A., & Blanton. R. L. (1988). Oddity learning in developmentally delayed children: Facilitation by means of familiar stimuli. Americun Journal on Mental Retardution, 93, 138-143. Campione. J. C.. & Brown, A. L. (1978). Toward a theory of intelligence: Contributions from research with retarded children. Intelligence. 2, 279-304. Cha, K.-H.. & Merrill. E. C. (1990). The effects of memory load on the development .f' automaricitv. Manuscript submitted for publication. Cody, W. J.. & Borkowski, J. G. (1977). Proactive interference and its release in short-term memory of mildly retarded adolescents. Americun Journal of Mental Deficiency, 82, 305-308. Davies, D.. Sperber. R. D., & McCauley. C. (1981). Intelligence-related differences in semantic processing speed. Journal of Experimental Child Psychology, 31, 307402. Dugas, J. L.. & Kellas. G. (1974). Encoding and retrieval processes in normal children and retarded adolescents. Journal of Experimental Child Psychology. 17, 177-185. Duncan. J. ( 1980). The demonstration of capacity limitation. Cognitive Psychology. 12,7596. Ellis, N. R. (1978). Do the mentally retarded have poor memory? Intelligence. 2, 41-54. Ellis, N. R., & Allison. P. (1988). Memory for frequency of occurrence in retarded and nonretarded persons. Intelligence. 12, 61-75. Ellis. N. R.. Katz, E., & Williams, J. E. (1987). Developmental aspects of memory for spatial location. Journal of Experimwtal Child Psychology, 44, 401-412. Ellis. N. R.. & Meador. D. M. (1985). Forgetting in retarded and nonretarded persons under conditions of minimal strategy use. Intelligence, 9, 87-86. Ellis, N. R.. Meador, D. M.. & Bodfish, J. W. (1985). Differences in intelligence and automatic memory processes. Intelligence, 9, 265-273. Ellis, N. R.. Palmer, R. L., & Reeves, C. L. (1988). Developmental and intellectual differences in frequency processing. Developmental Psychology, 24, 38-45. Ellis. N. R., Woodley-Zanthos, P., & Dulaney. C. L. (1989). Memory for spatial location in children, adults. and mentally retarded persons. Americun Journal uf Mental Retordarion. 93, 521-527. Ellis. N. R., Woodley-Zanthos. P., Dulaney, C. L., & Palmer, R. L. (1989). Automaticeffortful processing and cognitive inertia in persons with mental retardation. American Journal of Mental Retardation. 93, 412423. Fish. M. A., & Zeaman. D. (1973).An attention-retention theory of retardate discrimination learning. In N. R. Ellis (Ed.). International review of research in mental retardation (Vol. 6. pp. 169-256). New York: Academic Press.

ATTENTIONAL RESOURCE ALLOCATION

85

Fisk. A. D., & Schneider, W. (1983). Category and word search: Generalizing search principles to complex processing. Journal of Experimental Psychology: Learning. Memory, and Cognition. 9, 177-195. Fisk, A. D., & Schneider, W. (1984). Memory as a function of attention, level of processing and automatization, Journal of Experimental Psychology: Learning. Memory, and Cognition. 10, 118-197. Hagen, J. W.. & Huntsman, N. (1971). Selective attention in mental retardation. Developmental Psychology. 5 , 151-160. Hagen, J. W.. Meacham, J. A., & Mesibov. G. (1970). Verbal labeling, rehearsal, and shortterm memory. Cognitive Psychology, 1, 47-58. Hams, G. J.. & Fleer, R. E. (1974). High speed memory scanning in mental retardates: Evidence for a central processing deficit. Journal of Experimental Child Psychology, 17, 452-459. Hasher, L., & Zacks. R. T. (1979). Automatic and effortful processes in memory. Journal of Experimental Psychology: General, 108, 356-388. Hasher, L., & Zacks, R. T. (1984).Automatic processing of fundamental information: The case for frequency of occurrence. American Psychologist, 39, 1372-1388. Heal, L. W., & Johnson, J. T. (1970). Inhibition deficits in retardate learning and attention. In N. R. Ellis (Ed.), International review of research in mental retardation (Vol. 4, pp. 107-144). New York: Academic Press. Hunt, E. (1978). Mechanics of verbal ability. Psychological Review, 85, 109-130. Hunt, E., Frost, N., and Lunneborg, C. (1975). Individual differences in cognition: A new approach to intelligence. In G. Bower (Ed.). The psychology of learning and motivation (Vol. 7, pp. 87-122). New York: Academic Press. Jensen, A. R. (1982). The chronometry of intelligence. In R. J. Sternberg (Ed.), Advances in the psychology of human intelligence (Vol I ) (pp. 255-310). Hillsdale, NJ: Erl baum. Kahneman, D. ( 1973). Attention and effort. Englewood Cliffs, NJ: Prentice-Hall. Kahneman, D., & Treisman, A. (1984). Changing views of attention and automaticity. In R. Parasuraman & D. R. Davies (Eds.), Varieties of attention (pp. 29-61). New York: Academic Press. Kahneman, D., Treisman, A., & Burkell, J. (1983). The cost of visual filtering. Journal of Experimental Psychology: Human Perception and Performance. 9, 5 10-522. Keating. D. P.. & Bobbitt. B. L. (1978). Individual and developmental differences in cognitive processing components of mental ability. Child Development, 49, 155-177. Keele, S. W. (1973). Attention and human performance. Pacific Palisades, CA: Goodyear. Kerr, B. (1973). Processing demands during mental operations. Memory & Cognition, I, 40 I4 12. Krupski, A. (1975). Heart rate changes during a fixed reaction time task in normal and retarded adult males. Psychophysiology, 12, 262-267. Krupski, A. (1977). The role of attention in the reaction time performance of mentally retarded adolescents. American Journal of Mental Deficiency, 82, 79-83. LaBerge, D. (1973). Attention and the measurement of perceptual learning. Memory & Cognition, I, 268-276. LaBerge, D. (1975). Acquisition of automatic processing in perceptual learning and attention. In P. M. A. Rabbit & S . Dornic (Eds.), Attention and performance (Vol. 5 , pp. 50-64). New York: Academic Press. LaBerge, D., & Samuels. S. J. (1974). Toward a theory of automatic information processing in reading. Cognitive Psychology, 6, 293-323. Lipps Birch. L. (1976). Age trends in children's time-sharing performance. Journal of Experimental Child Psychology. 22, 33 1-345.

86

Edward C . Mrrrill

Lipps Birch, L. ( 1978). Baseline differences. attention, and age differences in time-sharing performance. Journal of Experimental Child Psychology, 25, 505-5 13. Logan. G. D. (1979). On the use of a concurrent memory load to measure attention and automaticity. Journal of Experimentul Psychalogv: Human Perception and Performcince. 5 , 189-207. Logan, G. D. (1985). Skill and automaticity: Relations, implications, and future directions. Cancidian Joitrnol of Psychology, 39, 367-386. Luria, A. (1963). The mentally retarded child. New York: Pergamon. Maccoby. E. E.. & Hagen, J . W. (1965). Effects of distraction upon central versus incidental recall: Developmental trends. Journal of Experimentul Child Psychology, 2, 280-289. Maisto. A. A.. & Jerome, M. A. (1977). Encoding and high speed memory scanning of retarded adolescents. American Joimul of Mental Deficiency, 82, 282-286. Manis. F. R., Keating. D. P., & Morrison. F. J. (1980). Developmental differences in the allocation of processing capacity. Journal of Experimental Child Psychology, 29, 1 5 6 169. McCauley, C., Kellas, G., Dugas. J.. & DeVellis, R. F. (1976). Effects of serial rehearsal on memory search. Journal of Educational Psychology, 78, 474481. Meador. D. M., & Ellis, N. R. (1987). Automatic and effortful processing by mentally retarded and nonretarded persons. American Journal of Mental Deficiency, 91,613-619. Merrill. E. C. ( 1982). Intelligence-related dqferences in the ejJciency .f verbal processing. Unpublished doctoral dissertation. Vanderbilt University. Nashville, TN. Merrill, E. C. (1985). Differences in semantic processing speed of mentally retarded and nonretarded persons: Comparison of short- and long-term memory processing. Americun Journal of Mental Deficiency, 90, 71-80. Merrill. E. C. ( 1990). Intelligence-related dj‘ferences in the resource requirements of siitnu/its encoding and decision processes. Manuscript submitted for publication. Merrill. E. C.. Goodwyn. E. H.. & Gooding. H. L . (1990). Intelligence-related dqferences in the development of autrimafic processing. Manuscript submitted for publication. Merrill, E. C., & McCauley. C. (1988). Phasic alertness and differences in encoding speed. American Joiirnal on Mental Returdation. 93, 245-249. Merrill, E. C., Sperber, R. D., McCauley. C., Littlefield, J.. Rider. E. A., & Shapiro. D. (1987). Picture encoding speed and mental retardation. Intelligenre. 11, 169-191. Mulhern, T., & Baumeister, A. A. (1971). Effects of stimulus-response compatibility and complexity upon reaction times of normals and retardates. Journal of Cotnpirruiive und Physiological Psychology. 75, 459463. Navon, D.. &Gopher, D. (1979). On the economy of the human processing system. Psyclrological Review. 86, 214-255. Nettelbeck, T., & Brewer, N. (1981). Studies of mild mental retardation and timed performance. In N. R. Ellis (Ed.). Internationul review of research in mental retardation (Vol. 10) (pp. 61-106). New York: Academic Press. Neuman, 0. (1984). Automatic processing: A review of recent findings and a plea for an old theory. In W. Pring & A. F. Sanders (Eds.), Cognition and motor processes (pp. 255293). Heidelberg: Springer-Verlag. Norman. D.. & Bobrow, D. (1975). On data limited and resource limited processing. Journal of Cognitive Psychology, 7 , 44-60. Ogden. W. C., Martin, D. W., & Paap. K. R. (1980). Processing demands of encoding: What does secondary task performance reflect? Journal of Experimental Psychology: Human Perception and Performance, 6 , 355-367. Parasuraman. R., & Davies. D. R . (Eds.). (1984). Vcirieties of aftention. New York: Academic Press. Pew, R. W. (1974). Levels of analysis in motor control. Brain Research. 71, 393400.

A7TENTIONAL RESOURCE ALLOCATION

87

Pick. A. D.. Frankel, D. G., & Hess, V. L. (1975). Children's attention: The development of selectivity. In E. M. Hetherington (Ed.). Review of childdevelopment research (Vol. 5 . pp. 325-383). Chicago: University of Chicago Press. Posner. M. I . (1978). Chronometric explorations of mind. Hillsdale. NJ: Erlbaum. Posner. M. I.. & Boies. S. J. (1971). Components of attention. Psychological Review, 78, 391408.

Posner. M. I.. & Mitchell, R. F. (l%7). Chronometric analysis of classification. Psycliokogical Review, 74, 392409. Posner. M. I.. & Rogers, M. G. K.(1978). Chronometric analysis ofabstraction and recognition. In W. R. Estes (Ed.), Handbook of learning and cognitive processes. Vol. 5: Human information processing (pp. 143-188). Hillsdale, NJ: Erlbaum. Posner, M. I., & Snyder, C. R. R. (1975). Facilitation and inhibition in the processing of signals. In P. M. A. Rabbitt & S. Dornic (Eds.). Attention and performance (Vol. 5 , pp. 669-682). New York: Academic Press. Rosch, E. (1975). Cognitive representations of semantic Categories. Journal of Experimental Psycliology: General, 104, 192-234. Rosch, E., Mervis, C. B.. Gray, W. D.. Johnson, D. M., & Boyes-Braem, P. (1976). Basic objects in natural categories. Cognitive Psychology, 8, 382439. Runcie, D., & O'Bannon. R. M. (1975). Relationship of reaction time to deceleration and variability of heart rate in nonretarded and retarded persons. American Journal of Mental Dyficienc.v. 79, 553-558. Schneider, W. (1985). Toward a model of attention and the development of automaticity. In M. I. Posner & 0. S. Marin (Eds.), Attention and performance (Vol. 11, pp. 473492). Hillsdale, NJ: Erlbaum. Schneider. W.. & Fisk. A. D. (1982a). Concurrent automatic and controlled visual search: Can processing occur without cost'? Journal of Experimental Psychologv: Learning, Memory. and Coxnition. 8, 261-278. Schneider, W.. & Fisk. A. D. (1982b). Degree of consistent training and the development of automatic processing. Perception und Psychophysics, 31, 160-168. Schneider. W., & Shiffrin, R. M. (1977). Controlled and automatic human information processing: I. Detection, search, and attention. Psychological Review. 84, 1-66. Schneider, W.. Shiffrin. R. D.. & Dumais. S. T. (1984). Automatic and control processing and attention. In R. Parasuraman & D. R. Davies (Eds.), Varieties of attention (pp. I27). New York: Academic Press. Shiffrin. R. M., Dumais, S. T.. & Schneider, W. (1981). Characteristics of automation. In J. B. Long & A. D. Baddeley (Eds.). Attention and Performance IX (pp. 223-238). Hillsdale, NJ: Erlbaum. Shiffrin. R. D., & Schneider, W. (1977). Controlled and automatic human information processing: I I . Perceptual learning, automatic attending. and a general theory. Psychologicul Review, 84, 127-190. Silverman, W. P. ( 1974). High speed scanning of nonalphanumeric symbols in cultural-familially retarded and nonretarded children. American Journal of Mental Deficiency. 79, 44-51.

Smith, E. E.. Shoben, E. J.. & Rips, L. J. (1974). Structure and process in semantic memory: A featural model for semantic decision making. Psychological Review. 81,214-241. Soraci. S. A.. Barlean, J. L., Haenlein, M., & Baumeister, A. A. (1986). Lower sensitivity to alterations of auditory relational information in mentally retarded than in nonretarded adults. Psysiological Psychology, 14, 146-149. Sperber. R. D., Davies. D., Merrill, E. C., & McCauley. C. (1982). Cross-category differences in the processing of subordinate-superordinate relationships. Child Development. 53, 1249-1253.

88

E d w a r d C. M e r r i l l

Sperber, R. D., & McCauley. C. (1984). Semantic processing efficiency in the mentally retarded. In P. Brooks. R. Sperber, & C. McCauley (Eds.), Learning and cognition in the mentally retarded (pp. 141-163). Hillsdale, NJ: Erlbaum. Sperber, R. D., Ragain. R. D.. & McCauley, C. (1976).Reassessment of category knowledge in retarded individuals. American Journal of Mental Deficiency, 81, 221-234. Sternberg, S. (1%9). The discovery of processing stages: Extensions of Donders’ method. Acta Psychologica. 30, 276-315. Sternberg, R. J. (1984).Toward a triarchical theory of human intelligence. Brain andEehavioral Sciences. 7 , 269-315. Stroop, J. R. (1935).Studies in interference in serial verbal reactions. Journal of Experimental Psychology. IS, 643-661. Vipond, D.(1980).Micro- and macroprocesses in text comprehension. Journal of Verbal Learning and Verbal Behavior, 19, 2762%. Wickens. C. D. (1984).Processing resources in attention. In R. Parasuraman & D. R. Davies (Eds.), Varieties of attention (pp. 63-102).New York: Academic Press. West, R. F.. & Stanvovich, K. E. (1978).Automatic contextual facilitation in readers of three ages. Child Development. 49, 717-721. Woodley-Zanthos, P., & Ellis, N. R. (1989).Memory for frequency of occurrence: Intelligence level and retrieval cues. Intelligence, 13, 53-61, Zeaman. D. (1978).Some relations of general intelligence and selective attention. Intelligence, 2,55-73. Zeaman, D.. & House, B. J. (l%3). The role of attention in retardate discrimination learning. In N. R. Ellis (Ed.), Handbook of mental deficiency (pp. 159-223). Hillsdale. NJ: Erlbaum. Zearnan. D., & House, B. J. (1979).A review of attention theory. In N. R. Ellis (Ed.), Handbook of mental deficiency. psychological theory and research (2nd ed., pp. 63120). Hillsdale. NJ: Erlbaum.

Individual Differences in Cognitive and Social Problem-Solving Skills as a Function of Intelligence ELIZABETH J. SHORT DEPARTMENT OF PSYCHOLOGY CASE WESTERN RESERVE UNIVERSITY CLEVELAND, OHIO 44106

STEVEN W. EVANS WESTERN PSYCHIATRIC INSTITUTE AND CLINIC PITTSBURGH. PENNSYLVANIA 15213

1.

INTRODUCTION

The study of problem solving and other higher order processes of human thought began despite powerful assertions that it could not be studied. Wilhelm Wundt suggested that the only psychological phenomena that could be reliably examined were basic physiological responses such as reflexes, sensation, and perception (Mayer, 1983). Partially due to this opinion, the study of problem solving, along with other higher order processes, did not truly get underway until the 20th century. Studies examining individual differences in problem-solving strategies and abilities between mentally retarded and nonretarded individuals did not begin until well into the second half of this century. Generally speaking, mentally retarded individuals do not perform as well as nonretarded individuals on all cognitive tasks (see Woodley-Zanthos & Ellis, 1989, for a noteworthy exception). While this main effect is not surprising, it provides us with very little information about the nature of the deficit in retarded individuals. Although there have been dramatic increases in the number of studies examining problem-solving differences in the past 20 89 INTERNATIONAL REVIEW OF RESEARCH IN MENTAL RETARDATION. Vol. 16

Copyright 0 1990 by Academic Press. Inc. All rights of reproduction in any form reserved.

90

Elizabeth J . Short arid Strvrn W.E w n s

years, there has been only minimal progress toward understanding the causes underlying these differences. Progress in the field of problem solving has been limited for several reasons, including imprecise definitions of problem solving, imprecise specification of task parameters, and the lack of uniformity in assessment methodology. Definitions of problem solving have been quite varied. Most authors assume problem solving is a higher order skill that is largely dependent on the efficiency of lower order or basic cognitive processes, including perception, attention, working memory, and background knowledge. In addition, researchers and educators alike assume that problem solving involves an unknown route to a goal and the search process employed by the problem solver to achieve this goal (Bereiter & Scardamalia, 1989). Despite this consensus in the field, a point of contention has been where to place the primary emphasis of study. As seen in the definitions to follow, possible candidates have included the problem formulation, goal attainment, and strategy invention phases of problem solving. For example, Polya (1968, p. ix) defined problem solving as "finding a way out of a difficulty, a way around an obstacle, attaining an aim that was not immediately attainable." Implicit in Polya's definition of problem solving is both devising a method of solving a problem and the successful solution of the problem. In contrast, Belmont (1983) emphasized the removal process or method designed to eliminate barriers to goals, not the actual attainment of problem-solving goals. He stated, "a problem is a novel thing that establishes a barrier. Problem solving is not the removal of that barrier, but rather the process by which a method is devised to remove it" (p. 3). Finally, Chi and Bassock (1989) viewed problem solving as "the organization of a derivation of a set of actions which are guided by a general domain principle" (p. 254). Chi and Bassock's (1989) definition focuses our attention on the problem-formulation phase and argues for domain specificity in problem-solving skills. Since all three definitions are meritorious in their own right, this article presents research findings aimed at understanding deficits in all three phases of the problem-solving process. Besides inconsistencies in definitions. there appears to be lack of consensus regarding what tasks involve problem solving. Problem solving is an abstract concept that may encompass many cognitive and social events. A problem may vary greatly in terms of the demands it places on an individual. As a result, researchers studying problem solving have used a variety of tasks as stimuli. Cognitive tasks include such things as anagrams, Tower of Hanoi (Minsky, Spitz, & Bessellieu, 1985; Spitz, 1982), lock box (Tzuriel & Klein, 1983, math problems (Bilsky & Judd, 1986), balance scale (Hall & Day, 1982), and analogies (Short, Schatsch-

COGNITIVE A N D SOCIAL PROBLEM SOLVING

91

neider, Cuddy e f d.,1990; Sternberg, 1982). Social tasks have included such things as means-end problem solving (Evans & Short, 1990). appreciation of humor (Basili & Short, 1988), making friends (Camp & Bash, 1981), and serving as a talk-show host (Donahue, Pearl, & Bryan, 1980). Despite this variety, however, few researchers have compared problem-solving performance across tasks within either the cognitive or social domain (Vernon & Strudensky. 1988), nor have many comparisons been made between the social and cognitive domains (Evans & Short, 1990). In an attempt to understand the variability of problem types and t h e demands inherent in these problems, researchers have begun to classify problems on a continuum from those that are well defined to those that are ill defined (Simon, 1974). In a well-defined problem, the initial state, goal state, and the rules designed to minimize the distance between the two states are clearly specified. By specifying the problem parameters completely. well-defined problems enable the problem solver to evaluate systematically the accuracy of his or her solution (Kahney, 1986). In contrast, ill-defined problems either lack clear-cut initial states, goal states. and/or rules to minimize the distance between the two states ( M a t h . 1989).Given the fuzzy nature of ill-defined problems, particularly the goal state, there appears to be no systematic way to evaluate the accuracy of one's solution (Reitman, 1964). Most problems are neither completely well defined nor completely ill defined, but rather have both well-defined and ill-defined elements. Given this variability in task demands among problem-solving tasks, it is not surprising that the conclusions regarding problem-solving capabilities of the mentally retarded have proved to be inconsistent. A third factor that has contributed to confusion regarding the specific nature of the problem-solving deficit in mentally retarded populations is the lack of consensus regarding the appropriate methodology for studying problem solving. Researchers have employed a variety of methodologies, including chronometric analysis, protocol analysis, error analysis, and observational analysis. The methodology employed in these studies often dictates which phase of the problem-solving process is emphasized. For example, the use of think-aloud protocols during problem solving focuses our attention on the process of inventing a strategy to overcome obstacles, whereas error analysis focuses our attention on goal attainment. While all four methodologies are useful for studying problem solving and will be discussed at some length later in this article, differences in results obtained should be interpreted in light of these methodological discrepancies. The purpose of this article is to provide a comprehensive review of

92

Elizabeth J . Short

titid

Stetpen

W.Evtitis

problem-solving performance in mentally retarded individuals. We begin by presenting a brief overview of problem solving, from a theoretical and practical standpoint. This overview discusses several models of problem solving and their implications for mentally retarded individuals. The major portion of this article is devoted to individual differences in problem solving, methods of assessment, and remedial techniques. Finally, the limitations of current research and future directions are discussed. II.

MODELS OF COGNITIVE PROBLEM SOLVING

Theoretical models of cognitive development approach the study of problem solving from either a macroscopic level (i.e., Piaget and Vygotsky) or a microscopic level (i.e., Sternberg). Piaget and Vygotsky were concerned with developmental changes in cognitive structures and the global processing strategies employed by the learner. In contrast, Sternberg and other information-processing theorists are concerned with isolat”. ing basic cognitive skills responsible for differences in global problemsolving skills. All three points of view are briefly discussed as they pertain to mental retardation. Piagetian psychology has been based, in part, on the developmental study of problem-solving abilities. According to Piagetian theory, children initially solve problems using simple sensorimotor schemes, with problem-solving skills evolving to enable the use of concrete schemes, and eventually abstract schemes. Gallagher and Reid (1981) reviewed the work on mental retardation from a Piagetian perspective and suggested that mentally retarded children progress through the same stages as nonhandicapped children, but at a slower pace. In addition, mentally retarded individuals show more variability among domains of functioning. This developmental difference position has been long argued by Zigler ( 1969). From a Piagetian point of view, individual differences in problem solving and other mental abilities are often seen as stemming from “disturbances in the functioning of the central nervous system” which place limits on the ability of the organism to learn from interactions with the environment (Gallagher & Reid, 1981). Vygotsky, like Piaget, acknowledged a biological influence in problemsolving ability (especially in cases of mental retardation; Wertsch, 1985). The majority of his effort, however, was directed toward uncovering the influence of culture and language on higher order thinking. Societal influences were thought to provide the models for solving problems: strategies that begin on an interpsychological plane (external events) are gradually shifted to an intrapsychological plane where they become the basis of

COGNITIVE AND SOCIAL PROBLEM SOLVING

93

higher level thought (i.e., problem solving). Such a transformation is accomplished through semiotic mediation and is reflected in changes not only in the intrapsychological plane but in the interpsychological plane as well. That is, language serves as a mediator between cognition and behavior. Vygotsky’s emphasis on the role of the environment and social interaction in problem-solving ability has had a significant impact on the education of retarded individuals. For example, Vygotsky suggested that one of the reasons retarded children demonstrate concrete and rigid problem solving is that their environment is deficient in modeling effective strategies for a variety of problem-solving situations. This point was reiterated by Feuerstein (1979), who remarked on the importance of mediated learning experiences; this notion, in fact, lies at the heart of his instrumental enrichment program (IEP) (Feuerstein, 1980). The assumption is that retarded children behave in a concrete and rigid fashion during problem solving largely because these are the strategies that have been modeled for them by teachers and parents. The process is a transactional one. The teacher and parent model concrete strategies because they believe that mentally retarded persons, given their limited cognitive capacity, cannot benefit from observing complex and flexible strategies. The mentally retarded student then engages in concrete strategic behavior because this is his or her only model for behaving. With limited expectations for strategic potential, few attempts are made to model a variety of task-appropriate strategies for mentally retarded persons. This vicious cycle dramatically affects the retarded student’s subsequent level of achievement. Given that inadequate models of strategy selection and application are presented, it is not surprising that mentally retarded individuals appear to be rigid and inflexible in their strategic approach. As noted by Haywood (l989), “the experience of being mentally retarded makes one more so.” Vygotsky further suggested that presenting adequate models for strategic behavior should encourage retarded children to solve more abstract problems (Vygotsky, 1978). Although useful from an educational standpoint, this macroscopic approach to the study of problem solving has not facilitated a clear delineation of the specific deficits responsible for poor performance. Cognitive psychologists, in contrast, approach the study of problem solving at a microscopic level. This microscopic approach has been quite useful for highlighting the basic cognitive processes responsible for differences in competence. An example of this approach can be drawn from Sternberg’s Triarchic theory. According to Sternberg ( 1985), intelligence comprises three information-processing components: knowledge acquisition components, performance components, and metacomponents. The

94

Elizabeth J . Short und StcwcJn W . E w n s

knowledge acquisition process involves the selective encoding, selective combination, and selective comparison of both new and known information. Intelligent people are typically successful problem solvers. They are quite adept at selectively attending to relevant information while ignoring irrelevant information (e.g., encoding). In addition, they are quite proticient at integrating new information in a meaningful way (e.g., combination) and relating new information to known information (e.g., comparison). According to the theory, the performance components are used to implement a decision to solve a problem (Siegler, 1986). Four performance components are typically employed during the problem-solving process, including encoding, inferring, mapping, and application. For example, in solving an analogy, the first step involves encoding or identifying the defining attributes of each term in the analogy (A:B::C:D). Step two involves inferring the relationship between the first (A) and second ( B ) terms in the analogy. The third step involves mapping the relationship between the first (A) and third terms (C). Finally, step four involves applying the relationship observed between the A:B pair to the C:D pair. Although novice and experts alike appear to use the same processes in their problem solving, individual differences in the amount of cognitive resources devoted to each performance component have been obtained (Sternberg & Rifkin, 1979; see section on individual differences in cognitive problem solving for a complete description). Expert problem solvers spend more time on the encoding phase of problem solving than do novices. Given that they have thoroughly represented the problem and have grasped its structure, experts proceed quickly through the subsequent steps of the problem-solving process. In contrast. novice problem solvers spend less time encoding the stimuli and more time on subsequent components in the problem-solving process (Sternberg & Rifkin, 1979). One explanation for this differential deployment of cognitive resources by experts and novices is that novices spend less time encoding in an attempt to minimize the initial processing load on memory. Unfortunately, this incomplete encoding lengthens the processing time devoted to other components and decreases the probability of successful problem solving. The orchestration of the performance and knowledge acquisition components of problem solving into goal-oriented processes is completed by the metacomponents. According to Sternberg (1983, the metacomponents serve the function of the higher order executive which directs the performance and knowledge acquisition components in the “how tos” of problem solving. With sufficient understanding about how to solve a problem, only the metacomponents and performance components are needed (Siegler. 1986). That is, the metacomponents determine which

COGNITIVE A N D SOCIAL PROBLEM SOLVING

95

performance components are needed to perform the task and in which order to execute them, whereas the performance components are actually used to complete the task. Without sufficient knowledge to complete the task, all three components are employed. Under conditions of insufficient knowledge, the knowledge acquisition components are employed to obtain new information necessary for the metacomponents to construct a strategy. The utility of the Triarchic theory for understanding problem-solving performance in the mentally retarded is quite clear. Although all three components may be implicated in the inferior problem-solving performance of mentally retarded persons, Sternberg (1984) argues these deficits are thought to arise primarily from the inferior use of metacomponents. Three scenarios are possible. First, mentally retarded persons’ choice of which performance component and what knowledge acquisition component to use on a particular task may be inadequate due to their inferior metacomponents. Second, inferior metacomponents lead to faulty coordination of controlled and automatic problem-solving strategies. Finally, inferior metacomponents make the monitoring process tedious; therefore it becomes difficult to make corrections in processing strategies midstream. Experimental evidence in support of these three predictions is presented in the next section.

111.

INDIVIDUAL DIFFERENCES IN COGNITIVE PROBLEM SOLVING

Gestalt psychologists identified a cognitive pattern of problem solving in the 1940s which decreased the likelihood of a successful solution. Characterized by repetition of past strategies to solve current problems without adapting to new stimuli or new task demands (Mayer, 1983), this pattern is commonly called cognitive rigidity and is frequently manifested by an inability to generalize. In the past, this inflexible pattern of problem solving has been called functional fixedness (Duncker, 1945), problemsolving set (Luchins, 1942), and negative transfer (Bartlett, 1958). Pertinent to our discussion is the observation that cognitive rigidity has been found in mentally retarded individuals (e.g., Bray, Goodman, & Justice, 1982; Ellis, Woodley-Zanthos, Dulaney, & Palmer, 1989; Ferretti & Butterfield, 1989). Despite the presence of a global problem-solving deficit in mentally retarded persons, few researchers have been able to pinpoint precisely the causes of the problem-solving difficulties encountered by the mentally retarded. Spirited debates have occurred between researchers arguing from the developmental difference perspective (Zigler, 1969) and those

96

Elizabeth J . Short m d Steven W . Eiwtis

championing the deficit position (Ellis, 1969; Ellis & Cavalier, 1982). As noted previously, the developmental difference approach assumes that mentally retarded individuals develop more slowly and reach a lower level of performance than their normally achieving counterparts. In contrast, the deficit position holds that intelligent behavior is based on a small set of cognitive processes, one or more of which are deficient in the retarded population (Detterman, 1987). Although these positions seem to be incompatible, Detterman (1987) has argued that the differences are largely a function of the type of measurement employed. Specifically, the former is based on molar measurement and the latter is based on molecular measurement. Most research on the problem-solving skills of the mentally retarded has used molecular measurement. The following review of intellectual differences in cognitive problem solving examines differences in both academic and experimental problem-solving skills among mentally retarded, learning disabled, and normally achieving children and adults. A.

Academic Problem Solving

The study of academic problem solving has focused largely on one area: the solving of mathematics problems (Bilsky & Judd, 1986; Judd & Bilsky, 1989; Mayer, 1989; Siegler. 1989). The study of mathematical problem solving has primarily centered on arithmetic computational problems, arithmetic word problems, and computer programming. The methodology employed by these mathematical problem solving studies has been quite diverse, ranging from chronometric analysis to protocol analysis (Mayer, 1989). Despite methodological diversity, all researchers studying mathematical problem solving have assumed a multidimensional perspective. The assumption is that problem-solving difficulties may result from a variety of dimensions, including person, task, and strategy variables (Flavell, 1979). While research on mathematical problem solving has been quite extensive to date, the data on mathematical problem solving by mentally retarded subjects has been specifically tied to verbal arithmetic problems. It has been suggested that the difficulties in verbal arithmetic performance experienced by mentally retarded individuals arise largely from problems in a variety of areas, including memory, logical structure, semantic structure, and syntactic structure (Nesher. 1982). Bilsky and Judd (1986) attempted to tease out the importance of these factors to the problem-solving performance of retarded adolescents and nonretarded 10 year olds using verbal arithmetic problems. The methodology employed was that of error analysis. The variable of logical structure was explored in two ways, by problem type (addition vs. subtraction)

COGNITIVE A N D SOCIAL PROBLEM SOLVING

97

and by amount of extraneous information (present vs. absent). The semantic component was assessed by systematically varying the verbal content of the problem. Problems could either be viewed as dynamic and involving active change (i.e., you have three hamburgers on the grill and you take four out of the freezer and put them on the grill, how many hamburgers do you have?) or static and involving no active change (i.e., you have three hamburgers on the grill and you have four in the freezer, how many hamburgers do you have?). Finally, memory was also assessed in two ways: problems were presented once or twice and memory aids (i.e., number cards) were available or not. As expected, retarded subjects performed poorly on all arithmetic problems compared to their nonretarded counterparts. All groups performed better on addition than on subtraction problems. This difference was more pronounced for the retarded students than for the normally achieving students. Based on these findings, the authors concluded that retarded subjects had more trouble representing the problem space and appeared to adopt rote computational strategies (Goodstein, Cawley , Gordon, & Helfgott, 1971) in response to unfamiliar problems. This is consistent with Sternberg’s (1984) notion that mentally retarded subjects’ inferior metacomponents resulted in the improper orchestration of performance and knowledge acquisition components. Thus pattern recognition skills, strategic flexibility, and insufficient background knowledge could all be implicated as potential factors in the inferior performance of mentally retarded individuals on verbal arithmetic problems. Further support for both background knowledge and strategic flexibility as causal factors in the problem-solving difficulties experienced by mentally retarded individuals can be drawn from the problems varying in semantic structure. Mentally retarded subjects had more difficulty with static problems than with dynamic problems. One possible explanation for this can be drawn from the work of Briars and Larkin (1983). Their data suggest that skilled problem solvers are able to draw on their background knowledge in an attempt to re-represent the problem-solving task. In contrast, mentally retarded subjects appear to be bound to strategies that mimic the action of the problem. Given that the subsets of “cooked” and “frozen” were retained as separate entities in the static problem, the retarded students left them as such and therefore failed to solve the problem correctly. For all subjects, memory seemed to affect problem-solving performance in that two presentations of the problem yielded higher solution rates than did one presentation. I n addition, extraneous information had a deleterious effect on both retarded and nonretarded subjects’ performance. The authors conclude by suggesting that perhaps some of the dif-

98

Elizcibrtli J . Short arid S t t v c t i W . Eviitis

ficulties encountered by retarded students on arithmetic problem-solving tasks result from comprehension failures. These failures in turn make problem definition and representation unlikely, which dramatically limits the probability of selecting the appropriate strategy designed to accomplish the goals of the task. Thus, several basic cognitive skills appear to predispose the mentally retarded person to failure in mathematical problem solving. Even on well-defined arithmetic problems, mentally retarded students appeared to experience difficulties in pattern recognition, memory capacity, and strategic flexibility. 8.

Experimental Problem Solving

The experimental work in the area of problem solving has been conducted using well-defined tasks in which the initial and goal states of the problem are specified. Perhaps the classic experimental work on problem-solving skills and the mentally retarded has been done using the Tower of Hanoi task (Spitz, 1982; Spitz, Minsky & Bressellieu, 1985). The Tower of Hanoi problem is a transformation task involving size-graduated disks. Disks are arranged from smallest to largest on a peg board in some predetermined configuration. The subject’s task is to rearrange the disks on the peg board such that they conform to a second contiguration, with the stipulation that a larger disk may never be placed on a smaller disk. The level of difficulty in the Tower of Hanoi problems can be manipulated by changing the initial and/or goal states and the number of disks (Spitz, Webster, & Borys, 1982). In addition, successful solution of this task is dependent on planning and place-keeping operations (Spitz et al., 1982).Search capacity is typically measured in the number of nodes the subject transverses on the way to the goal state, while overall capacity is measured by increasing the difficulty of the problem until failure results. The results obtained for the mentally retarded on this well-defined task have been fairly consistent. Retarded children perform far below the expectations calculated for their mental age. In fact, lags of almost three years have been found (Borys, Spitz, & Dorans, 1982). Spitz et al. (1982) found that although retarded learners appear to behave strategically and to profit somewhat from practice on the task, their limited search skills predispose them to failure. A similar pattern of findings has been observed with young children; but they mature around the third grade and appear to overcome these limitations. These findings nicely complement both the developmental difference perspective (Zigler, 1969)and the deficit perspective (Ellis, 1969). In an attempt to improve the problem-solving performance of the men-

COGNITIVE AND SOCIAL PROBLEM SOLVING

99

tally retarded on this task, Minsky et (11. (1985) developed a training procedure designed to encourage a more systematic search of the problem space. Training consisted of three parts. First, subjects were given assistance in representing the problem space: the “chunking hint” given was “you need to figure out how to get the large disk from this peg (pointing) to this peg (pointing)” (p. 192). This chunking hint was designed to facilitate the encoding phase of the problem-solving process. Second, if subjects failed to profit from the hint, the experimenter modeled the optimal solution for the subject. Modeling was designed to determine whether the subjects could exercise strategic flexibility. Finally, if the subjects still did not succeed after the demonstration, step-by-step assistance was provided in a collaborative fashion. This transactional approach capitalizes on mediated learning experiences and is the approach advocated by Vygotsky (1978) and Feuerstein (1979). Experimental subjects improved their problem-solving performance dramatically on the maintenance task, but were not capable of transferring this skill to a novel task. Although transfer of training was not obtained, mentally retarded subjects appeared to be more confident, more persistent, and less prone to rule violation than were control subjects. Training appeared to improve pattern recognition skills and strategic flexibility, while minimizing the demands placed on memory skills. Although the effects of training suggest some hope for remediating the poor problem-solving skills of the mentally retarded in a limited task situation, no attempt was made to tease out which aspect of training was responsible for changes in task performance. Another task commonly used to examine individual differences in problem-solving skills is analogies. Much of the credit for current interest in the analogy is owed to Sternberg and his componential method (Sternberg, 1977, 1984). Five processes are employed to solve analogy problems: encoding, inference, mapping, application, and preparation-response. Using a precueing procedure, separate estimates of each component can be obtained using the chronometric and error analysis approaches. The analogy is presented in two parts. Initially the subject is presented with zero to three components of the analogy as a precue for the upcoming analogy. The cue is immediately followed by the presentation of the full analogy in the second half of the problem. The assumption is that initial processing time will allow a corresponding reduction in processing time on the second preview. The analogy task employed is the People Pieces task which comprises four two-dimensional elements: height (tall-short), weight (fat-thin), sex (male-female), and color (redblue). McConaghy and Kirby (1987) were the first to explore problem-solving

differences between retarded and nonretarded subjects using this analogy task. Their results suggest that mentally retarded subjects took more time to solve the analogies. made more processing errors, spent less time encoding, and spent more time on other components than their nonretarded counterparts. In addition, McConaghy and Kirby found that mentally retarded and average-achieving subjects both adopted an exhaustive search strategy for the encoding and inference components, whereas they adopted a self-terminating search strategy for the mapping and application components. Despite similarities in search strategies adopted by average-achieving and mentally retarded subjects, little of the variance in problem-solving performance was accounted for by the models employed by the retarded subjects. The authors argue that one potential reason for the failure of this model to predict variance in problem-solving performance was that the retarded subjects seemed not to take advantage of the precued information. In order to assess this notion further, McConaghy and Kirby (1987) trained their subjects to make better use of the cued trial to aid them in their solution of the test analogies. Training consisted of an elaboration strategy which required subjects to specify verbally the dimensions that had been changed and those that had not on the A and B elements. After doing so, they were instructed to compare these verbalized changes and nonchanges from the A and B elements with those changes and nonchanges that were present on the C and D elements. Positive feedback and reinforcement were provided to the experimental group only. In all instances, both experimental and control subjects were instructed to make use of the precued items. Results indicated that trained subjects were less prone to make errors and were faster in their solution times than were nontrained subjects. More specifically, trained subjects spent more time encoding the stimuli than did nontrained subjects. Thus, like the expert problem solver, trained subjects spent more of their processing capacity on the initial step in the problem-solving process, in an attempt to speed u p further processing. Although extensive training appeared to encourage better representation of the problem, retarded subjects appeared to make less effective use of the precued information than had been shown by normally achieving students in previous studies (Sternberg, 1977).This finding again points to both their strategic inflexibility and their capacity limitations as potential factors responsible for their problem-solving difficulties. Individual differences in the ability to solve analogies were further explored by Short, Schatschneider, Cuddy et af. (1990) using both verbal and nonverbal analogies. Rather than adopting the componential method, Short, Schatschneider, Cuddy et al. (1990) used a protocol analysis ap-

COGNITIVE AND SOCIAL PROBLEM SOLVING

101

proach. Bright, average achieving, learning disabled, and mentally retarded students completed verbal and nonverbal analogies on their own (IND) and while thinking aloud (TA). There were five types of analogies: simple matching, addition, subtraction, alteration, and progression. With the exception of simple matching, learning disabled and mentally retarded students were less able to solve all forms of analogies than their normally achieving counterparts. Protocol analysis revealed that mentally retarded students had trouble defining the problem and selecting the appropriate strategy designed to meet the task demands. That is, they were less able to differentiate among problem types than their normally achieving counterparts. As a result of their inability to represent the problem adequately. mentally retarded subjects appeared uncertain about the appropriate strategy for each problem. They often resorted to random guesses and simple matching strategies, regardless of the type of analogy presented. In contrast, learning disabled students were aware of the strategy designed for each task, but had difficulty defining the problem. Thinking aloud proved to be a formidable task for many of the mentally retarded and learning disabled students, with amount of adult prompting for verbalizations significantly elevated in both handicapped groups as compared to their normally achieving peers. In addition, mentally retarded students were deficient compared to normally achieving students in the lexical diversity observed in their verbal protocols. Taken together, these findings suggest protocol analysis is a useful vehicle for examining the factors responsible for differential problem-solving performance. Unfortunately, the protocol analysis did not reveal whether the deficit observed was a function of the mentally retarded person’s limited strategic repertoire or simply poor task analysis. C.

Scientific Problem Solving

Scientific problem solving is a third area of problem-solving research. Two tasks have been employed to examine problem-solving differences between mentally retarded and nonretarded subjects: the balance scale task and the inclined plane task. These tasks attempt to minimize the influence of verbal skills by sdopting the rule assessment method (Siegler, 1976, 1981). Strategies or binary decision rules are measured by administering problems that lend themselves to unique performance patterns dependent on the rule adopted (Ferretti & Butterfield, 1989). The assumption is that brighter students should adopt more sophisticated problem-solving strategies than their retarded counterparts. It is the implementation of these strategies that results in successful problem-solving performance.

I02

Elizdwtli J . Short arid Steven W . Ewim

In a recent study by Ferretti and Butterfield (1989), the scientific problem-solving skills of mentally retarded and gifted children were explored using both the balance scale and inclined plane task. Children were assumed to use one of five rules to solve both problems. Rule I involves prediction from a single dominant dimension (i.e., the weight or angle). Rule 2 involves prediction from a single dominant dimension, yet consideration is also given to the lesser dimension (i.e., distance from either the fulcrum or vertex). Rule 3 involves the simultaneous consideration of both dimensions. Rule 4 involves the integration of both dimensions and the adoption of the addition rule. Rule 5 involves the integration of both dimensions and the adoption of the multiplication rule. Although physical models for each task were initially presented to the subjects, Ferretti and Butterfield assessed the problem-solving skills of their 10-year-old subjects using a paper and pencil format. Results from the balance scale problem suggest that strategy usage was associated with intelligence. That is, mentally retarded children were more apt to rely on a single dimension (Rule 1) in this problem-solving task than were normally achieving or gifted students. I n addition. gifted children were more apt to integrate dimensions by addition (Rule 4)than were their mentally retarded or normally achieving counterparts. Although in general gifted children presented a pattern of more sophisticated rule usage than their mentally retarded counterparts, significant variability in rule use was observed within all groups. Gifted students were not only more sophisticated in their strategic approach to the task but were also more accurate than their normally achieving counterparts, who, in turn, were more accurate than their mentally retarded peers. Results from the inclined plane task suggest that strategy usage was also associated with intelligence. The findings on this task paralleled those obtained for the balance scale problem. Mentally retarded students were more apt to focus on a single dimension of the stimulus array in solving problems, whereas gifted students were more apt to integrate dimensions using addition. Again, variability in rule usage was the norm for these groups. It appears that mentally retarded subjects rarely consider multiple dimensions of the problem as they are solving it. Interestingly, within the mentally retarded group intelligence did not differentially predict rule usage on either task. The potential reasons behind their strategy deficits could be quite numerous, including problem definition, limited memory capacity, limited strategic repertoire, deficient background knowledge or familiarity, discrepant task definition, and motivational problems to name a few. Future research on these tasks should examine each factor systematically. More recently, Day and Hall (1988) attempted to examine how both

COGNITIVE A N D SOCIAL PROBLEM SOLVING

103

background familiarity and task variables influenced performance on the balance scale problem. Their training study was designed to evaluate problem-solving performance in a dynamic fashion by employing the graduated prompt assessment approach. Pretest, training. and maintenance were assessed on a 4-peg balance scale, near transfer was assessed on a 10-peg balance scale, and far transfer was assessed using a doll-sized teeter-totter with a movable fulcrum. Four groups of subjects participated in this study: above average, average, and two groups of educable mentally retarded students. Two groups of retarded students were selected in order to examine the effects of extended training or increased task familiarity on performance. Similar findings to those obtained by Ferretti and Butterfield were obtained based on pretest performance. That is, mentally retarded children typically performed at a Rule I level, focusing on one dimension of the stimulus. Mentally retarded children needed more assistance or hints than did normally achieving children on the balance task. Even when mentally retarded children were trained to mastery, they differed from their nonretarded counterparts in their unaided maintenance and transfer. Above average children maintained what they were taught and demonstrated some spontaneous transfer of this skill. Average achievers maintained training but did not transfer. Finally, retarded children had difficulty maintaining the training. One interesting finding was that, after extended training, retarded subjects were able to re-achieve mastery with less assistance. Day and Hall argue that transfer propensity may be better achieved by equating retarded and nonretarded subjects on their familiarity and facility with the training task. Thus, their study highlights the important role of background knowledge in strategic performance. Static measures were useful for differentiating average from above average students, while dynamic measures were sensitive to posttraining differences. Both measures provide useful information about individual differences in problem-solving performance. D.

Summary

While there has been a conspicuous absence of models of cognitive problem solving to date, researchers have drawn heavily on models of intelligence to guide their research. The one exception to this statement is the model of problem solving proposed by Hayes (1989). This model draws heavily from information-processing theory and suggests that experts differ from novices in a variety of ways, including recognition of the problem, representation of the problem, planning of the solution, implementation of the plan, and evaluation of the effectiveness of the plan

(Hayes, 1989). This information-processing approach has been most useful for illuminating the possible causes of the mentally retarded person’s problem-solving inefficiency. A variety of methodologies designed to study cognitive problem solving point to differences between mentally retarded and nonretarded persons in three areas, including ( I ) the ability to represent the problem, (2) the flexible use of strategies, and (3) the utilization of background knowledge on well-defined tasks. To date, few training studies targeting cognitive problem solving have been conducted. The few noteworthy exceptions (Day & Hall, 1988; McConaghy & Kirby, 1987; Minsky rt id., 1985), while promising. have been primarily directed at remediating all three areas of deficits simultaneously. Systematic studies are needed to tease o u t how training directed at each component separately would affect the problem-solving process. Many questions remain unanswered at present. Would training in problem representation alone better enable problem solvers to select or construct appropriate strategies designed to achieve their task goals‘?Support for this contention can be derived from the memory strategy literature (Campione & Brown, 1978). Would training in specific strategies better enable the problem solver to represent the task adequately? The extensive strategy training literature would suggest that, while strategies designed to achieve specific task goals are teachable, their generalizability is quite limited (Brown, Bransford, Ferrara, & Campione, 1983; Pressley. Borkowski, & Schneider, 1987).

IV.

MODELS OF SOCIAL PROBLEM SOLVING

In contrast to the data presented on cognitive problem solving, the data obtained on social problem solving involves ill-defined tasks. Unlike the domain of cognitive problem solving, numerous models have been generated about social problem solving (Dodge. Pettit, McClaskey, & Brown, 1986; D’Zurilla & Goldfried. 1971; Spivak & Shure, 1974). One of the most useful models of social problem solving has been proposed by D’Zurilla and Goldfried ( 1971). This prescriptive model delineates the way in which individuals should effectively solve problems in the world. According to the model, effective problem solving consists of five components: (a) problem orientation, (b) problem definition and formulation, (c) generation of alternative solutions, (d) decision making, and ( e ) solution implementation and verification. Although these five components are assumed to be orderly, the process is not always a unidirectional one. While D’Zurilla and Goldfield (197 I ) focused their model on problem

COGNITIVE AND SOCIAL PROBLEM SOLVING

I05

solving in social situations, the utility of this model for understanding cognitive problem solving is quite clear. Problem orientation is fundamental to effective problem solving. For every learner, problem orientation encompasses perception of the problem (i.e., recognition and labeling of the problem), attribution of causal responsibility for problem outcome, appraisal of the significance of the problem, feeling of personal control or self-efficiency (Bandura, 1977), and time/effort commitment (i.e., accurate time appraisal and recognition that effort pays off). Unless problem solvers are accurate in their recognition and labeling of the problem type, they will not successfully formulate or implement a strategy designed to solve the problem. I n addition, the realization that the learner has control over the problem outcome and that personal effort leads to successful solution is of paramount importance to effective problem solving. Problem definition and formulation are also essential factors in successful problem solving (Short, Schatschneider, Cuddy et al., 1990). As noted previously, both learning disabled and retarded learners have been shown to be deficient in these data-gathering skills. That is, they were deficient in the data-gathering process and therefore were not as capable of establishing problem-solving parameters. This failure to define and formulate the problem successfully has a dramatic impact on the third step in the problem-solving process-the ability to generate alternative task-appropriate strategies. Successful problem solvers must not only define the problem clearly but must also select from an extensive repertoire of strategies the one that is most apt to yield successful solution of the problem. Thus, the successful problem solver must first generate a list of possible strategies designed to achieve the task goal and must then make a decision about which strategy is most apt to lead to success. This decision-making skill is predicated on relational strategic knowledge (Pressley et al., 1987). The final step in the problem-solving sequence involves solution verification (i.e., monitoring). The ability to monitor solution outcome is central to our notion of intelligence (Pressley, in press). The ability to monitor strategic attempts is essential for the development of relational strategic knowledge. Successful monitoring enables the problem solver to broaden his or her strategic knowledge to include not only the strategy but knowledge concerning the types of tasks for which a strategy is appropriate.

V.

INDIVIDUAL DIFFERENCES IN SOCIAL PROBLEM SOLVING

Despite the proliferation of models of the social problem-solving process, few systematic studies of social problem-solving skills in mentally

I06

Elizabeth

J.

Short

ond

Steven

W .E w n s

retarded individuals have been undertaken to date. When differences in social problem-solving skills have been considered, these differences have been explored in an attempt to understand differences in interpersonal competence (Herman & Shantz, 1983). Two studies have been selected to depict the diversity of problems presented and methodologies employed in the social problem-solving arena. According to Herman and Shantz (1983), “very little study of retarded children’s social knowledge and social reasoning” has been conducted compared to nonretarded children (p. 225). In an attempt to understand the social problem-solving skills of educable mentally retarded children better, Herman and Shantz ( 1983) examined mother-child interaction patterns in mentally retarded and nonretarded children. Social problem-solving skills were assessed using the Alternative Solutions to Problems Task (Spivak & Shure, 1974). Children are presented eight common social situations in which a conflict exists with either a peer, mother, or teacher. The child must generate as many solutions to resolve the conflict as possible. An example would be, What strategies might you use to remove a child from a swing so that you can have a turn? Possible solutions included asking for adult intervention, sharing, and hitting. The results on social problem-solving skills on ill-defined tasks are somewhat consistent with those obtained for cognitive problem-solving skills on well-defined tasks. Mentally retarded children generated fewer solutions to the problems than did their nonretarded counterparts. Although some evidence for diversity of strategy selection was observed within the retarded group, mentally retarded children were less flexible in their application of strategies than the nonretarded children. Herman and Shantz (1983) further demonstrated that social problem-solving skills were related to both the intellectual competence of the child and the directiveness of the mother. They argued the mothers of the mentally retarded children appeared to adopt a “moment-by-moment” monitoring style that did not encourage reflectivity in problem solving on the part of their child. Thus, the presentation of inappropriate strategic models is not specific to cognitive problem solving, but also appears to be common in social problem-solving situations as well. In both cases, these inappropriate models appear to discourage active self-monitoring of problem-solving skills by mentally retarded persons. While this study of social problem-solving differences between retarded and nonretarded children is consistent with the findings on cognitive problem solving, no attempt was made to tease out differences in problem representation. One might legitimately question whether educable mentally retarded children represented or defined the social problem-solving situation in the same manner as did their nonret arded peers.

COGNITIVE AND SOCIAL PROBLEM SOLVING

I07

In a second study designed to assess social problem-solving differences between mentally retarded and normally achieving elementary school children, Basili and Short (1988) explored the ability of mentally retarded and nonretarded children to appreciate and comprehend humor in nonverbal Ziggy cartoons. To assess appreciation of humor, children were asked to rate humorous and neutral cartoons on a four-point scale, from not funny at all to very, very funny. To assess comprehension of humor, children were asked to isolate what aspect of the cartoon results in the humor elicited. Striking differences emerged for both appreciation and comprehension of humor. Mentally retarded subjects appeared to be less sensitive to differences in types of cartoons than their nonretarded peers. That is, their ratings of humor were comparable across both the neutral and humorous cartoons. In addition, they were less able to indicate what aspect of the cartoon was responsible for eliciting the humorous response. The finding that mentally retarded children tend to be differentially less sensitive to cartoons than their normally achieving peers may provide important insights into their deficiencies in social problem-solving situations. In order to construe a cartoon as humorous, the perceiver must encounter an incongruity (punch linekaption) and then be motivated to resolve the incongruity, either by retrieval of information in the jokekartoon or from his or her own personal information. Several explanations can be offered for the appreciation and comprehension differences observed. First, mentally retarded subjects experienced more difficulty representing the problem than did their nonretarded counterparts. Second, mentally retarded subjects may have lacked the background knowledge to resolve the incongruity in the cartoons. Third, given that they were unable to recognize that some of the cartoons were funny while others were not, they uniformly or rigidly applied the strategy of moderute luughter across all situations. Finally, mentally retarded subjects may have been unmotivated to resolve the incongruity in the cartoons. These findings suggest an intimate connection exists between problem definition, background knowledge, strategic flexibility, and motivation in social problem solving. It should be noted that the social problem-solving task employed in this study was of a hypothetical nature. Naturalistic assessments of humor as a social problem-solving task should be conducted in future research. If one were to generalize these findings to social interactions with peers, deficiencies in social interactions may arise in part due their inappropriate shared affect. Perhaps their difficulties in social interactions in large part may be due to their deficiencies in social problemsolving skills. Although there has been relatively little research on the social problem-

I08

Elizrrbetli J . Short orid S t e w n W . Ewins

solving skills of mentally retarded individuals, training programs have proliferated nonetheless. Social problem-solving training has focused largely on both problem-solving skills and social skills. Problem-solving skills help us to decide what the most appropriate response is in a given situation and social skills enable us to deliver an effective response (Castles & Glass, 1986). Mentally retarded persons have been shown to be deficient in both areas (Matson, Kazdin, & Esveldt-Dawson, 1980). Given this fact, both problem-solving and social skills have been targeted for intervention. Most successful training studies have been done in experimental settings. In a recent naturalistic study, Castles and Glass (1986) examined whether the poor community adjustment of mentally retarded adults was simply a function of social skills or whether it was also a function of poor problem-solving skills. Subjects were assigned to one of four groups: social skills training only, problem-solving training only, both social skills and problem-solving training, or a no-treatment control. Problem-solving training was conducted in small groups and involved the role-playing of the following four steps: (a) generation of alternative solutions, (b) evaluation of possible consequences, (c) selection of best alternatives, and (d) enumeration of means to implement a solution. Social skills training involved the therapist modeling a predetermined solution to a problem, with group members practicing the solution. Examples of social problemsolving situations included refusing unreasonable requests, handling disagreements, and dealing with criticism. The results indicated that mentally retarded subjects profited from the problem-solving training, but not from the teaching of specific solutions to the problem. That is, problem-solving training increased the likelihood of the mentally retarded subjects generating multiple solutions to a problem. As with cognitive problem solving, mentally retarded students often perform inadequately in social problem-solving situations due to strategic inflexibility and inadequate problem representation or encoding. As a result of these difficulties, hasty solutions are executed in an unsystematic manner. Summary Models of social problem solving have been quite useful for elucidating the potential problems experienced by mentally retarded individuals in social situations. These models are only now beginning to be systematically tested. To date, the general findings from research on social problem-solving skills in mentally retarded individuals suggest that deficiencies exist in strategic flexibility, with potential reasons for this difference

COGNITIVE AND SOCIAL PROBLEM SOLVING

109

due to poor problem representation, underdeveloped background knowledge, and unutilized background knowledge. Thus the picture for social problem solving and cognitive problem solving seems somewhat consistent, despite the fact that few studies have systematically compared the two domains. The relationship between social problem solving and general measures of intelligence remains unclear. There is evidence for a close connection between social and cognitive problem-solving skills (Evans & Short, 1990). This research suggests that individual differences in means-end problem solving may in large part be attributable to differences in intelligence. The relationship between intelligence and social problem solving, while not surprising, has been largely overlooked in the literature. Despite this apparent link, why is it that few studies have systematically compared cognitive and social problem-solving skills? One reason for this failure may be because most studies of cognitive problem solving employ well-defined tasks, whereas most studies of social problem solving employ ill-defined tasks. To better understand how one might draw important conclusions across domains of problem solving, it would be helpful to examine specifically the methods of assessment available to date and the information gleaned from each.

VI.

METHODS OF ASSESSMENT

During the last decade, much debate has centered around pinpointing the most effective means of examining problem-solving differences between handicapped and nonhandicapped learners. Traditional or static assessment methods have been the norm for the past century. Static assessment techniques, involving measurement of subjects’ decontextualized knowledge, are important for establishing current levels of knowledge and performance that typically reflect past learning goals. For example, the vocabulary subtest of the WISC-R (Wechsler, 1974) requires children to define words in the absence of a sentence context and without adult assistance. While they are effective for assessing demonstratable competence, static techniques fall short in isolating the cognitive processes responsible for differences in competence. For this reason, educators and researchers have turned to more dynamic methods of assessment, particularly with handicapped populations (Vye, Burns, DelClos, & Bransford, 1987). According to Sattler ( 1988). traditional assessment and dynamic assessment both focus on the evaluation of cognitive functions, however dynamic assessment also addresses their modifiability. Researchers and

110

Elizabeth J . Short m d S t w e n W .Evcins

practitioners who argue for dynamic assessment of mentally retarded students’ performance contend that the demonstrated level of competence obtained from static assessment (i.e.. the observed performance) is not always predictive of potential for learning (Tzuriel & Klein, 1985). Rather than measuring existing skill level and predicting future learning potential. the dynamic approach focuses on the individual’s ability to profit from instruction. Whether dynamic assessment is any better at predicting learning potential than static assessment is an empirical question that has not yet been resolved. Dynamic assessment adopts a process approach to the measurement of problem solving. The assumption is that simply failing to solve a problem may not always indicate the same deficit. Coming close to the solution of a problem represents greater skill than does producing a totally irrelevant response, yet both would be scored as wrong under the static assessment method. Methods have been developed to quantify the steps a person takes to solve a problem and to look for differences between individuals of varying mental ability (Brown et u l . , 1983; Short, Cuddy, Friebert. & Schatschneider, 1990). Unfortunately. these studies often rely on selfreport measures which may be influenced by language ability, motivation, and introspective skills (Short & Weissberg-Benchell, 1989). Nonetheless, such studies hold promise for isolating critical problem-solving differences among handicapped groups. Two approaches have been taken with regard to dynamic assessment: the mediational approach (Feuerstein, 1979; Vye et d., 1987) and the graduated prompt approach (Brown et al., 1983; Campione, Brown, & Ferrara. 1982). Both methods argue for the importance of assessing the modifiability of cognition and hold that individual differences in response to instruction provide important diagnostic information. The purposes of the graduated prompt and mediation approaches, however, are somewhat different. The mediational approach was developed as an educational intervention technique, while the graduated prompt approach arose as an alternative assessment technique in research. Each approach is briefly described. The mediational approach assumes that cognitive functions are impaired or deficient due to “the absence, paucity, or ineffectiveness of the adult-child interaction that produce in the child an enhanced capacity to become modified, that is to learn” (Feuerstein, 1979, p. 70). Possible impairments are proposed to occur at the input phase, the elaboration phase, and/or t h e output phase of learning. Proponents of the mediational approach hypothesize an interactional relationship between an intentioned adult and a novice child. Functioning as an intermediary between the child and the outside world, the adult serves to highlight appropriate

COGNITIVE A N D SOCIAL PROBLEM SOLVING

111

information from the environment such that learning takes place. The aim of this approach is to follow a train-test-train-test model to change the child’s level of knowledge through mediated experiences. The support that the child receives is dependent on his or her level of previously demonstrated knowledge, and assessments occur only after training so as to avoid the mind-set of failure. The graduated prompt approach begins by assessing the student’s independent performance in a static fashion prior to the initiation of the prompting procedure. After a baseline level of performance is established, the graduated prompting procedure is instituted. This procedure involves test items which have all undergone extensive task analysis. Based on this analysis, a set of prompts varying in explicitness is generated. When failure occurs on an item, the experimenter delivers the scripted prompts in order from least explicit to most explicit until successful performance is achieved. Using this approach, it is possible to assess not only the current performance level of the subject, but also ease of learning or learning speed (i.e., number of prompts to criterion). In addition to dynamic assessment, protocol analysis has also been seen as a useful way to obtain information about the cognitive processes responsible for differences in cognitive competence (Ericsson & Simon, 1980). The key component of protocol analysis that offers great promise for elucidating the problem-solving difficulties of mentally retarded individuals is the “think-aloud” technique (Short, Cuddy, Friebert, & Schatschneider, 1990). As the name implies, this technique requires learners to verbalize their thoughts, feelings, and actions while solving the problem. Thus, the protocol provides a window to their covert cognitive, metacognitive, and motivational approaches to the task (Ericsson & Simon, 1985; Short & Weissberg-Benchell, 1989). Thus it should be clear that focusing on the static outcome-that is, the correctness or incorrectness of an answer-may be quite uninformative. The problems with this approach were noted in Bray and Turner’s (1987) discussion of “production anomalies and strategic competence.” Rather than identifying static deficits, Bray and Turner argue that we should place the emphasis on the range of task conditions that enable the learner to use a strategy and those conditions that do not. When problem-solving strategies vary with task variables, such as memory load or task complexity, poor performance represents a production anomaly, not a static deficit. This concept of production anomaly thus argues for a n examination of the relationship between strategic flexibility and task conditions or the point at which the learners’ orchestration of strategies is facilitated or impeded. Dynamic assessment and protocol analysis offer two methodologies that would enable such an examination across both cognitive and

I12

Elizabeth J . Short rind Slewti W . E w m

social problem-solving situations. Further refinement in the area of assessment of problem-solving skills is certainly needed, especially in order to evaluate the effectiveness of interventions designed to improve problem-solving performance.

VII.

METHODS OF FOSTERING EFFECTIVE COGNITIVE AND SOCIAL PROBLEM SOLVING

Although the study of both cognitive and social problem-solving performance in the mentally retarded is still in its infancy, interventions are currently being developed quite rapidly. Several techniques have been shown to be quite effective for improving the problem-solving skills of mentally retardcd individuals, as well as of their normally achieving counterparts. These methods include self-instructional techniques, self-questioning techniques, reciprocal teaching, and thinking aloud. The method that has received the most attention to date is the selfinstructional technique (Meichenbaum, 1977; Whitman. 1987). Effectiveness of the self-instructional method has been documented in the following areas with diverse populations: attention with both learning disabled and normally achieving children (Loper, Hallahan, & Ianna. 1982). handwriting with learning disabled children (Robin, Armel, & O’Leary. 1975), reading with poor readers (Bommarito & Meichenbaum, 1978; Short & Ryan, 1984), mathematics with mentally retarded children (Johnston. Whitman, & Johnson, 1980), impulsivity with preschoolers and elementary schoolers (Bornstein & Quevillon, 1976; Meichenbaum & Goodman, 1971), list recall with mentally retarded children (Brown & Barclay, 1976), and social skills with mentally retarded adults (Castles & Glass, 1986). The self-instructional method has not only been effective for increasing thc target behaviors but has also been successful at promoting some transfer of training across situations and tasks (Whitman. 1987). The assumption is that by employing self-verbalizations, the learner will become more active and self-regulated in the problem-solving process (Ryan, Short, & Weed, 1986). Although self-instructional training techniques are quite diverse in terms of content, a common focus and set of key elements are normally retained. Self-instructional approaches usually focus on problem identification, solution strategies, and support strategies. That is, the self-instructional package usually enables the learner to target a problem for solution, to identify a strategy for solving the problem, and to use strategies for maintaining attention and motivation to perform the task. Key elements of self-instructional packages include verbal directions, model-

COGNITIVE AND SOCIAL PROBLEM SOLVING

I I3

ing, feedback, prompts, systematic fading, and often physical demonstrations. Through the interaction with skilled others, the student gains control over his or her learning. Despite the widely accepted notion that self instructions greatly improve the problem-solving skills of the learner, few studies have explored systematically the way individual difference variables influence the efficacy of self-instructional training. Whitman (1987) notes that relationship of individual difference variables to the self-instructional process, with special attention devoted to concerns for mentally retarded individuals. Two variables thought to impact directly on the usefulness of self-instructional training were knowledge base and linguistic ability of the learners. A second method designed to place the locus of control for feedback and learning in the hands of the student is the self-questioning technique (Wong, 1985), which is a variant on the self-instructional theme. This approach was designed to promote active strategic learning through the use of self-verbalizations (Meichenbaum, 1977). Students can be taught to employ self- or experimenter-generated questions as a means of monitoring their own comprehension (Miller, 1985, 1987; Miller, Giovenco, & Rentiers. 1987; Short & Ryan, 1984; Wong &Jones, 1982). By employing the self-questioning techniques, students are able to assess first-hand the quality of their understanding of new material. Students profit from the self-generated feedback obtained from testing their comprehension in that they recognize either test-readiness or need for further study. An alternative to the self-instructional and self-questioning approaches is the socioinstructional method (Belmont, 1989). The two underlying assumptions of the socioinstructional approach are that strategy instruction should be social in nature and that the ultimate goal of training should be the trunsfer ofresponsibility from the other to the self. A recent example of a strategy based on the socioinstructional approach is reciprocal teaching (Brown & Palincsar, 1982). Reciprocal teaching has been quite successful for delivering process feedback, increasing monitoring skills, and promoting strategic problem solving for diverse populations, including poor readers, normally achieving elementary students, and learning disabled students. The reciprocal teaching technique (Palincsar & Brown, 1984) alters the nature of the teaching relationship. The technique capitalizes on the method of Socratic dialogue for the acquisition of strategic problem solving. In a 20-day training program, four strategies designed to promote self-regulated learning are presented, including summarization, prediction, question generation, and clarification. Initially, the teaching technique is other regulated in that it is modeled by the teacher. Eventually, students assume both the role of the learner and the role of the teacher

I I4

Elizabeth J . Short irnd S t i w ~ nW . t?wm

during the teaching process, and therefore the process becomes self-regulated. Teaching segments of the text to the class provides learners with clear feedback about the depth of their understanding. If students encounter difficulty teaching the lesson, the point at which comprehension difficulties exist becomes readily apparent to both the teacher and the learners. Although monitoring skills were not formally evaluated by Palincsar and Brown (1984) in their reciprocal teaching studies, these skills are critical to teaching effectiveness. Unless an evaluation of the quality of the message occurs, no corrective action can be taken. Reciprocal teaching techniques have been shown to be effective for improving the comprehension performance of less skilled learners. Broadly construed, reciprocal teaching can be thought of as a problem-solving task. Future studies should evaluate whether reciprocal teaching techniques improve problem-solving performance of retarded learners. Finally, thinking aloud has been recently shown to be quite effective for improving the problem-solving performance of normally achieving, learning disabled, and mentally retarded elementary school students (Short, Schatschneidcr, Cuddy ct al., 1990; Short, Evans, Friebert, & Schatschneider, 1990). By requiring subjects to verbalize their thoughts as they solve problems, attention appears to be focused on the task at hand. Thinking aloud thus seems to encourage a more thorough task analysis on the part of the learner. The major use of the think-aloud training technique in social problems has been conducted with aggressive boys by Camp and her colleagues (Camp & Bash, 1981; Camp, Blom, Hebert. & van Doorninck, 1977). Effective social problem-solving skills have been taught by Bash and Camp (1985) through a program of 30 lessons covering such topics as making friends, predicting consequences, and recognizing different perspectives. The think-aloud technique is employed as a vehicle to model effective strategies with young aggressive boys. Their studies have been very effective in reducing aggression and improving prosocial behaviors of kindergarten through third grade males. These studies suggest that thinking aloud during problem solving improves not only academic performance but social skills as well. All four methods of training are designed to enhance learners’ self-regulation or self-control. To date, these four methods of self-control training have proved effective in improving the problem-solving performance of skilled, learning disabled, and mentally retarded learners, as well as their specific memory, attention, and language skills. As has been noted repeatedly in the literature, these newly acquired skills have limited generalizability for all three populations (Brown rt al., 1983). Efforts must be made to teach students how to generalize their skills to a variety of con-

COGNITIVE AND SOCIAL PROBLEM SOLVING

I I5

texts and tasks (Borkowski & Cavanaugh, 1979; Brown et ul., 1983). Clearly, it is important to train flexible application of newly acquired strategies, but students also need to be taught to analyze both the task and the environment to glean information about where and when to apply which strategies. In order to foster optimal strategy selection by learners, educators should build in relational metacognitive knowledge (Borkowski, Weyhing, & Turner, 1986). Learners need to recognize that some strategies are more effective at improving specific performances than are others (Clifford, 1984; Pressley et ul., 1987). Through extensive practice and the provision of explicit performance feedback, children will learn to recognize which strategies are more effective and to employ the better ones objectively (Pressley, Borkowski, & O’Sullivan, 1984).

VIII.

LIMITATIONS OF CURRENT RESEARCH

Perhaps the most serious limitation of the problem-solving research to date is that studies are limited in scope. That is, there are not only few controlled studies on problem-solving skills in the mentally retarded, but in addition many of these studies are of questionable ecological validity. Specifically, cognitive problem solving has been somewhat limited to a small set of obscure experimental tasks, while social problem solving has typically employed hypothetical situations to assess performance. A second limitation of the problem-solving literature is the failure to acknowledge the heterogeneity of mentally retarded populations. Most researchers assume that mentally retarded learners are uniformly deficient. Again, as noted by Bray and Turner (l987), researchers interested in problem-solving performance should set out in search of production anomalies. In particular, they should be looking for situations that foster problem-solving competence and situations that result in problem-solving failures. To assume that all retarded learners are operating at a uniform deficiency level in all situations is misleading. A third limitation of this research is the failure to employ common methodologies across studies. As noted previously, both static and dynamic methods of assessment are critical for understanding the cognitive competence of learners in particular task situations and the processes responsible for differences in competence. Finally, although training studies have shown some promise for improving both the cognitive and social problem-solving skills of mentally retarded individuals, the generalizability of treatment effects has been sorely lacking (Brown et d . , 1983). This finding is not surprising, given that generalizability of treatment effects has been difficult to achieve in

I I6

Elizabeth J . Short (ind Steven W . Ewns

skilled populations as well. Nonetheless, if our understanding of problemsolving skills is to be enhanced, more carefully controlled experimental and educational studies must be conducted with precisely defined populations using a multidimensional approach. The multidimensional approach has been conspicuously absent in most studies of problem solving. Problem-solving performance, while perhaps largely dependent on cognitive factors, has been shown to be dependent on metacognitive. motivational, and emotional factors as well.

IX.

FUTURE DIRECTIONS AND CONCLUSIONS

This multidimensional approach was recently employed by Short and Weissbcrg-Benchell (1989) in their Triple Alliance Model. This modcl appears to be useful for illustrating why individual differences in problem solving and learning may arise. They argue that successful problem solving and academic learning are largely dependent on the formation of an effective alliance between the cognitive, metacognitive, and motivational skills of the learner. That is, expert problem solvers delicately balance and/or coordinate their cognitive, metacognitive, and motivational skills such that goals become easily attainable. By definition, an alliance is “a formal pact or confederation of nations in a common cause.’’ If one substitutes the word skills for nations, it becomes clear that these three skills (i.e., cognition, metacognition, and motivation) are designed to fortify one’s position toward a common goal. Consistent with this multidimensional view, ineffective problem solving by mentally retarded students would result from the formation of a faulty alliance between one or more of the three domains of functioning. For example, failure of mentally retarded persons in cognitive problem solving may result from faulty motivational skills. The research on attributional profiles of the mentally retarded addresses this issue. The perception of lack of control over outcome of performance engenders in the learner feelings of helplessness and reduces the probability of building new cognitive skills (Kanfer & Hagerman, 1981). Metacognitive processes are also important to cognitive skill development. Faulty alliances between cognitive and metacognitive domains may develop as well. On the one hand, given limited cognitive capacity, mentally retarded persons may have difficulty devoting additional cognitive resources to higher order metacognitive planning (see Merrill, this volume). On the other hand, given limited metacognitive knowledge and strategic inflexibility, mentally retarded persons may have difficulty utilizing their cognitive resources. According to Brown (1987). effective

COGNITIVE A N D SOCIAL PROBLEM SOLVING

I17

learners/problem solvers need to adjust and fine tune their cognitive actions continuously using their metacognitive knowledge. This adjustment and fine-tuning process are at the heart of the self-regulation issue (Whitman, 1990). As noted by Whitman (1990), self-regulation is a complex response system comprising such skills as self-reinforcement, self-monitoring, and self-evaluation. Self-regulated learners must form effective alliances between their cognitive, metacognitive, and motivational skills such that successful problem solving occurs (Short & Weissberg-Benchell, 1989). An attempt to examine the utility of the Triple Alliance Model for explaining individual differences in problem-solving performance of normally achieving and mentally retarded students was undertaken by Short, Schatschneider, Basili, and Evans (1989). Using a regression model, they examined whether metacognitve and motivational skills improved the prediction of problem-solving performance (i.e., verbal and nonverbal analogies) above and beyond the prediction obtained from cognitive ability alone (i.e., the Kaufman Assessment Battery for Children-Mental Processing Composite Score). Their data indicate quite clearly that the relationships formed between these three domains of functioning are especially important for handicapped learners. That is, individual differences in problem-solving performance within the handicapped group were best explained by an examination of both the cognitive and metacognitive data. Prediction of problem-solving performance by mentally retarded subjects appeared to be minimal using only the mental processing component of the Kaufman Assessment Battery for Children. However, when the self-regulation component (i.e., strategic awareness) was added to the equation, prediction was dramatically improved for the handicapped learners. Thus, the relationship between cognitive and metacognitive domains appeared to be especially important for mentally retarded students. Individual differences in problem-solving performance for normally achieving students appeared to be largely a function of cognitive ability. Motivational skills did not factor into the prediction of problemsolving performance in these data. What insights can be offered to researchers interested in problem solving from the Triple Alliance Model and the literature on self-regulated learning'? Both viewpoints argue strongly for a more dynamic view of problem solving. Problem solving should not be seen as a static entity, but rather as an evolving process. Self-regulated problem solvers must make decisions about which skills to employ in which situations. This complex process involves an examination of the task at hand, an examination of their strategic repertoire, a construction of a working plan of action, the execution of the plan, and an evaluation and revision of the

I18

Elizuhcth J . Short ctnd S t e ~ w iW . Evnris

plan if necessary. Not only must mentally retarded subjects be given the opportunity to engage in self-regulated learning so that effective alliances can be formed, but also opportunities must be presented so that their selfregulatory skills can be gradually shaped in order to ensure that faulty alliances do not develop.

ACKNOWLEDGMENTS Preparation of this article was supported by the Research Incentive Fund from Western Reserve College. Special thanks are extended to Dr. Jane Kessler. Sarah Friebert. and Patty McKinney for their thoughtful comments on earlier versions of the manuscript.

REFERENCES Bandura. A. (1977). Social learning theory. Englewood Cliffs. NJ: Prentice-Hall. Bartlett, F. C . (1958). 7hinkitrg. London: Allen & Unwin. 1958. Bash. M. A., & Camp, B. W. (1985). Think ciloiid: Incwcising soc~irileind co,gnirit*t,skillsA problem solving progrunr J)r c-hikefren. Champaign, 1L: Research Press. Basili. L. A.. & Short. E. J. ( 1988). Sociul prr,blon-.soli'ing .skills in the rnerrtdly retcirded. Paper presented at the 21st Annual Gatlinburg Conference on Mental Retardation and Developmental Disabilities. Gatlinburg. TN. Belmont. J . M. ( 1983). Concerning Hunt's new way of assessing intelligence. Intc~lli,gc~nc~c~, I, 1-7. Belmont. J . M. ( 1989). Cognitive strategies and strategic learning: The socio-instructional approach. Amuriccin Psychologis/, 44, 142-148. Bereiter. C., & Scardamalia, M. (1989). Intentional learning as a goal of instruction. I n L. B. Resnick (Ed.). K n o t t i n g . lecirning. und instruction: Essuvs in honor c?f'RobertGlciser (pp. 361-392). Hillsdale. NJ: Erlbaum. Bilsky. I,. H.. & Judd. T. (1986). Sources of difficulty in the solution of verbal arithmetic problems by mentally retarded and nonretarded individuals. American Journul of Merrreil Deficiency. 90, 395-402. Bommarito, J., & Meichenbaum, D. (1978). Enhuncing reading compreliensiorr by tirentis of .sc,/flirtstrirc,tional training. Unpublished manuscript, University of Waterloo, Ontario. Borkowski, J. G . , & Cavanaugh, J. (1979). Maintenance and generalization of skills and strategies by the retarded. In N. R. Ellis (Ed.), Handbook of mental dc,ficiencv: Psychologicul theory und ruseorch (pp. 120-220). Hillsdale, NJ: Erlbaum. Borkowski, J. G . , Weyhing. R. S.. & Turner, L. A. (1986). Attribution retraining and the teaching of strategies. Excvprionul Children. 53, 130-137. Bornstein, P., & Quevillon, R. (1976). The effects of a self-instructional package on overactive preschool boys. Journul of Applied Behuvior Anal.vsis. 9, 179-188. Borys. S. V., Spitz, H. H.. & D o r m s , B. A. (1982). Tower of Hanoi performance of retarded young adults and nonretarded children as a function of solution length and goal state. Journal of Experinrental Child Psyctrology. 33, 87-1 10. Bray, N.. Goodman, M. A,, &Justice, E. M. (1982).'l'ask instructions and strategy transfer in the directed forgetting performance of mentally retarded adolescents. Intelligence. 6, 187-200.

COGNITIVE AND SOCIAL PROBLEM SOLVING

I I9

Bray, N. W., & Turner, 1.. A. (1987). Production anomalies (not strategic deficiencies) in mentally retarded individuals. Intelligence, 11, 49-60. Briars, D. J . , & Larkin. J. H . (1983). An integruted model of skill in solving c4einentur~~~ word problems. Paper presented at the annual convention of the American Educational Research Association. Montreal. Brown. A. L. (1987). Metacognition. executive control, self-regulation, and other mysterious mechanisms. In F. E. Weinert & R. H. Klewe (Eds.), Mefircognition. m o t i w t i o n . und understunding (pp. 65-1 16). Hillsdale, NJ: Erlbaum. Brown, A. L.. & Barclay, C. R. (1976). The effects of training specific mnemonics on the metamnemonic efficiency of retarded children. Child Development, 47, 7 1-80. Brown, A. L., Bransford, J. D., Ferrdra, R. A., & Campione, J. C. (1983). Learning, remembering, and understanding. In P. H. Mussen (Ed.), Hundbook ofc/iildp.syc/lology: Cognitive diw4opment (Vol. 3, pp. 77-166). New York: Wiley. Brown. A. L., & Palincsar, A. S. (1982). Inducing strategic learning from texts by means of informed. self-control training. Topics in Leurning und Leurning Disabilities. 2, Iin. Camp. B. W., & Bash. M. A. S. (1981). Think ulortd: Increusing sociulund ~ ~ ~ g n i t i ~ ~ ~ ~ . s k i l l . s A problem-sohing progrum for cliildren (Primury level). Champaign, IL: Research Press. Camp, B. W.. Blom, G. E., Hebert. F., & van Doorninck. W. J. (1977). Think aloud: A program for developing self-control in young aggressive boys. Joumul of Ahnorinul Child Psvi~liology.5, 157-169. Campione J. C.. & Brown, A. L. (1978). Toward a theory of intelligence: Contributions from research with retarded children. Intelligence. 2, 279-304. Campione, J. C.. Brown, A. L., & Ferrara, R. A. (1982). MR and intelligence. In R. J . Sternberg (Ed.), Hundbook qfliumun intelligence (pp. 392-490). London & New York: Cambridge University Press. Castles. E. E., & Glass, C. R. (1986). Training in social and interpersonal problem-solving skills for mildly and moderately mentally retarded adults. Amerirun Journd of Mentul Dqficii~ncy,91, 3 5 4 2 . Chi. M. T. H.. & Bassock. M. (1989). Learning from examples via self-explanations. In L. B. Resnick (Ed.), Knowing. learning. und instruction: Essuys in honor of Robert Gluser (pp. 251-282). Hillsdale. NJ: Erlbaum. Clifford, M. M. ( 1984). Thoughts o n a theory of constructive failure. Educutionul Psychologist. 19, 108-120. Day, J. D., & Hall, L. K . (1988). Intelligence-related differences in learning and transfer and enhancement of transfer among mentally retarded persons. Americun Jourriul of Mentol Returdution. 93, 125-137. Detterman, D. K. (1987). Theoretical notions of intelligence and mental retardation. Americun Jorrrnul of Mentul Dejiciency. 92, 2-1 I . Dodge, K . A., Pettit. G . S.. McClaskey. C. L., & Brown, M. M . (1986). Social competence in children. Monogruphs oftlie Socirtyfiir Reseurch in Child Development. 51(2. Serial NO. 213), 1-85. Donahue, M.. Pearl, R., & Bryan, T. (1980). Learning disabled children’s conversational competence: Responses to inadequate messages. Applied P.~yrho/ingrri.stic.s.1, 387403. Duncker. K. (1945). On problem solving. Psychological Monogruphs. 58(5, Whole No. 270). D’Zurilla. T. J.. & Goldfried, M. R. (1971). Problem solving and behavior modification. Journol o j A hnormul Psychology. 78, 107- 126. Ellis, N . R. (1969). A behavioral research strategy in mental retardation: Defense and critique. American Journul of Mental Dejiciency, 73, 557-567.

I20

Elizabeth J .

Short rind S t i v , n

W .Evrrns

Ellis, N. R.. & Cavalier. A. R. (1982). Research perspectives in mental retardation. In E . Zigler & D. Bailer (Eds.). Mentul returdution: Tlie dervlopmi~ntuld$fi.renc.e controversy (pp. 121-152). Hillsdale. NJ: Earlbaum. Ellis. N. R., Woodley-Zanthos, P.. Dulaney, C. L., & Palmer, R. L. (1989). Automaticeffortful processing and cognitive inertia in persons with mental retardation. Americrrri Journul of Mentul Returdution, 93, 412423. Ericsson, K. A,, & Simon, H. A. (1980). Verbal reports as data. Psycliologic~erlRiwicw. 87, 215-25 I . Ericsson, K. A,, & Simon, H. A. (198s). Protocol analysis. In T. A. Van Dijk (Ed.). Hirndbook ofdiscoitr.se crncr1y.si.s (Vol. 2). London: Academic Press. Evans, S., & Short, E. J. (1990). Individrrul d8ference.s in tnennslend probleni sol,'Ing us 11 .fiinc.tion of irggression irnd intc~lligc~ncx~. Manuscript submitted for publication. Ferretti. R. P., & Butterfield, E. C. (1989). Intelligence as a correlate of children's problem solving. Americun Jortrnul of Mentul Returdution. 93, 424-433. Feuerstein, R. ( 1979). The dynriinic nssessnient of retarded perfi)riners: The lecrrriirig potenl i d iissesstnent device. theory, instruments, und techniques. Baltimore, MD: University Park Press. Feuerstein. R. ( 1980). Instrrtmentul enriclimmt: An intervention progruin for cognitiiv m i d ifiuhilitv. Baltimore, MD: University Park Press. Flavell, J. H. (1979). Metacognition and cognitive monitoring. Americirn Psvc/iologi.st. 34, '

9obYII.

Gallagher. J. M., & Reid, D. K. (1981). The Ieurning theory i f f i u g e t uridInlielder. Austin. TX: Pro-Ed. Goodstein, H. A.. Cawley. J. F.. Gordon, S., & Helfgott, J. (1971). Verbal problem solving among educably mentally retarded children. Americun Joitrncrl of Mc,ntul Delficicwc.y. 76, 238-24 I . Hall, L. K., & Day. J. D. (1982). A cwnpurison o f t h e ziine of proxitnul dcvelopn~rntin leurning disublc,d, educuble mcwtully retarded, und norniul children. Paper presented at the annual meeting of the American Education Research Association, New York. Hayes. J. R. (1989). The complete problem solver (2nd ed.). Hillsdale. NJ: Erlbaum. Haywood. H. C. (1989). Multidimensional treatment of mental retardation. Psychology in Mental Rcturdution und L)cveli~pmc~ntul Disuh

Herman, M. S . . & Shantz, C. U. (1983). Social problem-solving and mother-child interace~ifcrl tions of educable mentally retarded children. Journirl of Applied D e ~ i ~ ~ l o i p ~ ~ iPsychology, 4, 217-226. Johnston, M. B., Whitman, T. L., & Johnson, M. (1980). Teaching addition and subtraction to mentally retarded children: A self-instructional program. Applied Re.seurc/i in Matitul Retardution, 1, 141-160. Judd, T. P., & Bilsky, L. H. (1989). Comprehension and memory in the solution of verbal arithmetic problems by mentally retarded and nonretarded individuals. Jotrrncrl ofEditcutionul Psychology, 81, 541-546. Kahney, H. (1986). froblern solving: A cognitive upprouch. Milton Keynes. England: Open University Press. Kanfer, F. H., & Hagerman, S. (1981). The role of self-regulation. In L. P. Rehm (Ed.). Behavior therapy for depression: Present status andfittitre directions. New York: Academic Press. Loper, A. B., Hallahan. D. B., & lanna, S . 0. (1982). Metaattention in learning disabled and normal students. Lrurning Disubilitirs Quurterly, 5, 29-36. Luchins, A. S. (1942). Mechanization in problem solving. fsychologicul Monogruphs. 54(6. Whole No. 248). M a t h . M. W. (1989). Cognition. New York: Holt. Rinehart & Winston.

COGNITIVE AND SOCIAL PROBLEM SOLVING

121

Matson. J. L.. Kazdin, A. E., & Esveldt-Dawson, K. (1980). Training interpersonal skills among mentally retarded and socially dysfunctional children. Behuviour Reseurch und Therupy, 18, 4 1 9 4 2 7 . Mayer, R. E. (1983). Thinking, problem solving, cognition. New York: W. H. Freeman. Mayer, R. E. (1989). Introduction to the special section: Cognition and instruction in mathematics. Jorrrnol of Educational Psvc'hology. 81, 4 5 2 4 5 6 . McConaghy, J., & Kirby, N. H. (1987). Using the componential method to train mentally retarded individuals to solve analogies. American Journul of Mental Deficiency. 92, 12-23, Meichenbaum. D . M. ( 1977). Cognitive behavior modificution: An integrutive upprouch. New York: Plenum Press. Meichenbaum, D. M.. & Goodman, J. (1971). Training impulsive children to talk to themselves: A means of developing self-control. Journal of Abnormul Psychology. 77, 115126. Miller, G. E. (1985). The effects of general and specific self-instruction training on children's comprehension monitoring performances during reading. Reuding Reseurch Quurterlv. 20, 616-628. Miller, G. E. (1987). Influence of self-instructions on the comprehension monitoring performance of average and above average readers. Journul of Reading Behavior. 19, 303316. Miller, G . E., Giovenco, A.. & Rentiers, K. A. (1987). Fostering comprehension monitoring in below average readers through self-instructional training. Journal of Reuding Behuvior. 19, 379-393. Minsky. S. K., Spitz. H. H., & Bessellieu, C. L. (1985). Maintenance and transfer of training by mentally retarded young adults on the Tower of Hanoi problem. Americun Journu/ of Mentul Deficiency, 90, 190-197. Nesher, P. (1982). Levels of description in the analysis of addition and subtraction ard problems. In T. P. Carpenter, J. M. Moser, & T. A. Romberg (Eds.). Addition, crnd subtraction: A cognitive perspective. Hillsdale, NJ: Erlbaum. Palincsar, A. S., & Brown, A. L. (1984). Reciprocal teaching of comprehension-fostering and comprehension-monitoring activities. Cognirion und Instruction. I, 117-175. Polya, G. ( 1968). Mathetnaticul discovery. Volume 11: On understunding, leurning, und reaching problem solving. New York: Wiley. Pressley, M. (in press). Can learning disabled children become good strategy users'?: How can we find out? In L. Feagans, E. J. Short, & L. Meltzer (Eds.), Subrypes of leurning disubilities. New York: Erlbaum. Pressley, M., Borkowski, J. G.. & O'Sullivan, J. T. (1984). Memory strategy instruction is made of this: Metamemory and durable strategy use. Educutionul Psychologist. 19, 94-107. Pressley, M., Borkowski. J. G., & Schneider, W. (1987). Cognitive strategies: Good strategy users coordinate metacognition and knowledge. In R. Vasta & G. Whitehurst (Eds.), Annuls of child development (Vol. 4, pp. 89-129). Greenwich, CT: JAl Press. Reitman. W. R. (1964). Heuristic decision procedures, open constraints, and the structure of ill-defined problems. In M. W. Shelly & G. L. Bryan (Eds.), Human judgments und oprimulity. New York: Wiley. Robin, A., Armel, S., & O'Leary, K. D. (1975). The effects of self-instructions on writing deficiency. Behuvior Therupy. 6 , 178-187. Ryan. E. B., Short, E. J . . & Weed, K. A. (1986). The role of cognitive strategy training in improving the academic performance of learning disabled children. Journal oflearning Disabilities. 19, 52 1-529. Sattler, J. M. (1989). Assessment of children (3rd ed). San Diego: Jerome Sattler.

I22

Elizuheth J . Short und Steven W .E w n s

Short, E. J.. Cuddy, C. L., Friebert, S. E., & Schatschneider, C. W. (1990). The diagnostic and educational utility of thinking aloud during problem solving. In H. L. Swanson & B. Keogh (Eds.), Leiirning disubilities: l%eoreticuI irnd reseurch is.site.s (pp. 93-109). Hillsdale. N J : Erlbaum. Short, E. J., Evans. S. W.. Friebert, S. E., & Schatschneider, C. W. (1990). Thinking uloitd during problem solving: Fucilitution effects. Manuscript submitted for publication. Short, E. J., & Ryan, E. B. (1984). Metacognitive differences between skilled and less skilled readers: Remediating deficits through story grammar and attributional training. Journal of Edurutionul Psyclrology. 16, 225-235. Short. E. J.. Schatschneider, C. W., Basili, L. A., & Evans, S. W. (1989).I n d i v i d r r d d ~ f i ~ r ences in problem-solving performance: The compensutory role of metucognitive trnd motiiwtionul prrformirnce. Paper presented at the E n d Annual Gatlinburg Conference on Mental Retardation/Developmental Disabilities, Gatlinburg, TN. Short, E. J.. Schatschneider, C. W., Cuddy. C. L.. Evans, S. W., Dellick, D. M.. & Basili, L. A. (1990). Individrrul dgyerences in problem solving us ufunction of thinking ulorrd. Manuscript submitted for publication. Short, E. J., & Weissberg-Benchell, J. (1989). The triple alliance for learning: Cognition. metacognition, & motivation. In C. McCormick. G. Miller, & M. Pressley (Eds.), Cognitive strutegv reseorch: 1mplicution.s f b r the curriculum (pp. 33-63). New York: Springer-Verlag. Siegler, R. S. (1976). Three aspects of cognitive development. Cognitive Psychology. 4, 481-520. Siegler, K . S. (1981). Developmental sequences within and between concepts. Monogruphs oftlre Society Jhr Reseurch in Child Development. 4 6 ( 2 , Serial No. 189). Siegler, R. S. (1986). Cliildren’s thinking. Englewood Cliffs, NJ: Prentice-Hall. Siegler. R. S. ( 1989). Hazards of mental chronometry: An example from children’s subtraction. Jortrniil of Educirtionul Psychology. 81, 497-506. Simon, H. A. (1974). How big is a chunk’? Science, 183, 482-488. Spitz. H. H . (1982). Intellectual extremes. mental age, and the nature of human intelligence. Mrrrill-Pulmer Qiturter1.y. 28, 167- 192. Spitz. H. H.. Minsky, S. K.,& Bressellieu. C. L. (1985). Influence of planning time and first move strategy on Tower of Hanoi problem solving performance of retarded young adults and nonretarded children. American Journul of Mentnl Deficiency. 90, 46-56. Spitz, H. H . , Webster. N. A.. & Borys. S. V. (1982). Further studies of the Tower of Hanoi problem-solving of retarded young adults and nonretarded children. Dei~elopmetrtd Psyrliology, 18, 922-930. Spivack, G., & Shure. M. €3. (1974). Social udjustmcnt of young children: A cognitive upprouch to solving red-life problems. San Francisco: Jossey-Bass. Sternberg, R. J. ( 1977). Intelligence, irformution processing. und unulogicul riwsoninp. Englewood Cliffs, N J : Prentice-Hall. Sternberg, R. J. (1982). Reasoning, problem solving, and intelligence. In R. J. Sternberg (Ed.), Hutidbook of humun intc4igence. London: Cambridge University Press. Sternberg, R. J. ( 1984). Mechanisms of cognitive development: A componential approach. In R. J. Sternberg (Ed.), Mechonisms ofcognitive development. New York: Freeman. Sternberg, R. J. (1985). Bevond IQ: A triurchic theory of intelligence. London: Cambridge University Press. Sternberg, R. J., & Kifkin, B. (1979). The development of analogical reasoning processes. Journal of Experimentul Child Psychology. 27, 195-232. Tzuriel, D., & Klein, P. S. (1985). The assessment of analogical thinking modifiability among regular, special education. disadvantaged, and mentally retarded children. Journu1 of Ahnormul Child Psychology, 13, 539-552.

COGNITIVE AND SOCIAL PROBLEM SOLVING

123

Vernon, P. A , , & Strudensky, S. ( 1988). Relationships between problem-solving and intelligence. Intelligence. 12, 4 3 5 4 5 3 . Vye, N. J.. Burns. M. S., DelClos. V. R., & Bransford, J. (1987). A comprehensive approach to assessing intellectually handicapped children. In C. Schneider-Lidz (Ed.), Dynutnic ussessment: A n intertrctionul iipprouch to evuluuting learning porenriul. New York: Guilford. Vygotsky. L. S. (1978). Mind in society: l h e development of higher psychologicul processes. Cambridge, MA: Harvard University Press. Wechsler. D. ( 1974). Mtrnrtul ,for the Wechsler Inrelligence Scale for Children-Revised. San Antonio, TX: Psychological Corporation. Wertsch, J. V. (1985). Vvgotskv und the socicrl~irtncrrionof mind. Cambridge. MA: Harvard University Press. Whitman. T. L. (1987). Self-instruction, individual differences, and mental retardation. Americun Jorirnul qf Mentiil Deficiency. 92, 213-223. Whitman, T. 1,. ( 1990). Self-regulation and mental retardation. Americun Journul ofMi.titu1 Retardution. 94, 347-362. Wong. B. Y. (1985). Metacognition and learning disabilities. In D. L. Forrest-Pressley, G. E. MacKinnon, & T. G. Waller (Eds.). Merucognition. cognition, und hrrmun petjbrtnunce (Vol 2, pp. 137-180). New York: Academic Press. Wong, B. Y. L.. & Jones. W. (1982). Increasing metacomprehension in learning disabled and normally achieving students through self-questioning training. Lrurning Disuhi1it.v Quurrerlv. 5 , 228-240. Woodley-Zanthos, P.. & Ellis, N . R. (1989). Memory for frequency of occurrence: Intelligence level and retrieval cues. Intelligence. 13, 53-62. Zigler. E. (1969). Developmental vs. difference theories of mental retardation and the problem of motivation. A m e r i c m Journal c!fMentul Dejiciencv. 73, 5 3 6 5 5 6 .

This Page Intentionally Left Blank

Social Intelligence, Social Competence, and Interpersonal Competence JANE L. MATHIAS DEPARTMENT OF PSYCHOLOGY UNIVERSITY OF ADELAIDE ADELAIDE, SOUTH AUSTRALIA 5001, AUSTRALIA

1.

INTRODUCTION

Attempts to define mental retardation have involved a gradual process of revising and refining the set of conditions that are thought to be necessary concomitants to this form of disability. The purpose of these efforts is to determine the type and extent of deficits in competence, which together define mental retardation in a way that distinguishes it from other types of mental and physical handicap. To this end, the American Association on Mental Retardation (AAMR) has repeatedly defined mental retardation in terms of diminished intellectual and social competence (i.e., IQ and adaptive behavior) (Grossman, 1973, 1977, 1983; Heber, 1959, 1961). The use of dual diagnostic criteria to define mental retardation has provided a balance between the assessment of social and nonsocial skills. Nevertheless, Greenspan (1979, 1981a, 1981b) has argued that adaptive behavior scales fail to consider some of the skills that are necessary for competent social interactions. In general, these scales assess a range of skills that are requisite to independent living, together with a variety of maladaptive behaviors which interfere with this process (Meyers, Nihira, & Zetlin, 1979). Thus, measures of adaptive behavior define social competence in terms of practical living skills and an absence of socio-emo125 INTERNATIONAL REVIEW OF RESEARCH IN MENTAL. RETARDATION. Vol. 16

Copyright 0 1990 by Academic Press. Inc. All rights of reproduction in any form reserved.

126

Jane L . Mathias

tional problems, but overlook a whole range of “interpersonal” skills that have been extensively documented within the psychological literature. To date, the implications of this research for the study of mental retardation remain largely unexplored. In response to this, Greenspan (1979, 1981a, 1981b)has argued that the additional consideration of interpersonal compelence (labeled sociul infelligence) would provide a better understanding of how individuals cope in their social environment. This argument has introduced the possibility that existing concepts like social intelligence might usefully be transposed to the investigation of mental retardation. Developing on this, Greenspan (1979, 1981a) has proposed a model of social intelligence that is based on existing research on social intelligence and social and interpersonal competence. Within this model, social intelligence is defined by seven social-cognitive variables that are hierarchically organized according to their underlying psychological processes. In addition, Greenspan (1979, 1981a) has formulated a more general model of personal competence that integrates physical competence, “adaptive intelligence,” and “socio-emotional adaptation” (personality). The adaptive intelligence subsection of this model further combines cognitive intelligence (conceptual intelligence), adaptive behavior (practical intelligence), and interpersonal skills (social intelligence) using existing research to define the first and second of these constructs, and Greenspan’s own model to define the third. By selecting a set of social-cognitive variables to define social intelligence, and combining these with IQ and adaptive behavior in a way that reflects previous theories of intelligence (e.g., E. L. Thorndike, 1920), Greenspan has brought together what have otherwise been quite discrete areas of research. At the same time, he has provided two models that encourage the systematic investigation of social intelligence and its relationship to other forms of competence. The purpose of this article is to (a) review the literature relating to the investigation of socially based competencies, including social intelligence and social and interpersonal competence; (b) outline Greenspan’s models of social intelligence and personal competence with an awareness of how they relate to existing research; and (c) describe research which considers the implications of Greenspan’s models for the definition and diagnosis of mental retardation.

II. A.

INVESTIGATIONS OF SOCIALLY BASED COMPETENCIES

Social Intelligence

Not surprisingly, social intelligence research began as part of a more general investigation of intelligence. The intuitive belief that there exists

SOCIALLY BASED COMPETENCIES

I27

a range of abilities specific to social content and separate from cognitive abilities has prompted researchers to seek empirical evidence for the existence of this construct. To date, there are divided views as to whether social intelligence exists independently of the cognitive skills measured by standard psychometric measures of intelligence. On the whole, critics have focused on the degree of overlap between social and cognitive (especially verbal) measures of intelligence (Marlowe, 1986; Shanley, Walker, & Foley, 1971) and the lack of suitable measures to assess the social domain (R. L. Thorndike & Stein, 1937; Walker & Foley, 1973). In 1960, Cronbach concluded that social intelligence had not yet been adequately defined or measured, while Walker and Foley (1973) stated that there was greater acceptance of this construct than there was scientific evidence for its existence. More recently, Marlowe (1986) has echoed these sentiments, stating that a separate construct of social intelligence had not been demonstrated, despite its theoretical and intuitive appeal. Although there have been numerous attempts to define and describe social intelligence, there are only two theories (ignoring, for the moment, Greenspan’s model) that consider the relationship between social intelligence and other types of intellectual skills. These theories, formulated by Thorndike and by Guilford, have clearly influenced much of the research in this area. 1 . THORNDIKE’S MODEL OF INTELLIGENCE

As early as 1920, E. L. Thorndike introduced the dimension of social intelligence to a model which included social, abstract, and mechanical intelligences. Social intelligence was defined as “the ability to understand and manage men and women, boys and girls-to act wisely in human relations” (E. L. Thorndike, 1920, p. 228). This definition emphasizes both the ability to understand people and the ability to act upon this understanding. In contrast, mechanical intelligence referred to an ability to understand and use mechanisms (e.g., knife, gun, lathe), while abstract intelligence described an ability to understand and use symbols (e.g., words, numbers, scientific laws). At the time of Thorndike’s publication, abstract intelligence had been the most thoroughly investigated, with Thorndike himself recognizing that measures of social intelligence were difficult to devise. It was his belief that such measures would need to involve real-life situations which would allow an individual to respond to facial expression, gesture, voice, etc. 2. EARLY DEFINITIONS AND STUDIES

The George Washington Test of Social Intelligence (Moss, Hunt, Omwake, & Ronning, 1927) was the most widely used measure of social intel-

128

Jane L . Mathias

ligence following the publication of E. L. Thorndike’s (1920) definition. Social intelligence was described as an ability to get along with others and was thought to be an important variable in determining one’s success in the world (Moss & Hunt, 1927). The George Washington Test contained subtests designed to assess, among other things, judgment in social situations, memory for names and faces, and recognition of the mental state of a speaker. Although an early study showed low average intercorrelations between the different parts of this test (ranging from .22 to .35), the finding that these parts showed low correlations (.02 to .43) with a test of abstract intelligence was seen as promising (Moss, 1931). A wellknown measure at the time, this test was the subject of much interest and featured in numerous attempts to establish the validity of a social intelligence construct. Hunt (1927) elaborated on the psychometric properties of the George Washington Test in her article titled “What social intelligence is and where to find it.” Here, Hunt’s enthusiasm for the test was apparent, but far exceeded the amount of statistical evidence that was presented to support her case. Although Hunt claimed that the measure had “a high degree of reliability” (p. 607), neither the reliability coefficients nor the method of calculation was reported. The criterion validity of this measure was assessed by comparing scores from the George Washington Test with indices of social ability (i.e., ability to get along with people and participation in campus activities). These measures were found to correlate positively with social intelligence scores, prompting Hunt to conclude that the George Washington Test assessed abilities thought to require social intelligence. The validity of the George Washington Test was further investigated by Broom (1928), who compared the scores from this measure with those from the Thorndike Intelligence Examination. A correlation of .56 (which increased to .ti4 when corrected for attenuation) was interpreted to indicate that the two tests measured the same variable(s) and that, contrary to their purposes, the two tests validated one another. Broom (1930) subsequently argued that the variable common to both measures was reading comprehension, and that when the effect of this variable was controlled for, a much lower correlation of .21 resulted. Social and academic intelligence are, he concluded, different variables and any failure to distinguish between the two was thought to be the result of poor measurement techniques. In 1933, Vernon described social intelligence as incorporating an ability to get along with people, a knowledge of social concerns, a sensitivity to the communications of others, and an ability to perceive accurately both the moods and personality of other individuals. Recognizing the breadth of this construct, Vernon focused on the intellectual, social, and artistic characteristics of people who were good and bad judges of other people’s

SOCIALLY BASED COMPETENCIES

129

personalities. Social intelligence was assessed by means of the five subtests of the George Washington Test. Using correlational analyses, he found no evidence to suggest the presence of a general trait which represented “intuitive ability” (Vernon, 1933). Doubts about the validity of the George Washington Test reemerged in 1936, when Thorndike factor-analyzed 10 measures of social and abstract intelligence. Three factors were derived from this analysis and were thought to represent comprehension and use of words, abstract ability, and a speed component. The social intelligence measures loaded highly on the first factor, negatively on the second, and tended to have very low loadings on the third. From this, it was concluded that the George Washington Test of Social Intelligence largely assessed the ability to understand and work with words. Such evidence prompted Thorndike to question seriously the validity of this measure in a review published in the 1940 Mental Measurements Yearbook (R. L. Thorndike, 1941). In 1939, Woodrow confirmed R. L. Thorndike’s (1936) findings when he administered 52 different tests of mental abilities (including social intelligence; attention; musical, verbal, and spatial abilities; and general intelligence) in an attempt to investigate the factorial structure of these variables. Although a factor analysis yielded 10 factors, none of these represented a distinctly social cluster of variables. In fact, the subscales of the George Washington Test loaded highly on the verbal factor. O’Sullivan, Guilford, and de Mille (1965) subsequently attributed this finding to the highly verbal nature of the tests, arguing that it may simply have been an artifact of the measures. When reviewing the research on the George Washington Test, R. L. Thorndike and Stein (1937) concluded that it did not adequately measure an individual’s ability to deal with people, as intended. The test content seemed to have tapped what E. L. Thorndike (1920) would have labeled abstract intelligence, rather than social intelligence. Thorndike and Stein’s review again raised doubts about the validity of this measure and was subsequently followed by a decline in the use of this test. In 1947, Wedeck published the final results of a 16-year study into the construct psychologicul ability, which he defined as “an ability to judge correctly the feelings, moods, [and] motivations of individuals” (Wedeck, 1947, p. 133). Thus, psychological ability very closely resembled what others have labeled social intelligence. Interestingly, despite the similarity between the two constructs, Wedeck did not refer to Thorndike or other social intelligence researchers in his review of the literature (Walker & Foley, 1973). It was Wedeck’s intention to investigate whether psychological ability consisted entirely of general intelligence, or whether a separate ability was involved. Using eight specially constructed tests of

I30

Jane L . Mafhias

psychological ability, together with three nonverbal and four verbal tests, Wedeck found evidence for the existence of three clusters of variables. These he labeled general intelligence, verbal ability, and psychological ability. Some years later, O’Sullivan et al. (1965) reanalyzed Wedeck’s correlation matrix using factor analytic methods. This additional analysis revealed five factors which included a verbal and a figural or spatial cluster, plus three factors on which the tests of psychological ability loaded. 3. GUILFORD’S MODEL OF INTELLIGENCE The next major theoretical development came some years after E. L. Thorndike’s publication when, in 1959, Guilford published his Structureof-Intellect Model. Here, Guilford provided a consolidated account of the nature of human intelligence. In this model, it was argued that intellectual abilities could be broadly classified in terms of ( I ) the mental operation involved, (2) the content or information upon which the operations are performed, and (3) the products into which the information is processed.It was further hypothesized that there were four kinds of content, five kinds of operation, and six kinds of product, which combined to form a total of 120 different abilities, each having its own combination of these factors. Guilford has since proposed that there are in excess of 120 distinct skills (Guilford & Hoepfner, 1971). Of interest here are the content areas identified in this model. These included semantic, symbolic, figural, and behavioral domains, with the behavioral domain forming the basis of Guilford’s notion of social intelligence. While the first three of these content areas were empirically supported, the fourth was not; its inclusion was based on the need to add a category that formed the basis of the mental operations that occur during interactions with other people. The addition of the behavioral content area was, therefore, motivated primarily by logical considerations, but was also influenced by E. L. Thorndike’s (1920) earlier work suggesting a distinct kind of social intelligence (Guilford, 1967). Guilford’s notion of behavioral intelligence embraced such things as feelings, thoughts, attitudes, and other variables that affect a n individual’s social behavior, and was defined as the range of intellectual skills necessary to our interactions with other individuals (Hendricks, Guilford, & Hoepfner, 1969). The Structure-of-Intellect Model originally predicted a minimum of 30 separate social or behavioral intelligence factors (O’Sullivan et al., 1965). Guilford and his associates set about devising measures of each of the 120 abilities predicted by the model and attempted, with the use of factor analysis, to demonstrate the existence of different forms of intelligence. For instance, when looking at the six hypothesized behavioral-cognition abilities, O’Sullivan et al. (1965) were able to demonstrate the existence

SOCIALLY BASED COMPETENCIES

13 I

of six separate factors as predicted by the model using tests that minimized verbal content. Similarly, Hendricks et al. (1969) devised measures that allowed them to confirm the existence of six behavioral-divergent production abilities. Research such as this ultimately led O’Sullivan and Guilford (1966) to publish the Six Factor Tests of Social Intelligence in an effort to provide a standard measure of the behavioral domain of Guilford’s Structure-of-Intellect Model. 4. RECENT DEVELOPMENTS

In more recent years, Keating (1978) has continued the search for empirical evidence to confirm the existence of a set of skills which could be labeled social intelligence. Using measures of moral reasoning, interpersonal reasoning, social functioning, and verbal and nonverbal intelligence, Keating was not able to demonstrate the existence of a separate social intelligence domain. The measures of social intelligence were not found to correlate more highly with each other than with measures of general intelligence. Similarly, a factor analysis failed to reveal a separate factor on which the measures of social intelligence loaded. Finally, when the social functioning variable was chosen as a criterion measure in a regression analysis, the measures of general intelligence provided the best predictors. Together, these results provided no evidence to suggest that Keating’s measures were assessing social intelligence. Two explanations were offered by Keating (1978) to account for his findings. First, he suggested that the measures selected did not adequately assess the social intelligence domain. Second, he noted that the format of the tasks (paper-and-pencil)acted as a constraint and that alternative methods of assessment may have been more appropriate to the measurement of social skills. In response to Keating’s second remark, Marlowe and Bedell (1982) undertook a study to investigate the usefulness of paper-and-pencil tests in the measurement of social intelligence. This work was prompted by C. L. Williams’ (1981) finding that role playing and a paper-and-pencil test (social introversion scale of the Minnesota Multiphasic Personality Inventory) were equally good predictors of actual social behavior (i.e., social intelligence). By demonstrating that this measure of social intelligence did not correlate with verbal or abstract intelligence, Marlowe and Bedell not only found evidence for the existence of a separate social intelligence factor, but they also negated Keating’s hypothesis that paper-and-pencil tests are not suited to the assessment of social functioning skills. From this, they concluded that while the format of the measure may be important, it was the high level of verbal ability required in Keating’s (1978) tasks that masked the presence of social intelligence.

132

Jane L . Marhins

Ford and Tisak (1983) have also attempted to isolate a social intelligence construct using statistical techniques comparable to those described by Keating (1978). Social intelligence was defined as an individual's ability to achieve social objectives, thereby emphasizing the outcome of one's behavior. Four measures of academic ability were correlated with six measures of social intelligence, with the result that the social intelligence measures intercorrelated more highly with one another than with the measures of general intelligence. A subsequent factor analysis revealed two factors equivalent to academic and social intelligence, while a regression analysis found that the social intelligence variables were better predictors of the behavioral criterion variables. Thus, unlike Keating, Ford and Tisak found support for the existence of a separate cluster of social intellectual skills. Marlowe (1985) used the term social intelligence to describe an ability to understand the feelings, thoughts, and behaviors of both ourselves and others in interpersonal situations, and to use this understanding in guiding our own behavior. No distinction was made between the terms social intelligence and social competence (Marlowe, 1986). It was Marlowe's (1984, 1986) contention that social intelligenceis a multidimensionalconstruct and that it can be shown to exist independently of general intelligence when it is defined in terms of social effectiveness and when it is measured using instruments that do not overlap with academic abilities. In keeping with this, Marlowe used measures of social interest, social self-efficacy, empathic skills, and social skills to assess social intelligence,and measures of verbal and abstract-thinkingskills to assess general intelligence. A factor analysis of the five measures of social intelligence extracted a total of five factors (prosocial attitude, social skills, empathy skills, emotionality, and social anxiety) which were considered to be different dimensions of social intelligence. A further factor analysis of both the social and general intelligence measures investigated the independence of these constructs. The resulting three-factor solution consisted of a general intelligence, a social self-perception (i.e., combined prosocial and emotionality factors), and a social and empathy skills factor. From this, Marlowe concluded that not only is general intelligence distinct from social intelligence but that the latter is not a unitary construct.

B.

Social Competence

The term social competence is most often associated with the child development literature and is characterized by an approach which differs from that used in the investigation of social intelligence. While investigators of social intelligence have attempted to demonstrate the existence of two independent constructs associated with social and academic skills, social competence researchers have sought to define the skills that are

SOCIALLY BASED COMPETENCIES

I33

associated with a competent adulthood and the means by which these can be fostered in the preceding years. This difference is significant because it has meant that the study of social competence has derived from less specific goals. Not only is the research in this area more varied, but it is also more prolific. Moreover, there is no clear consensus as to the relationship between the different approaches. As Walker and Foley (1973) state, “the present failure to recognize that one person’s social intelligence is another person’s interpersonal competence or role-taking and communication is apparently based not only on the diversity of terminology but the different theoretical origins of the various approaches” (p. 842). Within the literature, terms are used inconsistently and often interchangeably as in the case of Kleck (1975) who, while discussing Weinstein’s (1969) theory of interpersonal competence, refers to it as social competence. There are also instances where terms are used indiscriminately, with either a failure to define their meaning (e.g., Connolly & Doyle, 1981) or a definition implied only by means of the measures used (e.g., Cole, 1976). A number of researchers (Anderson & Messick, 1974; Goldfried & D’Zurilla, 1969; Greenspan, 1981a; Kleck, 1975; O’Malley, 1977; Waters & Sroufe, 1983) have attempted to formulate broad categories into which different theoretical approaches to the study of competence and social competence can be grouped, in an effort to reduce the complexities of the area. Even here there is little agreement about the categories that are most appropriate. For present purposes, research on social competence is divided into three main categories which differ in the breadth of their focus. First, there are investigators who adopt a comprehensive view of social competence and who incorporate an assortment of variables, including both social and nonsocial (e.g., cognitive) skills, in their description of this construct. This approach has been described by Odum and McConnell(l985) as the all-inclusive approach. Second, a composite approach equates social competence with other relatively broad constructs, such as social maturity, adaptive behavior, and social skills, which are themselves composed of a variety of different skills. Finally, a social-cognitive approach considers specific aspects of social competence, such as role-taking, moral judgment, and referential communication. These variables provide us with the social knowledge that is a prerequisite to effective interpersonal interactions. 1. COMPREHENSIVE APPROACHES The most general approach to the definition of social competence is based on a list of descriptors that sample from a broad range of behaviors

134

Jane L . Mathias

in an attempt to represent the different components of social competence. It is a perspective that is more commonly associated with the child development than the developmental disability literature. The work summarized by Anderson and Messick (1974) is one of two very clear examples of this style. In 1973, a panel of experts attempted to define social competence in young children using a total of 29 statements which dealt with a range of variables, including personal maintenance, feelings of personal worth, positive personal relationships, role appreciation, morality, perceptual skills, fine and gross motor skills, language ability, and memory skills. These statements were intended to form the basis of a comprehensive system of assessment. Anderson and Messick, therefore, describe social competence in terms of a set of variables that tease out the various components of competence which are offered in place of a single definition. Perhaps one of the most obvious criticisms of this approach is summarized by Greenspan (1981a, 1982), who questions the use of the term social competence in a situation where physical and intellectual competence are given equal consideration. While the importance of these variables is not disputed, it is argued that the term competence would be more appropriate where there is nothing uniquely social about the skills in question. Essentially, some criteria are needed for clearly demarcating the areas of social and general competence. In addition, Kleck (1975) notes that it is difficult to achieve general agreement concerning both the variables that should be included in such a list, and the relative importance of each of these items. Finally, Odum and McConnell (1985) state that the very breadth of these descriptions makes assessment difficult. A second notable example of a comprehensive approach to the definition of social competence is the work of Zigler and Trickett (1978). Not satisfied with previous descriptions of social competence, they claimed that measures of this construct need to assess the success with which an individual has met societal demands, while also providing information about personal development. They then list four areas that are considered essential to an index of social competence, these being physical health, cognitive skills, achievement, and motivational and emotional variables. Although this list is less inclusive than Anderson and Messick’s, it continues to stress intellectual and physical competence at the expense of truly social abilities, perhaps making the term social competence a misnomer. Greenspan (1980) is critical of Zigler and Trickett’s description of social competence for precisely this reason, stating that it neglects a whole body of literature which deals with the latter construct. Seitz and Zigler (1980), however, respond to this by pointing out that Greenspan has chosen to disregard the noncognitive elements (i.e., physical health, motivation, and emotion) in his evaluation of their description.

SOCIALLY BASED COMPETENCIES

I35

2. COMPOSITE APPROACHES

A second perspective that has emerged is one that equates social competence with another construct which is itself a composite of skills. Studies of social maturity, adaptive behavior, and social skills are examples of such an approach to the study of social competence. Interestingly, this approach seems to be more a feature of developmental disability research than child development research. Within the disability literature, research on social competence has often centred on the study of social maturity, adaptive behavior, and social skills. A burgeoning interest in these constructs has largely resulted from the reinstatement of social competence as a defining feature of mental retardation. To accommodate the needs of practitioners working in the area, the study of social competence was directed toward skills that are of central importance to the lives of mentally handicapped individuals. As a consequence, the term social competence has come to be aligned with the study of practical living skills when used in the context of developmental disability. a . Social Maturity. Social maturity is a term that was first used by Doll ( l953), who strongly advocated the assessment of social competence in mentally retarded populations. Social competence was defined as “the functional ability of the human organism for exercising personal independence and social responsibility” (Doll, 1953, p. 10). The term social maiuriiy came from Doll’s view that social competence could be viewed as a gradual process of maturing social skills. In order to asses this variable, Doll developed the Vineland Social Maturity Scale, which consisted of items measuring self-help, self-direction, locomotion, occupation, communication, and socialization. For present purposes, it is important to note that Doll equated social competence with a set of skills that would allow a person to achieve a degree of responsibility and independence appropriate to his or her age. b. Adaptive Behavior. Adaptive behavior is a concept that has evolved from the work of Doll and has primarily, although not exclusively, been investigated in the context of mentally retarded populations. [For more detailed discussions of adaptive behavior, refer to Coulter and Morrow (1978a, 1978b, 1978c), Kennett (1977), Mercer (1978), Meyers et al., (l979), and Nihira (1977).] Like the variable social maturity, adaptive behavior has largely achieved status through a desire to include a social component in the definition and diagnosis of mental retardation. According to the American Association on Mental Retardation (AAMR), “adaptive behavior refers to the quality of everyday performance in coping with environmental demands” (Grossman, 1983, p. 42). Thus, the emphasis is on an individual’s ability to cope with the demands

136

Jane L. Mathias

of the environment (Meyers et al., 1979), with a range of adaptive behavior scales having been developed to assess current observable functioning (Leland, 1972). Probably the best known of these measures is the AAMR’s Adaptive Behavior Scale, which was first published in 1969 (Nihira, Foster, Shellhass, & Leland, 1969). Although there are many adaptive behavior scales, Meyers et al. have noted that these scales consistently sample from a range of behavioral domains, including self-help skills, physical development, communication skills, cognitive functioning, domestic and occupational activities, self-direction and responsibility, and socialization. There is also an increasing tendency to provide a measure of maladaptive behavior, thereby taking into account the socioemotional problems experienced by mentally retarded persons in their day-to-day lives. Essentially, adaptive behavior subsumes a range of variables that are critical to maintaining personal independence and, although largely concerned with practical living skills, this construct also includes physical and cognitive variables. c. Social Skills. A third and related construct is that of social skills. This term has broader application than either social maturity or adaptive behavior, being used in the context of child development, developmental disability, and psychiatric research. [More detailed discussions can be found in Gresham (1981a, 1981b), Gresham and Elliot (1984, 1987), Hops (1983), and McFall (1982).] Sarason (1981) defines social competence in terms of a person having and using appropriate social skills, while Gresham (1981a) describes a person’s social skills as his or her behavioral repertoire. Once again. this is a generic term that incorporates a range of variables, many of which show considerable overlap with the previous two constructs. Existing research on social skills has covered a range of variables, including assertiveness, heterosocial skills, aggressive behavior (Matson, Kazdin, & Esveldt-Dawson, 1980), interaction skills (Holley, 1980),work-related skills (La Greca, Stone, & Bell, 1983), discrimination of social cues (Minskoff, 1980a, 1980b), and interpersonal deficits (Bornstein, Bach, McFall, Friman, & Lyons, 1980). to name only a few. While the term social skills is very broad in its usage, it is always used to refer to “social” behaviors, in contrast to assessments of adaptive behavior, which usually also contain measures of cognitive abilities. 3. SOCIAL-COGNITIVE APPROACHES The investigation of social competence can also be approached from a much narrower perspective. Here, specific variables that are considered to be important contributors to a person’s level of competence are studied

SOCIALLY BASED COMPETENCIES

I37

in isolation from other variables. The systematic investigation of these variables is thought to provide a useful direction for the study of social competence (Monson, Greenspan, & Simeonsson, 1979). Such research has focused on a diversity of areas, including role taking, moral judgment, referential communication, and person perception. Collectively, these variables fall within the rubric of social cognition, which is a term used to refer to a “child’s intuitive or logical representation of others, that is, how he characterizes others and makes inferences about their covert, inner psychological experiences” (Shantz, 1975, p. 258). Also labeled social understanding, this construct refers to the knowledge and reasoning that mediate our social behavior (Simeonsson, Monson, & Blacher, 1984). The list of social-cognitive variables is extensive (Damon, 1979), and the defining feature seems to be that these variables contribute to our social knowledge. Among the variables described as forming part of the study of social cognition are role-taking skills, person perception, moral judgment, and referential communication (Simeonsson, 1978). To this list Taylor (1982) adds social inference and social problem solving, while Simeonsson et ul. (1984) also include social comprehension. Although Taylor also discusses goal-directedness and motivational orientation as components of social competence, they are not strictly social-cognitive variables and so, for present purposes, will not be considered further. Finally, Muuss (1982) describes social cognition as including role taking, empathy, moral reasoning, interpersonal problem solving, and social comprehension. Together, these authors have identified seven separate areas of social-cognitive functioning. u. Role Tuking. The term role taking refers to our ability to understand the feelings, perspectives, and thoughts of others (Simeonsson et al., 1984) and has been described as a social-interpersonal skill (Selman, 1971). Competent role taking requires an individual to assume the position of another person and to make inferences about that person’s perspective in any given situation (Shantz, 1975). This variable is multidimensional and can be divided into perceptual, cognitive, and affective role taking, thereby distinguishing between a person’s knowledge about what others are seeing, thinking, and feeling, respectively (Taylor, 1982). Both cognitive and affective role taking are of particular interest here because they are more strictly social in content, although it has been suggested that role-taking skills, in general, are likely to be important cognitive components of social competence (Meichenbaum, Butler, & Gruson, 1981). 6 . Person Perception. Person perception refers to the way in which we perceive another person’s qualities, traits, and characteristics (Si-

I38

Jane L. Mathias

meonsson, 1978), and is a social-cognitive variable that shows considerable overlap with role taking. Included within the realm of person perception research is the study of how a child perceives and describes other people. This entails descriptions of both overt variables such as physical appearance and possessions, together with less visible, covert variables, referring to descriptions of another person’s attitudes and abilities (Shantz, 1975). Person perception is a skill that seems to develop rapidly from around the age of 7 to 10 years. During this time frame, children demonstrate an improved ability to make inferences about the thoughts, feelings, personality attributes, and behavior of other people (Barenboim, 1977). It is this knowledge about others that allows inferences to be made which subsequently guide behavior and improve the effectiveness of interactions (Youniss, 1975). c . Moral Judgment. Moral judgment involves an assessment of the ethical and moral principles associated with a particular act. It entails deciding whether an act is good or bad (Simeonsson et al., 1984) and is composed of a set of beliefs about the nature of justice (Johnson, 1962a, 1962b). As Mischel and Mischel (1976) write, moral judgment is concerned with the evaluation of what one should do, or what is right and wrong, in contrast to moral conduct, which is related to actual performance. Of the social-cognitive variables under consideration here, moral judgment is one of the most widely researched. At the heart of these investigations is the work of Piaget and Kohlberg, who have both developed research paradigms for the study of moral reasoning. [More detailed discussions of their work can be found in Keasey (1977), Kohlberg (1969), Kurtines and Greif (1974), Piaget (1932), and Rest (1974).] d . Social Comprehension. The term social comprehension was first used by Greenspan (1979) and later by Muuss (1982) and Simeonsson et al. (1984) to refer to what had previously been labeled social reasoning (Selman, 1977) and interpersonal awareness (Selman, Jacquette, & Lavin, 1977). The work of Selman and his co-workers is central to this concept, with attention being focused on a child’s developing conceptions of interpersonal role-relations such as peer-friendship, parent-child, and peer-group relations (Selman, 1976b). Over the years, Selman has published a number of articles in which he has elaborated on this topic (Selman, 1975, 1976a, 1976b, 1977, 1980, 1981; Selman & Jacquette, 1977; Selman et al., 1977). Basically, Selman uses a structural-developmental approach which assumes that social reasoning develops through an ordered set of qualitatively different stages (Selman, 1977). Each stage is associated with an underlying logic that influences a person’s thoughts on a range of interpersonal issues (Selman & Jacquette, 1977).

SOCIALLY BASED COMPETENCIES

139

e . Social Inference. Another social-cognitive variable is social inference, which has been defined as the ability to attend to and utilize cues that communicate information about purpose, attitude, relationships, and roles, etc. (Edmonson, de Jung, Leland, & Leach, 1974). Also labeled social perception (Minskoff, 1980a, 1980b), it has been argued that a sensitivity to social cues and an ability to interpret these accurately are essential prerequisites to competent interpersonal interactions (Taylor, 1982). What is important, then, is an ability to perceive the meaning of a social situation; that is, “to note the cues that are relevant, to infer from them what is on-going and what and why certain behaviors are most appropriate” (Edmonson, de Jung, & Leland, 1965, p. 7). Any deficits in social inference skills will necessarily affect an individual’s interpretation of a social situation, and deny that person access to the information that is needed in order to respond effectively to the particular situation. Research on social inference is largely associated with Edmonson and associates who, in their work with mentally retarded persons, have described the importance of social inference skills to social competence, the means by which social inference can be assessed, and its responsiveness to training (de Jung & Edmonson, 1972; de Jung, Holen & Edmonson, 1972, 1973; Edmonson et al., 1965, 1974; Edmonson, Leland, & Leach, 1968, 1970).

f. Referential Communicafion. The study of referential communication emphasizes the use of language in interpersonal interactions (Cohen & Klein, 1968) and draws a distinction between linguistic and communicative competence (Glucksberg, Krauss, & Higgins, 1975). Referential communication has been defined as “the process by which a person attempts to provide sufficient descriptive information regarding a particular object . . . such that a second person has a clear idea of what the first person has in mind” (Greenspan, Burka, Zlotlow, & Barenboim, 1975, p. 97). Thus, successful communication is dependent not only on linguistic skills, but also on an ability to accommodate both the needs and processing capacities of the listener (Hoy & McKnight, 1977; Rueda & Chan, 1980). Essentially, it is a person’s task to construct a message that will allow the listener to identify a referent from among the available nonreferents (Glucksberg et al., 1975), where the term referent is used to describe the target stimuli (e.g., objects, properties, events, and relationships) (Rosenberg & Cohen, 1966). g . Social and Interpersonal Problem Solving. The terms social and interpersonal problem solving have been used to describe the ability to generate solutions in order to resolve interpersonal problems. The study of this variable is most closely linked with the work of Spivack and coworkers, whose research has been summarized in a number of publica-

140

Jane L . Mathias

tions dealing with deficits in these skills, as well as their measurement and response to training (Platt, Siegel, & Spivack, 1975; Platt & Spivack, 1972, 1975a, 1975b; Platt, Spivack, Altman, Altman, & Peizer, 1974; Shure & Spivack, 1972, 1981; Shure, Spivack, & Jaeger, 1971; Spivack, Platt, & Shure, 1976; Spivack, Shure, & Platt, 1981). Spivack rt ul. (1976) have developed a theory of cognitive problem solving which describes five separate interpersonal cognitive problem-solving skills that are considered to be important contributors to a person’s level of social competence. Thus, rather than a single ability, interpersonal problem solving is thought to be a composite of skills, including sensitivity to problems, alternative solution thinking, means-ends thinking, consequential thinking, and causal thinking. C.

Interpersonal Competence

As noted, social intelligence and social competence are concepts that have originated from different research traditions. While research on social intelligence has attempted to distinguish between social and general intellectual functioning, social competence research has focused on the development of skills that are associated with social adaptation in adulthood. Research on interpersonal competence is more closely related to social competence than to social intelligence. Certainly, the term interpersonal competence is used less frequently than social competence, although the importance of interpersonal skills to our interactions is widely recognized (Affleck, 1975, 1976, 1977; Bates, 1980; Bornstein r t ul., 1980; Castles & Glass, 1986; Cheney & Foss, 1984; Foss & Peterson, 1981; Greenspan & Shoultz, 1981; Greenspan, Shoultz, & Weir, 1981; La Greca et ul., 1983; Longhurst, 1974; Weiss & Weinstein, 1967). There are two main theories which deal specifically with interpersonal competence. Foote and Cottrell (1955) define interpersonal competence in terms of six components, each of which is important to effective interpersonal relations. These they identified as (a) health, or good physical condition; (b) intelligence; (c) empathy, the ability to perceive accurately another’s viewpoint; (d) autonomy, essentially referring to one’s identity; (e)judgment, the ability to evaluate alternative courses of action; and (0 creativity, the ability to devise appropriate responses to problem situations. Together, these qualities were thought to identify the defining features of competent interpersonal relations. A second theory was proposed by Weinstein (1969), who defined interpersonal competence as “the ability to manipulate other’s responses” (p. 755). It was Weinstein’s belief that interpersonal competence consists of establishing and maintaining identities for one’s self and also for others.

SOCIALLY BASED COMPETENCIES

141

The ability to do this is related to three variables: (a) a person’s empathic skills, referring to an ability to take the role of another person accurately; (b) the possession of an extensive response repertoire; and (c) the possession of intrapersonal resources which allow a person to use appropriate response tactics as required. The first and second of these variables have been investigated under the respective labels of role taking and interpersonal problem solving (O’Malley, 1977), while the third variable, interpersonal tactics, has been researched by Weinstein himself (Weinstein, 1966; Weiss & Weinstein, 1967; Wood, Weinstein, & Parker, 1969). D.

Conclusion

In the absence of any general consensus regarding the definition and assessment of social intelligence, social competence, and interpersonal competence, it is useful to seek trends which, in some way, summarize the existing research in each of these areas. Investigators of social intelligence, for example, have adopted a general research paradigm whereby separate measures of social and cognitive skills are employed in an attempt to demonstrate the existence of a distinctly social domain of skills. It is only when one looks at how this paradigm is made operational that the differences in views become apparent. Thus, the term social intelligence has been used to refer to the type of information being processed, in order to distinguish it from nonsocial or cognitive content. Social competence research, in contrast, lacks a common paradigm, although it is possible to identify three theoretical approaches which summarize the majority of studies. These three approaches equate social competence with a cross section of social and nonsocial skills (comprehensive approach), a conglomerate of associated variables (composite approach), or separate aspects of social knowledge (social-cognitive approach). The term social is, therefore, used to convey both very general and highly specific meaning, making it difficult to reconcile current trends in the study of social competence. Finally, interpersonal competence research is concerned with the psychological aspects of interpersonal interactions and overlaps with the last of the three social competence paradigms. Although no clear definition exists, this term is generally used to refer to the social knowledge which influences our interactions with other people. Thus, the terms sociul intelligence, sociul competence, and interpersonal competence are used to refer to three overlapping areas of research. One feature that is common to all of these perspectives is the endeavor to tap the uniquely social component of functioning. What is needed, given this overlap, is some way of integrating the social intelligence, social com-

I42

Jane

L. Mathias

petence, and interpersonal competence literature in a way that would utilize existing research in these areas and stimulate a more systematic investigation of the relationship between social and nonsocial competencies. Arguably, it is also desirable to restrict the use of these terms to particular meanings, thereby reducing the confusion in this area. 111.

GREENSPAN’S MODELS

Social competence has reemerged as a defining feature of mental retardation, with adaptive behavior scales being used to assess this construct. While these measures have proved useful to clinicians, there is growing concern that adaptive behavior has come to be equated with daily living skills, to the exclusion of interpersonal skills (Greenspan, 1981b). In short, adaptive behavior scales may be interpreting the social competence criterion too narrowly. Despite the proliferation of studies dealing with social functioning, the theoretical and experimental findings in this area remain largely unconsolidated. Indeed, the study of socially based competencies has, in the main, remained quite separate to the investigation of mental retardation. A framework is, therefore, needed for the systematic investigation of this area. In addition, there is a need to examine the relationship between socially based skills and other skills in order to develop a more general model of human competence. A comprehensive model that incorporates a broad range of abilities would then be useful for understanding differences between more and less competent individuals (by considering profiles of skills) and for defining particular types of disability (Greenspan & Javel, 1982). Greenspan (1979, 1981a) has formulated three different models dealing with social intelligence, personal competence, and social competence to address these needs. Although his models have broad applicability, Greenspan focuses on the implications of these models for mentally retarded persons, arguing for a broader definition of mental retardation. His models of social intelligence and personal competence (as shown in Sections 111, A and B, Figs. 1 and 2) are central to this argument. A.

Greenspan’s Model of Social Intelligence

In 1979, Greenspan published the first of a number of papers dealing with social intelligence and its implications for the study of mentally retarded populations. At this time, Greenspan was clearly influenced by E. L. Thorndike when he defined social intelligence as “ a person’s ability to understand and to deal effectively with social and interpersonal objects

143

SOCIALLY BASED COMPETENCIES

I*

COMMUNICATION

TAKING

INFERENCE

FIG. 1. Greenspan’s (1979) proposed model of social intelligence. From Greenspan, S. (1979). Social intelligence in the retarded. In N . R. Ellis (Ed.), Handbook of mental dejiciency: Psychological rheory and research (p. 487). Hillsdale, N.J: Erlbaum. Reprinted by permission.

and events” (Greenspan, 1979, p. 483). By 1981, he had substituted the term social awareness for social intelligence in an attempt to avoid the undesirable connotations often linked with the term intelligence. Social awareness he then defined as “the individual’s ability to understand people, social events, and the processes involved in regulating social events” (Greenspan, 1981a, p. 18). According to Greenspan, interpersonal understanding is central to the notion of social awareness. A further change in terminology occurred in 1982 when the term social awareness was replaced by the term social judgement, although the other elements of the definition remained unchanged. Throughout these alterations, Greenspan’s model has remained constant, his changing use of terms apparently reflecting a desire to locate a more accurate descriptor with fewer preconceptions about its meaning. For present purposes, the term social intelligence is used hereafter to refer to what Greenspan has variously named social intelligence, social awareness, and social judgement. Greenspan’s (1979, 1981a) hierarchical model of social intelligence is presented in Fig. 1. Social intelligence is divided into three main constructs labeled social sensitivity, social insight, and social communication, with each subdivided into its constituent variables. According to Greenspan: 1. Social sensitiviry is described as an ability to make accurate inferences on the basis of cues provided both by other individuals and by the social situation itself, and comprises two variables. First, role taking deals with an individual’s ability to utilize personal cues by being aware of other people’s thoughts (cognitive role taking) and feelings (empathy

144

June L. Muthius

or affective role taking). Social inference, on the other hand, is concerned with an ability to attend to the situational or contextual cues that provide information about what is happening in any particular situation. 2. Social insight refers to an individual’s understanding of the processes underlying social interactions. Whereas social sensitivity requires a responsiveness to relatively overt cues, social insight demands an understanding of more covert processes. Three variables collectively define this construct. The first of these, social comprehension, is concerned with different types of relationships and the features that are unique to them, while psychological insight describes an awareness of the personal characteristics of others (person perception) and of the motives underlying their behavior (psychological causality). Finally, morul judgment is concerned with ethical considerations. 3. Sociul communication focuses on the ability to communicate effectively in interpersonal interactions (Greenspan, 1981a). This construct is divided into separate components, the first being referential communication, which involves an ability to anticipate and subsequently impart the information needed by a listener. The second component, social problem solving, is described as “the ability to deal effectively with situations in which there is a divergence of needs between the actor and one or more people” (Greenspan, 1979, p. 506). The actual achievement of goals and the intermediate processes involved in realizing these goals are essential components of this variable. Greenspan approached social intelligence from a social-cognitive perspective, emphasizing the knowledge that mediates interpersonal interactions. In effect, Greenspan selected seven variables which he considered to be central to the study of social cognition and defined social intelligence on the basis of these. Collectively, Simeonsson (19710, Taylor (l982), and Simeonsson et al. (1984) refer to the same basic set of socialcognitive variables as Greenspan. The only exception is Greenspan’s psychological insight, which is somewhat broader than the variable person perception. While Greenspan’s choice of variables is representative of current listings of social-cognitive variables, an absence of existing research has led him to divide these into three categories on the basis of what are thought to be common underlying psychological processes. Clearly, research is needed not only to investigate the validity of the social intelligence construct, but also to investigate its internal structure. 6.

Greenspan’s Model of Personal Competence

Greenspan’s (1979, 1981a) model of personal competence is more general than his model of social intelligence. Personal competence is divided

145

SOCIALLY BASED COMPETENCIES

c PERSONAL COMPETENCE

SOCIO-EMOTIONAL

CONCEPTUAL INTELLIGENCE

PRACTICAL

INTELLIGENCE

SOCIAL INTELLIGENCE

r - l r - l mf - l m FIG. 2. Greenspan's (1979) model of personal competence. From Greenspan, S . (1979). Social intelligence in the retarded. In N . R. Ellis (Ed.), Handbook of mental deficiency: f'sycholo~icultheory rind reseurch (p. 5 10). Hillsdale, N.J: Erlbaum. Reprinted by permission.

into physical, intellectual (adaptive intelligence), and emotional (socioemotional adaptation) competence (see Fig. 2). More specifically, the adaptive intelligence subsection of this model is divided into three main areas of intellectual functioning in a way that resembles E. L. Thorndike's (1920) views. These include conceptual, practical, and social intelligence, the latter having been dealt with more extensively in Greenspan's model of social intelligence. The term conceptual intelligence is used to refer to an ability to solve abstract problems and to use language (Greenspan, 1981a). In keeping with this, conceptual intelligence is divided into two variables, language ability and thinking ability, and is thought to be assessed using standard IQ tests. Pructical intelligence is on a par with Thorndike's concept of mechanicul intelligence and includes skills which allow us to solve problems that arise in our everyday work and recreational pursuits (Greenspan, 1981a). Practical intelligence is further divided into maintenance skills and interchange skills, consistent with the first two factor scores (personal selfsufficiency and community self-sufficiency) which can be derived from Part I of the Adaptive Behavior Scale (Nihira, Foster, Shellhaas. & Leland, 1974). Practical intelligence is, therefore, equated with a subset of

146

Jane

L. Mathias

the skills that are assessed using currently available measures of adaptive behavior. Finally, social intelligence is divided into three constructs which are defined by a total of seven social-cognitive variables, as described previously. Greenspan’s model of personal competence combines different areas of functioning in an effort to provide a more holistic view of the range of factors which contribute to an individual’s level of competence. The adaptive intelligence subsection is reminiscent of E. L. Thorndike’s ( 1920) classification of intellectual functioning into three different areas, while also developing upon current trends which distinguish between conceptual and practical intelligence when diagnosing mental retardation. The addition of a social intelligence construct was intended to complement IQ and adaptive behavior measures by considering those aspects of “intellectual” functioning that are not currently being assessed by either of these variables. Personality variables are referred to as socio-emotional adaptation, and are portrayed as forming a separate area of competence. According to Greenspan (1979) this portion of the model was based on Guilford’s work on personality, although some renaming of variables has occurred. Finally, while physical competence is included in this model, this construct is given only cursory attention. C.

Definition of Mental Retardation

Greenspan’s provisional models of social intelligence and personal competence were intended to schematize the relationship between different forms of competence. These two models present the view that an individual’s competence can be defined in terms of a profile of abilities and, although they have general application, they are of particular interest to the investigation of developmental disabilities. By identifying and organizing some of the distinguishing features of a competent individual, Greenspan’s models provide a context for the evaluation and refinement of the criteria used to define various disabilities. More particularly, Greenspan (198 1b) has argued that mental retardation should be defined as a developmental disability marked by diminished intellectual competence (adaptive intelligence). Thus, “in addition to deficits in ‘conceptual intelligence’ (IQ) and ‘practical intelligence’ (part 1 of adaptive behavior criterion), an individual would be required to show deficits in ‘social intelligence’ (social awareness) in order to be classified as retarded” (Greenspan, 1981b. p. 54). At present, The AAMR defines mental retardation in terms of only subaverage intellectual functioning (IQ) and impaired adaptive behavior (Grossrnan, 1983). The main

SOCIALLY BASED COMPETENCIES

I47

difference, then, between the two definitions is that Greenspan also includes social intelligence. According to Greenspan (l979), available measures of adaptive behavior tend to overlook interpersonal (social intelligence) variables, despite mounting evidence that they contribute to the competence of both retarded and nonretarded persons. The AAMR’s definition was, therefore, reformulated by Greenspan to redress this anomaly. The adoption of Greenspan’s definition would, in practical terms, involve the replacement of the “personality-type’’ items of adaptive behavior scales (e.g., ABS Part 2 items) with measures of social intelligence (Greenspan, 1981b). From a theoretical point of view, Greenspan’s (1979, 1981a) model of adaptive intelligence implies the existence of three distinct areas of intellectual functioning, in contrast to the two areas implicit in the AAMR definition. Therefore, before contemplating the use of this definition, evidence is needed (a) for the existence of a set of reliably measured interpersonal/social intelligence skills and (b) to confirm that these measures assess something over and above what is already being assessed by existing tests of intelligence and adaptive behavior. IV.

INVESTIGATIONS OF SOCIAL AND ADAPTIVE INTELLIGENCE IN MENTALLY RETARDED ADOLESCENTS

Only a very limited number of studies have investigated the validity of Greenspan’s models. Among them is a study by Greenspan (1984) in which he assessed the factorial structure of competence in mentally retarded adults. Twelve items, representing each of the variables in his model of personal competence, were combined to form a 12-item rating scale (Personal Competence Index). Six general areas were assessed by the items within this scale, including physical, conceptual, practical, judgmental, temperamental, and characterological competence. A principal components analysis of the data extracted three factors which were labeled intellectual, emotional, and physical competence. On the whole, these constructs were in keeping with the adaptive intelligence, socioemotional adaptation, and physical competence factors which were predicted on the basis of Greenspan’s model of personal competence. This investigation was, therefore, seen as providing preliminary support for Greenspan’s model, using a single instrument that assessed both social and nonsocial forms of competence. In addition, mentally retarded subjects were found to perform least well on the conceptual, practical, and social intelligence variables that collectively defined the adaptive intelligence factor.

148

Jane L. Mathios

Mathias (1989) and Mathias and Nettelbeck (1990a) describe an investigation in which they used a sample of mentally retarded adolescents (mean age 17 years 2 months) to evaluate the internal, inter-rater, and test-retest reliabilities of a battery of measures which were selected to assess the seven variables included in Greenspan’s (1979, 1981a) model of social intelligence. This study was predicated on the fact that the square root of a measure’s reliability acts as a ceiling to any correlation calculated between that measure and any other variable (Carmines & Zeller, 1979). Consequently, the assessment of reliability has direct implications for any investigation of validity. The importance of such a study was further highlighted in a preliminary investigation reported by Mathias (1988) which found that two commonly used measures of social-cognitive ability had poor inter-rater or internal reliability when used with a mentally retarded sample. Reliability properties were calculated for the Projective Role-Taking Task (Feffer, 1959; Schnall & Feffer, 1968), Interpersonal Understanding Interview (Selman, 1979), Test of Psychological Causality (Whiteman, 1970a. 1970b, 1976), Test of Moral Reasoning (Porter & Taylor, 1972). Matrix Test of Referential Communication (Greenspan & Barenboim, 1975), and Means-Ends Problem-Solving Procedure (Platt & Spivack, 1975a; Spivack et al., 1981). The resulting coefficients were largely in the order of .7 and above when calculated using methods which provide conservative estimates of test reliability (i.e., an intraclass correlation was used to calculate interrater reliability instead of a Pearson r ) . These estimates were subsequently compared to the reliability coefficients reported for the subtests of the Wechsler lntelligence Scale for Children (WISC-R; Wechsler, 1974) to provide a standard of comparison. Internal reliability coefficients for the WISC-R range between .70 and .86, compared to estimates of between .66 and .90 for the measures used by Mathias and Nettelbeck (1990a). Test-retest coefficients for the WISC-R subtests following a Imonth interval varied from S O to .92, while the social intelligence measures achieved coefficients between .54 and .74 after approximately 3 months. Inter-rater reliabilities were not provided for the subjectively scored subtests of the WlSC-R, but ranged between .76 and .98 for the Mathias and Nettelbeck measures. It was concluded from this that the seven measures selected to assess Greenspan’s social intelligence variables have good internal, test-retest, and inter-rater reliabilities when used with mentally retarded adolescents. Following on from this, Mathias and Nettelbeck (1990b) investigated the validity of a social intelligence construct by examining the relationship between conceptual (IQ), practical (adaptive behavior), and social

SOCIALLY BASED COMPETENCIES

149

(interpersonal skills) intelligence in a group of 75 mildly and moderately retarded adolescents (mean age 17 years 2 months). These three constructs collectively define what Greenspan (1979, 1981a) labeled adaptive intelligence. If, as predicted, there is an independent social intelligence construct which is identifiable through its uniquely interpersonal content, then a factor analysis of measures chosen to represent these domains would yield three separate factors corresponding to each form of intelligence. Based on their analysis, Mathias and Nettelbeck (l990b) extracted three factors which partially supported Greenspan’s model of adaptive intelligence. The factor which represented interpersonal competence was more accurately described as practical-interpersonal competence than social intelligence (interpersonal competence). This factor combined a general measure of adaptive behavior or practical intelligence with five of the seven social intelligence variables (i.e., role taking, social comprehension, psychological insight, moral judgment, and social problem solving). The second factor was labeled uccurucy of inference, and seemed to represent either the accuracy with which an individual interpreted visually presented tasks (social and nonsocial) or nonverbal intelligence. Finally, a third factor was likened to verbal intelligence because the measures defining this construct assessed different aspects of language ability. Thus, Mathias and Nettelbeck (1990b) found support for the existence of a practical-interpersonal competence construct that was distinct from other more cognitive forms of intelligence. Such a construct was shown to exist using verbally administered measures with a group of subjects for whom verbal deficits could conceivably interfere with the expression of other forms of competence. Though not completely independent of one another, there was a high proportion of unique variance that derived from either social or nonsocial sources. The common variance, on the other hand, was thought to be a consequence of the fact that both the “social” and “nonsocial” measures were verbally administered and required subjects to provide verbal responses. Two additional issues that are closely related to the validity of a social intelligence construct have been addressed by Mathias and Nettelbeck (199Ob) and Mathias (1988). These concern (1) the validity of Greenspan’s model of social intelligence and (2) the factor structure of the practicalinterpersonal competence construct extracted by Mathias and Nettelbeck (1990b) in their analysis of Greenspan’s model of adaptive intelligence. Greenspan’s model of social intelligence not only defines this construct in terms of seven different variables, but also tentatively divides these variables into the three domains of social sensitivity, social insight, and social communication. When separately examining the measures of social

150

June L . Mathius

intelligence, Mathias and Nettelbeck (l9Wb) revealed the existence of two factors. The first of these was a general social intelligence or interpersonal competence factor, defined by the measures of role taking, social inference, social comprehension, psychological insight, moral judgement, and social problem solving. The second factor, on the other hand, was defined by the two scores derived from the measure of social inference, and appeared to be analogous to the accuracy of inference factor extracted in the more general analysis of Greenspan’s model of adaptive intelligence. It was thought that this factor was a methodological artifact resulting from the inclusion of a single measure which used a picture interpretation format. The only measure which failed to load on either factor was the referential communication task, this result having been attributed to the nonsocial content of the test. Thus, Mathias and Nettelbeck ( l9Wb) provided no evidence to substantiate the hierarchical organization of the social intelligence construct, but did support the definition of social intelligence in terms of six of the seven variables described by Greenspan (1979, 1981a). An additional analysis, reported by Mathias (l988), factor-analyzed the practical-interpersonal competence construct (which comprised one practical intelligence and five social intelligence variables) and extracted only one factor. This factor was more strongly defined (i.e., had higher loadings) by the social intelligence (interpersonal) variables than the practical intelligence (social) variable. Thus, there was no evidence to suggest the existence of any superordinate structure to the practical-interpersonal competence construct. Mathias and Nettelbeck ( 1990b) not only isolated a practical-interpersonal construct but also defined it empirically in terms of a specific set of variables. Having done this, they evaluated the criterion validity of the practical-interpersonal competence construct by attempting to demonstrate that such measures provide better predictors of independent assessments of practical and interpersonal competence than measures representing either the verbal intelligence or accuracy of inference factors. Three different criterion measures were used: the first involved teacher ratings of a set of behaviors considered important to social interactions (Bay Area Functional Performance Evaluation, Social Interaction Scale; S. L. Williams & Bloomer, 1987); the second required teachers to provide an intuitive rating of a person’s social intelligence based on a brief description of this concept (Social Intelligence Rating Scale; Mathias, 1988); the third directly tested an individual’s understanding of appropriate versus inappropriate interpersonal strategies (Test of Interpersonal Competence; based on Foss, Cheney, & Bullis, 1986). Nine measures, which were selected to represent the three factors described previously, were

SOCIALLY BASED COMPETENCIES

I5 I

correlated with these three criterion measures using a sample of 20 mentally retarded adolescents (mean age 16 years 8 months). The results of this analysis revealed that the Test of Interpersonal Competence and the Social Intelligence Rating Scale provided evidence for the validity of both the practical-interpersonal competence and verbal intelligence factors. The Social Interaction Scale, in contrast, validated the measure of adaptive behavior but did not provide a good criterion measure for any of the factor scores. The failure to demonstrate criterion validity was attributed to (a) the use of indirect assessments of practical and interpersonal competence; (b) substantial correlations between factors, making it difficult to find measures which would selectively validate constructs; together with (c) an absence of data establishing the validity of the criterion measures. Although the results of this investigation did not provide evidence for the validity of a practical-interpersonal competence construct, they were also insufficient to negate this hypothesis. They did, however, highlight the difficulties associated with finding suitable measures and the need to develop more direct and intricate assessments of the target behaviors. Based on their findings, Mathias and Nettelbeck (1990b) concluded that the practical-interpersonal competence construct and its measures had construct and face validity, but that criterion validity had yet to be confirmed. Although variables which define a “social” construct (as distinct from “nonsocial” or cognitive competence) have been isolated, their actual contribution to a person’s observable level of practical and interpersonal competence remains to be established. A.

Implications for the Definition of Mental Retardation

The findings reported by Mathias and Nettelbeck (1990b) support a compromise between the current AAMR definition and Greenspan’s definition of mental retardation. Instead of three separate factors corresponding to conceptual, practical, and social intelligence, there were two factors (in addition to one that appeared to be a methodological artifact) which were likened to conceptual or verbal intelligence and a construct which combined the social and practical intelligence variables. Contrary to a prediction based on Greenspan’s model, there was not a distinct social intelligence factor. Similarly, there was not the adaptive behavior factor that would be expected on the basis of the AAMR’s distinction between cognitive intelligence and adaptive behavior. Instead, the resulting practical-interpersonal competence construct represented a combination of practical and social intelligence variables which was broader than the current AAMR definition of adaptive behavior. At the same time, it repre-

I52

June L . M a t h i a s

sented a domain that was less clearly distinguished than the practical and social intelligence constructs portrayed in Greenspan’s model of intelligence. On the whole, Mathias and Nettelbeck’s (19Wb) research substantiates the distinction between social and nonsocial competence when defining and describing mental retardation. In their present form, measures of adaptive behavior seem to provide a restricted interpretation of social competence and may need to be supplemented with measures of social intelligence (or interpersonal competence), the latter having proved to be more central to the practical-interpersonal competence construct than was the measure of adaptive behavior. 6.

Conclusion

The fact that the measures of interpersonal competence (social intelligence) combined with adaptive behavior (practical intelligence) to form a single construct suggests that the AAMR definition of mental retardation may be interpreting the social criterion used in the diagnosis of mental retardation too narrowly. Instead, it may be more appropriate to consider not only the skills that are prerequisites to independent living but also the subtle interpersonal skills which may ultimately determine the quality of our interactions with others. If substantiated by replication and additional refinements, this view would tie the definition of mental retardation to E. L. Thorndike’s classic model of intelligence by providing a clearer differentiation between social and academic intelligence. ACKNOWLEDGMENTS The work described here forms part of a project receiving financial support from the Apex Foundation for Research into Mental Retardation and the Channel 7 Children’s Research Foundation of South Australia Inc. I am indebted to Dr. Ted Nettelbeck for his assistance and encouragement throughout this project.

REFERENCES Ameck. G . G. (1975). Role-taking ability and the interpersonal competencies of retarded children. American Journal of Mental Deficiency. 80, 3 12-316. Ameck, G . G. (1976). Role-taking ability and the interpersonal tactics of retarded children. American Journal of Mental Deficiency. 80, 6-70. Ameck, G. G. (1977). Interpersonal competencies of the mentally retarded. A Piagetian perspective. In P. Mittler (Ed.), Research to practice in mental retardation. Editcation and Training (Vol. 2. (pp. 85-91). Baltimore, MD: University Park Press. Anderson, S., & Messick, S. (1974). Social competency in young children. Developmental Psychology. 10, 282-293.

SOCIALLY BASED COMPETENCIES

153

Barenboim, C. (1977). Developmental changes in the interpersonal cognitive system from middle childhood to adolescence. Child Development, 48, 1467-1474. Bates, P. (1980). The effectiveness of interpersonal skills training on the social skill acquisition of moderately and mildly retarded adults. Journal of Applied Behavior Analysis, 13, 237-248. Bornstein, P. H., Bach, P. J., McFall, M. E., Friman, P. C., & Lyons, P. D. (1980). Application of a social skills training program in the modification of interpersonal deficits among retarded adults: A clinical replication. Journal of Applied Behavior Analysis, 13, 171-176. Broom, M. E. (1928). A note on the validity of a Test of Social Intelligence. Journal of Applied Psvcholog.y, 12, 426-428. Broom, M. E. (1930). A further study of the validity of a Test of Social Intelligence. Journal of Educational Reseurch. 22, 403-405. Carmines, E. G., & Zeller, R. A. (1979). Reliability and validity assessment. Quantitative Applications in the Sociul Sciences. 17, 1-71. Castles, E. E.. & Glass, C. R. (1986). Training in social and interpersonal problem-solving skills for mildly and moderately mentally retarded adults. American Journal of Mentul Deficiency. 91, 3 5 4 2 . Cheney. D., & Foss, G. (1984). An examination of the social behavior of mentally retarded workers. Educution and Training qf the Mentully Retarded, 19, 216-221. Cohen, B. D., & Klein, J. F. (1968). Referent communication in school age children. Child Dc vclopmenI , 39, 597-609. Cole, P. G. (1976). Parents’ and teachers’ estimates of the school competence of handicapped and normal children. Ausfrrilian Journal of Mental Retardation. 4, 1-8. Connolly, J., & Doyle, A. (1981). Assessment of social competence in preschoolers: Teachers versus peers. Developmental Psychology. 17, 454-462. Coulter, W. A., & Morrow, H. W. (197th). A contemporary conception of adaptive behavior within the scope of psychological assessment. In W. A. Coulter & H. W. Morrow (Eds.), Adaptive behuvior: Concepts and measurements (pp. 3-20). New York: Grune & Stratton. Coulter. W. A., & Morrow, H. W. (1978b). A collection of adaptive behavior measures. In W. A. Coulter & H. W. Morrow (Eds.), Adaptive behavior: Concepis und meiisurements (pp. 141-156). New York: Grune & Stratton. Coulter. W. A., & Morrow, H. W. (1978~).The future of adaptive behavior: Issues surrounding the refinement of the concept and its measurement. In W. A. Coulter & H. W. Morrow (Eds.), Adaptive behavior: Concepts rind measurements (pp. 215-225). New York: Grune & Stratton. Cronbach, L. J. (1960). Essentials of psychological testing (2nd ed.) New York: Harper & Row. Damon, W. (1979). Why study social-cognitive development? Human Development, 22, 20621 I . de Jung, J . E., & Edmonson, B. (1972). Measurement ofsocial incompetency in adolescents (Final Report, Project No. 15-P-55298/0-03).Eugene: University of Oregon. (NTIS No. PB 234 638) de Jung, J. E., Holen, M. C., & Edmonson, B. (1972). Review of studies using the Test of Social Inference 10 measure social cue perception in retarded adolescents (Supplement to Final Report, Project No. RD IS-P-55298). Eugene: University of Oregon. (NTIS No. PB 236 487) de Jung, J. E., Holen, M. C., & Edmonson, B. (1973). Test of Social Inference for retarded adolescents: Measuring social-cue perception. Psychological Reports, 32, 603-618.

154

Jane L . Mothins

Doll, E. A. (1953). Measurement of social competence. A monual for the Vinelond Social Maturity Scale. Pines, MN: American Guidance Service. Edmonson. B.. de Jung. J. E.. & Leland, H. (1965). Social perceptual (nonverbal communication) training of retarded adolescents. Mental Retardation, 3, 7-9. Edmonson, B., de Jung, J. E., Leland, H., & Leach, E. M. (1974). The Test of Sociul Inference. Demographic and validity dutri. Administration and scoring giiidc. New York: Educational Activities. Edmonson. B., Leland, H., & Leach, E. M. (1%8). Social inference [ruining of retarded adolescents at the pre-vocationul /eve1 (Final Report, VRA Project RD- 1388-P). Kansas City: University of Kansas Medical Center. Edmonson, B., Leland, H., & Leach, E. M. (1970). Social inference training of retarded adolescents. Education and Training of the Mentally Retarded, 5, 169-176. Feffer, M. (1959). The cognitive implications of role-taking behavior. Journal of Personality, 27, 152-168. Foote, N., & Cottrell, L. S., Jr. (1955). Identity and interpersonal competenc~e.Chicago: University of Chicago Press. Ford, M. E., & Tisak, M. S. (1983). A further search for social intelligence. Journal of Educational Psychology, 75, 196-206. Foss, G., Cheney. D., & Bullis, M. (1986). Test of Interpersonal Competencejor Employment. Part of the working I and I1 series. Santa Monica, CA: James Stanfield. Foss, G., & Peterson, S. L. (1981). Social-interpersonal skills relevant to job tenure for mentally retarded adults. Mental Returdation, 19, 103-106. Glucksberg, S., Krauss. R., & Higgins, E. T. (1975). The development of referential communication skills. In F. D. Horowitz (Ed.), Review of child development research (Vol. 4. pp. 305-345). Chicago: University of Chicago Press. Goldfried, M. R., & D’Zurilla, T. J. (1969). A behavioral-analytic model for assessing competence. In C. D. Spielberger (Ed.), Current topics in clinical find community psycholo g y (Vol. I., pp. 151-196). New York: Academic Press. Greenspan, S. (1979). Social intelligence in the retarded. In N. R. Ellis (Ed.), Handbook of mental deficiency: Psychological theory und research (2nd ed., pp. 483-53 I). Hillsdale NJ: Erlbaum. Greenspan, S . ( 1980). Is social competence synonymous with school performance? Americon Psychologist. 35, 938-939. Greenspan, S. (198Ia). Defining childhood social competence: A proposed working model. In B. K. Keogh (Ed.), Advances in special education. Vol. 3: A research annual (pp. 1-39). Greenwich, CT: Jai Press. Greenspan, S. (1981b). Social competence and handicapped individuals: Practical implications of a proposed model. In B. K. Keogh (Ed.), Advances in special education. Vol. 3: A research annual (pp. 41-83). Greenwich, CT: Jai Press. Greenspan, S. (1982). On the trail of social competence: The case of mental retardation. Newsletter fbr Division 7 o j the American Psychological Association. pp. 19-23. Greenspan, S. (1984). Factor structure of competence in retcirded adults: Support ,for (1 narrower definition of mental retardation. Paper presented at the Gatlinburg Conference on Research in Mental Retardation, Gatlinburg. TN. Greenspan, S., & Barenboim, C. (1975, June). A Matrix Test ofRqferential Commrtnic~ation. Paper presented at the Fifth Annual Symposium of the Jean Piaget Society. Philadelphia. (ERIC Documentation Reproduction Service No. ED 125 784) Greenspan, S., Burka, A., Zlotlow, S., & Barenboim, C. (1975). A manual of referential communication games. Academic Therapy. 11, 97-106. Greenspan, S., & Javel, M. E. (1982). Personal competence profiling of exceptional chil-

SOCIALLY BASED COMPETENCIES

155

dren: Proposed ulternative to traditional categorization. Unpublished manuscript. (Available from S. Greenspan, University of Connecticut, Box U-64, Storrs. CT 06268) Greenspan, S., & Shoultz, B. (1981). Why mentally retarded adults lose their jobs: Social competence as a factor in work adjustment. Applied Research in Mental Returdution. 2, 23-38. Greenspan, S., Shoultz, B., & Weir, M. M. (1981). Social judgement and vocational adjustment of mentally retarded adults. Applied Research in Mental Returdation, 2, 335346. Gresham, F. M. (1981a). Social skills training with handicapped children: A review. Review of Educutionul Reseurch, 51, 139-176. Gresham, F. M. (1981b). Assessment of children's social skills. Journal of School Psvchology. 19, 12&123. Gresham, F. M., & Elliot, S. N. (1984). Assessment and classification of children's social skills: A review of methods and issues. School Psychology Review, 13, 292-301. Gresham, F. M., & Elliot, S. N . (1987). The relationship between adaptive behavior and social skills: Issues in definition and assessment. Journal of Special Education. 21, 167- I8 I . Grossman, H. J. (Ed.). (1973). Munual on terminology and classification in mentul returdution. Washington, DC: American Association on Mental Deficiency. Grossman, H. J. (Ed.). (1977). Munuul on terminology and classification in mentul returdution. Washington. DC: American Association on Mental Deficiency. Grossman, H. J . (Ed.). (1983). CIassiJication in mental retardation. Washington, DC: American Association on Mental Deficiency. Guilford, J. P. (1959). Three faces of intellect. American Psychologist, 14, 469479. Guilford. J. P. (1967). The nature of human intelligence. New York: McGraw-Hill. Guilford, J. P., & Hoepfner, R. (1971). The analysis of intelligence. New York: McGrawHill. Heber, R. (Ed.). (1959). A manual on terminology und classification in mental returdution (Monograph Supplement to the American Journal of Mental Deficiency). Washington, DC: American Association on Mental Deficiency. Heber, R. (Ed.). (1961). A manual on terminology and classification in mental returdution (Monograph Supplement to the American Journal of Mental Deficiency). Washington, DC: American Association on Mental Deficiency. Hendricks, M., Guilford, J. P., & Hoepfner, R. (1969). Measuring creative social intelligence. Reports from the Psychological Laboratory of the University of Southern Calijornia. 42, 1-35. Holley, S. (1980). A social skills group with mentally retarded subjects. British Journal of Sociul and Clinical Psychology, 19, 279-285. Hops, H. (1983). Children's social competence and skill: Current research practices and future directions. Behavior Therupy, 14, 3-18. Hoy, E. A.. & McKnight, J. R. (1977). Communication style and effectiveness in homogeneous and heterogeneous dyads of retarded children. American Journal of Mentul Deficiencv. 81, 587-598. Hunt, T. (1927). What social intelligence is and where to find it. Industriul Psychologv Monthlv. 2, 605-612. Johnson, R. C . (1962a). A study of children's moral judgments. Child Development, 33, 327-354. Johnson, R. C . (1962b). Early studies of children's moral judgements. Child Development, 33, 603-605. Keasey, C. B. (1977). Children's developing awareness and usage of intentionality and mo-

156

Jrine L . Mrithias

tives. In C. B. Keasey (Ed.), Nebraska Symposium on Motivation (Vol. 25. pp. 219-260). Lincoln: University of Nebraska Press. Keating, D. P. (1978). A search for social intelligence. Joitrnal of Educational Psyclrology. 70, 218-223. Kennett, K. F. (1977). Adaptive behavior and its assessment. In P. Mittler (Ed.). Research to practice in mental retardation: Educution und truining (Vol. 2, pp. 147-150). Ballimore, MD: University Park Press. Kleck. R. E. (1975). Issues in social effectiveness: The case of the mentally retarded. In M. J. Begab & S. A. Richardson (Eds.), The mentally retarded and societv: A sociol science perspective (pp. 181-195). Baltimore, MD: University Park Press. Kohlberg. L. (1969). Stage and sequence: The cognitive-developmentalapproach to socialization. In D. A. G o s h (Ed.), Handbook of socializatirin theory and reseurch. Chicago: Rand McNally. Kurtines, W., & Greif, E. B. (1974). The development of moral thought: Review and evaluation of Kohberg's approach. Psychological Bulletin, 81, 453-470. La Greca, A. M., Stone, W. L.. & Bell, C. R. (1983). Facilitating the vocational-interpersonal skills of mentally retarded individuals. American Journal of Menttrl Dy/iiienc.y, 88, 270-278. Leland, H. (1972). Overview of adaptive behavior concepts. In H. Leland, K. Nihira, & M. Husseini (Eds.), Measurement of adaptive behuvior IV: Consumer conference (pp. 7-8). Sponsored by Nisonger Centre, Ohio State University, Columbus and American Association on Mental Deficiency, Washington. DC. Longhurst, T. M. (1974). Communication in retarded adolescents: Sex and intelligence level. American Journal of Mental Deficiency, 78, 607-618. Marlowe, H. A., Jr. ( 1984).The structure of social intelligence. Dissertation Abstracts International, 45, 1993A. (University Microfilms No. DA 84-21.043) Marlowe, H. A,, Jr. (1985). Competence: A social intelligence perspective. In H. A. Marlowe. Jr. & R. 8. Weinberg (Eds.), Competence, development, theory and practice in special populations (pp. 50-82). Springfield, IL: Charles C. Thomas. Marlowe, H. A,, Jr. (1986). Social intelligence: Evidence for multidimensionality and construct independence. Journal of Educational Psychology. 78, 52-58. Marlowe, H. A., & Bedell. J. R. (1982). Social intelligence: Evidence for independence of the construct. Psychologicul Reports. 51, 461462. Mathias, J. L. (1988). Social intelligence and personal competence in mentally retarded adolescents. Unpublished doctoral dissertation, University of Adelaide. Adelaide, South Australia. Mathias, J . L. (1989). Social intelligence as a factor contributing to the personal competence of mentally retarded adolescents. In M.A. Luszcz & T. Nettelbeck (Eds.), Psychological development: Perspectives across the lqe-span (pp. 281-286). New York: NorthHolland/Elsevier. Mathias, J., & Nettelbeck, T. (199Oa). Internal, inter-rater and test-retest reliabilities of seven measures of social intelligence. Manuscript submitted for publication. Mathias. J., & Nettelbeck, T. (199Ob). Construct and criterion validity of Greenspan's model of adaptive intelligence. Manuscript submitted for publication. Matson. J. L., Kazdin. A. E., & Esveldt-Dawson, K. (1980). Training interpersonal skills among mentally retarded and socially dysfunctional children. Behuvbur Research and 7herapy. 18, 419-427. McFall, R. (1982). A review and reformulation of the concept of social skills. Behavioral Assessment, 4, 2-33. Meichenbaum, D., Butler, L.. & Gruson, L. (1981). Toward a conceptual model of social

SOCIALLY BASED COMPETENCIES

I57

competence. In J. D. Wine & M. D. Smye (Eds.), Social competence (pp. 36-60). New York: Guilford Press. Mercer, J. R. (1978).Theoretical constructs of adaptive behavior: Movement from a medical to a social-ecological perspective. In W. A. Coulter & H. W. Morrow (Eds.), Adaptive hehavior: Concepts und measurements (pp. 59-82). New York: Grune & Stratton. Meyers, C. E., Nihira, K.. & Zetlin, A. (1979). The measurement of adaptive behavior. In N. R. Ellis (Ed.), Handbook of mental deficiency: Psychological theory and research (2nd ed., pp. 431481). Hillsdale, NJ: Erlbaum. Minskoff, E. H. ( 1980a). Teaching approach for developing nonverbal communication skills in students with social perception deficits. Part 1. The basic approach and body language cues. Journal qf’Learning Disabilities, 13, 118-124. Minskoff, E. H. (l980b). Teaching approach for developing nonverbal communication skills in students with social perception deficits. Part 2. Proxemic, vocalic and artifactual cues. Journal of Learning Disuh Mischel, W., & Mischel, H. N. (1976). A cognitive social-learning approach to morality and self-regulation. In T. Lickona (Ed.), Moral development and behavior: Theory. reseurch und sociul issues (pp. 84-107). New York: Holt, Rinehart & Winston. Monson, L. B., Greenspan, S.,& Simeonsson, R. J. (1979). Correlates of social competence in retarded children. American Journul of Mental Deficiency, 83, 627-630. Moss, F. A. (1931). Preliminary report of a study of social intelligence and executive ability. Public Personnel Studies, 9, 2-9. Moss, F. A., & Hunt, T. (1927). Are you socially intelligent? Scientific American, 137, 108110. Moss, F. A., Hunt, T., Omwake, K. T., & Ronning, M. M. (1927). Social intelligence /est. Washington, DC: Center for Psychological Services. Muuss. R. E. (1982). Social cognition: Robert Selman’s theory of role taking. Adolescence. 67, 499-525. Nihira. K. (1977). Development of adaptive behavior in the mentally retarded. In P. Mittler (Ed.), Research to practice in mental retardation: Education and training (Vol. 2. pp. 157-168). Baltimore, MD: University Park Press. Nihira, K., Foster. R., Shellhaas, M., & Leland, H. (1969). AAMD Adaptive Behuvior Scale. Washington, DC: American Association on Mental Deficiency. Nihira, K., Foster, R., Shellhaas, M., & Leland, H. (1974). AAMD Adaptive Behavior Scale. 1974 Revision. Washington, DC: American Association on Mental Deficiency. Odum. S. L., & McConnell, S. R. (1985). A performance-based conceptualization of social competence of handicapped pre-school children: Implications for assessment. Topics in Eurly Childhood Special Educution, 4, 1-19. O’Malley, J. M. (1977). Research perspective on social competence. Merrill-Palmer Quarterly, 23, 2 9 4 4 . O’Sullivan, M., & Guilford, J. P. (1966). Sirfactor tests of social intelligence: Manual of instructions and interpre/utions. Beverly Hills, CA: Sheridan Psychological Services. O’Sullivan, M., Guilford, J. P.. & de Mille, R. (1965). The measurement of social intelligence. Reports from the Psychologicol Laboratory of the University of Southern Californiu. 34, 1-40, Piaget. J. (1932). The morol judgement of the child. London: Routledge & Kegan Paul. Platt, J. J., Siegel, J. M., & Spivack, G. (1975). Do psychiatric patients and normals see the same solutions as effective in solving interpersonal problems? Journal of Consulting und Clinical Psychology, 43, 179. Platt, J . J., & Spivack, G. (1972). Problem-solving thinking of psychiatric patients. Journal of Consulring und Clinical Psychology, 39, 148-151.

158

Jane L . Mmtliiw

Platt. J. J.. & Spivack. G. (197Sa). The means ends problem solving procedrtre N ~ ~ I I I I I N ~ . Philadelphia: Department of Mental Health Sciences, Hahnemann Medical College and Hospital. Platt. J. J., & Spivack, G. (1975b). Undimensionality of the means-ends problem-solving (MEPS) procedure. Jortrnnl rf' Clinical Psvchologv, 31, 15-16. Platt. J. J . , Spivack. G . . Altman, N . , Altman. D.. & Peizer, S. 9. (1974). Adolescent problem-solving thinking. Jortrnal of Consrtlting and Clinical Psvchology. 42, 787-793. Porter, N., & Taylor, N . (1972). H o w to ii.ssess the moral reasoning c~f'.stitdent.s.A rrochcw' guide to the NSC of' L a n w n c e Kolilberg's s~age-de~~elopmentcrl mettrod. Toronto: Ontario Institute for Studies in Education. Rest. J . R. (1974). The cognitive developmental approach to morality: The state of the art. Coun.seling ond Vtrlrres, 18, 64-78. Rosenberg, S . . and Cohen. B. D. (1966). Referential processes of speakers and listeners. Ps.vc~lio/~~giccil Review. 73, 208-23 I . Rueda, R.. & Chan. K. S. (1980). Referential communication skill levels of IIIOderdkly mentally retarded adolescents. American Jortrnal of Mentcil Deficiency. 85, 45-52, Sarason. B. (1981). The dimensions of social competence: Contributions from a variety of research areas. In J. D. Wine & M. D. Smye (Eds.). Socicrl competence (pp. 100-122). New York: Guilford Press. Schnall, M., & Feffer, M. ( 1968). Role-taking lirsk scorina criterio (American Documentation Institute Document No. 9010). Washington, DC: Library of Congress Photoduplication Service. Seitz. V.. & Zigler, E. (1980). Measure for measure. Americun Psychologist. 35, 939. Selman, R. I,. (I97 I ). Taking another's perspective: Role-taking development in early childhood. Child Development. 42, 1721-1734. Selman, R. L. (1975). Level of social perspective taking and the development of empathy in children: Speculations from a social-cognitive viewpoint. Jorrrria/ of M o r a l E d r c c ~ r i m , 5 , 35-43. Selman, R. L. (l976a). Social-cognitive understanding. A guide to educational and clinical practice. In T. Lickona (Ed.), M o r a l development und behavior: 7heory. resenrclr and socicil issues (pp. 299-315). New York: Holt, Rinehart & Winston. Selman, R. L. (3976b). Toward a structural analysis of developing interpersonal relations concepts: Research with normal and disturbed pre-adolescent boys. In A. I). Pick (Ed.), Minnesota Symposia on Child P.s.ycliologv (Vol. 10, pp. 156-200). Minneapolis: University of Minnesota. Selman, K. L. (1977). A structural-developrnental model of social cognition: Implications for intervention research. Corrnseling Psychologist. 6, 3 4 . Selman, R. L. (Ed.). ( 1979). Assessing intcrpersonal understanding: A n i n t e r i i e a ~und scoring mcrnrtul. Cambridge, MA: Harvard-Judge Baker Social Reasoning Project. crnd cliniSelman, R. L. ( 1980). 7 l i e growtlr of interpersonal understanding: Dc~i~elopmentcrl c ( ~analysis. / New York: Academic Press. Selman. R. L. (1981). The development of interpersonal competence: The role of understanding in conduct. Developmento/ Review. I , 401-422. Selman, R. L., & Jaquette. D. (1977). Stability and oscillation in interpersonal awareness: A clinical-developmental analysis. In C. 9. Keasey (Ed.), Nebrcrsko Svmposirtrn o n Motitvirion (Vol. 25. pp. 261-304). Lincoln: University of Nebraska Press. Selman. R. L.. Jaquette, D., & Lavin, D. R. (1977). Interpersonal awareness in children: Toward an integration of developmental and clinical child psychology. American Journal of Orthopsvchicrtry. 41, 264-274. Shanley. L. A , . Walker, R. E., & Foley. J. M. (1971). Social intelligence: A concept in search of data. Psvcliologicul Reports, 29, 1123-1 132.

SOCIALLY BASED COMPETENCIES

I59

Shantz, C. U. (1975). The development of social cognition. In E. M. Hetherington (Ed.). Review of child de\~elopmcwt reseurch (Vol. 5 , pp. 257-323). Chicago: University of Chicago Press. Shure, M. B., & Spivack, G. (1972). Means-ends thinking, adjustment, and social class among elementary-school-aged children. Joirrnal of Consulting und Clinicul Psycho/o ~ V , 38, 348-353. Shure, M. B., & Spivack, G. (1981). The problem-solving approach to adjustment: A competency-building model of primary prevention. Prevention in Human Services, 1, 87103. Shure, M. B., Spivack, G., & Jaeger, M. (1971). Problem-solving thinking and adjustment among disadvantaged pre-school children. Child Development, 42, 1791-1803. Simeonsson, It. J. (1978). Social competence. In J. Wortis (Ed.), Mento/ returdution and di~veloptnentuldisubilities: An unnitul review (Vol. 10, pp. 130-171). New York: Bruner/Mazel. Simeonsson, R. J.. Monson, L. B.. & Blacher. J. (1984). Social understanding and mental retardation. In P. H. Brooks R. Sperber, & C. McCauley (Eds.), Leurning und cognirion in the m e n t d v returded (pp. 389417). Hillsdale, NJ: Erlbaum. Spivack. G., Platt, J., & Shure, M. B. (1976). The problem solving upprouch ro udjusrmenr. San Francisco: Jossey-Bass. Spivack. G., Shure, M., & Platt, J. J. (1981). Meuns-ends problem solving. Srimuli and scoring procmfures supplement. Philadelphia: Department of Mental Health Sciences, Hahneman Medical College and Hospital. Taylor, A. R. (1982). Social competence and interpersonal relations between retarded and nonretarded children. In N.R. Ellis (Ed.), Inrernutionul review of reseurih in mental returdution (Vol. 11, pp. 247-283). Orlando, FL: Academic Press. Thorndike, E. L. (1920). Intelligence and its uses. Harper’s Muguzinc. 140, 227-235. Thorndike. R. L. (1936). Factor analysis of social and abstract intelligence. Journal of Educutionul Psvchologv. 27, 23 1-233. Thorndike, R. L. (1941). Social intelligence test. I n 0. K. Buros (Ed.), The IY40 mental rneusrtrement.s yearbook (p. 1253). Highland Park, NJ: Gryphon Press. Thorndike. R. L.. & Stein, S. (1937). An evaluation of the attempts to measure social intelligence. Psvchologicul Bulletin, 34, 275-285. Vernon, P. E. (1933). Some characteristics of the good judge of personality. Journul of Sociul Psvihlogv, 4, 42-57. Walker, R. E., & Foley, J. M. (1973). Social intelligence: Its history and measurement. P.sychologicul Reports. 33, 839-864. Waters, E., & Sroufe, L. A. (1983).Social competence as a developmental construct. Developmentul Review, 3, 79-97. Wechsler, D. ( 1974). Wech.sler Intelligence Scule for Children-Revised. New York: Psychological Corporation. Wedeck, J. (1947). The relationship between personality and ‘psychological ability’. Urirish Journul I$ Psychology. 37, 133-151. Weinstein, E. A. (1966). Toward a theory of interpersonal tactics. In C. W. Backman & P. F. Secord (Eds.). Problems in sociul psycho/ogy (pp. 394-398). New York: McGrawHill. Weinstein, E. A. (1969). The development of interpersonal competence. In D. A. G o s h (Ed.), Hundhook of sociulizuiion theory und research (pp. 753-775). Chicago: Rand McNally College Publishing Co. Weiss, D.. & Weinstein. E. (1967). Interpersonal tactics among mental retardates. Americon Journul of Mentul Lkficiencv, 72, 653461. Whiteman, M. ( 1970a). Development of conceptions of psychologicul cuctsuliry (Final Re-

I60

June L . Mathios

port, PHS Grant No. HD-02574-02). Washington, DC: U.S. Public Health Service. National Institute of Child Health and Human Development. Whiteman, M. (1970b). The development of conceptions of psychological causality. In J. Hellmuth (Ed.), Cognifive studies (Vol. I , pp. 339-360). New York: Bruner/Mazel. Whiteman. M. (1976). Children's conceptions of psychological causality as related to subjective responsibility, conservation, and language. Journul of Generic Ps.ycliologv. 128, 215-226. Williams. C. L. (1981). Assessment of social behavioral role play compared with Si of the MMPI. Behavior Therapy. 12, 578-584. Williams. S. L., & Bloomer. J. (1987). Buy ureu frtnctional perforniunce cvulrtcrfiort (2nd ed.). Palo Alto, CA: Consulting Psychologists Press. Wood, J. R.,Weinstein, E. A., & Parker, R. (1969). Children's interpersonal tactics. Sociologirul Inquiry, 37, 129-138. Woodrow, H. (1939). The common factors in fifty-two mental tests. Psychomefriku, 4,99108. Youniss, J . (1975). Another perspective on social cognition. In A. D. Pick (Ed.), Minnesotu S-vmposiu on Child Psychology (Vol. 9, pp. 173-193). Minneapolis: University of Minnesota Press. Zigler, E.. & Trickett, P. K. (1978). 1.0..Social competence. and evaluation of early childhood intervention programs. American Psychologist. 33, 789-798.

Conceptual Relationships between Family Research and Mental Retardation ZOLINDA STONEMAN UNIVERSITY AFFILIATED PROGRAM FOR PERSONS WITH DEVELOPMENTAL DISABILITIES AND DEPARTMENT OF CHILD AND FAMILY DEVELOPMENT THE UNIVERSITY OF GEORGIA ATHENS, GEORGIA 30602

1.

INTRODUCTION

The field of mental retardation research may be viewed as multiple subdisciplines in which advances in one area seldom are drawn upon to inform research in other, potentially related areas (Zigler & Hodapp, 1986). Yet, all mental retardation researchers share common theoretical, conceptual, and methodological challenges. Research specialization seems necessary, but it may result in narrowness and lack of integration of ideas and approaches (McCall, 1981). In order to counteract this tendency, conserted efforts must be made by investigators to tap the richness of information and conceptualization emerging from other areas of mental retardation research. The purpose of this article is to focus on one subspecialty, family mental retardation research, and to examine how research and theory in the general field of mental retardation can be drawn upon to enrich and strengthen research on families with mentally retarded members. In addition, contributions that this family research can make to advancing our 161 INTERNATIONAL REVIEW OF KESEARCH IN MENTAL RETARDATION. Vol. 16

Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

162

Zolindu Stonrrniitr

understanding of the condition of mental retardation are explored. There is no intent to provide a review of the literature on families with mentally retarded members; comprehensive reviews may be found in several recent volumes (e.g.. Blacher, 1984; Fewell & Vadasy, 1986; Gallagher & Vietze, 1986; McConachie; 1986; Mori, 1983; Seligman, 1983; Thurman. 1985). Rather, it is hoped that this article will encourage the building of conceptual bridges between family researchers and those studying other aspects of mental retardation.

A.

Theoretical Isolation

Theories guiding most mental retardation research have their roots in experimental and operant psychology (Brooks & Baumeister, 1977a). Theories guiding family mental retardation research, on the other hand, have been derived primarily from family sociology. Drawing from these rather discrete, nonoverlapping theoretical traditions has separated the models, methodologies, and constructs of mental retardation family research from most other mental retardation research. This, in part. is responsible for the subdiscipline isolation lamented by Zigler and Hodapp (1986). All research on mentally retarded individuals focuses on people who are unable to cope with the demands of schools, community, and family life and who have been singled out for special attention and study. Thus, although the modern-day cognitive researcher, for example, may believe he or she shares little common ground with the family researcher, and vice versa, researchers in the multiple subdisciplines of mental retardation are inextricably linked by the characteristics and circumstances of the population they study.

B.

Family Research Goals

Brooks and Baumeister ( 1977a) suggested that those conducting cognitive research on mentally retarded persons typically have one of two goals: to refine general cognitive theory or to understand the condition of mental retardation. They assert that, although researchers may claim that their work has the dual purposes of testing general cognitive models and of explicating the phenomenon of mental retardation, the two goals often do not lead to the same research questions or designs. The same can be said of family mental retardation research. When a similar perspective is applied to family mental retardation research, three main research goals can be identified: (a) to understand general family processes, (b) to under-

FAMILY RESEARCH AND MENTAL RETARDATION

163

stand the impact of mental retardation on families, and (c) to use family research as a means to understand better the condition of mental retardation. Pursuing the first goal, some family researchers have approached mental retardation as a potentially illuminating exemplar of a family stressor, which, when understood, can assist in explaining the generalized effects of a variety of stress factors on families (see Minnes, 1988; Singer & Irvin, 1989; Wikler, 1986, for reviews). Most family researchers have followed the second goal and examined the impact of mental retardation on families (see Bryne & Cunningham, 1985; Marfo, 1984; McConachie, 1986, for reviews). Unfortunately for the field of mental retardation, few family researchers have pursued the third goal. C.

Why Does Mental Retardation Affect Families?

Implicit in the family mental retardation literature is an assumption that a mentally retarded family member has a pervasive effect on families. In fact, the majority of mental retardation family research has been devoted to understanding that effect. Jacobsen and Humphry (1979) suggested that the assumption guiding this research has been that mental retardation is an unusual or deviant situation confronted by families, and, as such, would be expected to have a psychosocial impact. It may seem like calling into question the obvious to ask, Why would mental retardation be expected to have a significant impact on families? This seemingly simple question proves to be unexpectedly difficult to answer. As with the deceptively simple question What is mental retardation?, which has plagued the field for years, the question concerning why mental retardation should affect families seems self-evident, but has never been clearly conceptualized. In reality, these two questions are inseparably linked. The impact of mental retardation on families cannot be understood without clearly understanding the nature of mental retardation. Baumeister ( 1984) noted that the most fundamental theoretical consideration facing mental retardation researchers can be framed by the question “How is it that we shall conceptualize retarded behavior” (p. 18). Similar questions have stimulated strong debate across the decades and still defy a clear, unified response. Zigler and Hodapp (1986) caution that, although knowledge about mental retardation is accumulating quickly, scientific progress will remain limited until these basic definitional issues are resolved. Family researchers have often attempted to study the impact of mental retardation on families without joining the field in the common conceptual struggle to define clearly the population under study.

I 64

Zolincln Stonemnn

Yet, without providing an answer to the basic questions What is mental retardation? and Who is mentally retarded?, the question How does mental retardation affect families? is unanswerable. The remainder of this article is divided into four major sections, each devoted to seeking avenues for cross-fertilization between the fields of mental retardation and family research. The first is a more detailed discussion of the major conceptual issues surrounding the nature and conceptualization of mental retardation. The second addresses the implications of classification of subgroups of mentally retarded individuals for research on families with a mentally retarded member. The third focuses on the implications of personality, motivation, and other individual differences for family mental retardation research. The final section provides concluding comments.

II.

ISSUES SURROUNDING THE CONCEPTUALIZATION OF MENTAL RETARDATION

It is a truism that no phenomenon can be reliably studied unless it can be operationally defined and consistently identified. Thus, any research endeavor must first define key constructs. As noted previously, for the field of mental retardation research, this has resulted in an ongoing attempt to provide a clear, unambiguous set of definitional criteria for mental retardation. This has been an arduous task, and controversy remains. Discussion of the implications of these issues for family mental retardation research focuses on the following: (a) mental retardation as a cognitive deficit, (b) mental retardation as a psychometric construct, (c) adaptive behavior, and (d) mental retardation as a socially defined phenomenon. The importance of each issue for the family researcher will be explored, as well as contributions that family research could make to resolution of the issue. A.

Mental Retardation as a Cognitive Deficit

Farber (1968) described the intuitive societal view of mentally retarded persons as “dull-witted, deficient in vocabulary, slow to understand, unable to follow an argument logically, inattentive, poor in memory, and unable to manipulate symbols readily” (p. 3). These characteristics are predominately cognitive in nature. From a scientific viewpoint, Baumeister (1984), Zigler, Balla, and Hodapp (l984), and others have suggested that the fundamental property of retarded persons is that their cognitive

FAMILY RESEARCH AND MENTAL RETARDATION

165

processes develop at a slower rate, asymptote earlier, and are less efficient than those of other persons in society. 1. RELEVANCE OF COGNITION FOR THE

FAMILY RESEARCHER One of the most extensive literatures in the field of mental retardation is devoted to understanding cognitive functioning. At first glance, this large literature would seem to be of little relevance to family researchers. This impression, however, would be incorrect. Brooks and Baumeister (1977a) have a telling question embedded in their paper calling for increased ecological validity in mental retardation research. This question can be paraphrased as follows: How do deficits in frequently studied cognitive processes, such as short-term memory and discrimination learning, contribute in any meaningful way to understanding day-to-day functioning of mentally retarded individuals? Although this question was framed as a challenge to cognitive researchers to defend the ecological validity of their work, it serves to highlight an important interface between cognitive and family research. Families share a home with an individual who, among other characteristics, learns slowly, has impaired memory and problem-solving skills, and often does not transfer learning from one task to another. Specific cognitive deficits associated with mental retardation are so pervasive that Detterman (1987) coined the term the “everything deficit” to capture the breadth of cognitive domains in which the functioning of mentally retarded persons has been found to be impaired (p. 3). It is surprising that day-to-day adaptations which result from living with a family member with such pervasive cognitive deficits have not received more research attention. It is common to read reviews of mental retardation family research (or research articles) without the words “cognitive” or “cognition” appearing in text. If one accepts the premise that cognitive deticits lie at the heart of an understanding of mental retardation, then how can family researchers propose to study mental retardation while ignoring cognition? Consider a frequent finding in the family literature, namely, that mentally retarded children limit the freedom of family members to go out into the community (e.g., Jacobsen & Humphry, 1979; Salisbury, 1987). A mentally retarded child with no secondary handicapping conditions (e.g., physical handicaps) may be precluded from remaining at home unsupervised because of the child’s cognitive competencies (or lack thereof). The parent cannot trust that the child will have sufficient cognitive efficiency (e.g., perceptual encoding, short and long-term memory) and strategic cognitive processing abilities (Brooks, McCauley, & Merrill, 1988) to solve problems and respond to unexpected events safely while home

166

Zolirido Storiemori

alone. Such examples suggest that an examination of cognitive underpinnings of findings in the family literature can advance our knowledge about mental retardation. One example of a more integrated approach has been offered by Cicchetti and Pogge-Hesse (1982), who posited developmental interdependence between cognitive and affective processes. In the case of mental retardation, this implies that cognitive delays will be mirrored by the child’s affective responses to others in his or her world. Thus, the study of emotions and affectivity brings the researcher to consider the roles of memory, relational abilities, person-other discriminations, and other cognitive processes as they relate to affective responsivity, and, ultimately, to the nature of the parent-child relationship. Other researchers (e.g., Dunst, 1984; Simeonsson, Monson, & Blacher, 1984) have also used a cognitive developmental/Piagetian framework to link social and cognitive domains, thus targeting issues of interest to family researchers. These models, as well as recent work following a Vygotskian perspective (see Belmont, 1989, for review), provide the possibility of a plausible response to the challenge set forth by Brooks and Baumeister (1977a) to demonstrate how mental retardation research in cognition is socially relevant. Sternberg and Spear (1985) suggested that the common impression that mentally retarded persons differ from each other only on one single cognitive continuum is mistaken. Heterogeneity probably exists in the levels and patterns of cognitive deficits experienced by mentally retarded persons (e.g., Edgerton & Edgerton, 1973; Sternberg & Spear, 1985). Family researchers seem very well positioned to examine differential influences of specific constellations of cognitive skills on family interaction and adaptation, and to investigate longitudinally parameters of family experience which may help to shape different patterns of cognitive abilities. The interface between cognition and family process exists across the life cycle; it remains an open challenge to modcl these interactive processes across development in families with a mentally retarded member. 2. RELEVANCE OF FAMILIES FOR THE COGNITIVE RESEARCHER Less surprising in their absence than studies examining the influences

of cognitive deficits on families, but nonetheless important, are studies which attempt to abstract important information about cognitive processing from the family context. House (1977) described the difficulty in deriving information about cognition from the ebb and flow of naturalistic contexts, such as the family, and argued that basic research and theoretical refinement require the experimental control only possible in labora-

FAMILY RESEARCH AND MENTAL RETARDATION

I67

tory settings. Although this view has been challenged (Brooks & Baumeister, 1977a, 1977b),most researchers seem to behave in ways consistent with House's belief. In their report on the state of the science in mental retardation, the National Institute of Child Health and Human Development ( 1986) called for a reconciliation between cognitive researchers working in laboratory settings and those few cognitive researchers using methods of naturalistic observation. The report concluded that little integration exists between the two approaches, to the detriment of the field. Contrary to the view expressed by House ( 1977),theoretically relevant questions about cognition can be addressed in the natural milieu; studying families is one excellent means to this end. Family research can never replace traditional laboratory methods, but can complement knowledge gained in the laboratory and expand its meaning. One of the strongest arguments for linking the social milieu with the study of cognition was made by Borkowski and Turner (1988): Inner, cognitive lives are inseparably linked to the social events that surround and define outer lives. Hence, the study of social contexts and social interchanges needs to be interrelated with the analysis of cognitive development. (p. 253) Researchers need to emphasize the fact that multi-component theories will be developed and research undertaken, both within and across psychological systems, only when theoreticians become bold enough to include social, familial, and personal constructs-which are so critical to impaired development-in their models of cognition. Only then can the causes and consequences of memory, cognition, and problem-solving in mentally retarded persons be fully understood. (p. 263)

Turnure (1985) also argued for consideration of the social aspects of cognition, proposing language and communication as primary vehicles through which social agents, such as family members, influence cognitive development. Borkowski and Turner (1988) urged investigators to identify how interactions between mentally retarded persons and their families affect cognition, metacognition, and motivation over the course of development. Such studies would require investigators to cross subdisciplinary boundaries by drawing upon what is known about families, as well as what is known about the development of cognition. This theoretical integration would not be easily accomplished. Neither current models of cognition nor models of social influence (or family influence) are sufficient to guide such research. Borkowski and Turner (1988) issued a charge to the field for the next decade to develop multisystem transactional models which incorporate

168

Zolirida Stoneman

motivational and social variables into a framework focusing on the development of cognition in mentally retarded persons. That charge was delivered to cognitive researchers, but is relevant for family mental retardation researchers as well. A multisystem family process model must be able to account for the direct and indirect influences of mothers, fathers, and siblings, as well as the influences of extended family and significant extrafamilial others. Additionally, qualitative aspects of key family subsystems (including marital, parent-child, and sibling dyads) and reciprocal transactions among family subsystems must be considered. Finally, contextual aspects of family life (i.e., divorce, single-parenthood, family size, poverty, major life stresses), parent emotional and physical health, and life-cycle transitions cannot be overlooked. Understanding how the development of specific cognitive skills by mentally retarded individuals is facilitated or compromised by the complex, dynamic pattern of influences present in the family presents a major challenge for the next decade, and the years beyond.

3. SOCIAL MEANING OF COGNITIVE DEFICITS With competencies often far below those expected for their age, mentally retarded persons have been conceptualized as children dwelling in adolescent or adult bodies, a notion that Wolfensberger ( 1972) termed the "eternal child" (p. 23). Although this view has been discredited, the meaning to be attributed to the discrepancy between the mental and chronological ages of mentally retarded individuals remains a murky and uncharted domain (Baumeister, 1984). Farber (1959, 1960; Farber & Jenne, 1963) conceptualized families of mentally retarded children as experiencing an arrest in family life cycle as the children matured, with the mentally retarded family member forever locked in childlike role relationships with parents and siblings. Although this conceptualization captures an important aspect of family role relationships and has proved useful to u s in our family research (e.g., Stoneman, Brody, & Abbott, 1983; Stoneman, Brody, Davis, & Crapps, 1987. 1988, 1989), few would suggest that sharing a home with an 12-year-old child functioning at a 4-year level would resemble having a normally developing 4-year-old child. Differences in physical maturity, personal experience, and family history make older mentally retarded children unlike their younger nonretarded counterparts. As adolescents move through puberty, these differences become even more acute. The accumulation of life experience which accompanies greater chronological age can give mentally retarded persons an advantage in some competencies, including age-appropriate social behavior and sex roles (Zigler & Hodapp, 1986). Some living skills may be more directly related to chronological than mental age. Zigler and Hodapp (1986) use the exam-

FAMILY RESEARCH A N D MENTAL RETARDATION

I69

ple of parents considering a 7-year-old child too young to travel alone on a bus, while a mentally retarded adult with a mental age of 7 might be encouraged to do so. The family context is a primary arena for multiple dynamic tensions resulting from interplays between mental age deficits and advancing chronological age, such as the tension between a mentally retarded adolescent’s desire for age-appropriate independence and continued need for close family supervision, between puberty and an inability to exercise responsible judgment, and between increased body size and weight and need for physical caretaking. Family researchers are logical candidates for tackling the difficult task of disentangling the effects of chronological and mental age. As with other major questions concerning cognition and families. however, family researchers have yet to address these issues with the intent of contributing to a general understanding of chronological and mental age-related cognitive deficits. 4. SUMMARY While cognitive researchers have been criticized for ignoring the social aspects of mental retardation (Brooks & Baumeister, 1977a), family researchers could be criticized for ignoring the cognitive aspects. What is needed is an understanding of the causative relationships between deficient cognitive abilities and family processes. Several barriers stand on the path to this goal. Methodological and theoretical allegiances divide, rather than unite, researchers. Cognitive mental retardation researchers focus most energy on older children who are mildly mentally retarded (Baumeister, 1984), while family researchers tend to work in the area of infancy and early childhood, often with moderately or severely retarded children (Stoneman, 1989). Although differences and obstacles exist, compelling issues potentially unite family and cognitive science, to the ultimate benefit of the field of mental retardation research. Family researchers are in a position to discern the cognitive deficits that make a difference in the child’s ability to function within the family system from those that are either irrelevant or easily circumvented and to chronicle the cognitive demands placed on mentally retarded persons by multiple subenvironments within the same family, as well as across different families. Possibilities for amalgamating cognitive and family models to understand mental retardation better are intriguing, but will require researchers to reach across subdisciplinary boundaries to share research methods and conceptual frameworks. B.

Mental Retardation as a Psychometric Construct

Identification of persons as mentally retarded rests on a psychometric foundation, objectified by the IQ test. IQ tests are not theoretically

I70

Zolindu S t o t i m m n

driven; their construction and content reflect measurement and statistical considerations (Weinberg, 1989). Even though IQ scores carry no explanatory power, the behavior of those labeled as mentally retarded is often interpreted as caused by their low IQ. Zigler and Burack (1989) note that his logic creates a tautology: that mentally retarded individuals behave like they do because they are mentally retarded. Family researchers sometimes have been captured by a similar circular argument. Finding differences between families with and without mentally retarded members on some dimension of interest, the researcher attributes the difference to the presence of an individual with low IQ in the family. To paraphrase Zigler and Burack's (1989) tautology, families with mentally retarded members are different from other families because they have a family member who is mentally retarded. Any psychometric definition relies on a set of arbitrarily determined criteria, which can change quite dramatically over time. In 1961 (Heber, 19611, the defining IQ cutoff for a diagnosis of mental retardation was changed from two to one standard deviation below the mean, resulting in ajump in prevalence from 2 to 16 % of the population. from 6 to 20 million persons (Zigler & Hodapp, 1986). As a result, an additional 14 million families were placed in a position of coping with a diagnosis of mental retardation in a child. The competencies and characteristics of these children presumably remained stable across this transitional period: only the process through which the children were labeled changed. Conversely, in 1973, the criteria reverted to two standard deviations below the mean, once again constricting the number of persons identified. In addition to definitional inconsistencies across time, the criteria used by schools and other agencies for classifying children as mentally retarded vary across states, creating variability in prevalence rates (Frankenberger & Harper, 1988;Janicki, 1988; Utley, Lowitzer & Baumeister, 1987). This variability lead Ulley er ul. (1987) to recommend that researchers report detailed information on study samples rather than rely on global data obtained from extant records. Unfortunately, however, family mental retardation researchers do not always provide a detailed description of their study samples (Stoneman, 1989). These definitional inconsistencies have important ramifications for interpreting findings from family mental retardation research. First, cohort effects, corresponding to definitional changes, are evident in family mental retardation studies examined across time. These cohort effects compromise comparisons among studies and can distort the meaning of study findings. For example, families of mildly retarded children in studies from 1961 to 1973 plausibly could have found their way into nonretarded comparison groups in 1974 and thereafter. The severely mentally retarded

FAMILY RESEARCH AND MENTAL RETARDATION

171

children in Farber's classic family studies (Farber, 1959, 1960; Farber & Jenne, 1963) had IQs as high as 50, clearly inconsistent with current definitions. Second, differences in definitions introduce sampling variability across studies, since family researchers often rely on service agencies for identification of study populations. Variations in definitions may translate directly into sampling heterogeneity in research implemented in different geographic regions or using different service agencies for subject identification. C.

Adaptive Behavior

Persons with the same IQ can have large differences in adaptive and social competence (Brooks & Baumeister, 1977a; Zigler ef a / . , 1984). Although adaptive deficits are the most salient aspect of mental retardation to the general public (Leland, 1973) and add conceptual meaning to an otherwise psychometrically based definition (Barnett, 1986), adaptive behavior has been of only secondary interest to mental retardation researchers (Baumeister, 1984) and has received even less attention from those studying families. The failure of adaptive behavior to capture the imagination of researchers can be attributed to measurement dilemmas which have plagued the construct. When the markers commonly used to assess adaptive behavior are examined, it is possible to see direct ramifications for the study of the family. Using the Interview Edition of the Vineland Adaptive Behavior Scales (Sparrow, Balla, & Cicchetti, 1984) as an example, one finds items focusing on eating, self-care, food preparation, general household tasks, home and community safety, restaurant skills, money management, and time orientation. Other subdomains assess interpersonal relationships, play and use of leisure time, social communication, responsiveness, expression of emotion, friendships, sharing and cooperation, independence, sensitivity to others, and impulse control. These competencies are central to family living. Conceptualized as the interaction between the person and his or her environment (Brooks & Baumeister, 1977a), adaptive behavior takes on importance for understanding the interface between the mentally retarded person and the family environment. Social adaptation is culturally relative. It is, in essence, a question of goodness-of-fit between a person and the demands of a social context. "A person can be neither adaptive nor maladaptive in a vacuum" (Nihira, 1973, p. 111). Landesman and Ramey (1989) suggested that day-to-day adaptation can be more closely tied to contextual variables than to an individual's skills or abilities. Unfortunately, little is known about the characteristics of the social environments in which mentally retarded chil-

172

Zolindri Stonemtin

dren and adults live (Charlesworth, 1984). Few studies have examined the daily life experiences of mentally retarded persons. This is particularly true for persons living at home with their families. LandesmanDwyer and Butterfield (1983) note that one cannot generalize from environments of nonretarded persons, since the disparity between mentally retarded children’s mental and chronological ages may create atypical, nonnormative patterns of experiences unique to them. The triarchic theory of mental retardation, presented by Sternberg and Spear (1985), provided elaboration of the adaptive behavior construct, making it a primary criterion for intelligence assessment. Adaptive behavior (the contextual subtheory), one of three proposed components, or subtheories, of intelligence, is conceptualized as the process through which persons obtain optimal fit between themselves and their environments. Adaptation strategies include changing one’s behavior to fit the demands of an existing environment, selecting another better-suited environment, and modifying the environment to achieve better fit. Sternberg and Spear (1985) asserted that research on adaptive behavior is the most needed area of mental retardation research (and, by extension, of family mental retardation research as well). Family researchers can contribute to the further refinement of the adaptive behavior construct by providing important empirical information about the contextual demands that family life places on mentally retarded children and adults, the modifiability of those demands, and the family’s response to adaptive inadequacies. Research on person-setting interactions is needed (Brooks & Baumeister, 1977a; Haywood, 1976; Landesman-Dwyer & Butterfield, 1983). By the intense study of the family, specific components of adaptive behavior critical to successful functioning as a family member can be identified and quantified and variability across families can be examined. Valid studies of adaptive behavior must be observed across multiple social contexts (Weinberg, 1989). including the family home. For the family researcher, adaptive behavior may be a mediating factor (or set of factors) that can assist in understanding the processes through which mental retardation affects families. Consideration of adaptive behavior can contribute to the advancement of family research, and, reciprocally, family research paradigms can speak to important conceptual and definitional issues relative to the construct itself.

D.

Mental Retardation as a Socially Defined Construct

The previous definitional and conceptual issues all presume that mental retardation is a condition or phenomenon that resides in an individual, to be identified and measured. Not all agree with the conceptualization of mental retardation as an “objective entity” (Zigler & Hodapp, 1986, p.

FAMILY RESEARCH AND MENTAL RETARDATION

173

12) or trait which exists independently of societal identification or labeling. I . SOCIAL RELATIVISM

It has been argued that mental retardation is a socially defined condition, and, as such, cannot be understood without examining cultural and societal variables (e.g., Barnett, 1986; Brooks & Baumeister, 1977a; Farber & Rowitz, 1986). Scheerenberger (1983) began his comprehensive volume on the history of mental retardation with the comment “Mental retardation is primarily a socioculturally determined phenomenon . . .” (p. 3). Research documents cross-cultural variation in the meaning of mental retardation (Barnett, 1986), reinforcing the notion that the identification of mentally retarded persons rests on culture-specific criteria. Perhaps the most radical statement of a social-relativist position came from Wolfensberger (1972), “Since deviancy exists by social definition, it can also be prevented or reversed by social redefinition” (p. 25). For persons in sympathy with this more extreme view, mental retardation exists only when it is socially defined; should a social definition cease to exist, mental retardation would disappear. A conclusion follows from this social-relativist argument that mental retardation would have little or no impact on family life in a society which did not acknowledge its existence or define it. Conversely, the impact of mental retardation on families in modern society results, in large part, from external societal factors, rather than from characteristics of the mentally retarded family member. If a given individual “escaped” diagnosis, and thus were never formally labeled, his or her family would be spared any negative (or positive) effects arising from the person’s mental retardation. Should the psychometric definition of mental retardation become more conservative, identifying fewer persons, fewer families would be affected. Hodapp and Zigler (1986) used a common analogy to frame the key issue: like the proverbial tree falling in the forest with no witnesses, would mental retardation exist (and would it have an impact on families) if it were not identified or labeled? Advocates of a radical socialrelativist position would answer, No. Most researchers would adopt the more moderate view that identification of persons as mentally retarded emerges through an interaction between society and the individual (e.g., Barnett, 1986; Baumeister, 1987; Edgerton, 1967; Farber, 1968). The Surplus Population Model (Farber, 1968) provides the most detailed and in-depth presentation of an interactive view. For Farber, the societal act of defining certain individuals as mentally retarded helps to maintain the existing social structure by creating a group of persons who are marginal in the workforce. These marginal

workers serve as a buffer as the economy of a country expands and constricts, flowing into and out of the workforce as the need arises. Persons labeled as mentally retarded are not arbitrarily singled out. Certain characteristics bring this group to the attention of societal institutions. According to Farber (1968), these persons cannot succeed in ordinary school situations, must be supported in daily life by members of the general population, and have difficulty in maintaining marital relationships. Thus, the label of mental retardation results from persons with certain characteristics being unable to fill specific “organizational slots” (Farber, 1968, p. 10) within a given society. Definitional criteria for mental retardation reflect the political, social, and economic climate existing at a specific geographic location at a given point in history (Rarnett, 1986; Landesman & Ramey, 1989; Lowitzer, Utley, & Baumeister, 1987; Utley et d.,1987). Family researchers are in a strong position to address questions about social relativism. Barnett (1986) described the challenge as one of isolating effects which result from societal influences from those that result from medical, genetic, and individual differences. Societal forces may be especially potent when the mental retardation is mild, while characteristics and demands of the retarded family member may have overriding influence when the mental retardation is more severe. Models are needed that simultaneously consider effects on families of societal forces, effects of living day to day in the same home with a mentally retarded individual, and interactions between these two sources of influence. 2. MENTAL RETARDATION AS A PHENOMENON BASED IN THE SCHOOL CONTEXT The initial need to determine who was and who was not mentally retarded came with the advent of universal education (Farber, 1968; Scheerenberger, 1983). The first intelligence tests were designed to assist in the identification of children experiencing school-related learning difficulties. The evolution of the definition of mental retardation has been closely linked to the needs and realities of the educational system. For more severe mental retardation, often identified during infancy. prevalence is greatest in the birth to 5 years age group (MacMillan, 1982). Mild retardation. on the other hand, is most prevalent during the school years, peaking between the ages of 10 and 14 years, when school demands are greatest (Zigler & Hodapp, 1986). Children with mild retardation tend not to be detected until they enter school and are difficult to identify after leaving the educational system. Perceptions of a child as mentally retarded may differ across school and nonschool contexts (Barnett, 1986). This is reflected in the concept of the “six-hour mentally retarded child,”

FAMILY RESEARCH A N D MENTAL RETARDATION

175

who functions as mentally retarded in the classroom setting but is considered “normal” in all other aspects of his or her life (President’s Committee on Mental Retardation, 1969). Uneven detection and inconsistent application of definitions creates a discrepancy between a hypothetical 3% population incidence of mental retardation and the reported prevalence of I or 2% (Mercer, 1973; Zigler & Hodapp, 1986). The lack of correspondence between incidence and prevalence across the life span has several interpretations. A social relativist would suggest that the definition of mental retardation has age-specific meaning, reflecting developmental shifts in the fit between persons and their environments. The person-environment fit is most problematic during the school years, when demands for academic achievement are greatest. Difficulty experienced by many children in meeting these educational demands results in a greater prevalence of mental retardation among school-aged children. After school ends, many persons identified during the school years are able to meet the demands of adult society and are no longer considered mentally retarded. Thus, a person holding a social-relativist position would argue that developmental fluctuations in prevalence rates are real differences caused by systematic variations in environmental demands across the life span. There is strong disagreement with the social-relativist view of mental retardation. Zigler and Hodapp (1986), for example, cautioned that it is inappropriate to suggest that mildly retarded children become retarded at the onset of school and are cured when they leave. They focus on the stability of IQ across the life span and view prevalence rate fluctuations as due to differential detection rather than to differential incidence of mental retardation. The identification of mildly mentally retarded individuals may vary (50% disappear after school ends), but their mental retardation is a relatively stable trait which persists across time (Zigler & Hodapp, 1986; Zigler et d., 1984). Thus, Zigler and Hodapp’s (1986) response to their own question concerning the proverbial tree falling in the woods (Hodapp & Zigler, 1986) would be that mental retardation exists whether or not it is identified or labeled by others. Koegel and Edgerton (1982) provide some support for the notion that mildly mentally retarded young adults may not “disappear” when they leave school. They may vanish from agency records, since access to services often ceases with school termination, but their ethnographic data suggest that many continue to be viewed as handicapped by their families (and by themselves) and need extra support to negotiate life in the community. The chak of competence identified by Edgerton over 20 years ago (Edgerton, 1967) still seems to be used by at least some mildly mentally retarded adults to “pass” as nonretarded. Using the triarchic frame-

176

Zolitida Stonernan

work of Sternberg and Spear (19851, these persons demonstrate contextual intelligence by manipulating their environments to achieve better person-environment fit, but may not escape detection by close others, such as family members, who recognize and assist in creating the environmental modifications needed to sustain successful community adaptation. Little is known about individuals whose formal identification as mentally retarded is restricted to the school context. Even less is known about their families. Empirical evidence is lacking to document whether “sixhour mentally retarded children” actually exist, or exist in great numbers. Similarly, it has been questioned whether mildly mentally retarded adults actually disappear into their communities after school (e.g., Edgerton, 1988; Landesman-Dwyer & Butterfield, 1983). Life-span studies of families with mentally retarded members can test the validity of tenets of the social-relativist view of mental retardation. Edgerton ( 1988) called for increased efforts to study mentally retarded individuals who either were not diagnosed as children or who disappeared from agency roles as adults. This work falls squarely within the domain of family researchers: it is the family context which endures across time, providing an ideal vehicle for examining the functioning of these “invisible” individuals across the life span. The home is also the major noneducational context experienced by school-age children, and, as such, is critical to understanding the social ecology of those considered to be “six-hour retarded children.” Age fluctuations in the prevalence of mental retardation create a difficult and often neglected methodological issue. Uneven prevalence rates across development call into question the interpretability of cross-sectional research examining developmental or age-related differences in families of mentally retarded persons, because the characteristics of persons identified as mentally retarded fluctuate quite dramatically across the life span. Differences between age-stratified cohorts of families, in some cases, could be due to sampling distortions rather than “true” developmental or life-span patterns. 3. GENDER, CULTURE, AND RACE

Males and minority students are overrepresented among those labeled as mentally retarded. In their review of epidemiological studies, McLaren and Bryson ( 1987) concluded that male-to-female ratios averaged about 1.6:I , with evidence for an overrepresentation of males among mild mental retardation, as well as among certain more clearly organically caused conditions. Several causative factors have been proposed: Aggressive behavior may single boys out for testing and subsequent labeling (Farber, 1968: Shonkoff, 1982), society may expect more abstract and logical thought from males (Farber. 1968), and X-linked genetic or biological fac-

FAMILY RESEARCH AND MENTAL RETARDATION

I77

tors may predispose boys to mental retardation (Zigler & Hodapp, 1986). Plomin (1989) suggested that Fragile-X syndrome may be a major reason for the overrepresentation of boys. McLaren and Bryson (1987) concluded that gender ratios differ with level of retardation, etiology, social class, and age, thus implicating multiple social and biological determinants. Certain minority groups, particularly African-Americans, also are overrepresented among persons labeled as mentally retarded (Maheady, Algozzine, & Ysseldyke, 1984). Since family researchers primarily study majority culture children, they have shed little light on the relationship between minority group membership and mental retardation. There is little doubt that societal norms for behavior are related to gender, social class, ethnicity, and race. It follows that these factors also contribute to society’s identification of intellectual and adaptive inadequacy in its members (Nihira, 1973). It is unlikely that these social forces act in isolation. They undoubtedly interact with environmental and/or biological factors, as well as family factors, to place certain groups at risk for mental retardation. Family researchers bring unique skills and perspectives to the tasks of elucidating how these interactive processes cause certain individuals to be identified as mentally retarded and understanding how gender, race, and culture may influence family adaptation to mental retardation in a family member.

111.

CLASSIFICATION WITHIN MENTAL RETARDATION

The previous discussion dealt with a range of conceptual issues emanating from the diagnosis or identification of certain persons as mentally retarded. Baumeister (1987) noted that the simple bivariate dichotomy of individuals as mentally retarded or nonretarded fails to capture the diversity of persons labeled mentally retarded. Although most would agree with Baumeister (1987) that the nominal classification system of mild, moderate, severe, and profound mental retardation is only minimally useful, no clear consensus exists as to how best to subdivide the group of mentally retarded persons. Many family researchers have focused on families of Down syndrome children (e.g., Abramovitch, Stanhope, Pepler, & Corter, 1987; Berger & Cunningham, 1981, 1983, 1986; Buckhalt, Rutherford. & Goldberg, 1978; Cook & Culp, 1981; Crawley & Spiker, 1983; Gath, 1972, 1973; Gath & Gumley, 1984, 1986, 1987; Maurer & Sherrod, 1987; Smith & Hagan, 1984; Stoneman et al., 1987; Tannock, 1988), sometimes comparing these families to those with children experiencing other forms of mental retardation (e.g., Brooks-Gunn & Lewis, 1982; Strom, Wurster, & Rees, 1983). The following sections focus on

I78

Zolirido Stononrrri

issues associated with classification within groups of mentally retarded persons, examining their relevance for family mental retardation researchers. A.

Zigler’s Two-Group Model

A long-running debate was sparked by Zigler’s (1969; Zigler & Balla, 1982) contention that mentally retarded persons can be classified into two distinct groups, those with and without organic deficits. Support for this two-group model is derived from the greater-than-expected number of cases at the lower end of the distribution of intelligence. The overrepresented cases, causing the “bump” at the lower end of the distribution and believed to have organic causes, are placed into one group, and the remaining cases, believed to represent the expected lower end of the normal distribution of intelligence, are placed into a second group. A third undifferentiated group, cases who cannot be reliably classified, is also proposed (Zigler & Hodapp, 1986). Zigler argued that children whose mental retardation is due to nonorganic causes would have cognitive processes similar to nonretarded children, with only the rate of development differing, while the cognitive processes of children with clear organic etiologies might be expected to be idiosyncratic and nonnormative. Zigler’s two-group model has drawn criticism. It has been argued that persons with organic and nonorganic retardation are not two discrete groups and, even if they could be clearly divided, it is not necessary or desirable to postulate a separate set of explanatory constructs for children with mental retardation due to organic causes (Baumeister, 1984; Ellis, 1969; Scott & Carran, 1987; Sternberg & Spear, 1985). Ellis (1969) stressed that etiological heterogeneity does not imply behavioral heterogeneity, since diagnostic categories are not homogeneous. He further noted that only rarely have behavioral differences been found to be associated with specific etiological groups. Several authors have noted that new knowledge is rapidly being generated about mental retardation formerly viewed as having no identifiable cause, citing Fragile-X and fetal alcohol syndrome as two syndromes which had no known organic cause until quite recently (Landesman-Dwyer & Butterfield, 1983; Scott & Carran, 1987). Even Zigler admits definitional problems limit the extent to which individuals can be reliably assigned to his two groups, thus adding ambiguity to his model (Zigler & Hodapp, 1986). Recently, Zigler and colleagues have further elaborated their two-group model, arguing for fine-grained analysis of etiological subgroups (Burack. Hodapp, & Zigler, 1988; Zigler & Hodapp, 1986).

FAMILY RESEARCH A N D MENTAL RETARDATION

179

The intent here is not to enter into this debate, but, rather, to examine possible ramifications of tenets of the two-group model for the study of family process. Zigler has commented only briefly on differences in the family contexts experienced by children in his organic and nonorganic groups, suggesting that organically mentally retarded children have to cope with “not being like those around me at home,” while many of the children whose mental retardation is without organic cause must cope with stresses and deprivations which accompany poverty (Zigler et ul., 1984, p. 221). With the exception of these comments, however, Zigler provides little guidance as to how his model might be relevant for those interested in studying families. One of the few researchers to address directly the family issues involved in organic versus psychosocial (nonorganic) mental retardation has been Farber, who suggested that families of mildly and severely mentally retarded children were so different as to require distinct conceptual frameworks to guide the researcher (Farber, 1968; Farber & Rowitz, 1986). The appropriate focus of research on families of severely mentally retarded children, who tend to have clear organic causes for their retardation, was argued to be the impact of the child on family relationships, family role reorganization, community participation, and social mobility. Thus, Farber conceptualized the impact of an organically mentally retarded child in a manner similar to that currently favored by family researchers, namely, that the mentally retarded child represents an area of deviance in an otherwise normative family. Farber’s own research focused on elaborating the family impact of having an organically mentally retarded child (e.g., Farber, 1959, 1960; Farber & Jenne, 1963). The agenda which Farber described for researchers interested in mildly mentally retarded children, whose retardation is often familial or due to psychosocial causes, was quite different (Farber, 1968). For these families, often living in poverty, Farber perceived the important questions as centering on the family’s mobilization of resources to meet childcare burdens and other demands posed by the child’s limitations. Thus, while families of severely (organically) mentally retarded children face what Farber (1968) termed a “tragic crisis,” in which hopes and dreams go potentially unrealized, families of mildly (nonorganic) mentally retarded children face a “role reorganization crisis,” in which limited family resources must be marshaled to meet the demands of the situation (p. 155). Suggesting that mildly retarded children often come from unstable, insufficient, homes in which intellectual limitations are passed from one generation to another, Farber (1968) argued for a research agenda which would elucidate the mechanisms through which environmentally related mental

180

Zolindu Stonernun

retardation is created, or exacerbated, by the family context. Unfortunately, although these suggestions were made 20 years ago, they have had minimal impact in stimulating research on these issues.

B.

Mental Retardation Due to Psychosocial Disadvantage

Perhaps one of the greatest values for the family researcher of the debates over etiological subgroups is that they spotlight a large group of children whose mental retardation is intergenerational in nature, who live in a context of poverty and environmental deprivation, and whose families may, in some way, contribute to the intellectual limitations of their children. Understanding of these families is poorly served by existing family models. Persons with mental retardation due to psychosocial disadvantage comprise 70-75% of all mentally retarded persons, yet this group is often invisible in research on mental retardation (Zigler & Hodapp, 1986). In general, scientific advances have applied more to persons whose mental retardation arises from organic than nonorganic causes (Zigler et a l . , 1984). By definition, children with retardation due to psychosocial disadvantage (a large proportion of Zigler’s nonorganic group) have at least one parent and sibling (if siblings are present) who evidence subnormal intellectual functioning (Grossman, 1983). These families do not conform to the unidirectional family process models such as those which dominate mental retardation family research.

I . FAMILY MODELS OF PSYCHOSOCIAL DISADVANTAGE Although some still believe that psychosocial retardation is exclusively caused by heredity or by the environment, most now hold an interactionist view, with increased attention focused on malleability, the degree to which environmental modifications can influence development (Gallagher & Ramey, 1987; Weinberg, 1989). Ironically, Plomin (1989), who has contributed to the increased acceptance of genetic contributions to intelligence, now cautions that the influence of heredity has become so cornrnonly accepted that there is a risk that the important role of environmental factors may be overlooked. Zigler et (11. (1984) noted that, with a estimated 20-point reaction range for IQ scores, an optimally stimulating environment could raise a child’s IQ from the range of mild mental retardation to normal. There is a sizable body of research suggesting that the quality of the home environment can affect children’s development (see Gottfried, 1984; Nihira, Mink, & Meyers, 1984). A large share of what might loosely be called “family” research related to psychosocial disadvantage has

FAMILY RESEARCH AND MENTAL RETARDATION

181

arisen either out of a desire to document the impact of home environments on development (e.g., Bradley & Caldwell, 1984; Gottfried & Gottfried, 1984; Meyers. Nihira, & Mink, 1984) or to modify the family environment to decrease the risk of compromised development (e.g., Gray, Ramsey, & Klaus, 1982; Larnbie, Bond, Br Weikart, 1974). Few would disagree that families are important learning contexts for children or that deprived home environments can harm children’s development. Yet, little is known about the causative relationships between specific aspects of the family context and negative developmental outcomes, or about environmental factors that are critical at each developmental level (Nihira ef al., 1984). The general systems model proposed by Ramey and colleagues (Ramey, McPhee, & Yeates, 1982) is one of the few to provide a guiding conceptualization for the causative relationships between social disadvantage and compromised development. More recently, Ramey, (Ramey, Bryant, & Suarez, 1988) emphasized that interactions between the child and his or her environment have a cumulative influence on development. Another relevant conceptualization, developed by Whitman, Borkowski, Schellenbach, and Nath (l987), elaborates on Belsky’s (1984) model of the determinants of parenting, integrating social, psychological, and biological factors to explain the high incidence of mental retardation among offspring of adolescent mothers. Although both of these models include “family” factors (e.g.. care-giver teaching style), neither purports to be a family process model. Although these models represent movement toward more integrated approaches, no comprehensive family process model of psychosocial retardation currently exists. One factor which has been linked to psychosocial mental retardation is limited intellectual competence in a parent, usually the mother (e.g., Garber, 1988). Data collected during the initial survey phase of the Milwaukee Project found that among poor, high-risk families, mothers with IQs below 80 (45% of the sample) accounted for 78% of the children with LQs below 80; the lower a mother’s IQ, the greater probability that her child would also have limited intelligence (Garber, 1988; Heber & Dever, 1970). Recently, Ramey and Landesman (1988) offered data documenting that among children living in poverty whose mothers had IQs below 70, early intervention reduced children’s risk of mental retardation by a factor of 5.9; risk for other children reared in poverty homes was also reduced, but less dramatically. Unfortunately, the literatures on psychosocial mental retardation and on parenting by mentally retarded mothers have remained isolated from each other, with minimal overlap of issues or methodologies and few cross-citations of relevant studies (see, e.g., Thurman, 1985). Although

sensitive issues related to reproductive rights may play some part in keeping these literatures separate, integration is necessary. Limitations imposed by the decontextualization of the current definition of mental retardation, particularly as it relates to psychosocial retardation, have led to discussions concerning revised definitions of mental retardation. Landesman and Ramey ( 1989) have recommended abandoning the current AAMR definition in favor of a new classification system which would meld assessment of the individual and of the biosocial environment, obtaining measures of each over time to index rate of development. When faced with two same-aged children with similar competencies, one living in impoverished circumstances and one more advantaged, Landesman and Ramey 's approach would ascribe different meaning to the developmental lags of the two children depending on the stimulation value of their environments. Although there has been a recent upsurge in interest in families with mentally retarded members, it has taken place at a time when overall scholarly interest in mild, poverty-related mental retardation continues to be at a low ebb (Haywood, 1979, 1980). Much more work is needed to develop theoretical models explicating the transactional family processes associated with psychosocial mental retardation. McCall (1981) noted the need for the development of new methods and conceptualizations that integrate the contributions of genetics and environment to developmental change. Others (e.g., Borkowski & Turner, 1988) have suggested that there may be two types of mild retardation, one a result of environmental deprivation and the other representing polygenetic inheritance. When Mink and Nihira (1986) examined the direction of effects between family factors and development of educationally handicapped children, each family type studied had a different direction of effect, reinforcing the notion that causative processes may vary across families.

2. NONSHARED FAMILY ENVIRONMENTS Over 20 years ago, Farber (1968) posed a question as to why two siblings reared in the same poor, disadvantaged family often had very different levels of intellectual ability, with some siblings succumbing to mental retardation while others did not. Above and beyond the direct role of polygenetic inheritance and biologically mediated contributors, withinfamily influences probably account for a substantial portion of this variation. Behavioral geneticists have demonstrated that siblings evince greater differences than similarities in major developmental areas and that environments experienced by siblings living in the same home can differ as much as the environments experienced by two unrelated children living in different families (Plomin, 1989; Rowe & Plomin, 1981). Evidence

FAMILY RESEARCH A N D MENTAL RETARDATION

183

suggests that the most developmentally relevant variations in family environments do not occur between families, but, rather, are nonshared within-family variations which serve to make siblings different from each other (Plomin, 1986). Nonshared family influences can take numerous forms, including idiosyncratic events affecting only one sibling, birth order effects, differential caregiver attention or parent favoritism, or direct influences from a sibling. By selectively seeking out specific family contexts and activities (niche building), children create their own unique within-family environments (Scarr, 1982; Scarr & McCartney, 1983). Thus, siblings in the same deprived home may experience very different family environments. Exploration of within-family environments would seem to be a promising research direction to understand processes through which specific polygenetic predispositions of children and aspects of family environments interact to create, or exacerbate, psychosocial retardation. The study of nonshared family environments is a new area of family research which holds promise for understanding patterns of psychosocial retardation among siblings in disadvantaged families. C.

The New Morbidity

Haggerty, Roughman, and Pless (1975) coined the term new morbidity to refer to a constellation of negative child outcomes which are the end product of a variety of societal ills and which tend to occur among families living in poverty. More recently, Baumeister and colleagues (Baurneister & Kupstas, 1987; Baumeister, Dokecki, & Kuptas, 1988) expanded on this concept to provide a framework for understanding the interface between poverty and childhood disability. The new morbidity results from a transaction of environmental, behavioral, and biological forces, which act in concert, producing long-lasting effects which are cumulative and intergenerational (Baumeister & Kupstas, 1987). Citing Children’s Defense Fund data (1986), Baumeister and Kupstas (1987) concluded that children in poverty are I .5 to 2 times more likely to have a disability than those living in more fortunate circumstances. Associations between mental retardation and poverty are even stronger. Severe retardation, long thought to be equally distributed across social class, also seems to be more concentrated among poor families. Nichols (1984) documented racial differences in family patterns of mental retardation. For white families, siblings of mildly retarded children were 12 times more likely to be mentally retarded than were siblings of severely retarded children. For black families, however, sibling patterns were similar for families with mildly and severely retarded children: All

184

Zolindu Sronrmun

siblings were at risk for developmental disturbances. While 22% of American children live in poverty, 48% of African-American children live in poverty (Menolascino & Stark, 1988). It follows from these data that the sibling patterns identified by Nichols (1984) may be attributable more to poverty (new morbidity) than to race. Simple demographic data obtained from birth certificates has been demonstrated to be a reliable predictor of children's psychological and educational status as they enter school (Ramey, Stedman, Borders-Patterson, & Mengel, 1978). Infants at highest risk were later born, had an unmarried mother with minimal education who sought prenatal care late in pregnancy, had a live-born sibling now dead, and were black. Negative child outcomes are not directly caused by poverty, but, rather, are the result of multiple factors which tend to co-occur with poverty (e.g., poor nutrition, low birthweight). Recently, childhood AIDS (Crocker, 1989) and cocaine-addicted mothers have become additional major factors in the new morbidity. Rutter (1979) and others have demonstrated that concurrence of multiple risk factors dramatically increases probability of negative child outcomes; impoverished circumstances are characterized by a multiplicity of such risk factors. Negative environmental factors exacerbate preexisting biological vulnerabilities, creating less optimistic outcomes. Scott and Carran (1987) noted the lack of systematic information about the processes through which risks associated with poverty combine to produce mental retardation and other negative developmental outcomes. This knowledge is sorely needed. The new morbidity has profound ramifications for the family researcher. Affected children tend to live in families characterized by disorganization, multiple stressors, and severely limited resources (Baumeister & Kupstas, 1987). These families reside in substandard housing in neighborhoods where the crime rate is high and hope for the future is low. Researchers cannot attempt to understand these families without understanding the wider social system (Edgerton, 1988) and without crossing interdisciplinary boundaries that prevent the multiprofessional effort needed to address the new morbidity adequately (Rowitz. 1988b). These include the barriers that separate family researchers from the mainstream of mental retardation research. To date, family researchers have seldom focused on psychosocial retardation, preferring instead to study white, middle-class families coping with a child with Down syndrome, cerebral palsy, or other organically caused disability. It is unclear whether this results from a lack of awareness of the high proportion of mentally retarded persons whose mental retardation is associated with psychosocial deprivation, from the absence

FAMILY RESEARCH AND MENTAL RETARDATION

I85

of relevant conceptual models to guide family research efforts, from pragmatic difficulties inherent in studying these families, or whether it stems from other factors. Psychosocial retardation is an instance where the relative isolation of family mental retardation researchers has negatively affected theoretical and empirical advancement as compared to that which could occur if greater integration across research domains were more in evidence.

IV.

PERSONALITY, MOTIVATION, AND OTHER INDIVIDUAL DIFFERENCES

Like mental retardation researchers in general, family researchers have often ignored individual differences among persons labeled as mentally retarded. Incorrect implicit assumptions noted by others in critiques of the mental retardation literature include a belief that mentally retarded persons are more similar to each other than they are different (Haywood & Switzky, 1986), that the cognitive deficits of mentally retarded persons are so pervasive as to swamp all other sources of variance known to affect nonretarded persons (Zigler, 1969; Zigler & Burack, 1989), and that all atypical behavior of persons labeled as mentally retarded can be attributed to their cognitive deficits (Cohen & Bregman, 1988; Zigler, 1969). A.

Individual Differences in Personality and Motivation

Great variability exists among the personalities and motivational styles oimentally retarded persons (e.g., Haywood, 1989; Heal, 1970; Sternberg & Spear, 1985; Zigler & Hodapp, 1986). Yet, many mental retardation researchers and most family researchers have conceptualized these individuals as a homogeneous group, varying only in the degree of their intellectual deficits. Zigler and Hodapp (1986) termed this failure to view each mentally retarded individual as a unique person as a “common, but unforgivable oversight” (p. 115). The literature provides documentation that some mentally retarded individuals develop seemingly unhealthy motivational traits, such as overdependence on adult approval, fear of failure, and extrinsic motivational orientation (e.g., Edgerton, 1988; Haywood, 1989; Zigler & Hodapp, 1986). These traits may be more important determinants of behavior in everyday situations than cognitive or intellectual limitations (Zigler et al., 1984). Motivational factors may also interfere with the use of existing cognitive strategies, particularly on more difficult, challenging tasks (Borkowski & Turner, 1988).

186

Zolindci Stotiemcrn

The personalities and motivational styles of mentally retarded persons have been characterized as being quite sensitive to social and environmental influences (Haywood & Switzky, 1986). Haywood and colleagues (Haywood, 1989; Haywood & Switzky, 1986) have noted that mentally retarded children show performance deficits on cognitive tasks as compared with nonretarded peers of comparable mental ages. This discrepancy increases with age, suggesting that interactions between retarded children and persons in their immediate environments may accentuate, rather than eradicate, their learning problems (Haywood & Switzky. 1986). Edgerton ( 1988) suggested that overprotection by care-givers may deprive their charges of the opportunity to take risks and to sometimes fail (but also to sometimes succeed). Thus, mentally retarded persons may lack experience in confronting adversity and initial failure to secure eventual success, inhibiting the developmental of self-confidence. Haywood and Switzky (1986) speculate that parents’ responses to mastery behavior and exploration define success and failure for young children, resulting in task-intrinsic factors (e.g., challenge, the joy of learning becoming important for children who experience success, while task-extrinsic factors (e.g., adult approval) come to the fore for those who repeatedly receive messages from parents that they have failed. The attributions which parents make for the success and failure of mentally retarded children also may have an important impact on the children’s internalization of beliefs about their abilities (e.g., Borkowski & Turner, 1988). The success of nonretarded persons tends to be attributed to their effort, while success for mentally retarded persons is attributed to external factors, such as luck. Conversely, when mentally retarded people fail, they receive less blame than do nonhandicapped individuals, because their failure is expected (Gibbons, 1985; Gibbons, Sawin, & Gibbons, 1979). Gibbons et a/. (1979) suggested that not holding mentally retarded persons accountable when they fail, combined with not giving them credit for effort expended when they succeed, may reduce their internal sense of accomplishment and their motivation to try, thus decreasing the prospect of future success. These approaches share a general assumption that family members, particularly parents, play an important role in the development of children’s motivational styles, including motivational patterns developed by those who are mentally retarded. Unfortunately, this assumption is not founded on empirical evidence. Although motivational patterns have been of interest to mental retardation researchers for some time (e.g.. Haywood, 1968; Zigler, 1966), family researchers (who are in an excellent position to examine the development of motivational styles in the family context) have yet to demonstrate an interest in this important area.

FAMILY RESEARCH A N D MENTAL RETARDATION

I87

Sternberg and Spear (1985) suggested that seemingly negative motivational traits (e.g., dependence, low goal-setting, external motivation) might actually be adaptive for mentally retarded persons, given their limited competence and their need to shape environments (and others in their environment) to better meet their needs. Plausibly, the very motivational patterns assumed to be harmful and maladaptive may have been developed by mentally retarded individuals because these patterns are successful in manipulating the environment to achieve maximum adaptation and environmental fit. Naturalistic study of adaptation in the family context would clarify the functional utility of these behavior patterns and ascertain their impact on the family environment. I n this way, judgments about the adaptive or maladaptive nature of specific motivational styles could be made, based on their power to influence the family successfully to achieve a desired end. In addition to differences in motivational styles, researchers also have tended to overlooked personality and temperamental differences among those labeled as mentally retarded. Yet, our work, and the work of others, suggests that children’s temperamental characteristics can exert important influences on multiple aspects of family life, including marital happiness, family and sibling conflict, parent depression, and family emotional climate (e.g.. Brody, Stoneman, & Burke, 1987a, 1987b; Stoneman, Brody, & Burke, 1989). In the mental retardation literature, conflicting findings have emerged as researchers have attempted to determine whether temperaments of mentally retarded and nonretarded children differ or whether Down syndrome children have “easy” temperaments as compared to other disability groups (e.g., Greenberg & Field, 1982; Gunn & Berry, 1985a; Huntington & Simeonsson. 1987; Marcovitch, Goldberg, Lojkasek, & MacGregor, 1987; Marcovitch, Goldberg, MacGregor, & Lojkasek, 1986; Van Tassel, 1984). Those few researchers who have examined within-group differences have found that aspects of mentally retarded children’s temperaments are related to family functioning. Gunn and Berry (1985b), for example, found that young Down syndrome children show a range of temperamental styles and that children rated as having “easy” temperaments received more maternal teaching and fewer verbal negatives. In a study attempting to identify factors associated with stress in families of young children with mixed handicapping conditions, Beckman (1983)found temperamental difficulty and low social responsiveness to be more important factors than the child’s rate of developmental progress. Haywood and Switzky ( 1986) suggested that classic personality theories were not developed with mentally retarded populations in mind and may not fit these persons without modification. Standard assessment

188

Zolinda Stonernan

tools, developed on nonretarded populations, may not reflect the personality dimensions of retarded children which are most salient to their families. In general, neither mental retardation researchers nor family researchers have paid adequate attention to motivational and personality characteristics. Although the family context has been posited as a powerful influence on the motivational styles developed by mentally retarded children, the needed empirical research has not been conducted. Temperament has emerged as an area of interest in comparative studies of mental retardation, but few family researchers have broadened their focus to include personality and temperament factors in their research designs. Our knowledge about mental retardation and about families will be strengthened as these areas of study develop.

B.

Secondary Handicapping Conditions and Other Individual Differences

Persons labeled as mentally retarded differ from each other on numerous dimensions in addition to personality and motivation, including health problems, physical stigmata, motor and sensory disabilities, and the presence of secondary handicapping conditions. Persons having retardation attributable to organic causes are more likely to possess each of the above characteristics (Zigler et al., 1984) and to have multiple impairments (Landesman-Dwyer & Butterfield, 1983; Zigler & Hodapp, 1986). For some children, mental retardation is accompanied by life-threatening conditions. Families of such children must not only cope with mental retardation, but must also come to terms with fragility of life and fear of their child’s death. In family research samples, such as that of Berger and Cunningham (1981, 1983, 1986), it is not unusual for sample members, particularly infants, to die during the course of the study. Health problems force some families to cope with repeated hospitalization and surgery, painful medical procedures, and large medical expenses. Other families face reoccurring stresses of uncontrolled seizure disorders, entailing repeated emergency room visits and hospitalization. Additional demands placed on families by health factors can include tube-feeding, ventilator dependence, monitoring of shunts, susceptibility to infection, and severe allergies. It is almost impossible to study the effects of differing levels of mental retardation of families without addressing important confounding variables, such as the presence of multiple handicaps and chronic health problems in more severely retarded children. Burack et al. (1988) reported data suggesting that approximately half of all individuals with cere-

FAMILY RESEARCH AND MENTAL RETARDATION

189

bra1 palsy are also mentally retarded. Only a few studies (e.g., Hanzlik & Stevenson, 1986; Wasserman, Shilansky, & Hahn, 1986) have attempted to disentangle family effects attributable to mental retardation versus those attributable to physical handicaps. More family research which acknowledges the importance of secondary disabilities and associated health conditions is needed, as is vigilance among researchers for health and secondary handicapping conditions which act as confounding variables, making family studies open to multiple interpretations. Some individual characteristics covary with specific etiologies. Down syndrome infants, for example, tend to be hypotonic, resulting in delayed motor development and affective lags which make the child a less interactive partner in parent-child exchanges (Cicchetti & Pogge-Hesse, 1982).Other Down infants have heart defects which impede development and influence quality of the home environment (Barrera, Watson, & Adelstein, 1987). Studies comparing families of children with Down syndrome and those with other forms of mental retardation have seldom examined these potentially confounding differences between groups. Children with Fragile-X and certain other conditions tend to be high in activity and distractibility, factors with independently influence family functioning (e.g., Mash & Johnson, 1983a, 1983b; Stoneman, Brody, & Burke, 1989) but are seldom addressed by family mental retardation researchers. It has recently been recognized that mentally retarded persons are at high risk for a range of mental illnesses, including depression, schizophrenia, and severe behavioral disturbances (Cohen & Bregman, 1988; Cooke, 1988; Reid, 1989; Zigler & Burack, 1989). Persons with mental retardation are also at high risk for autism. Over 85% of all autistic persons are mentally retarded; autism occurs at a rate of 1:20,000 in the general population, but at a dramatically increased rate of 1:80 among those with mental retardation (Cohen & Bregman, 1988). Although there is a substantial literature addressing the impact of emotional illness and autism on families, a parallel literature addressing family members who have these diagnoses and are also mentally retarded does not exist. Given the high rates of these emotional disorders in the mentally retarded population, researchers have probably unwittingly included dually diagnosed children in many family mental retardation studies; it is plausible that at least some study findings have reflected the impact of emotional disturbances on families, rather than the impact of mental retardation. It is likely that mental illness and mental retardation are not simply additive in their effects on families, but, rather, interact multiplicatively in as yet unquantifiable ways.

190 C.

Zolinda Stoneman

Summary

The most significant point to be made here perhaps can be made best by paraphrasing Zigler and Hodapp (1986): One cannot attribute differences between families with and without mentally retarded members to the presence of mental retardation in one family member if the families or the mentally retarded individuals also differ in other ways that could plausibly be responsible for the group difference. Personality, motivational style, differential health status, secondary handicapping conditions, mental illness and autism, and other individual difference factors potentially serve as confounding variables, making clear attributions impossible. Farber (1968) noted that mental retardation tends to occur in combination with other characteristics which impair the ability of the individual to function in society (and, presumably, in the family as well). However, many presumed effects of mental retardation on families may reflect the influence of secondary conditions rather than mental retardation. Haywood and Switzky (1986) noted that, like talking about the weather, mental retardation researchers talk a lot about individual differences in motivation and personality, but do little in the way of actually studying them. Attending to differences among mentally retarded persons provides an important key to understanding family process variables, as well as to identify causative and mediating factors which are responsible for family outcomes. V. A.

CONCLUSIONS AND IMPLICATIONS

Causative Processes through Which Mental Retardation Affects Families

One important conclusion from this review is that by studying the causal relationships between mental retardation and the development of families with a mentally retarded member, researchers can make a signiticant contribution to our understanding of the condition of mental retardation. Haywood (1989) noted that mental retardation primarily manifests itself behaviorally; causative associations between mental retardation and family impact need to be related to the specific aspects of the family member’s behavior that is considered to be retarded (Brooks & Baumeister, 1977a). These behaviors are often social or adaptive in nature (Leland, 1973), but, as argued earlier, the direct impact of cognitive deficits on family functioning warrants closer examination. While it has been cautioned that cognitive deficits have been ascribed too much power in explaining aspects of the lives of mentally retarded people (Cohen & Breg-

FAMILY RESEARCH AND MENTAL RETARDATION

191

man, 1988; Zigler & Burack, 1989), the issue for family researchers is that cognition has been largely ignored, rather than overemphasized. Mentally retarded individuals differ in their pattern of abilities (Sternberg & Spear, 1985); there is a need to consider the differential and interactive effects of specific cognitive, social, and adaptive deficits on family functioning. There is also a need to look outside the mentally retarded individual to consider socially defined and mediated impacts which accrue to families. Wolfensberger (1972) conceptualized deviancy as a socially defined construct that categorizes an individual as different in some important way that has negative societal value. This social-relativist view suggests that mental retardation is created by society and exists primarily “in the eye of the beholder” rather than within the person. To date, few attempts have been made to examine the family impacts arising from society’s categorization of a family member as mentally retarded. The effects of labeling in education have been studied extensively (Horne, 1985; Jones, 1984; MacMillan, 1982), but the effects of labeling on the family have often been overlooked. Additionally, limited effort has been expended to understand how heredity and environment interact in the family context to minimize, or exacerbate, a child’s learning problems. The role of families in contributing to the mental retardation of children living in deprived circumstances is poorly understood, as are influences of nonshared family environments. Sternberg and Spear (1985) suggested that the complex interaction of genetics and environment probably differs from individual to individual, but this has yet to be theoretically modeled or empirically studied. The tendency of family researchers to study white, middle-class families with organically impaired children has minimized their ability to contribute to increased understanding of the causative processes that may underlie mental retardation due to psychosocial disadvantage. In a review of mental retardation cognitive research, Baumeister (1984) commented that researchers cannot hope to understand mental retardation unless they study mentally retarded persons. To paraphrase, it is not possible to understand the impact of mental retardation on families without studying families with mentally retarded members. Family researchers often have failed to describe study families adequately (see Stoneman, 1989), have studied heterogeneous families whose members have a variety of handicapping conditions (e.g., Bailey & Slee, 1984; Dyson & Fewell, 1986; Friedrich & Friedrich, 1981; Gallagher, Scharfman, & Bristol, 1984; Salisbury, 1987; Waisbren, 1980), or have used idiosyncratic definitions which are inconsistent with standard definitional criteria (e.g., McAllister, Butler, & Lei, 1973). These issues may be exacerbated by what Rowitz (l988a) has termed the homogenization of

192

Zolinda Stoneman

deviance, in which specific disabilities such as mental retardation are subsumed into a general classification of persons as developmentally disabled. Although an argument can be made for studying heterogeneous groups in order to examine the generality of family processes across handicapping conditions (Dunst, Trivette, & Cross, 1986), family research specifically addressing mental retardation is critically important. Burack et al. (1988) called for an increasingly precise science of mental retardation that would examine individual differences within specific etiologies, as well as differences between etiological groups. Yet, it is also legitimate to suggest, as did Ellis (1969), that findings that hold true across etiological groups have greater generality for the entire class of persons labeled as mentally retarded. Both generalizability and precision are important in deciphering which effects on families are specific to a certain etiology, which can be attributed to the more general phenomenon of mental retardation, and which exist across a wide range of handicapping conditions. It has been noted that few theoretical explanations of mental retardation exist (Detterman, 1987; Haywood, 1976). In their review of research, Haywood, Meyers, & Switzky (1982) stated that the most obvious conclusion to be drawn from their work was the need for theory specific to mental retardation. Most theories used by mental retardation researchers employ models and constructs derived from normal development (Haywood & Switzky, 1986). Family mental retardation researchers often have relied on theories developed by family sociologists and stress researchers. When these models have been adapted for use in studying mental retardation, the focus has tended to be on mediators of family coping, rather than on elucidating the processes through which specific behaviors or social influences associated with mental retardation impact family life (e.g., Byrne & Cunningham, 1985; Crnic, Friedrich, & Greenberg, 1983). These models have successfully guided research that has increased our understanding of stress and coping in families with mentally members. For family research to increase our understanding of the condition of mental retardation, however, models must be created that directly relate specific aspects of mental retardation (including societal influences) to family processes and outcomes across the life span. Such models must also account for the transactional influences (Sameroff & Chandler, 1975) between mentally retarded individuals and other family members, including reciprocal influences operating in homes characterized by intergenerational retardation due to psychosocial disadvantage. Although theoretical advancement is occurring at a rapid pace, few models which specifically

FAMILY RESEARCH AND MENTAL RETARDATION

193

link mental retardation with family process variables are currently available. B.

Understanding Mental Retardation through Family Research

A second conclusion is that family research brings many strengths to the study of mental retardation, but most of these strengths are yet to be fully developed. The social validity of research in naturally occurring contexts, such as the family home, is obvious and unquestioned. This is important because the ecological validity of many mental retardation studies is less than apparent. The importance of studying families, however, can obscure questions about whether family researchers are making a significant contribution to the understanding of mental retardation. Writing about cognitive mental retardation research, Brooks and Baumeister (1977a) asked, “How much greater is our understanding of mental retardation as a result of all this research activity?” (p. 407). Unfortunately, the recent explosion of interest in family research has advanced only minimally our understanding of mental retardation. To date, very few family researchers have focused their efforts on learning more about the phenomenon of mental retardation. The meaning and social validity inherent in studying families have captivated researchers, to the exclusion of research on more basic theoretical issues of interest to the field of mental retardation. Some might argue that such research is irrelevant, taking the field in the wrong direction, away from research with immediate applicability. Applied research is undeniably important, but so are studies that probe deeper, beyond the surface issues, to shed light on aspects of the phenomenon under study, namely, mental retardation. Fortunately, family researchers do not have to choose between social validity and theoretically relevant studies of mental retardation; the two approaches can complement, rather than compete with, each other. Research in general is plagued by what might be termed “scientific inertia,” the tendency for researchers to continue to study what they have studied in the past, using methods and techniques they have used before. To conduct research that advances our knowledge of mental retardation, family researchers must overcome this scientific inertia, working jointly with other mental retardation researchers to develop innovative approaches to achieve this goal. By so doing, the conceptual and methodological issues that divide family researchers from the mainstream of mental retardation research can be bridged, enabling family researchers to make more substantive contributions to knowledge about mental retardation and allowing family researchers to draw on theoretical and

194

Zolindu Stoneman

methodological advances in the general field of mental retardation to strengthen and enrich research on families. It is hoped that this article makes a contribution toward the achievement of these aims. ACKNOWLEDGMENT Preparation of this article was supported in part by Grant #04-DD-O(W-S8 from the Administration on Developmental Disabilities, U.S. Department of Health and Human Services. to the University AMiliated Program for Persons with Developmental Disabilities. The U niversity of Georgia.

REFERENCES Abramovitch, R., Stanhope, L., Pepler, D.. & Corter, C. (1987). The influence of Down's syndrome on sibling interaction. Jotrrnul of' Child Psychology und Psychitrtr? crnd Allied Disciplines, 28, 865-879. Bailey, L., & Slee. P. T. (1984). A comparison of play interactions between non-disabled and disabled children and their mothers: A question of style. A~.strtrlitr trnd N ~ w Zealund Jo~irnulof De~~elopmc~ntul Disabilities, 10, 5-10. Barnett. W. S. (1986). Definition and classification of mental retardation: A reply to Zigler. Balla, and Hodapp. American Jortrnul o f Mrntul Deficiency. 91, I 11-1 16. Barrera, M. E.. Watson. L. J., & Adelstein, A. (1987). Development of Down's syndrome infants with and without heart defects and changes in their caretaking environment. Child Care. Health. and Development. 13, 8 7 - 1 0 , Baumeister, A. A. (1967). Problems in comparative studies of mental retardates and normals. Americun Journal o f Mentul DiIficiencv, 71, 869-875. Baumeister, A. A. (1984). Some methodological and conceptual issues in the study of cognitive processes with retarded people. In P. H . Brooks, R. Sperber. & C . McCauley (Eds.), Learning and cognition in the mental1.v returded (pp. 1-38). Hillsdale, NJ: Erlbaum. Baumeister, A. A. (1987). Mental retardation: Some conceptions and dilemmas. Aniericun Psychologist. 44, 796-800. Baumeister, A. A., Dokecki, P. R..& Kupstas. F. D. (1988). Preventing the new rnorbidi/v: A guide for state planning for the prevention of mental retardation and reluted disuhilities associated with socioeconomic conditions. Nashville, TN: Vanderbilt University. Baumeister, A. A., & Kupstas, F. (1987). The new morbidity: Implications for prevention and amelioration. Presentation made to the Royal Society of Medicine, Social and environmental factors in the prevention and amelioration of mental handicup. Beckman, P. J. (1983). Influence of selected child characteristics on stress in families of handicapped infants. American Journal of Mental Deficiency, 88, 150-156. Belmont, J. M. ( 1989). Cognitive strategies and strategic learning. American Psychologist, 44, 142-148. Belsky, J. (1984). The determinants of parenting: A process model. Child Development, 55, 83-96. Berger, J., & Cunningham, C. C. (1981). The development of eye contact between mothers and normal vs. Down's syndrome infants. Developmental Psychology, 17, 678-689. Berger, J., & Cunningham, C. C. (1983). Development of early vocal behaviors and interac-

FAMILY RESEARCH AND MENTAL RETARDATION

195

tions in Down's syndrome and nonhandicapped infant-mother pairs. Developmental Psychology, 19, 322-331. Berger, J., & Cunningham, C. C. (1986). Aspects of early social smiling by infants with Down's syndrome. Child: Care, Health, and Development, 12, 13-24. Blacher, J . (1984). Severely handicapped young children and their families: Research in review,. Orlando, FL: Academic Press. Borkowski, J. G., &Turner, L. A. (1988). Cognitive development. In J. F. Kavanagh (Ed.), Understanding mental retardation: Research accomplishments and new frontiers (pp. 251-266). Baltimore, MD: Paul H. Brookes. Bradley, R. H., & Caldwell, B. M. (1984). 174 children: A study of the relationship between home environment and cognitive development during the first 5 years. In A. W. Gottfried (Ed.), Home environment and early cognitive environment (pp. 1-56). Orlando, FL: Academic Press. Brody, G. H., Stoneman, Z., & Burke, M. (1987a). Child temperaments, maternal differential behavior, and sibling relations. Developmental Psychology, 23, 354-362. Brody, G. H., Stoneman, Z., & Burke, M. (1987b). Family system and individual child correlates of sibling behavior. Americun Journal of Orthopsychiatry, 57, 561-569. Brooks, P. H., & Baumeister, A. A. (1977a). A plea for consideration of ecological validity in the experimental psychology of mental retardation: A guest editorial. American Journal of Mental Deficiency, 81, 407-416. Brooks, P. H., & Baumeister, A. A. (1977b). Are we making a science of missing the point? American Journal of Mental Deficiency. 81, 543-546. Brooks, P. H., McCauley, C. M., & Memll, E. M. (1988). Cognition and mental retardation. In F. J . Menolascino & J . A. Stark (Eds.), Preventive and curative intervention in mental returdation (pp. 295-318). Baltimore, MD: Paul H. Brookes. Brooks-Gunn, J . , & Lewis, M. (1982). Development of play behavior in handicapped and normal infants. Topics in Early Childhood Special Education, 2, 14-27. Buckhalt, J. A., Rutherford, R . B. & Goldberg, K. E. (1978). Verbal and nonverbal interaction of mothers with their Down's syndrome and nonretarded infants. American Journal of Mental Deficiency. 82, 337-343. Burack, J. A., Hodapp, R. M., & Zigler, E. (1988). Issues in the classification of mental retardation: Differentiating among organic etiologies. Journal of Clinical Psychology and Psychiatry, 29, 765-779. Byrne, E. A., & Cunningham, C. C. (1985). The effects of mentally handicapped children on families: A conceptual review. Journal of Child Psychology and Psychiatry and Allied Disciplines, 26, 847-864. Charlesworth, W. R. (1984). Ethology and sociobiology: Contributions t o the study of mental retardation. In P. H. Brookes. R. Sperber, & C. McCauley (Eds.). Learning and cognition in the mentally retarded (pp. 77-88). Hillsdale, N J : Erlbaum. Children's Defense Fund. (1986). Maternal and child health data book: The health of America's children. Washington, DC: Author. Cicchetti, D., & Pogge-Hesse, P. (1982). Possible contributions of the study of organically retarded persons to developmental theory. In E. Zigler & D. Balla (Eds.). Mental retardutionr The developmental-difference controversy (pp. 277-318). Hillsdale, N J : Erlbaum. Cohen. D. J . , & Bregman, J . D. (1988). Mental disorders and psychopharmacology of retarded persons. In J . F. Kavanagh (Ed.), Understanding mental retardation: Research accomplishments and new frontiers (pp. 319-329). Baltimore, MD: Paul H. Brookes. Cook, A. S ., & Culp, R. E. (1981). Mutual play of mothers with their Down's syndrome and normal infants. International Journal of Rehabilitation Research, 4, 542-544.

I96

Zolinda Stoneman

Cooke, R. E. (1988). Overview: Adulthood and old age. In J. F. Kavangh (Ed.), Understanding mental retardation: Research accomplishments and new frontiers (pp. 291295). Baltimore, MD: Paul H. Brookes. Crawley, S. B., & Spiker. D. (1983). Mother-child interactions involving two-year-olds with Down syndrome: A look at individual differences. Child Development, 54, 13 12-1323. Crnic, K. A., Friedrich, W. N.. & Greenberg, M. T. (1983). Adaptation of families with mentally retarded children: A model of stress, coping and family ecology. American Journal of Mental Deficiency, 88, 125-138. Crocker. A. C. (1989). Developmental disabilities and HIV infection: A symposium on issues and public policy. Mentul Relardation. 27, 197-198. Detterman, D. K. (1987). Theoretical notions of intelligence and mental retardation. American Journal of Mental Deficiency. 92, 2-1 1 1 . Dunst. C. (1984). Toward a social-ecological perspective of sensorimotor development among the mentally retarded. I n P. H. Brookes, R. Sperber, & C. McCauley (Eds.), Learning and cognition in the mentully retarded (pp. 359-387). Hillsdale, NJ: Erlbaum. Dunst, C., Trivette, C.. & Cross, A. (1986). Mediating influences of social support: Personal. family and child outcomes. American Journal of Mental Deficiency, 90, 403417. Dyson. L., & Fewell, R. R. (1986). Stress and adaptation in parents of young handicapped and nonhandicapped children: A comparative study. Journal of the Division for Early Childhood. 10, 25-34. Edgerton, R. B. (1967). The cloak of competence. Berkeley: University of California Press. Edgerton, R. B. (1988). Perspectives on the prevention of mild mental retardation. In F. J. Menolascino & J. A. Stark (Eds.). Preventive and curative intervention in mental retardation (pp. 325-342). Baltimore, MD: Paul H. Brookes. Edgerton, R. B.. & Edgerton, C. R . (1973). Becoming mentally retarded in a Hawaiian school. In G. Tarjan, R. K. Eyman, & C. Edward Meyers (Eds.), Sociohehavioral studies in mental retardation (pp. 21 1-233). Los Angeles: American Association on Mental Deficiency. Ellis, N. R. (1969). A behavioral research strategy in mental retardation: Defense and critique. American Journal of Mental Deficiency, 13, 557-566. Farber, B. (1959). Effects of a severely mentally retarded child on family integration. Monographs of the Society for Research in Child Development. 24(2, Serial No. 71). Farber, B. (1960). Family organization and crisis: Maintenance of integration in families with a severely mentally retarded child. Monographs of the Society for Research in Child Development, 25( I . Serial No. 75). Farber, B. ( 1968). Mental retardation: I t s social context and social consequences. Boston: Houghton Mimin. Farber, B., & Jenne. W. C. (1963). Family organization and parent-child communication: Parents and siblings of a retarded child. Monographs of the Society for Research in Child Development. U(7, Serial No. 91). Farber, B., & Rowitz, L. (1986). Families with a mentally retarded child. In N. R. Ellis & N. W. Bray (Eds.), International review of research in mental retardation (Vol. 14, pp. 201-224). Orlando, FL: Academic Press. Fewell, R. R.. & Vadasy, P. F. (1986). Families of handicapped children. Austin, TX: ProEd. Frankenberger, W.. & Harper, J. (1988). States’ definitions and procedures for identifying children with mental retardation: Comparison of 1981-1982 and 1985-1986 guidelines. Mental Retardation, 26, 133-136. Friedrich, W. N., & Friedrich, N (1981). Psychosocial assets of parents of handicapped and nonhandicapped children. American Journal of Mental Deficiency, 85, 55 1-553.

FAMILY RESEARCH AND MENTAL RETARDATION

197

Gallagher, J. J., & Ramey, C. T. (1987). The malleability of children. Baltimore, MD: Paul H. Brookes. Gallagher, J. J.. Scharfman, W., & Bristol, M. (1984). The division of responsibilities in families with preschool handicapped and nonhandicapped children. Journal of the Division for Eurly Childhood, 8, 3-12. Gallagher. J. J., & Vietze, P. M. (1986). Families of handicapped persons: Research, programs, and policy issues. Baltimore, MD: Paul H. Brookes. Garber, H. L. (1988). The Milwaukee Project: Preventing mental returdation in children at risk. Washington, DC: American Association on Mental Retardation. Gath, A. (1972). The mental health of siblings of a congenitally abnormal child. Journal of Child Psychology and Psychiatry, 13, 21 1-218. Gath, A. (1973). The school-age siblings of mongo1 children. British Journal ofPsychiatry, 123, 161-167. Gath. A,, & Gumley. D. (1984). Down’s syndrome and the family: Follow-up of children first seen in infancy. Developmental Medicine and Child Neurology. 26, 500-508. Gath, A , , & Gumley, D. (1986). Family background of children with Down’s syndrome and of children with a similar degree of mental retardation. British Journal of Psyrhiiitry. 149, 161-171. Gath, A,, & Gumley, D. (1987). Retarded children and their siblings. Journal of Child Psycholopv und Psychiatry and Allied Disciplines, 28, 715-730. Gibbons, F. X . (1985). A social-psychological perspective on developmental disabilities. Journirl of Soiiul und Clinicul Psychology, 4, 391404. Gibbons, F. X., Sawin, L. G., & Gibbons, B. N. (1979). Evaluations of mentally retarded persons: “Sympathy” or patronization? American Journal of Mental Deficiency. 2, 124-13 I . Gottfried, A. W. (1984). Home environment und ear1.v cognitive environment. Orlando, FL: Academic Press. Gottfried. A. W.. & Gottfried, A. E. (1984). Home environment and cognitive development in young children of middle-socioeconomic-status families. In A. W. Gottfried (Ed.), FL: Home environment und curly cognitive environment (pp. 57-1 IS). Orlando, FL: Academic Press. Gray, S. W., Ramsey, B. K.,& Klaus, R. A. (1982). From 3 to 20: The Early Truininp Pryject. Baltimore, MD: University Park Press. Greenberg, R., & Field, T. (1982). Temperament ratings of handicapped infants during classroom, mother. and teacher interactions. Journul of Pediatric Psychology, 7, 387405. Grossman. H. J . ( 1983). ClussiJiciition in mentul returdation. Washington, DC: American Association on Mental Deficiency. Gunn, P., & Berry, P. (1985a). The temperament of Down’s syndrome toddlers and their siblings. Journal of Child Psychology und Psychiutry, 26, 973-979. G u m , P., & Berry, P. (198Sb). Down’s syndrome temperament and maternal response to descriptions of child behavior. Development Psychology, 21, 842-847. Haggerty, R. J . , Roughman, K. J., & Pless, I. V. (1975). Child health und the community. New York: Wiley. Hanzlik, J.. & Stevenson, M. (1986). Interaction of mothers with their infants who are mentally retarded. retarded with cerebral palsy or nonrelarded. Americun Journal of Mentul Deficiency, 90, 513-520. Haywood, C. H. ( 1968). Motivational orientation of overachieving and underachieving elementary school children. Americiin Journal of Mental Dejkiency. 72, 662-667. Haywood, C. H . (1976). The ethics of doing research . . . and of not doing it. Americun Journal of Mental Deficienty, 81, 31 1-317.

198

Zolinda S t o n e m a n

Haywood, C. H. (1979). What happened to mild and moderate mental retardation’? American Journal of Mental Deficiency, 83, 42943 1. Haywood, C. H. (1980). Response to “On research with the lower levels of retardation.” American Journal of Mental Deficiency. 84, 405. Haywood, C. H. (1989). Multidimensional treatment of mental retardation. Psychology in Mental Retardation and Developmental Disabilities, 15, 1-10. Haywood, H. C., Meyers, C. E., & Switzky, H. N. (1982). Mental retardation. Annual Review of Psychology, 33, 309-342. Haywood, C. H., & Switzky, H. N. (1986). Intrinsic motivation and behavior effectiveness in retarded persons. In N . R. Ellis & N. W. Bray (Eds.), International review ofresearch in mental retardation (Vol. 14, pp. 2-46). Orlando, FL: Academic Press. Heal, L. W. (1970). Research strategies and research goals in the scientific study of the mentally subnormal. American Journal of Mental Deficiency, 75, 10-15. Heber, R. F. (l%l). Modifications in the manual on terminology and classification in mental retardation. American Journal of Mental Deficiency, 65, 499-500. Heber, R. F., & Dever, R. B. (1970). Research on education and habilitation of the mentally retarded. In C. H. Haywood (Ed.). Sorio-cultural aspects of mental retardation (pp. 395-427). New York: Appleton-Century-Crofts. Hodapp, R. M., & Zigler, E. (1986). Reply to Barnett’s comments on the definition and classification of mental retardation. American Journal of Mental Deficiency, 91, 117-1 19. Home, M. D. (1985). Attitudes toward handicapped students. Hillsdale, NJ: Erlbaum. House, B. J. (1977). Scientific explanation and ecological validity: A reply to Brooks and Baumeister. American Journal of Mental Deficiency, 81, 534-542. Huntington, G. S., & Simeonsson, R. J. (1987). Down’s syndrome and toddler temperament. Child Care, Health. and Development, 13, 1-11. Jacobsen, R. B., & Humphry, R. A. (1979). Families in crisis: Research and theory in child mental retardation. Social Casework: The Journal of Contemporary Social Work, 60, 597-60I . Janicki, M. P. (1988). The changing nature of the population of individuals with mental retardation: Historical artifacts and future trends. In J. F. Kavanagh (Ed.), Understanding mental retardation: Research accomplishments and new frontiers (pp. 297310). Baltimore, MD: Paul H. Brookes. Jones, R. L.(1984). Attitudes and attitude change in special eduration: Theory and practice. Reston, VA: Council for Exceptional Children. Katz, 1. (1981). Stigma: A social psychological analysis. Hillsdale, NJ: Erlbaum. Koegel, P., & Edgerton, R. B. (1982). Labeling and the perception of handicap among black mildly mentally retarded adults. American Journal of Mental Deficiency. 3, 266-276. Lambie, D. Z., Bond, J. T., & Weikart, D. P. (1974). Home teaching with mothers and infants: The Ypsilanti-Carnegie Infant Education Project- An experiment. Ypsilanti, MI: HigWScope Educational Research Foundation. Landesman, S., & Ramey. C. (1989). Developmental psychology and mental retardation. American Psychologist, 44, 409415. Landesman-Dwyer, S., & Butterfield, E. C. (1983). Mental retardation: Developmental issues in cognitive and social adaptation. In M. Lewis (Ed.), Origins of intelligence (2nd ed., pp. 479-519). New York: Plenum Press. Leland, H. (1973). Adaptive behavior and mentally retarded behavior. In G. Tajan, R. K. Eyman, & C. Edward Meyers (Eds.), Sociobehavioral studies in mental retardation (pp. 91-100). Los Angeles: American Association on Mental Deficiency. Lowitzer. A. C., Utley, C. A., & Baumester, A. A. (1987). AAMD’s 1983 Classification in mental retardation as utilized by state mental retardation/developmental disabilities agencies. Mental Retardation, 25, 287-291.

FAMILY RESEARCH AND MENTAL RETARDATION

199

MacMillan, D. L. (1982). Mental retardation in school and sociefy (2nd ed.). Boston: Little, Brown. Maheady, L., Algozzine, B., & Ysseldyke, J. E. (1984). Minority overrepresentation in. special education: A functional assessment perspective. Special Services in the Schools, 1, S-19.

Marcovitch, S., Goldberg, S., Lojkasek, M., & MacGregor, D. (1987). The concept of difficult temperament in the developmentally disabled preschool child. Journal of Applied Developmental Psychology, 8, 151-164.

Marcovitch. S.. Goldberg, S., McGregor, D. L., & Lojkasek, M. (1986). Patterns oftemperament variation in three groups of developmentally delayed preschool children: Mother and father ratings. Developmental and Behavioral Pediatrics, 7 , 247-252. Marfo, K. (1984). Interactions between mothers and their mentally retarded children: Integation of research findings. Journal of Applied Developmental Psychology, 5, 45-69.

Mash, E. J. &Johnson, C. (1983a). Parental perceptions of child behavior problems, parenting self-esteem, and mothers' reported stress in younger and older hyperactive and normal children. Journal of Counseling and Clinical Psychology, 51, 86-99. Mash, E. J., &Johnson, C. (1983b). The prediction of mothers' behavior with their hyperactive children during play and task situations. Child and Family Behavior Therapy, 5, 1-14.

Maurer, H., & Sherrod, K. B. (1987). Context of directives given to young children with Down syndrome and nonretarded children: Development over two years. American Journal of Mental Deficiency. 91, 576590.

McAllister, R.. Butler, E., & Lei, T. J. (1973). Patterns of social interaction among families of behaviorally retarded children. Journal of Marriage and the Family, 35, 93-100. McCall, R. B. (1981). Nature-nurture and the two realms of development: A proposed integration with respect to mental development. Child Development, 52, 1-12. McConachie, H. (1986). Parents and young mentally handicapped children: A review of research issues. Cambridge, MA: Brookline Books. McLaren, J . , & Bryson, S. E. (1987). Review of recent epidemiological studies of mental retardation: Prevalence, associated disorders, and etiology. American Journal of Mental Retardation. 92, 243-254.

Menolascino, F. J., & Stark, J. A. (1988). Introduction: Behavioral aspects of mental retardation. In F. J. Menolascino & J. A. Stark (Eds.), Preventive and curative intervention in mental retardation. Baltimore, MD: Paul H. Brookes. Mercer, J. (1973). The myth of 3% prevalence. In G. Tajan, R. K. Eyman, & C. Edward Meyers (Eds.), Sociobehavioral studies in mental retardation (pp. 1-8. Los Angeles: American Association on Mental Deficiency. Meyers, C. E. Nihira, K.. & Mink, I. T. (1984). Predicting retarded students' short-term growth from home environment. Applied Research in Mental Retardation, 5 , 137-146. Mink, 1. T., & Nihira, K. (1986). Family life-styles and child behaviors: A study of direction of effects. Developmental Psychology, 2, 610-616. Mimes, P. M. (1988). Family stress associated with a developmentally handicapped child. In N. W. Bray (Ed.). International review of research in mental retardation (Vol. 14, pp. 201-224). Orlando. FL: Academic Press. Mori. A. A. (1983). Families of children with special needs. Rockville. MD: Aspen. Natalicio. L. F. S. (1973). Dysfunctional reciprocal control in dyadidpolyadic interactions. In G. Tajan, R. K. Eyman, & C. Edward Meyers (Eds.), Sociabehavioral studies in mental returdation (pp. 117-122). Los Angeles: American Association on Mental Deficiency. National Institute of Child Health and Human Development. (1986). An evaluation and

200

Zolinda Stoneman

assessment of the state of the science: Mental retardation and development disabilities. Washington, DC: U.S. Department of Health and Human Services. Nichols, P. L. (1984). Familial mental retardation. Behavioral Genetics, 14, 161-170. Nihira, K. (1973). Importance of environmental demands in the measurement of adaptive behavior . In G. T a j a n , R. K. Eyman, & C. Edward Meyers (Eds.), Sociobehavioral strrdies in mental retardation (pp. 101-1 16). Los Angeles: American Association on Mental Deficiency. Nihira, K., Mink, I. T., & Meyers, C. E. (1984). Salient dimensions of home environment relevant to child development. In N. R. Ellis & N. W. Bray (Eds.), International review of research in mental retardation (Vol. 12, pp. 149-177). Orlando, FL: Academic Press. Plomin, R. (1986). Development. gene!ics. and psychology. Hillsdale, NJ: Erlbaum. Plomin, R. (1987). Behavioral genetics and intervention. In J. J. Gallagher & C. T. Ramey (Eds.), The malleability of children (pp. 15-24). Baltimore, MD: Paul H. Brookes. Plomin. R. (1989). Environment and genes. American Psychologist, 44, 105-1 1 I. President’s Committee on Mental Retardation. (l%9). The six-hour retarded child: A report on a conference on problems of education of children in the inner city. Washington. DC: Author. Ramey, C. T., Bryant, D. M., & Suarez, T . M. (1988). Early intervention: Why, for whom, how, at what cost? In J. F. Kavanagh (Ed.), Understanding mental retardation: Research accomplishments and new frontiers (pp. 277-288). Baltimore, MD: Paul H . Brookes. Ramey, C. T, & Landesman, S. (1988, March). Maternal retardation and children’s development. Paper presented at the Gatlinburg Conference on Research and Theory in Mental Retardation and Developmental Disabilities, Gatlinburg, TN. Ramey. C. T., McPhee, D., & Yeates, K. 0. (1982). Preventing developmental retardation: A general systems model. In L. Bond & J. Joffe (Eds.), Facilitating infant and eurlv childhood development (pp. 343401). Hanover, NH: University Press of New England. Ramey, C. T., Stedman, D. J., Borders-Patterson, A., & Mengel, W. (1978). Predicting school failure from information available at birth. American Jorrrnal of Mental Deficiencey , 82, 525-534. Reid, A. H. (1989). Schizophrenia in mental retardation: Clinical features. Research in Developmental Disabilities. 10, 241-249. Rowe. D. C.. & Plomin. R. (1981). The importance of nonshared (E,) environmental influences in behavioral development. Developmental Psychology. 17, 5 17-53 1. Rowitz, L. (1988a). Homogenization of deviance. Mental Retardation, 26, 1-3. Rowitz. L. (1988b). Multidisciplinary approaches to reversing mild mental retardation labels. In F. J . Menolascino & J. A. Stark (Eds.). Preventive and curative interifention in mental retardation (pp. 343-351). Baltimore, MD: Paul H. Brookes. Rutter, M. (1979). Maternal deprivation, 1972-1978: New findings, new concepts, new approaches. Child De\felopment, 50, 283-305. Salisbury, C. L. (1985). Internal consistency of the short-form of the Questionnaire on Resources and Stress. American Jortrnal of Mental Deficiency, 89, 610-616. Salisbury, C. L. (1987). Stressors of parents with young handicapped and nonhandicapped children. Jorrrnal of the Division f o r Early Childhood, 11, 154-160. Sameroff, A. J., &Chandler. M. J. (1975). Reproductive risk and the continuum of caretaker casuality. In F. D. Horowitz (Ed.), Review of child development research (Vol. 4, pp. 187-244). Chicago: University of Chicago Press. Scarr. S. (1982). On quantifying the intended effects of interventions: A proposed theory of

FAMILY RESEARCH AND MENTAL RETARDATION

20 1

the environment. In L. Bond & J. Joffe (Eds.), Facilitating infant and early childhood development (pp. 466484). Hanover, NH: University Press of New England. Scarr, S.. McCartney, K. (1983). How people make their own environments: A theory of genotype-environment effects. Child Development, 54, 424-435. Scheerenberger, R. C. (1983). A history of mental retardation. Baltimore, MD: Paul H. Brookes. Scott, K. G., & Carran, D. T. (1987). The epidemiology and prevention of mental retardation. American Psychologist. 42, 801-804. Seligman, M. (1983). The family with a handicapped child: Understanding and treatment. New York: Grune & Stratton. Seltzer, M. M. (1989, April). Aging families with mentally retarded children: Impact on lifelong caregiving. Invited paper presented at the University of Georgia, Athens. Shonkoff, J. P. (1982). Biological and social factors contributing to mild mental retardation. In K. A. Heller, W. H. Holtzman, & S. Messick (Eds.), Placing children in special education: A strategy for equity (pp. 133-181). Washington, DC: National Academy Press. Simmeonsson, R.J., Monson, L. B., & Blacher, J. (1984). Social understanding and mental retardation. In P. H. Brookes, R. Sperber, & C. McCauley (Eds.), Learning and cognition in the mentally retarded (pp. 389-417). Hillsdale, NJ: Erlbaum. Singer, H. S., & Irvin, L. K. (1989). Family caregiving, stress, and support. In H. S. Singer & L. K. lrvin (Eds.), Support for caregiving families. Baltimore, MD: Paul H. Brookes. Smith, L., & Hagen, V. (1984). Relationship between the home environment and sensorimotor development of Down syndrome and nonretarded infants. American Journal of Mentul Deficiency, 89, 124-132. Sparrow, S. S., Balla, D. A.. & Cicchetti, D. V. (1984). VinelandAdaptive Behavior Scales. Circle Pines, MN: American Guidance Service. Sternberg, R. J., & Spear, L. C. (1985). A triarchic theory of mental retardation. In N. R. Ellis & N. W. Bray (Eds.), International review of research in mental retardation (Vol. 13, pp. 301-326). Orlando. FL: Academic Press. Stoneman, Z. (1989). Comparison groups in research in families with mentally retarded members: A methodological and conceptual review. American Journal on Mental Returdution. 94, 195-2 15. Stoneman, Z., Broody, G. H., & Abbott, D. (1983). In-home observations of young Down syndrome children with their mothers and fathers. American Journal of Mental Deficiency, 87, 591-600. Stoneman. Z.. Brody. G. H., & Burke, M. (1989). Sibling temperaments and maternal and paternal perceptions of marital, family, and personal functioning. Journal of Marriage and the Fumilv. 51, 99-1 13. Stoneman, Z.. Brody, G. H., Davis, C. H., & Crapps, J. M. (1987). Mentally retarded children and their older same-sex siblings: Naturalistic in-home observations. Ameriran Journal of Mental Returdation, 92, 290-298. Stoneman. Z., Brody, G. H., Davis, C . H., & Crapps, J . M. (1988). Childcare responsibilities, peer relations, and sibling conflict: Older siblings of mentally retarded children. Americun Journul of Mentul Retardation, 93, 174-183. Stoneman. Z.. Brody, G. H., Davis, C. H., & Crapps, J . M. (1989). Role relations between mentally retarded children and their older siblings: Observations in three in-home contexts. Research in Developmentul Disabilities, 10, 61-76. Strom, R. D., Wurster, S., & Rees, R. J. (1983). Parents as teachers of their exceptional children. Journal of Instructionul Psychology, 10, 21-39.

202

Zolinda Stoneman

Tannock, R. (1988). Mothers’ directiveness in their interactions with their children with and without Down syndrome. American Journal of Mental Retardation. 93, 154-165. Thurman, S. K. (1985). Children of handicapped parents: Research and clinicul perspectives. Orlando, FL: Academic Press. Turnure, J. E. (1985). Communication and cues in the functional cognition of the mentally retarded. In N. R. Ellis & N. W. Bray (Eds.), International review of research in mental retardation (Vol. 13, pp. 43-78). Orlando, FL: Academic Press. Utley, C. A., Lowitzer, C., & Baumeister, A. A. (1987). A comparison of the AAMD’s definition, eligibility criteria, and classification schemes with State Departments of Education guidelines. Education and Training of the Mentally Retarded, 22, 35-43. Van Tassel, E. (1984). Temperament characteristics of mildly developmentally delayed infants. Developmental and Behavioral Pediatrics, 5 , 11-14. Waisbren, S. (1980). Parents’ reactions after the birth of a developmentally disabled child. American Journal of Mental Deficiency. 84, 345-351. Wasserman, G . A., Shilansky, M., & Hahn, H. (1986). A matter of degree: Maternal interaction with infants of varying levels of retardation. Child Study Journal. 16, 241-243. Weinberg, R. A. (1989). Intelligence and IQ. American Psychologist, 44, 98-104. Whitman, T. L., Borkowski, J . G., Schellenbach, C. J., & Nath, P. S. (1987). Predicting and understanding developmental delay of children of adolescent mothers: A multidimensional approach. American Journal of Mental Deficiency, 92,40-56. Wikler, L. M. (1986). Family stress theory and research on families of children with mental retardation. In J. J. Gallagher & P. M. Vietze (Eds.), Families of handicappedpersons: Research, programs, and policy issues (pp. 167-196). Baltimore, MD: Paul H. Brookes. Wolfensberger, W. (1972). The principle of normalization in human services. Toronto: National Institute on Mental Retardation. Yuker, H. E. (1988). Mothers’ perceptions of their disabled children: A review of the literature. Journal of the Multihandicapped Person, l, 217-232. Zigler, E. (1966). Research on personality structure in the retardate. In N. R. Ellis (Ed.), International review of research in mental retardation (Vol. I , pp. 77-108). New York: Academic Press. Zigler, E. (1%9). Developmental versus difference theories of mental retardation and the problem of motivation. American Journal of Mental Deficiency, 73, 536-555. Zigler, E., & Balla, D. (1982). Mental retardation: The developmental-dijjference controversy. Hillsdale, NJ: Erlbaum. Zigler, E., Balla, D., & Hodapp. R. M. (1984). On the definition and classification of mental retardation. American Journal of Mental Deficiency, 89, 215-230. Zigler, E., & Burack, J. A. (1989). Personality development and the dually diagnosed person. Research in Developmental Disabilities, 10, 225-240. Zigler, E., & Hodapp, R. M. (1986). Understanding mental retardation. London: Cambridge University Press.

Index

A

Abstract intelligence, social competence and, 127-129, 132, 145 Academic problem-solving, 96-98 Accuracy of inference, social competence and, 149, I50 Adaptation, family research and causative processes, 190. 192 conceptualization, 165. 166. 177 individual differences, 187 Adaptive behavior developmental delay and, 2 family research and, 171, 172, 176 social competence and, 125, 126, 133. 135. 136, 140 adolescents. 147, 149, 151, 152 Greenspan's model, 142, 145-147 Adaptive intelligence. social competence and, 126, 145. 146, 149, I50 Adolescents family research and, 168 social competence and, 147-152 Affective processes, family research and, 166, 189 Age family research and, 168, 169. 175-177. I86 neurotoxic risk factors and, 6. 17, 18 Aggression family research and, 176 problem-solving skills and, I14 social competence and, 136 Alcohol methylmercury toxicity and, 33, 34. 37. 42. 43. 47 neurotoxic risk factors and, 12. 13 Alerting interval, attentional resource allocation and, 66,67, 69 203

Alternative solutions, problem-solving skills and. 106, 108 American Association on Mental Retardation family research and, 182 social competence and, 125, 135, 136 adolescents, 151. 152 Greenspan's models, 146, 147 Analogy. problem-solvingskills and, 94,99101 Animal research methylmercury toxicity and, 45 neurotoxic risk factors and, 6, 25-28 Application, problem-solving skills and, 94, 99, 100 Attention, problem-solving skills and, 9092. 94 Attention deficits, neurotoxic risk factors and. 2. I5 Attentional resource allocation, 5 1-54, 8183 general framework, 60-63 hypothesis development, 54-60 methodology, 63-65 preliminary research, 65 automaticity, 78-81 cognitive processing, 65-69 cognitive resource requirements, 6973 individual differences, 73-78 Autism, family research and, 189, 190 Automatic processing, attentional resource allocation and, 57-61, 83 cognitive resource requirements, 70, 72, 73 development, 78-81 individual differences, 73-78 Autonomic nervous system, attentional resource allocation and, 53

204

INDEX

B Background knowledge. problem-solving skills and, 103, 104. 107 Behavior developmental delay and, 2 4 , 6. 8 lead, 16, 23, 24, 26, 28 Family research and. 170. 183. 186. 190 methylmercury toxicity and. 38 problem-solving skills and. 93 social competence and. 13C133 Biological plausibility. neurotoxic risk factors and. 10 Biological risk factors for developmental delay, I. 2, 7-9 dimension of risk. 4. 5 lead, 28 models. I I Biological systems model, neurotoxic risk factors and. 8. 9 Biology family research and. 176. 177, 181-184 methylmercury toxicity and, 35 problem-solving skills and, 92 Bivariate models, neurotoxic risk factors and, I2 Blind conditions. developmental delay and, 19

Blood level. methylmercury toxicity and, 36. 37. 40. 43, 44 Brain developmental delay and, 2 I , 26. 28 methylmercury toxicity and, 36, 38, 40. 45. 47 Breast-feeding, methylmercury toxicity and. 3 9 4 2

c Cadmium. developmental delay and, 2. 6, 18-22. 28 Canada. methylmercury toxicity and. 42. 43. 46 Capacity attentional resource allocation and, 52. 60. 61, 63, 64 problem-solving skills and, 93. 98. 100. 102. 116 social competence and. 139

Caregiving, neurotoxic risk factors and. 7. 8. 12. 25 Category. attentional resource allocation and automaticity. 74, 75. 78-80 hypothesis, 54-57 Causal relationship, neurotoxic risk factors and. 3 Causative processes. family research and, 190-193 classification. I 8 I , I82 conceptualization. 169, 176 Central nervous system methylmercury toxicity and, 33, 38. 39 neurotoxic risk factors and, 2. 6, 23. 24. 26-28 problem-solving skills and, 92 Cerebral palsy family research and, 184. 188. I89 methylmercury toxicity and, 34. 38. 41 Chronological age attentional resource allocation and. 65. 66 family research and, 168. 169. 172 Chunking hints. problem-solving skills and. 99 Clarification. problem-solving skills and, I I3 Classification, family research and. 177185, 192 Cloak of competence. family research and. I75 Cofxtors. neurotoxic risk factors and. 12, 18. 19

Cognitive deficits. family research and, 1 6 4 169, 185, 186, IN, 191 Cognitive problem-solving skills. 89-9 I . I 15-1 17 individual differences, 95-105 methods of assessment. 109-1 I I methods of fostering. 112-1 15 models. 92-95 social, 105, 106, 108, 109 Cognitive processes attentional resource allocation and. 5 I , 52. 54. 81-83 automaticity, 73, 78, 79 cognitive processing. 65-68 cognitive resource requirements. 69. 70. 72. 73 general framework. 60-62

INDEX

205

hypothesis development, 5 5 , 58. 59 methodology. 63, 64 developmental delay and. 2. 12. 13. IS lead. 21. 23, 26 family research and. 162. 178, 191 methylmercury toxicity and. 34 Cognitive rigidity. problem-solving skills and. 95 Cognitive skills, social competence and, 134, 136141 adolescents, 148. 149. IS1 Greenspan's model, 143, 144, 146 social intelligence. 127, 130 Coherence. neurotoxic risk factors and, 10. II

Cohort effects, family research and. 170 Comprehension, problem-solving skills and, 98. 107, 113. 114 Conceptual intelligence, social competence and. 145-149, 151 Conceptualization. family research and. 163, 164. 185. 193 adaptive behavior, 171, 172 classification, 179. 181, 182 cognitive deficit. 164-169 psychometric construct. 169- 17 1 socially defined construct, 172-177 Confounders family research and, 190 methylmercury toxicity and. 42. 43, 47 neurotoxic risk factors and, 6, 12, 18, 19, 21 Consistency, neurotoxic risk Factors and. 10. I I

Context. family research and, 193 classification, 180. 181. I83 conceptualization. 172, 174. 176 Coping family research and, 162, 188, 192 classification, 179, I84 conceptualization. 170 social competence and, 135, 136 Cree Indians. methylmercury toxicity and. 42, 43 Critical effect. neurotoxic risk factors and. 5 Cues problem-solving skills and. 99, 100 social competence and, 136. 139, 143. I44

Culture family research and, 171, 173. 177 problem-solving skills and. 92 Cumulative deficit hypothesis, 82 Cumulative incidence, neurotoxic risk factors and, 3

D Decision-making processes, attentional resource allocation and. 69, 70. 74, 75. 79. 82 Demographics, family research and, 184 Development family research and. 192 classification, 178, 180-182. 184 conceptualization. 166, 167, 176 individual differences, 187 social competence and, 132. 134-136. 146 Developmental delay methylmercury toxicity and, 47 neurotoxic risk factors for, see Neurotoxic risk factors Developmental differences. problemsolving skills and, 92, 95, 96 Deviance, family research and. 191. 192 Discrimination learning, family research and. 165 Dose-effect relationship, neurotoxic risk factors and, 5 , 17, 20-22 Dose-response relationship methylmercury toxicity and, 43 neurotoxic risk factors and, 5 , 6. 10 lead, 15-17, 22. 26, 27 Down's syndrome, family research and. 177. 184, 187. 189 Dual-task method, attentional resource allocation and, 61. 63, 64 Dynamic assessment, problem-solving skills and, 97, 109-1 I I , 115. I I7

E Early intervenlion, family research and. 181 Effect site, neurotoxic risk factors and, 5 Emotions family research and, 166, 168. 171 social competence and, 134, 145. 147

206

INDEX

Empathy, social competence and. 132, 137. 141

Familiarity, problem-solving skills and, 102, 103

Encephalopathy, neurotoxic risk factors and, 7. 15. 22, 23 Encoding attentional resource allocation and, 61, 65. 66. 68-74. 82 problem-solving skills and, 94, 99, 100. I08

Environment attentional resource allocation and, 52 family research and causative processes, 191 classification. 179-184 conceptualization, 171, 172. 175-177 individual differences, 186, 187, 189 methylmercury toxicity and. 33.35.39.47 neurotoxic risk factors and, 6,8.9, I I , 21, 28 lead, 16, 18, 25 pollution, I , 2. 14 social competence and. 135 Epidemiology, developmental delay and, 2, 3. 5

lead, 16, 17, 19 models, 8. 10 Established risk, neurotoxic risk factors and, 2 Experimental problem-solving, %. 98-101 Exposure methylmercury toxicity and, 34-36.38-44 neurotoxic risk factors and, 10, 14, 17-27 Exposure index methylmercury toxicity and, 37 neurotoxic risk factors and, 5. 16. 24 Exposure-effect relationships. developmental delay and. 15-17 External dose, neurotoxic risk factors and, 5. 16

Family research. 161-164. 193, 194 causative processes, 190-193 classification, 177, 178 new morbidity, 183-185 psychological disadvantages, 180- I83 Zigler’s two-group model, 178-180 conceptualization, 164 adaptive behavior, 171. 172 cognitive deficit, 164-169 psychometric construct, 169-17 I socially defined construct, 172-177 individual differences, 185-190 Feedback attentional resource allocation and, 52 problem-solving skills and. 100, 113-1 15 Fetal Alcohol Syndrome developmental delay and, 12. 13 family research and, 178 Fetus, methylmercury toxicity and, 33, 34. 47

clinical syndromes, 38. 39 exposure, 40.41. 43 sensitivity, 44, 45 Fish, methylmercury toxicity and, 34. 36, 4043.46-48

Fragile X-syndrome. family research and, 177, 178. 189

Frequency, methylmercury toxicity and, 42

G

Gender. family research and, 176. 177 General intelligence, social competence and, 130-132, 140 Generalization, problem-solving skills and. 114

Goals, problem-solving skills and, 90,91, I16

F Facilitation attentional resource allocation and, 57. 58, 65, 72

problem-solving skills and. 99. I I I Factor analysis, social competence and, 129-132. 150

cognitive, 94. 98, 104 social, 105 Greenspan, social competence and, 126, 148- I52 definition of mental retardation, 146. I47

personal competence, 144- I46 social intelligence, 142-144

207

INDEX

Guilford's model of intelligence, social competence and, 130, 131

H

Intuitive ability, social competence and, 129, 137 IQ family research and, 169-171. 175. 180. 181

Heredity, family research and. 180. 182, 191 Homogenization of deviance, family research and, 191, 192 Humor, comprehension of, problem-solving skills and. 107

neurotoxic risk factors and, 12, 19. 24, 27 social competence and, 125. 126. 145, 146 Iraq. methylmercury toxicity and, 36, 4043. 46-48 Irreversibility, developmental delay and. 24, 25. 28

I

J

Incidence rate, neurotoxic risk factors and, 3 Individual differences, family research and, 174, 185-190 Inference. problem-solving skills and. 94. 99, loo Information processing attentional resource allocation and, 53, 82. 83 automaticity, 74 cognitive processing. 65 cognitive resource requirements. 73 hypothesis development, 54. 56. 57 methodology, 64 problem-solving skills and, 92, 93, 103. 104 Inhibition. attentional resource allocation and, 58 Intelligence attentional resource allocation and. 53. 54, 59 family research and, 172, 176. 178 problem-solving skills and, 102. 103, 105,

Japan. methylmercury toxicity and. 39, 40. 45 Judgment family research and, 169 social competence and, 137, 138, 140. 147 comprehensive approach, 133 social intelligence, 128. 129

109

Interference. attentional resource allocation and, 63, 70. 72, 83 Internal dose. neurotoxic risk factors and. 5 , 16 Interpersonal competence, 126, 136. 139144, 147 adolescents, 149- I52 composite approaches, 136 social intelligence. 131, 132 Interpersonal problem solving. 139-141

K Kaufman Assessment Battery for Children, problem-solving skills and, 117 Knowledge acquisition, problem-solving skills and, 93-97

L Labeling, family research and, 173-177, 185, 191, 192 Language attentional resource allocation and, 74, 79 neurotoxic risk factors and, 8 problem-solving skills and, 92, 93, 110, I14 social competence and, 134, 139, 145, 149 Lead methylmercury toxicity and, 33. 34, 37 neurotoxic risk factors and, I . 14-28 Life cycle, family research and, 168 Logical structure, problem-solving skills and, %

208

INDEX

Long-term memory. attentional resource allocation and, 54, 56, 57

M Mapping. problem-solving skills and, 94.99, 100

Matching tasks, attentional resource allocation and. 57. 58 cognitive processing, 66. 68 cognitive resource requirements. 69, 72. 73 Mechanical intelligence. social competence and. 145 Memory family research and, 164-166 long-term, 54. 56, 57 problem-solving skills and, 90. I I I, I14 cognitive. 94. 96. 98. 99. 102, 104 search. 5 5 . 75 short-term, 54-57, 165 social competence and, 128. 134 Mental age. family research and, 168. 169. 172. 186 Mercury toxicity, see Methylmercury toxicity Metacognition attentional resource allocation and, 62, 83 family research and. 167 problem-solving skills and, 115. I16 Metacomponents, problem-solving skills and, 93-95. 97. I I I , I I7 Metals. methylmercury toxicity and, 34 Methylmercury toxicity, 33-35, 47, 48 clinical disorders, 43. 44 clinical syndromes. 37-39 exposure, 35, 36, 45-47 fetal sensitivity, 44. 45 measurement, 36, 37 outbreaks, 3 9 4 3 Microcephaly. methylmercury toxicity and. 38. 41 Minimata disease, methylmercury toxicity and. 39. 40. 45 Moderate laughter. problem-solving skills and. 107 Moral judgment, social competence and, 133. 149, 150

Greenspan’s models, 143, 144 social-cognitive approach, 137, 138 Motivation family research and. 167. 168. 185-188. 190 problem-solving skills and, 1 I I . 112. 116. I17 social competence and, 134. 137 Motor development, methylmercury toxicity and, 41. 42 Multiple regression models, neurotoxic risk factors and, 12, 13 Multivariate analysis, neurotoxic risk factors and. 19 Multivariate risk factor models. developmental delay and. 7

N Name identity, attentional resource allocation and. 67-73 Nervous system, methylmercury toxicity and, 34. 35. 47 Neurological abnormalities. methylmercury toxicity and, 41-44 Neurotoxic risk factors for developmental delay, I , 2, 28 cadmium, 14, 20-22, 28 dimensions of risk, 3-6 lead, 14. 19. 20 animal studies, 25-28 blind conditions, 19 confounding factors, 18, 19 exposure4Fect relationships, 15-1 7 health effects, 14, 15 internal exposure, 22-24 irreversibility, 24, 25 precision of measurement, 17, 18 selection bias, 18 statistics, 19 susceptibility, 25 literature. 6. 7 models, 7, 8 biosocial systems model, 8 new morbidity, 9. 10 probable causes. 10. I I transactional model, 8 statistical models bivariate models, 12

209

INDEX

multiple regression models. 12, 13 random effects models, 13, 14 structural equation models, 13 types of risk. 2, 3 New morbidity family research and, 183-185 neurotoxic risk factors and, 8-10 Niche building. family research and, 183

0

Organic causes. family research and, 178180, 184, 185. 191

Outcomes. neurotoxic risk factors and. 13

P Path diagram, neurotoxic risk factors and. 13

Pattern recognition. problem-solving skills and, 97-99 Performance deficits attentional resource allocation and. 51, 52. 62. 64 family research and, 186 Peripheral nervous system, neurotoxic risk factors and. 23, 24 Person perception, social competence and, 137. 138. 140. 144

Personal competence, 142. 144-147 Personality family research and. 185-190 social competence and. 126. 129, 138, 146. I47

Physical competence, 126. 134. 145. 147 Physical identity. attentional resource allocation and. 67-73 Piaget family research and. 166 problem-solving skills and, 92 Plasticity. neurotoxic risk factors and. 8. 9, 1 1 , 28

Poverty family research and. 179-184 neurotoxic risk factors and. 7-10 Practical intelligence. social competence and. 145-152

Practice attentional resource allocation and, 74, 75, 77-79. 81

problem-solving skills and, 98. I15 Precueing. problem-solving skills and. 99, 100

Predictors. neurotoxic risk factors and. 1113

Pregnancy methylmercury toxicity and, 33.34.37.38 exposure, 40, 42, 43, 47 neurotoxic risk factors and, 10 Problem definition, problem-solving skills and, 104. 105, 107 Problem orientation, problem-solving skills and. 104, 105 Problem-solving skills, 89-92, 116-1 18 cognitive academic problems. 96-98 experimental problems, 98-101 individual differences. 95. %, 103-105 models, 92-95 scientific problems. 101-103 current research, 115. 116 methods of assessment, 109-1 12 methods of fostering, 112-115 social individual differences, 105-109 models, 104, 105 Psychological ability, social competence and, 129. 130 Psychological insight. social competence and. 143. 144, 149. 150 Psychometric constructs. family research and. 169-171. 173 Psychosocial causes, family research and. 179-185, 191, 192

R Race, family research and. 176. 177. 183 Random effects model, neurotoxic risk factors and. 13, 14 Reaction time. attentional resource allocation and, 53. 54. 6S Reciprocal teaching. problem-solving skills and, 112-114 Reciprocal transactions. family research and. 168

210

INDEX

Referential communication. social competence and. 139. 143. 144. 150 Risk family research and. 177. 181, 184 methylmercury toxicity and. 47 Role taking. social competence and. 133. 134. 143. 149. 150 interpersonal competence, 141 social-cognitive approach, 137. 138 Roles Fdmily research and. 168 social competence and. 139. 141

S School context, family research and. 174176, 184 Scientific problem-solving, 101-103 Secondary handicapping conditions. family research and, 188-90 Selection, attentional resource allocation and, 52. 53 Selection bias methylmercury toxicity and. 47 neurotoxic risk factors and. 18 Self-help, social competence and, 135. 136 Self-instructional techniques, problemsolving skills and. 112. I13 Self-regulation. problem-solving skills and. 117, 118 Semantic processing, attentional resource allocation and. 54. 56-60. 72. 81. 83 Semantic structure. problem-solving skills and. 96, 97 Severity, neurotoxic risk factors and. 10, 1 I Sex roles. family research and, 168 Short-term memory attentional resource allocation and, 54-57 family research and, 165 Six Factor Tests of Social Intelligence. 131 Social ability, 128 Social awareness, 143. 146 Social behavior, family research and. 168 Social cognition. 137 Social communication. 143. 144, 149 Social competence, 125. 126. 132. 133, 141, I42 adolescents. 147-1 52 composite approaches, 134-136

comprehensive approaches. 133, 134 family research and. 171 Greenspan's model, 142 definition of mental retardation. 146. I47 personal competence. 144-146 social inlelligence. 142-144 interpersonal competence. 140. 141 social-cognitive approaches. 136-140 social intelligence. 126-132 Social comprehension. 137, 138, 143. 144. 149. 150 Social inferences. 138. 139. 143. 144, 150 Social insight, 143. 144. 149 Social intelligence. 126-132. 140. 141 adolescents, 148-152 Greenspan's model, 142-147 Social judgment. 143 Social maturity. 133. 135. 136 Social problem-solving skills, 91. 112-1 I5 individual differences. 105-109 models. 104. 105 social competence and. 139. 143. 144. I49 Social relativism. family research and 173175. 191 Social sensitivity. 143. 144. 149 Social skills, social competence and, 33, 135, 136. 141. 142 Social understanding. - 137 Socialization. social competence and, 135. 136 Socio-emotional adaptation, social competence and, 125, 126, 136, 146, I47 Socioeconomic status methylmercury toxicity and, 47 neurotoxic risk factors and, 7, 12 Socioinstructional method, problem-solving skills and, 113 Solution verification, problem-solving skills and. 104, 105 Specificity. neurotoxic risk factors and, 10. II

Static assessment, problem-solving skills and. 97. 109-111. 115, 117 Statistical models, neurotoxic risk factors and, 10-14 Statistical regression analysis, neurotoxic risk factors and, 19

21 I

INDEX

Stimulus attentional resource allocrtion and, 52. 61, 82 automaticity, 7678, 81 cognitive processing. 65-69 cognitive resource requirements, 69-73 hypothesis development, 5 5 , 56, 58 problem-solving skills and, 90. 100. 102 social competence and, 139 Stimulus onset asynchrony, attentional resource allocation and, 66-68. 70,71 Strategic flexibility. problem-solving skills and, 97-99. 104, 107, I08 Strategy attentional resource allocation and, 58, 59 problem-solving skills and, I 11-1 13, I IS, I I6 cognitive, 92, 93, 95. 98, 101. 102 social. 105. 106. 108 Stress, family research and. 163. 179, 184. 187, 188, 192 Structural equation model, neurotoxic risk factors and, 13 Structure-of-Intellect model, social competence and, 130, 131 Surplus Population Model, family research and. 173 Susceptibility, neurotoxic risk factors and, 4, 5. 25

Transactional processes, family research and, 182. 192 Transfer family research and. 165 problem-solving skills and, 99. 103. I12 Transfer of responsibility, problem-solving skills and, I13 Triarchic theory, problem-solving skills and. 92, 95 Triple Alliance Model. problem-solving skills and, 116. 117 Two-group model, family research and, 178-180

V Verbal ability. social competence and. 129131. 14%151 Vineland Adaptive Behavior Scales, family research and, 171 Vineland Social Maturity Scale, social competence and. 135 V ygot sk y family research and, 166 problem-solving skills and, 92, 93. 99

W T Task-intrinsic factors, family research and, I86 Temperament, family research and, 187, 188 Temporal sequence, neurotoxic risk factors and. 10. I1 Thinking aloud, problem-solving skills and. 101. 1 1 1 . 112, 114 Thorndike. social competence and, 127130. 142. 145. 146 Toxicity, methylmercury, see Methylmercury toxicity

Washingtion Test of Intelligence. social competence and, 127-1 29 Wechsler Intelligence Scale for Children problem-solving skills and. 109 social competence and, 148

Z

Zigler. family research and, 162. 190 classification, 178-1 80 conceptualization. 170, 175

This Page Intentionally Left Blank

Contents of Previous Volumes

Volume 1

The Role of Input Organization in the Learning and Memory of Mental Retardates HERMAN H. SPITZ

A Functional Analysis of Retarded Development SIDNEY W. BIJOU Classical Conditioning and Discrimination Learning Research with the Mentally Retarded LEONARD E. ROSS

Autonomic Nervous System Functions and Behavior: A Review of Experimental Studies with Mental Defectives RATHE KARRER

The Structure of Intellect in the Mental Retardate HARVEY F. DINGMAN AND C. EDWARD MEYERS

Learning and Transfer of Mediating Responses in Discriminative Learning BRYAN E. SHEPP AND FRANK D. TURRISI

Research on Personality Structure in the Retardate EDWARD ZIGLER

A Review of Research on Learning Sets and Transfer of Training in Mental Defectives MELVIN E. KAUFMAN AND HERBERT J. PREHM

Experience and the Development of Adaptive Behavior H. CARL HAYWOOD AND JACK T. TAPP A Research Program on the Psychological Effects of Brain Lesions in Human Beings RALPH M. REITAN Long-Term Memory in Mental Retardation JOHN M. BELMONT The Behavior of Moderately and Severely Retarded Persons JOSEPH E. SPRADLIN AND FREDERIC L. GIRARDEAU Author Index-Subject Index

Volume 2 A Theoretical Analysis and Its Application to Training the Mentally Retarded M. RAY DENNY

213

Programming Perception and Learning for Retarded Children MURRAY SlDMAN AND LAWRENCE T. STODDARD Programming Instruction Techniques for the Mentally Retarded FRANCES M. GREENE Some Aspects of the Research on Mental Retardation in Norway IVAR ARNLJOT BJORGEN Research on Mental Deficiency During the Last Decade in France R. LAFON AND J. CHABANIER Psychotherapeutic Procedures with the Retarded MANNY STERNLICHT Author Index-Subject Index

214 Volume 3 Incentive Motivation in the Mental Retardate PAUL S. SIEGEL Development of Lateral and ChoiceSequence Preferences IRMA R. GERJUOY AND JOHN J . WINTERS, JR. Studies in the Experimental Development of Left-Right Concepts in Retarded Children Using Fading Techniques SIDNEY W. BIJOU Verbal Learning and Memory Research with Retardates: An Attempt to Assess Developmental Trends L. R. GOULET Research and Theory in Short-Term Memory KEITH G. SCOTT AND MARCIA STRONG SCOTT Reaction Time and Mental Retardation ALFRED A. BAUMEISTER AND GEORGEKELLAS Mental Retardation in India: A Review of Care. Training. Research, and Rehabilitation Programs J. P. DAS Educational Research in Mental Retardation SAMUEL L. GUSKIN AND HOWARD H. SPICKER Author Index-Subject Index

Volume 4 Memory Processes in Retardates and Normals NORMAN R. ELLIS A Theory of Primary and Secondary Familial Mental Retardation ARTHUR R. JENSEN Inhibition Deficits in Retardate Learning and Attention LAIRD W. HEAL AND JOHN T . JOHNSON. JR.

CONTENTS OF PREVIOUS VOLUMES

Growth and Decline of Retardate Intelligence MARY ANN FISHER AND DAVID ZEAMAN The Measurements of Intelligence A. B. SILVERSTEIN Social Psychology and Mental Retardation WARNER WILSON Mental Retardation in Animals GILBERT W. MElER Audiologic Aspects of Mental Retardation LYLE L. LLOYD Author Index-Subject Index

Volume 5 Medical-Behavioral Research in Retardation JOHN M. BELMONT Recognition Memory: A Research Strategy and a Summary of lnitial Findings KEITH G. SCOTT Operant Procedures with the Retardate: An Overview of Laboratory Research PAUL WEISBERG Methodology of Psychopharmacological Studies with the Retarded ROBERT L. SPRAGUE AND JOHN S. WERRY Process Variables in the Paired-Associate Learning of Retardates ALFRED A. BAUMEISTER AND GEORGE KELLAS Sequential Dot Presentation Measures of Stimulus Trace in Retardates and Normals EDWARD A. HOLDEN. JR. Cultural-Familial Retardation FREDERIC L. GIRARDEAU German Theory and Research on Mental Retardation: Emphasis on Structure LOTHAR R. SCHMIDT AND PAUL B. BALTES Author Index-Subject Index

215

CONTENTS OF PREVIOUS VOLUMES

Volume 6

Volume 8

Cultural Deprivation and Cognitive Competence J. P. DAS

Self-Injurious Behavior ALFRED A. BAUMEISTER AND JOHN PAUL ROLLINGS

Stereotyped Acts ALFRED A. BAUMEISTER AND REX FORE H AN D

Toward a Relative Psychology of Mental Retardation with Special Emphasis on Evolution HERMAN H. SPITZ

Research on the Vocational Habilitation of the Retarded: The Present. the Future MARC W. GOLD Consolidating Facts into the Schematized Learning and Memory System of Educable Retardates HERMAN H. SPITZ An Attention-Retention Theory of Retardate Discrimination Learning MARY ANN FISHER AND DAVID ZEA M A N Studying the Relationship of Task Performance to the Variables of Chronological Age, Mental Age, and IQ WILLIAM E. KAPPAUF Author Index-Subject Index

Volume 7 Mediational Processes in the Retarded JOHN G. BORKOWSKI AND PATRICIA B. WANSCHURA The Role of Strategic Behavior in Retardate Memory ANN L. BROWN Conservation Research with the Mentally Retarded KERl M. WILTON AND FREDERIC J. BOERSMA Placement of the Retarded in the Community: Prognosis and Outcome RONALD B. McCARVER AND ELLIS M. CRAIG Physical and Motor Development of Retarded Persons ROBERT H. BRUININKS Subject Index

The Role of the Social Agent in Language Acquisition: Implications for Language Intervention GERALD J . MAHONEY AND PAMELA B . SEELY Cognitive Theory and Mental Development EARL C . BUTTERFIELD AND DONALD J . DICKERSON A Decade of Experimental Research in

Mental Retardation in India ARUN K. SEN The Conditioning of Skeletal and Autonomic Responses: NormalRetardate Stimulus Trace Differences SUSAN M. ROSS AND LEONARD E ROSS Malnutrition and Cognitive Functioning J. P. DAS AND EMMA PIVATO Research on Efficacy of Special Education for the Mentally Retarded MELVINE E. KAUFMAN AND PAUL A. ALBERT0 Subject Index

Volume 9 The Processing of Information from ShortTerm Visual Store: Developmental and Intellectual Differences LEONARD E. ROSS AND THOMAS B. WARD Information Processing in Mentally Retarded Individuals KEITH E. STANOVICH Mediational Processes in the Retarded: Implications for Teaching Reading CLESSEN J. MARTIN

216

CONTENTS OF PREVIOUS VOLUMES

Psychophysiology in Mental Retardation J. CLAUSEN

Rumination NIRBHAY N . SINGH

Theoretical and Empirical Strategies for the Study of the Labeling of Mentally Retarded Persons SAMUEL L. GUSKIN

Subject Index

The Biological Basis of an Ethic for Mental Retardation ROBERT L. ISAACSON AND CAROL VAN HARTESVELDT Public Residential Services for the Mentally Retarded R. C. SCHEERENBERGER Research on Community Residential Alternatives for the Mentally Retarded LAIRD W. HEAL, CAROL K. SIGELMAN, AND HARVEY N. SWITZKY

Volume 11

Cognitive Development of the LearningDisabled Child JOHN W. HAGEN, CRAIG R. BARCLAY. AND BETTINA SCH WETHELM Individual Differences in Short-Term Memory RONALD L. COHEN Inhibition and Individual Differences in Inhibitory Processes in Retarded Children PETER L. C . EVANS

Mainstreaming Mentally Retarded Children: A Review of Research LOUISE CORMAN AND JAY COTTLIEB

Stereotyped Mannerisms in Mentally Retarded Persons: Animal Models and Theoretical Analyses MARK H. LEWIS AND ALFRED A. BAUMEISTER

Savants: Mentally Retarded Individuals with Special Skills A. LEWIS HILL

An Investigation of Automated Methods for Teaching Severely Retarded Individuals LAWRENCE T. STODDARD

Subject Index

Social Reinforcement of the Work Behavior of Retarded and Nonretarded Persons LEONIA K. WATERS

Volume 10

The Visual Scanning and Fixation Behavior of the Retarded LEONARD E. ROSS AND SUSAN M. ROSS Visual Pattern Detection and Recognition Memory in Children with Profound Mental Retardation PATRICIA ANN SHEPHERD AND JOSEPH F. FAGAN 111 Studies of Mild Mental Retardation and Timed Performance T. NETTLEBECK AND N. BREWER Motor Function in Down’s Syndrome FERIHA ANWAR

Social Competence and Interpersonal Relations between Retarded and Nonretarded Children ANGELA R. TAYLOR The Functional Analysis of Imitation WILLIAM R. McCULLER AND CHARLES L. SALZBERG Index

Volume 12

An Overview of the Social Policy of Deinstitutionalization BARRY WILLER AND JAMES INTAGLIATA

217

CONTENTS OF PREVIOUS VOLUMES

Community Attitudes toward Community Placement of Mentally Retarded Persons CYNTHIA OKOLO AND S A M U E L GUSKIN Family Attitudes toward Deinstitutionalization AYSHA LATIB, JAMES CONROY. AND CARLA M. HESS Community Placement and Adjustment of Deinstitutionalized Clients: Issues and Findings ELLIS M. CRAIG A N D RONALD B. McCARVER Issues in Adjustment of Mentally Retarded Individuals to Residential Relocation TAMAR HELLER Salient Dimensions of Home Environment Relevant to Child Development KAZUO NIHIRA. IRIS TAN MINK, A N D C. EDWARD MEYERS Current Trends and Changes in Institutions for the Mentally Retarded R. K. EYMAN. S. A. BORTHWICK. A N D G . TARJAN Methodological Considerations in Research on Residential Alternatives for Developmentally Disabled Persons LAIRD W. H E A L A N D G L E N N T. FUJI U RA A Systems Theory Approach to Deinstitutionalization Policies and Research ANGELA A. NOVAK AND TERRY R. BERKELEY Autonomy and Adaptability in Work Behavior of Retarded Clients JOHN L. GIFFORD. FRANK R. RUSCH. JAMES E. MARTIN, AND DAVID M. WHITE Index

Volume 13

Communication and Cues in the Functional Cognition of the Mentally Retarded JAMES E. T U R N U R E Metamemory: An Aspect of Metacognition in the Mentally Retarded ELAINE M. JUSTICE Inspection Time and Mild Mental Retardation T. NETTELBECK Mild Mental Retardation and Memory Scanning C. J. PHILLIPS A N D T. NETTELBECK Cognitive Determinants of Reading in Mentally Retarded Individuals KEITH E. STANOVICH Comprehension and Mental Retardation LINDA HICKSON BILSKY Semantic Processing, Semantic Memory, and Recall LARAINE MASTERS GLIDDEN Proactive Inhibition in Retarded Persons: Some Clues to Short-Term Memory Processing JOHN J. WINTERS, JR. A Triarchic Theory of Mental Retardation ROBERT J. STERNBERG A N D LOUISE C . SPEAR Index

Volume 14

Intrinsic Motivation and Behavior Effectiveness in Retarded Persons H. CARL HAYWOOD A N D HARVEY N . S WlTZKY The Rehearsal Deficit Hypothesis NORMAN W. BRAY A N D LISA A. TURNER

Molar Variability and the Mentally Retarded Sustained Attention in the Mentally STUART A. SMITH A N D PAUL S. Retarded: The Vigilance Paradigm JOEL B. WARM AND DANIELB. BERCH SIEGEL

218

CONTENTS OF PREVIOUS VOLUMES

Computer-Assisted Instruction for the Mentally Retarded FRANCES A . CONNERS. DAVID K. CARUSO, A N D DOUGLAS K. DETTERMAN

Volume 15

Procedures and Parameters of Errorless Discrimination Training with Developmentally Impaired Individual5 GIUI.IO E. 1,ANCIONI A N D PAU12 M. SM EEI'S

Developmental Impact of Nutrition on Pregnancy. Infancy, and Childhood. Public Health I \ w e \ in the United State\ ERN ESIO POI~L.I'T'1

Reading Acquisition and Remediation in the Mentally Retarded NIRBHAY N . SINGH A N D J U D Y S I N G H

T h e Cognitive Approach to Motivation in Retarded Individual5 S H U L A M I T H KREITLER A N D H A N S KKEITLER

Families with a Mentally Retarded Child BERNARD FARBER AND LOUIS ROWITZ Social Competence and Employment of Retarded Perwnb CHARI.ES I.. SAI.ZBERG, MARILYN LIKINS. E. KATHRYN McCONAUGHY. A N D BENJAMIN LIGNUGARIS/KRAFI Toward ;I Taxonomy o f llome Environnients SHARON L.ANI>ESMAN Behavioral Treatment of the Sexually Deviant Behavior of Mentally Retarded Individuals R . M. FOXX. R . G . B1TTI.E. D. R . BECHTEL. AND J . R. LIVESAY

Mental Retardation a s a Thinking Disorder: The Rationalist Alternative to Empiricism HERMAN H . S P l T Z

Mental Retardation. Analogical Reawning. and the Componential Method J . McCON AG H Y Application of Self-Control Strategies t o Facilitate Independence in Vocational and Instructional Settings J A M E S E. MARTIN. I ~ O N A L I )I.. BURGER. SUSAN ELIAS-BURGER. A N D D E N N I S E. MITHAUG Family Stress Associated with a Developmentally Handicapped Child PATRICIA M. M I N N E S

Behavioral Approaches to Toilet Training for Retarded Persons S. BETTISON

Physical Fitness of Mentally Retarded Individuals E. KATHRYN McCONAUGHY A N D C H A R L E S L. SALZBERG

lndex

lndex