International Review of Research in Mental Retardation, Volume 18: Volume 18 (International Review of Research in Mental Retardation)

International Review of RESEARCH IN MENTAL RETARDATION VOLUME 18

Consulting Editors Ann M. Clarke T H E UNIVERSITY OF HULL

J. P. Das T H E UNIVERSITY OF ALBERTA

H. Carl Haywood VA N D E R B ILT U N I V E R S I T Y

Ted Nettelbeck T H E U N I V E R S I T Y OF A D E L A I D E

International Review of RESEARCH IN MENTAL RETARDATION

EDITED BY

NORMAN W. BRAY CIVITAN INTERNATIONAL RESEARCH CENTER AND DEPARTMENT OF PSYCHOLOGY THE UNIVERSITY OF ALABAMA AT BIRMINGHAM BIRMINGHAM, ALABAMA

VOLUME 18

ACADEMIC PRESS, INC. Harcourt Brace Jovanovich, Publishers

San Diego New York Boston London Sydney Tokyo Toronto

This book is printed on acid-free paper. @

Copyright 0 1992 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. 1250 Sixth Avenue, San Diego, California 92101-4311 United Kingdom Edition published by Academic Press Limited 24-28 Oval Road, London NW1 7DX Library of Congress Catalog Number: 65-28627 International Standard Book Number: 0- 12-366218-4 PRINTED IN THE UNITED STATES OF AMERICA 929394959697

QW

9 8 7 6 5 4 3 2 1

Contents

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

ix

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi

Contributors

Perceptual Deficits in Mildly Mentally Retarded Adults Robert Fox and Stephen Oross 111

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

I . Introduction

rams ..................... Random Element Kinematograms . . ................................. Ecological Significance of Random Element Stimuli . . . . . . . . . . . Basic Research on Random Element Stimuli ................................ General Methodological Considerations .............................. Research Progress . . . . .......................... General Discussion . . . . References ..................... ................

1

11. Random Elem

111. IV. V. VI. VII. VIII.

5

9

15 25

Stimulus Organization and Relational Learning Sal A. Soraci, Jr., and Michael T. Carlin I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

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

11. Stimulus Organization . . . . . . . . .

Ill. IV. V. V1.

Approaches Emphasizing Early St Neurophysiological Correlates . . . . . . . . . . . . . . . . . . . The Connectionist Perspective and Relational Learning ....................... Conclusion .............................. References .....................

29 32

48

Stimulus Control Analysis and Nonverbal Instructional Methods for People with Intellectual Disabilities William J. Mcllvane 1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Overview of Analytical Framework ....................................... 111. Major Research Issues and Findings ....................................... V

55

58 62

vi

CONTENTS

IV. General Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

101

I02

Sustained Attention in Mentally Retarded Individuals Phillip D. Tomporowski and Lisa D. Hager 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............. ...

11. Methodological Issues . . 111. Vigilance in Mentally Ret

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

IV. Theories of Vigilance and ........................ References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Ill 112 1 I5 I27 131

How Modifiable Is the Human Life Path? Ann M. Clarke and Alan D. B. Clarke I . Introduction

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

111. Recovery from Perinatal Problems

IV. V. VI. VII.

Adjustment Difficulties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Escape from Disadvantage . ............................. Discussion . . . . . . . . . . . . . . ........................ Conclusions . . . . . . . . . . . . ........................ References . . . ...............................................

I37 142 I44 146 149 I52 I54 I55

Unraveling the “New Morbidity”: Adolescent Parenting and Developmental Delays John C.Borkowski. Thomas L. Whitman, Anne Wurtz Passino, Elizabeth A. Rellinger, Kristen Sommer, Deborah Keogh, and Keri Weed

...

I59

11. Adolescent Parenting: High Risks for Mothers and Their Children . . . . . . . . . . . . . . 111. Influence of Environmental and Personal Factors . . . . . . . . ...

I60 I67

IV. Componential Analyses: The Relative Importance of Cognitive Readiness . . . . . . . . ... V. Influence of Early Attachment on Later Cognitive Development . . . . . VI. Unraveling the “New Morbidity” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . ........................

182 1nx I90

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

181

Longitudinal Research in Down Syndrome Janet Can

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Nonintervention Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Intervention Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I97 I98 199

vii

CONTENTS IV . The Surrey Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

201 218 221

Staff Training and Management for Intellectual Disability Services Chris Cullen I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1 . A Model of Staff Training and Management ................................ 111. Concluding Remarks: Social Validity ...................................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

225 227 240 242

Quality of Life of People with Developmental Disabilities Trevor R . Parmenter

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

........ Approaches to the Conceptualization of Quality of Life ....................... Theoretical Models of Quality of Life in the Area of Developmental Disability . . . . Recent Empirical Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methodological Issues in Measuring the Quality of Life of People with Developmental Disabilities ................................ VI . Conclusions . . . . . . . . . . . . . . . . . . .......... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1 . 111. IV. V.

Index . . . . . . . . . . . . . . . . . Contents of Previous Volumes . . . .

247 248 255 210 278 279 281

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

289

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

303

This Page Intentionally Left Blank

Contributors

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

John G . Borkowski (159), Department of Psychology, University of Notre Dame, Notre Dame, Indiana 46556 Michael T. Carlin (29), Department of Psychology and Human Development, Vanderbilt University, Nashville, Tennessee 37240 Janet Can- ( 1 97), Department of Psychology, St. George’s Hospital Medical School, University of London, London SW17 ORE, England Alan D . B . Clarke (137), Department of Psychology, The University of Hull, Hull HU6 7RX, England Ann M . Clarke (137), School of Education, The University of Hull, Hull HU6 7RX, England Chris Cullen (225), Psychological Laboratory, University of St. Andrews, St. Andrews, Fife KY16 9JU, Scotland Robert Fox ( l ) , John F. Kennedy Center and Department of Psychology, Vanderbilt University, Nashville, Tennessee 37240 Lisa D . Hager ( 1 1 l), Department of Psychology, University of Denver, Denver, Colorado 80208 Deborah Keogh (159), Department of Psychology, University of Notre Dame, Notre Dame, Indiana 46556 William J. McIlvane ( 5 3 , Behavioral Sciences Division, E. K . Shriver Center, Waltham, Massachusetts 02254; and Northeastern University, Boston, Massachusetts 02115 Stephen Oross I11 ( I ) , John F. Kennedy Center and Department of Psychology, Vanderbilt University, Nashville, Tennessee 37240 ix

X

CON TRI BUT0 RS

Trevor R . Parmenter (247), Unit for Rehabilitation Studies, Macquarie University, New South Wales 2109, Australia Anne Wurtz Passino ( 159). Department of Psychology, University of Notre Dame, Notre Dame, Indiana 46556 Elizabeth A. Rellinger (159), Department of Psychology, University of Notre Dame, Notre Dame, Indiana 46556 Kristen Sommer (159), Department of Psychology, University of Notre Dame, Notre Dame, Indiana 46556 Sal A. Soraci, Jr. (29), Department of Psychology, The University of Alabama, Tuscaloosa. Alubama 35487 Phillip D. Tomporowski ( 1 1 l), Department of Psychology, The University of Alabama, Tuscaloosa, Alabama 35487 Keri Weed (159), Department of Psychology, University of South Carolina-Aiken, Aiken, South Carolina 29801 Thomas L. Whitman (159), Department of Psychology, University of Notre Dame, Notre Dame, Indiana 46556

Preface

Current research in mental retardation focuses on a variety of psychological, developmental, and service delivery issues. In this volume four articles deal with various psychological aspects of mental retardation, including perception, learning, and sustained attention. Developmental issues are addressed in three articles reviewing longitudinal research on the modifiability of the life path, adolescent parenting, and Down syndrome. Service delivery issues are reviewed in two articles on management of services in residential settings and quality of life of mentally retarded individuals. In the first article Robert Fox and Stephen Oross review their program of research on perceptual deficits in mildly mentally retarded adults. These deficits apparently occur in very early stages of visual information processing and cannot be attributed to visual anomalies, motivation, inattention, or comprehension failures. Fox and Oross may have discovered an underlying neural impairment that may influence up to 90% of mildly mentally retarded individuals. They discuss the possibility that the intellectual impairment of mildly mentally retarded individuals may arise from such neurologically based deficits. If this proves to be correct, the view that mild mental retardation is due, largely, to cultural-familial variables may need to be revised. The relationship of perception and intellectual functioning is also the focus of the article by Sal Soraci and Michael Carlin. They review their research on relational learning and how the detection of interstimulus relationships such as similarity and dissimilarity influences learning. Soraci and Carlin suggest that the locus of “attention deficits” may be in the detection and utilization of relational information. This possibility would broaden the domain of attentional processes and challenge the mediational view of attention. William McIlvane is concerned with learning in mentally retarded persons, but uses a behavior-analytic rather than a perceptual learning approach. In his article, McIlvane describes a program of research on discrimination learning in moderately and severely mentally retarded individuals designed to discover the stimulus conditions that result in stimulus control of learning. The stimulus control research is concerned specifically with methods for teaching the prerequisites for symbol use necessary for reading. The implications for education of moderately and severely retarded individuals are far-reaching. xi

xii

PREFACE

In the next article, Phillip Tomporowski and Lisa Hager provide an update of research on sustained attention in mentally retarded individuals. They review their own program of research, which fails to support the developmental-lag hypothesis advanced by other investigators in this area. Their research also incorporates memory deficits as one aspect of sustained attention in mentally retarded individuals. Tomporowski and Hager find that information processing approaches to understanding sustained attention may provide the most promising theoretical approach in this area. The first of three articles reviewing longitudinal studies is by Ann Clarke and Alan Clarke. Their concern with the life-course of mentally retarded individuals began with their studies reported in the early 1950s. Their subsequent contributions and those from many different areas are reviewed. The evidence points to a conclusion that the trajectory of adaptation and adjustment across the life-span is not fixed; rather, the life-course is a result of a complex set of processes that influence the experience of the individual. These experiences, in turn, influence those processes. This view requires investigators in the area of mental retardation to reconsider their views regarding the influence of early experience on development. A quite different perspective is described in the article by John Borkowski, Thomas Whitman, Ann Wurtz Passino, Elizabeth Rellinger, Kristen Sommer, Deborah Keogh, and Keri Weed. They review longitudinal studies related to adolescent parenting and child outcome and provide an overview of their own longitudinal research, emphasizing the importance of the cognitive readiness of the mother for the task of parenting. Their findings suggest the interesting possibility that attachment and the resulting infant-mother interactions may influence the development of attention, which may, in turn, influence the development of information processing impairments in children of adolescent mothers. Janet Carr reviews several longitudinal studies of Down syndrome individuals and provides an overview of her own prospective study that spans a period of 2 1 years. Although there was a decline in IQ across age in her population-based study, the effects of out-of-home placement became less evident when the children reached adolescence. Her results also indicate that Down syndrome children make steady and demonstrable progress in the development of self-help skills and in other areas of independence. Carr's review integrates the results from several long-term longitudinal studies of Down syndrome individuals and provides an empirical framework for understanding their development. The last two articles in this volume are concerned with service delivery. Chris Cullen reviews current practice and research on staff management and training in residential settings for mentally retarded adults, including the literature on determining performance standards, methods of staff training, and maintenance of staff behavior. Cullen develops an argument that the social validity of manage-

PREFACE

...

XI11

ment procedures needs increased attention if the quality of life of mentally retarded residents is to be improved. In the final article, Trevor Parmenter provides a broad overview of the research on quality of life issues for adults with developmental disabilities. One contribution of this article is the inclusion of literature on quality of life areas not specifically related to developmental disabilities, thereby providing a more general basis for the consideration of these issues. This article also includes various theoretical models of quality of life and a discussion of the problems associated with measurement of this concept. The article ends with a consideration of predictions concerning the conceptual and practical problems that must be addressed in order to understand quality of life for mentally retarded individuals. As illustrated by these articles, this series will continue to publish integrative reviews on a wide range of current topics, including 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 are written by invitation, but unsolicited manuscripts will be considered. Manuscripts for this series are reviewed by the Editor, the Contributing Editors, and by outside reviewers. We thank Ann M. Clarke, Alan D. B. Clarke, Norman R. Ellis, Beate Hermelin, Robert L. Schalock, Marten Soder, Joseph Spradlin, and Joel S. Warm for their reviews of earlier versions of the articles included in this volume.

NORMAN W. BRAY

This Page Intentionally Left Blank

Perceptual Deficits in Mildly Mentally Retarded Adults ROBERT FOX AND STEPHEN OROSS Ill JOHN F. KENNEDY CENTER AND DEPARTMENT OF PSYCHOLOGY

VANDERBILT UNIVERSITY

NASHVILLE, TENNESSEE 37240

1.

INTRODUCTION

This article is both a tutorial and a progress report because it describes a new line of inquiry directed toward understanding mild mental retardation. The starting point is our investigation of the ability of mildly retarded adults to perceive certain classes of visual stimuli. We have found that these individuals encounter considerable difficulty in perceiving correctly the forms generated by arrays of random dots or elements that are produced by either stereoscopic depth or motion (Fox & Oross, 1988, 1990). These deficits in perception are quite substantial; under certain stimulus conditions, the discrimination performance of the retarded observers can be at chance levels, while nonretarded observers perform without error. This result is surprising in that the stimuli are registered at an early stage of the visual system and are perceived automatically, without effort or attention. Because the accruing evidence is rendering alternative interpretations (e.g., peripheral visual anomalies, low motivation, inattention, and failure to comprehend) increasingly less probable, we are compelled to entertain the hypothesis that the perceptual deficits reflect an underlying neural impairment heretofore unsuspected in mildly retarded persons. An examination of the evidence converging on that hypothesis constitutes the central theme of this article. It is organized in the following way. First, a description of the random element stimuli and their major attributes is given, together with a summary of what is known about underlying mechanisms. Next, the results of the investigations we have pursued using these stimuli are summarized, with an emphasis on those not yet published or those presented at meetings. Finally, the implications of these data with respect to the genesis of mild mental retardation are considered. INTERNATIONALREVIEW OF RESEARCH IN MENTAL RETARDATION. Vol. 18

I

Copyright 0 1992 by Academic Press. Inc. All rights of reproduction in any form reserved.

Robert Fox and Stephen Oross Ill

2

II.

RANDOM ELEMENT STEREOGRAMS

It is helpful to begin by reviewing some basic aspects of binocular vision. All evidence from evolutionary biology indicates that binocular vision, which refers to the existence of two horizontally separated eyes that view the same segment of visual space, serves to provide animals with the ability to make very precise discriminations of depth. The elaborate mechanisms that maintain binocularity, present in predatory animals representing diverse lines of descent, demonstrate the adaptive significance of depth perception based on two eyes. The stimulus conditions requisite for depth perception were identified by Wheatstone ( 1838) who integrated conceptually the geometric relationships that describe the lines of sight of two mobile, horizontally separated eyes. Because of the separation, each eye sees the world from a slightly different angle. Thus, if you fixate some point in space (hold up a pencil) and alternately close each eye, the difference in views become apparent. The pencil appears to shift laterally as the eyes open and close. But, when both eyes are open, the views are unified into a single binocular view, a condition known as binocular fusion. When both eyes fixate some object straight ahead, it is seen as a single fused object. Further, the object stimulates the same retinal area in each eye, a concept referred to as corresponding retinal points. For any convergence angle of the eye, there are a set of locations in visual space where a stimulus will appear fused and will stimulate corresponding retinal points. These locations, called the horopter, serve as a baseline depth position. Stimuli displaced in depth relative to the depth position of the horopter stimulate slightly different retinal positions in each eye, a condition known as retinal disparity. Thus, if you fixate a pencil, it defines the horopter; if you then hold up a finger next to the pencil and position it closer to you, the finger stimulates different retinal areas and is said to be retinally disparate. Wheatstone (1 838) devised two-dimensional drawings that simulate the stimulus conditions present when objects are at different depth positions. In such drawings, one for each eye, different horizontal separations of the stimuli simulate retinal disparity. When the drawings are viewed simultaneously, the stimulus with disparity will appear in depth displaced from the surface of the stereogram. Wheatstone demonstrated that retinal disparity is a sufficient condition for producing the depth percept and it is a cue that can be used only by the simultaneous operation of both eyes. He dubbed the capacity to perceive such percepts stereopsis, which means seeing solid. Hence, the drawings are known as stereograms . Wheatstone’s discovery of stereopsis provided a rationale for the evolutionary pressure devoted to maintaining binocular vision. But, it also provided the basis for such applications as stereophotography and three-dimensional movies. Stereopsis has attracted considerable research interest over the years and a detailed, quantitative description of the stimulus conditions governing it has been developed. One classic summary of much of this knowledge is given by Ogle (1959).

E

3

PERCEPTUAL DEFICITS

M3IA 3A3 l H 9 W

M31A 3A3 1331

FIG. I . Example of a random element stereogram. When the two matrices are viewed stereoscopically, a percept of a center square can be seen in depth above the background.

That date is noteworthy because, in the next year, a rather different kind of stereogram, the random element stereogram, was described (Julesz, 1960) that changed several conceptions about stereopsis. One way to appreciate the characteristics of a random element stereogram is to view one, such as the example given in Fig. 1. If viewed stereoscopically (consider making a copy and placing it in a stereoscope) a form will appear in depth relative to the background. The matrices for each eye are composed of 10,OOO minute cells. In each, half of the cells are filled on a random basis with white dots; the empty cells are black. When a subset of dots in one matrix is shifted laterally, an amazing thing happens. The observer with stereopsis perceives a form in depth, the shape of which corresponds to the configuration of the shifted dots. All dots appear at a single depth plane, forming a solid surface with distinct edges. The depth percept is triggered by the lateral shift of the dots, which gives them retinal disparity with respect to the fused background dots. This is the same process identified by Wheatstone, and, indeed, the primary geometrical relationships associated with stereopsis that have been studied since Wheatstone also apply to random element stereograms. Nevertheless, there are important differences between this kind of stereogram and those devised by Wheatstone. Three that merit attention involve the relationship between depth and form, the detection of disparate elements, and the locus of binocular interaction.

A.

Depth and Form

In the classic Wheatstone stereogram including its more sophisticated variants such as three-dimensional films, the scenes depicted can be perceived with one

4

Robert Fox and Stephen Oross 111

eye. The absence of binocular viewing eliminates only the stereoscopic depth aspect of the display. This observation is consistent with the idea that the visual system first processes the stimulus information about the contours and edges that constitute the pattern or scene. Then, the disparity information presumably is processed to produce a depth percept. But, the sequence is reversed with the random element stereogram. The initial percept is always seen in depth followed by the emergence of the stereoscopic form. The reversal of the processing sequence has significant implications for theories of perception. It also has an important, practical consequence for determining whether or not an observer has stereopsis. In the Wheatstone stereogram, various nonstereoscopic. cues (e.g., familiar size) are ever present and can be used to make correct depth discriminations. Moreover, responding to the depth percept requires attention and cooperation on the part of the observer. These considerations have impeded the investigation of stereopsis in animals, children, and infants. With a random element stereogram, however, the impediments to testing are eliminated because only observers who possess stereopsis can perceive or react to the stereoscopic form. Therefore, if a response (e.g., eye movements) is elicited that is correlated with the form, it is necessary to conclude that the observer possesses stereopsis (see Fox, 1981, for a review).

B.

Detection of Disparity

It is important to emphasize that the examination of any single matrix of a random element stereogram reveals only the array of random elements. No form can be seen nor is the subset of shifted dots detectable. Note, however, if an observer had sufficient time to note the position of every dot in one matrix and compare it with corresponding cells in the other matrix, the position of the shifted dots could in principle be discovered. Yet, even this lengthy, cumbersome strategy would not work with dynamic random element stereograms in which matrices are generated dynamically and arrays changed randomly 30 or more times per second. Under this condition the shifted elements are completely camouflaged and cannot be discerned by any kind of conscious analysis. Nevertheless, the stereoscopic form does appear, which means that the visual system has carried out some kind of rapid, automatic comparison of the matrices and detected the shifted dots, thereby triggering the generation of the form, which appears at a specific position in depth in accord with the geometry of stereopsis. The induction of stereoscopic percepts by random element stereograms has revealed that the mechanisms underlying binocular vision are more elaborate than previously expected and that they are closely linked to mechanisms involved in pattern recognition. For that reason, binocular vision has attracted considerable research interest at both the experimental and the theoretical levels. One

PERCEPTUAL DEFICITS

5

aspect that has attracted special attention concerns the process by which the dots in each matrix are matched appropriately, an issue known in the literature as the correspondence problem. At the level of individual neuronal interactions the probability for false matches increases exponentially as a function of the number of elements in the matrices, and when that number is large the correspondence problem is theoretically not resolvable. Nevertheless, the emergence of the stereoscopic form provides an existence proof that the visual system can solve the problem. An effort to learn just how the visual system does this motivates many current researchers, many of whom are working in the areas of artificial intelligence and computational vision. Understanding how a biological system can solve the correspondence problem facilitates development of physical systems with comparable properties. Indeed, the formidable problems posed by random element stereograms, from a computational perspective, led us to conjecture whether retarded persons who may be regarded as possessing limited computational resources might encounter difficulty in perceiving this class of stimuli.

C.

Locus of Binocular Interaction

A unique feature of random element stereograms is that at the retinal level the stimuli are simply the arrays of random dots. Comparison of the arrays by the visual system cannot start before the information conveyed by the monocular pathways is combined at the cortical level. This means that the stereoscopic forms are generated exclusively by cortical mechanisms. Because processing of the form by retinal and thalamic stages does not occur, it is possible to pose questions about the extent to which such stages participate in the generation of any particular perceptual phenomenon, an approach that has been dubbed psychoanaromy by Julesz (197 I). Many perceptual phenomena, observed originally in the luminance domain, wherein the stimuli are defined by physical variations in luminance and contrast, can be replicated in the stereoscopic domain. There, contours, in a very real sense, are fabricated within the visual system and cannot be specified physically. Nevertheless, contours from both domains can interact to produce such phenomena as aftereffects and backward masking (e.g., Patterson & Fox, 1990; Tyler, 1975).

Ill.

RANDOM ELEMENT KINEMATOGRAMS

There are many similarities between random element stereograms and random element kinematograms, both in terms of their physical characteristics and in the way they are processed by the visual system. In the same way that perception of the stereoscopic form depends on stereopsis, perception of a kinetic form depends on the capacity to perceive motion. As in the case of the stereogram, the

6

Robert Fox and Stephen Oross I l l

FRAME 1

FRAME 2

FIG.2. Schematic representation of a random element kinematogram. The elements located in the background are displaced in an uncorrelated manner across frames; the elements located within the outlined region are displaced in a correlated manner across frames. Note that the outline of the E form is included for illustrative reasons only. It is included to depict the phenomenal report of nonretarded adults. In the actual stimuli, forms are not apparent in any given frame.

starting point for a kinematogram is an array of dots in a matrix; over time the dots move over distances in random directions. The display resembles the visual noise seen on an untuned television channel. If a subset of dots begin to move in the same direction, say jump to the left and back to the right, a moving form with a connected surface and well-defined edges will be seen. Once this synchronized motion stops, the form disappears immediately. Because motion is required, Fig. 2 can show only in a schematic way the construction of a kinematogram. Dependence of the form on motion demonstrates that motion is a fundamental dimension of perception rather than being derived indirectly through spatial changes over time. Kinematograms have also attracted considerable research interest at both the experimental and the theoretical levels (for a review, see Nakayama, 1985). One reason is that they pose the same kind of computational problems as random element stereograms. To perceive the kinetic or moving form, the moving elements in one temporal frame must be matched with corresponding elements in the next frame. This is the correspondence problem in the temporal domain, and it carries with it the same possibilities for false or incorrect matches that exist for the spatial matching required for random element stereograms. Both kinds of stimuli are also subject to, or influenced by, the same kind of stimulus manipulations. These are described below.

7

PERCEPTUAL DEFICITS

A.

Correlation

In a standard kinematogram, all of the elements that constitute a kinetic form in one temporal frame have corresponding elements in successive frames. That is, the correlation of the elements is 100%. Similarly, in a stereogram, the elements in one matrix have corresponding elements in the other matrix. That is, the spatial correlation between the eyes is 100%. But, if the correlation is decreased (i.e., some elements are not in correspondence), the forms (stereo and kinetic) become more difficult to see. This reduction in perceptibility as a function of reduction in correlation is similar to a reduction in signal strength because of an increase in noise. B.

Density Reduction

In the standard stereogram and kinematogram, 50% of the cells in the matrices are filled with dots or elements. When density is reduced the missing elements are replaced phenomenally by a perceptual filling-in process that maintains the edges and surfaces of the forms. This makes it possible to perceive the shape or orientation of the forms with reductions in density down to levels on the order of 5 to 1%. In Fig. 3, the density of the stereogram is 5% yet the form is readily seen. The filling-in process, which is similar to the formation of subjective contours, is referred to as interpolation within the domain of computational models (e.g., Grimson, 1981; Marr, 1982).

LEFT EYE VIEW

RIGHT EYE VIEW

FIG. 3. Example of a reduced-density random element stereogram. This example illustrates the stereogram with only 5% element density.

Robert Fox and Stephen Oross 111

8

C.

Frame Number

The motion in kinematograms is apparent motion, which means that the dots jump from one X , Y position to another in discrete steps over time. A minimum of two jumps are required to produce apparent motion. As an example, consider a dot positioned to the left of a point of fixation on some display; the dot jumps first to the right and then back to the initial position. Repetition of this cycle produces a percept in which the dot appears to move continuously between the two positions. In most displays, each of the dot positions is presented within a discrete period of fixed duration called a frame. For example, in many video systems images are presented at the rate of 30 frames per second, which is sufficiently high to produce percepts that appear to have continuous motion. In kinematograms, the salience of the kinetic form can be enhanced if the number of frames is increased beyond the minimum of two. The additional frames permit the dot to jump in the same direction for more than one frame. For example, with four frames, the dot would jump to four successive positions in the same direction. The perceptibility of the kinetic form is enhanced because the stimulation provided by jumps in the same direction is integrated through some kind of neural summation that operates over time. The effect of frame number is unique to kinematograms; there is no stimulus variable associated with random element stereograms that is strictly comparable. But, under certain conditions, increases in retinal disparity will enhance the perceptibility of the stereoscopic form.

IV.

ECOLOGICAL SIGNIFICANCE OF RANDOM ELEMENT STIMULI

Because random element stereograms and kinematograms are both constructed from arrays of random elements, almost invariably generated by a computer, it is sometimes thought that they are artificial stimuli unrelated to real-world conditions. This is not the case. Although we now live in rectilinear environments in which many nonstereoscopic cues to distance exist, the environment of our distant ancestors consisted of trees, bushes, and other foliage. These stimuli, composed of many small elements such as leaves, contain few nonstereoscopic cues to depth and are the natural counterparts of random element stimuli. If one views a foliage-filled scene with one eye and compares it with the view provided by two eyes, a dramatic difference in depth information becomes apparent. One eye yields a flat two-dimensional scene, yet with two eyes, the scene now appears solid and three-dimensional with many gradations of depth. The same difference occurs when the task involves detection of camouflaged objects. These considerations make it difficult to discount the view that binocular vision evolved

PERCEPTUAL DEFICITS

9

to help predatory animals perceive depth in foliage-laden, probably arboreal environments and detect or discover camouflaged prey. The need to navigate successfully through such environments undoubtedly provided pressure for the development of neural mechanisms for processing the motion attendant to kinematograms. Indeed, a central theme of the theoretical approach developed by James J. Gibson was the great quantity of information provided by motion. It is interesting that to demonstrate the power of motion to specify surfaces, Gibson devised a kind of random element kinematograrn. His kinematograms were produced laboriously by film and animation methods, the only techniques available at that time. Subsequent advances in computer technology, of course, have greatly facilitated the generation of random element stimuli and this has played a significant role in promoting such research on these topics. A second factor of even greater significance in promoting such research was the realization that learning more about how these stimuli are processed would contribute to an understanding of the general problem of object and pattern recognition. In natural environments, visual scenes are quite complex, containing many surfaces and edges that must be yoked correctly to the objects they represent. At the level of physical stimulation, this can be very difficult. For example, surfaces often conceal or occlude other surfaces, and the determination of where one surface ends and another begins can be ambiguous. Nevertheless, the perceptual system solves the problem rapidly and correctly. The fact that biological perceptual systems are so effective and so superior to any physical perceptual or image processing system devised to date has attracted the interest of computer scientists working in the areas of computational vision and artificial intelligence. This has also served to promote research on the topic. As a consequence, the literature on random element stereograms and kinematograms, already extensive, continues to grow rapidly. That portion of the literature that bears most directly on research on mental retardation is summarized in the next section.

V.

BASIC RESEARCH ON RANDOM ELEMENT STIMULI

Basic research on random element stimuli proceeds at several interactive levels of analysis-behavioral, developmental, physiological, combined behavioral and physiological, and modeling. It is helpful to consider very briefly the key features of each of these paradigms and the kind of data they generate. The behavioral level, which is often referred to in the literature as psychophysics because psychophysical methods are used, involves the collection of large quantities of data from a relatively small number of well-trained observers. This approach yields quantitative data that specify the effects of manipulating

10

Robert Fox and Stephen Oross 111

various kinds of stimulus conditions. The developmental paradigm is essentially a subset of the psychophysical approach. The goal, however, is to determine when during development sensitivity to certain stimulus dimensions emerges. The physiological paradigm employs anatomical and electrophysiological techniques that determine the effect of stimulus conditions on neuronal activity and, via anatomical tracing methods, those portions of the brain that are maximally excited by such stimuli. Technical advances, particularly those at the molecular level, are producing almost overwhelming increases in information about neuronal processes. A powerful new paradigm combines psychophysics with physiology. In this approach, various species of monkeys, who have visual systems similar to that of humans, are taught to make precise psychophysical discriminations. At the same time, responses from individual cortical neurons are recorded, making it possible to correlate psychophysical behavior with the underlying electrophysiological activity. And finally, the wealth of new information coupled with enhanced computing power has rendered feasible quantitative modeling by means of computer-based simulation.

A.

Psychophysical Data

A fundamental characteristic of both random element stereograms and kinematograms revealed by phenomenal or psychophysical observation is that the stimuli are perceived rapidly and without effort. The rapid perception and the concomitant involvement of large numbers of spatially extensive stimulus elements mean that processing is parallel rather than serial. In the literature this kind of processing is often referred to as preattentive because perception of the stimuli does not engage attention or other volitional processes. The stimuli are regarded as exemplars of the modularity of perception as defined by Fodor (1 983). Although the effect of many stimulus variables has been examined psychophysically, we consider briefly only those that are directly related to our investigations of mild retardation. The relevant variables are spatial correlation in the case of stereograms and temporal correlation in the case of kinematograms, element density, and frame number. Reductions in spatial correlation in the case of stereograms and temporal correlation in the case of kinematograms serve to reduce the perceptibility of the stimuli. The parameters of these variables have not been examined extensively because it is thought that they represent straightforward cases of variations in signal/noise ratios and that specific values that change performance would probably vary as a function of other interacting variables. For instance, in Fox and Oross (1990) performance started to decline at around 50% correlation and fell to chance at 10% correlation. In that situation, only two frames were used; it is likely that if frame number was increased performance would improve, even at low correlation levels. Reductions in element density, on the other hand, do not produce reductions in

PERCEPTUAL DEFICITS

11

discriminability for either stereograms and kinematograms. As mentioned earlier, deleted elements are replaced by a compensatory perceptual filling-in process that maintains perceptual performance even at very low densities (e.g., Fox, Patterson & Langston, 1983). In general, reductions in density exert a negligible effect on perception of random element stimuli in a variety of conditions (Baker & Braddick, 1982; Julesz & Frisby, 1975; Regan & Hong, 1990; Williams & Phillips, 1987). Increases in frame number enhance the perceptibility of kinetic forms up to some limit that varies as a function of other stimulus conditions (Kramer & Todd, 1989; Snowden & Braddick, 1989). There is no strictly comparable manipulation applicable to random element stereograms. Considering now studies that bear on the question of ontogenetic development, sensitivity to random element stereoscopic stimuli, as well as the classic Wheatstone-type stereogram, emerges at about 4 months of age (Fox, A s h , Shea, & Dumais, 1980; Held, Birch, & Gwiazda, 1980). During that same period, sensitivity to kinetic forms emerges (e.g, Freedland & Dannemiller, 1987; Kramer, 1986, 1988). It is also noteworthy that by 5 to 7 months of age, infants become sensitive to the environmental cues that specify distance. These include familiar size (Granrud, Haake, & Yonas, 1985), relative size (Yonas, Granrud, & Pettersen, 1985), interposition (Granrud & Yonas, 1984), shading (Granrud, Yonas, & Opland, 1985), pictorial depth (Yonas, Cleaves, & Pettersen, 1978), and linear perspective and texture gradients (Yonas, Granrud, Arterbeny, & Hanson, 1986). Moreover, these cues are functionally integrated as demonstrated by sensitivity to the illusory motion of the Ames window (Oross, Frances, Mauk, & Fox, 1987). Although the information about distance that is specified by any single cue is ambiguous, when multiple cues are present (as they are in natural environments), the information from each is combined or integrated, thereby minimizing ambiguity. There is evidence that the integration process occurs in additive fashion, wherein the distance information becomes increasingly more precise as more cues are added (e.g., Bruno & Cutting, 1988; Cutting & Bruno, 1988). Models of the integration process are just beginning to be developed, yet it is clear that it is an automatic process that does not engage attention or other cognitive stages. Moreover, it is noteworthy that sensitivity to distance information is present in infants by 7 months of age. It is needed to calibrate the relative depth information provided by stereopsis and, in general, to locomote successfully in three-dimensional space. Thus, it seems that the requisite perceptual machinery is in place by the time walking commences at around 10 months of age.

B.

Physiological Data

Information about the function and structure of the brain has increased enormously in recent years as a consequence of anatomical and electrophysiological

12

Robert Fox and Stephen Oross 111

methods becoming increasingly refined and sophisticated. A major portion of the inquiry has been devoted to understanding the visual system of primates, with a special emphasis on the role of neocortex. These efforts have revealed that about 60% of cortex is involved in vision and that this area can be subdivided into at least 20 or more discrete cortical maps, many of which are retinotopically mapped (DeYoe & Van Essen, 1988; Kaas & Krubitzer, 1991). These areas interact by means of multiple pathways providing both feedback and feedforward connections. Neural signals from the thalamus are transmitted to primary visual cortex (VI) located in the occipital lobe and from there they are transmitted anteriorly to multiple visual areas in the parietal and temporal lobes. The optic nerve that transmits neural signals from the retina to V1 is divided into two separate parallel pathways, the magnocellular pathway and the parvocellular pathway. Each has distinct and different anatomical and electrophysiological characteristics, and the separation of the pathways is maintained in the projections from V1 to other cortical areas. The properties of the magnocellular pathway suggest that it may play a major role in processing stereoscopic and motion stimuli, but a definitive understanding of the psychophysical relevance of the pathways has not yet been reached (Livingstone & Hubel, 1988; Shapley, 1990). Results of electrophysiological recording from single cortical neurons driven by random element stereograms reveals the presence of disparity-sensitive cells in V1, V2, and V3 (Poggio, Motter, Squarito, & Trotter, 1985). Similar recordings from neurons driven by random element kinematograms have triggered neurons in V 1 , V2, and the adjacent area known as MT (for middle temporal). Many of the neurons of MT are especially sensitive to kinetic random element stimuli. In the behavioral/physiologicalparadigm monkeys have been trained to discriminate the directions of motions generated by random element kinematograms (their sensitivity is equivalent to that of humans). At the same time recordings have been made from cells in MT and a close correlation has been found between discrimination performance and neuronal activity (Newsome, Britten, 8t Movshon, 1989), but on ablation of MT by a rapidly acting chemical method, psychophysical performance fails as neuronal activity ceases (Newsome & Pare, 1988). This outcome clearly implicates MT as a major stage in the processing of kinetic stimuli. Yet, after a few days, the animals recover their ability to discriminate, suggesting that another cortical area, possibly V3, takes over MT’s function. In general the physiological data, which indicate that both motion and stereo stimuli are processed quite early in the visual system, are consistent with the rapid automatic preattentive processing of such stimuli present at the psychophysical level. Moreover, the processing of these stimuli in adjacent cortical areas that are extensively interconnected fits with the idea that they are processed by similar mechanisms. This is consistent with the similarities among stereograms and kinematograms seen psychophysically.

PERCEPTUAL DEFICITS

C.

13

Models

As mentioned earlier, the computational difficulties posed by the processing of random element stimuli have attracted the interest of computer scientists. Their efforts to develop models have also been stimulated by the wealth of psychophysical and physiological data that have accrued. Their general objective is to develop models that are realistic in terms of physiologically constrained assumptions and that perform successfully. Some reviews and representative examples are Grossberg and Mingolla (1983, Linsker (1990), and Sejnowski, Koch, and Churchland (1988). It is neither necessary nor possible to consider these efforts in any detail. Nevertheless, it is useful to consider a general feature of such models. A key metaphor is that of a neural network composed of individual hypothetical neurons interconnected such that activity in a local area or region can be transmitted throughout the network. For instance, each individual neuron would react to local stimulation such as the motion of a stimulus element in one direction. Through a series of excitatory and inhibitory connections, local activity can operate globally and influence the complete network. As an example, albeit a very speculative one, the continuing perceptibility of random element forms despite reductions in element density might be attributable to the global spread of excitation that, in some manner, substitutes for the absence of local stimulation. In any case, the existence of such models and their ongoing development are noteworthy because the models can also be modified to simulate various kinds of deficits. For instance, Kosslyn, Flynn, Amsterdam, and Wang (1990) are using a computational model to simulate the effect of brain damage at various sites on higher-order visual-cognitive processes. A similar kind of model has been developed to simulate the symptoms of dyslexia (Hinton & Shallice, 1991). It is possible that the deficits in perception present in the mildly retarded could be examined via simulation. The nature of those deficits is described in the next sections.

VI.

GENERAL METHODOLOGICALCONSIDERATIONS

Before describing the results of specific studies it is helpful to consider some methodological issues applicable to all experiments. The deficits in motion and stereopsis we have observed are surprising in that they are substantial ones occurring at perceptual stages quite distinct from cognitive processing. The question naturally arises as to whether these deficits are truly perceptual as opposed to being attributable to factors such as failure of comprehension, inattention, or response bias. It is important to consider such alternative interpretations and determine their likelihood. To that end, our approach has followed the converging operation strategy described by Garner, Hake, and Eriksen (1956). They point out that in psycho-

14

Robert Fox and Stephen Oross 111

physical experiments on perception, the objective is to make inferences, from stimulus and response relations, about where a process is localized within the perceiver. Such inferences are equivocal when the results of only a single experiment are available. Responses that ostensibly appear to reflect the operation of a perceptual stage may be attributable to memory or a response bias factor. A strategy they propose for identifying unequivocally perceptual phenomena so as to distinguish them from nonperceptual factors consists of successive elimination of possible alternative explanations by applying multiple independent measures or operations. Recently, Profitt and Bertenthal (1990) examined the concept of converging operations in the light of contemporary research on development. They endorse the validity of the logic and make it clear that it can be applied fruitfully to the investigation of perception in human infants and other nonverbal organisms. In addition to the application of converging operations, we have employed the objective response methods of contemporary psychophysics in our experiments. These methods require that the subject indicate or demonstrate the ability to make a discrimination by an overt response that can be linked to an external criterion. Whenever a measurement can be referenced to an external standard, the criterion for objective measurement is met. For example, a subject is required to point to the location of the stimulus, the correct location of which is known to the experimenter. The forced-choice methods are coupled with feedback as to the correctness of response that is given after every trial. In terms of signal detection theory, when target probabilities and payoffs are symmetric this procedure yields optimal discrimination performance. It places the response criterion at the intersection of the overlapping signal and noise distributions, that is, p = 1.0. To ensure that subjects understand the response requirements of an experiment they are first tested with stimuli produced by differences in luminance that are identical to those they will see in the experiment proper. These test stimuli appear as black forms against a background of random elements and are generated by simply suppressing elements in a specific area of the display. For example, if the task involves shape discrimination, the shapes to be discriminated in the experiment are first presented in the luminance domain. Only those subjects who correctly perceive these stimuli are used in the experiment. Once subjects demonstrate comprehension of the task with luminance-defined shapes they must then exhibit some capacity for perceiving the experimental stimuli such as kinetic forms. As will be described in a subsequent section, we have found some subjects who clearly comprehend the task yet apparently possess no ability to see random element stimuli. Random element stereograms and random element kinematograms can be seen even under conditions of considerable optical blur. Nevertheless, we have elected to measure the visual capacity of our subjects to forestall alternative interpretations in terms of peripheral visual anomalies. To measure contrast sensitivity we

15

PERCEPTUAL DEFICITS

use a commercial clinical test of sensitivity (VISTECH), which consists of photographs of patterns varying in spatial frequency and contrast. To measure visual acuity the Random E acuity test, which does not depend on language comprehension, is used, The subject is required to indicate the orientation of a form, in the shape of the letter E, which can assume any one of four orientations. The size of the form is progressively reduced until a threshold is reached. With respect to the characteristics of the retarded subjects, they are paid volunteers recruited from sheltered workshops in Nashville and surrounding counties who have been diagnosed as being mildly mentally retarded. The workshops are regulated by the State of Tennessee’s Department of Mental Health and Mental Retardation, which follows classification guidelines promulgated by the American Association on Mental Retardation. The workshops enroll individuals from 22 to 65 years of age. Within 30 days of joining a workshop, clients are evaluated by the staff to determine their suitability for participation in workshop programs. This includes administration of standard intelligence tests by a licensed psychologist and administration of the Vineland Adaptive Behavior Scales. These measures provide the basis for diagnosing the level of mental retardation. Most of the mildly retarded individuals have been so diagnosed previously by school systems. They are detected because of the academic difficulties they have encountered. The subjects in our studies have had IQs in the range 50 to 70 and have been approximately equally divided between men and women, with an average age of about 30 years.

VII.

RESEARCH PROGRESS

Our first set of experiments has already been published (Fox & Oross, 1988), so these studies are summarized only briefly. Three experiments examined the ability of mildly retarded adults to process correctly the information conveyed by random element stereograms. In the first experiment, subjects were required to discriminate correctly the left/right position of a stereoscopic form as a function of element density. Although the subjects clearly performed well above chance under the full density condition, performance was impaired as density was reduced. In contrast, the performance of nonretarded subjects was not impaired by density reduction. In the second experiment, the ability of retarded subjects to discriminate among global stereoscopic forms varying in shape was assessed under a full density condition. Although their overall performance was above chance, retarded subjects exhibited a substantial number of errors in discrimination, whereas the nonretarded subjects performed without error. The third experiment examined the ability of mentally retarded subjects to localize correctly the depth position of stereoscopic forms appearing in different positions in threedimensional visual space. The performance of the nonretarded subjects was

16

Robert Fox and Stephen Oross 111

accurate and consistent with the depth positions predicted by the geometry of visual space. That is, they showed what is called depth constancy, but the retarded subjects exhibited large errors in depth judgment, which suggests that constancy was not operative. These deficits were unexpected and they motivated us to consider very carefully the extent to which they could be attributable to the perceptual system. We decided to test their generality by using random element kinematograms. These stimuli also sidestepped methodological problems associated with questions about binocular alignment and the presence or absence of stereopsis. At the same time they allowed us to test the hypothesis arising from the results of the second experiment on form discrimination, which could be interpreted as indicating that the retarded individuals have some deficit in the filling-in process. These considerations led to the experiment described in Fox and Oross (1990) that examined the ability to detect kinetic forms as a function of variations in temporal correlation and element density. Subjects were required to make twochoice forced-choice discriminations of the left/right position of a kinetic form. The retarded subjects performed well when density and temporal correlation were high, but as the values of those variables were reduced performance was severely disrupted. Under many combinations of density and correlation retarded subjects performed at chance, whereas nonretarded subjects performed without error. These results indicate that the results found with stereograms in our first study are not restricted to stereoscopic stimuli. Further, they indicate that the density reduction variable generalizes across tasks. To examine further the viability of a perceptual interpretation of the deficits found by Fox and Oross (1990) we explored the degree to which the deficits could be removed or modified by training. In this study, which is described in Oross, Carlin, and Fox (1990a:1,two training methods were used. In one, called repeated exposure, a stimulus that was detected significantly above chance yet not perfectly was presented repeatedly under constant conditions for 160 trials. The second procedure, called fading, varied the salience of the stimulus over trials. During the initial trials, the salience of the stimulus was quite high because of differences in element density between it and the background. Then salience was gradually reduced as density differences became smaller. The stimulus proved to be no more detectable after the training period than it had been before training. This outcome certainly does not settle the general question of the extent to which deficits might be modified by various kinds of training procedures, a topic that merits further research. Yet the absence of large immediate improvement in performance, after training, is consistent with the hypothesis that deficits originate within the visual system. The second study (Oross, Carlin, & Fox, 1990b), with implications for alternative interpretations of the results of Fox and Oross (1990), examined the effects of optical blur on the detection and discrimination of kinetic forms. The ability of

PERCEPTUAL DEFICITS

17

nonretarded subjects to detect and discriminate kinetic forms under varied conditions of dioptic blur produced by lenses was investigated. Overall, performance proved quite resistant to blur. This result is consistent with the hypothesis that the deficits observed among retarded persons cannot be attributed to impairments at the peripheral level. To test the hypothesis that mildly retarded persons have greater difficulty than nonretarded persons in discriminating the shape of kinetic forms relative to simply detecting them, thresholds for both discrimination and detection were assessed for the same subjects in the same experiment. The discrimination task required subjects to identify correctly the shape of any one of eight kinetic forms of the following shapes: H, T, E, F, P, U, I, +. Each of these forms was presented one at a time in a random order; subjects responded by pointing to physical replicas of the form. For the detection task, subjects were required to make two-choice forced-choice judgments about the left/right position of the form. The thresholds were obtained using a descending staircase method. To manipulate signal strength, the number of frames was varied. That is, sufficient frames were presented to produce a suprathreshold response and then, in accord with the dictates of the staircase method, frame number was reduced. A second independent stimulus manipulation was density, in which five values were used, ranging in 10% steps from 50 to 10%. The experiment tested 12 mildly retarded subjects, who ranged in age from 24 years, 8 months to 39 years, 1 month, with a mean age of 30 years, 3 months. The nonretarded subjects used as a control were 12 students enrolled at Vanderbilt University. They ranged in age from 22 years, 8 months to 25 years, 8 months, with a mean age of 24 years, 6 months. The detection thresholds, defined in terms of number of frames, were 2.70 for the mentally retarded subjects and 2.0 1 for the nonretarded subjects. Discrimination threshold for the mentally retarded subjects was 6.00, and for the nonretarded subjects, 2.42. The large difference between discrimination and detection thresholds is statistically significant and supports the hypothesis that the process responsible for form discrimination is substantially impaired in retarded persons. Support for this hypothesis also comes from the small difference between detection and discrimination thresholds present in the nonretarded subjects. In addition, the relatively good performance for detection displayed by the retarded subjects indicates that they understood the task and, in general, renders alternative interpretations in terms of such factors as motivation and attention less tenable. In a second experiment investigating form discrimination, two density values (50 and 10%)and two frame numbers (six and three) were combined factorially. The dependent variable was the number of forms correctly recognized. The subjects were the same ones that participated in the threshold experiment. Results are depicted in Fig. 4, in which it can be seen that both density and frame

18

Robert Fox and Stephen Oross III 0 Nonretarded Subjects

-

0 Mentally Retarded Subjects

JFRAMES

BFRAMES

100 90

100

80

t

40

-

20 ..--__--. 10 -

30

t

40

.-...

20

I._._---.-....

I

50%

10

I Element Density

10%

50%

Element Denslty

10%

FIG. 4. Percentage of motion-defined forms correctly discriminated by mentally retarded and nonretarded adults as a function of element density and number of frames. The error bars indicate one standard error.

number act jointly to impair the performance of retarded subjects. On the other hand, those variables have little effect on the performance of nonretarded subjects. The large differences between retarded and nonretarded persons seen in this experiment and in the other studies discussed so far have been based on comparisons with nonretarded subjects who are likely to be of superior intelligence. The question that arises naturally is how individuals of average intelligence would perform. A number of considerations, such as the sensitivity of infants to kinetic stimuli, suggest that performance of individuals of average intelligence does not differ substantially from performance of individuals with superior intelligence. The next experiments addressed this question directly. The key feature of this inquiry was the addition of a new group of subjects with low average intelligence. It was formed from a population of grade-school children ranging in age from 6 years, 7 months to 12 years, 9 months. Although formal indices of socioeconomic status were not available the catchment area of the school serves low-income families. Thirty-one children were tested with kinetic forms by a method to be described; in addition, estimates of intellectual capacity were obtained by means of the Peabody Picture Vocabulary Test (PPVTR); mean PPVT-R = 88.5, range = 66 to 108. The performance of this group was compared with that of two other groups, one, mildly retarded adults, and the other, nonretarded adults similar to those used in prior research. The characteristics of the mildly retarded subjects were as

PERCEPTUAL DEFICITS

19

follows: N = 15 (9 males, 6 females); mean age = 32 years, 4 months; mean PPVT-R = 51.7, range = 43 to 70. The characteristics of the nonretarded subjects, which reflect the average intelligence levels of college student populations, were as follows: N = 23 (9 males, 14 females); mean age = 24 years, 10 months, range = 18.1 to 30.3; mean PPVT-R = 110.8, range = 97 to 120. The method of testing the perceptibility of kinetic forms was designed to do two things: (1) reduce the time required for testing and (2) minimize the possibility that performance could be attributed to varying familiarity with alphabet letters. To fulfill the latter objective, a kinetic form was configured as the letter E and presented over trials in any one of four orientations: up, down, left, right. Subjects were required to make forced-choicejudgments of the orientation of the form on each trial, after which feedback was given as to the correctness of the response. To reduce the time required for testing, 42 stimuli varying in density and in correlation were generated by factorially correlating or combining seven levels of correlation with six levels of density. These can be thought of as analogous to paper-and-pencil test items varying in difficulty. From prior research it is known that a stimulus consisting of 10% correlation and 10% density would be quite difficult to perceive correctly, whereas a stimulus composed of 100% correlation and 50% density would be much easier to perceive. Each stimulus was presented once, in randomized order, and the dependent measure was number of stimulus items correct, that is, identifying correctly the orientation of the E form. The results are indicated in Fig. 5, where each point represents one subject. Performance of the retarded group is significantly depressed relative to the other two groups; there is no significant difference between the children and the nonretarded adults. Although the performance of the children is more variable, the correlation between age and performance is not significant; however, there is a significant correlation for the children ( r = .32) between performance and scores on the Peabody Picture Vocabulary Test. Another perspective on these results is provided by the three-dimensional graphs shown in Fig. 6. The results for the retarded group are clearly quite different from those for the other two groups. Because of poor performance, a second experiment was designed to determine if performance could be improved. The manipulation chosen to enhance performance was an increase in the number of frames. Using the same group of retarded subjects, we simply repeated the first experiment with the addition of a condition of six frames and then a second condition of eight frames. For comparison, we also tested a group of nonretarded adults. The results are depicted in Fig. 7. As can be seen, an increase in frames enhanced the performance of the nonretarded subjects relative to the four-frame condition, yet the addition of frames produced no significant change in the performance of the retarded subjects. The results from the eight-form discrimination experiments and the E-form

Robert Fox and Stephen Oross 111

20

40

3

35

fn

z

2fn

30

w

t

K

25 K IT

8

20

b 2

15

9

10

m

5

MENTALLY RETARDED ADULTS

NONRETARDED CHILDREN

NONRETARDED ADULTS

FIG. 5 . Plot of individual scores on the four-frame motion-defined E task for all three groups. Each point indicates the number of correct responses (out of 42 possible) for each subject. The horizontal bars indicate the median scores.

orientation experiments indicate that the retarded subjects had considerable difficulty in discriminating the shape of kinetic forms. This difficulty is exacerbated by reductions in density and correlation. This result is consistent with the hypothesis that there is some impairment in the process that defines the edges of forms, suggesting that the forms appear more as blobs than as sharply delineated shapes. The blob idea also fits with the fact that detection of forms, which is a task that does not require discrimination, is only moderately impaired in retarded subjects. The similarity of the performance of the grade-school children to that of college students, despite differences between the groups in terms of such variables as intelligence, age, and socioeconomic status, goes some way toward rejecting explanations of retarded subject performance in terms of such social or cultural factors. This, in turn, provides support for the hypothesis that the deficits involve the perceptual system. But at this stage of the inquiry such a conclusion must be considered provisional, for many questions remain unanswered. One such question concerns the origin of differences among individuals, which is discussed next.

PERCEPTUAL DEFICITS

21

FIG. 6. Three-dimensional representation of the group data obtained from mentally retarded adults, nonretarded adults, and nonretarded children in the four-frame motion-defined E task. Percentage of correct responses is displayed as a function of both temporal correlation and element density. The different shades correspond to 20% intervals. (Note: 25% correct is chance performance.)

Taken as a group the mildly retarded subjects are clearly impaired on the perceptual tasks used in our experiments. Yet differences in performance among individuals have appeared that are noteworthy. In virtually every experiment, a small number of subjects have performed at a level equivalent to that achieved by nonretarded subjects. A second set have given every indication that they understood the task when the stimuli were from the luminance domain, yet apparently could not perceive the experimental stimuli composed of random elements. The majority of the subjects, however, showed the predominant pattern of performance: successful discriminations were made when the random element stimuli were of high density and temporal correlation, followed by a decline in performance when density and correlation were reduced.

22

Robert Fox and Stephen Oross I l l

FIG.7. Three-dimensional representation of the group data obtained from mentally retarded adults and nonretarded adults in the six- and eight-frame motion-defined E task.

It is difficult to determine accurately the proportions of subjects that fall into these categories because of differences in task requirements and variations in the availability of subjects. For instance, some subjects have served in more than one study, and others have stopped attending the workshops. An impressionistic analysis, however, based on 99 subjects tested so far, is that 13%have performed at the nonretarded level, 15 to 20% appear not to be sensitive to kinetic forms, and the remainder show the variable pattern described earlier. We have been unable to identify factors that would account for these differences. Further, there appear to be no differences in performance between subjects who have served in multiple experiments and those who have served in only one. These data suggest that the population of mildly retarded individuals is not a homogeneous one, an idea by no means novel in the literature on mental retardation. Further research on the factors producing these individual differences is clearly required.

23

PERCEPTUAL DEFICITS

VIII.

GENERAL DISCUSSION

When we started our investigation the theoretical justification, couched in terms of computational difficulties posed by random element stimuli, was compelling. Yet, on an intuitive level, we were not at all confident that interesting differences would be found. Therefore, we were both surprised and skeptical about the substantial deficits that did emerge. This motivated us to assess the perceptual interpretation of the deficits using the converging operation strategy described earlier. To that end, we decided to concentrate on determining the parameters of detection and discrimination using kinetic forms because of the methodological advantages, described earlier, which this approach provided. This meant that we had to defer further evaluation of the ability of the mildly retarded individuals to process information about three-dimensional space. This is an important topic that will be examined in future research. The results from the experiments summarized in this article, taken together, provide some support for the hypothesis that the deficits are due to anomalies within the perceptual system. The effect of stimulus manipulations on both detection and discrimination performance is clearly consistent with that hypothesis. It is helpful to examine each of the three manipulations: density reduction, reductions in temporal correlation, and variations in frame number. Density reduction physically impoverishes the stimulus spatially. Similarly, reductions in temporal correlations reduce the proportion of elements that contribute to the stimulus over time. Increases in frame number strengthen the stimulus up to some limit. In our last experiment, an increase in frame number enhanced the performance of the nonretarded subjects, but did not change the performance of the retarded subjects. It is gratifying that the results we obtained with the nonretarded subjects (a large jump from four to six frames, with an asymptotic level between six and eight frames) was quite similar to the results obtained by Snowden and Braddick (1989) from a small number of sophisticated psychophysical observers. The effects of these stimulus variables can be conceptualized in terms of the neural network metaphor, in which local excitation from individual network nodes propagates globally throughout the network. Such propagation would act to compensate for the absence of stimulation produced by reductions in density and in temporal correlation. The effect of increases in frame number would be accounted for by some hypothetical yet plausible process of neural summation (e.g., Snowden & Braddick, 1989). It is conceivable that the results obtained from the mildly mentally retarded subjects could be modeled by a network with some parameter modified such that the network yields weak, rapidly decaying propagation. Under full density and high correlation conditions, such a system might function adequately. At the same time it would be vulnerable to reductions in those variables.

24

Robert Fox and Stephen Oross 111

This is all quite speculative, but it does provide a way of thinking about the results we have obtained. One implication is that there may be novel means of modifying network parameters and restoring them to normal values (for instance, through pharmacological manipulation of neurotransmitter levels). An additional implication is that the results compel us to posit a neural or structural factor associated with mild retardation in the bulk of the subjects we have tested. Indeed, we have found deficits in 87% of the subjects. Interestingly, it has been commonly assumed that approximately 80% of mildly mentally retarded individuals are neurologically or organically intact because no etiology can be specified (for a discussion of this point, see, e.g., Accardo & Capute, 1991; Scott & Carran, 1987). If a neural deficit is present, the question arising naturally concerns the relationship between it and diminished intellectual capacity. An insight about the way such a relationship might be forged derives from an analysis of the genesis of autism by Frith, Morton, and Leslie (1991). The starting point for their analysis is the various lines of evidence they marshalled, which suggest that autism is induced by abnormal biological conditions. The specific causative process is not known; one possible candidate is the dopamine system. But regardless of the character of the biological deficit, the next step is the assumption that it would impede the normal development of cognitive mechanisms relevant to autism. The core behavioral symptoms of autism are presumed to devolve from interactions with the environment that become flawed by the impaired cognitive mechanisms. Although this causal analysis linking basic biological variables to behavior reflective of higher cognitive processes involves a series of assumptions not yet established unequivocably, the general approach seems plausible. It is possible that the approach could be used to relate the kinds of perceptual deficits summarized in this report to the intellectual impairments attendant to mild mental retardation. The starting point would be the assumption that the perceptual deficits are attributable to neural factors of the kind described earlier. A further assumption is that the presumptive neural factors have been present for a long period during development, thereby producing correlative deficits in perception. The critical next step requires linking perceptual processes to cognitive processes, an issue that has received some attention. For instance, Shepard and Podgamy (1978) have described the genesis of cognitive mechanisms from perception. Shepard (1984) has suggested that perception defined in terms of the ecological perspective of J. Gibson could be related to the formation of internal mental representations, including those involved in creative thinking and dreaming. Freyd (1987) has contributed to this view by stressing the importance of dynamic mental representations derived from the perception of motion. It seems reasonable to suppose that factors related to intelligence fall within the domain of mental representations. Although such an analysis is speculative it does not appear unreasonable. Moreover, it provides a framework for further research.

25

PERCEPTUAL DEFICITS ACKNOWLEDGMENTS

This research was supported, in part, by Grant EY-00590 from the National Eye Institute and Grant HD-I5051 from the National Institute of Child Health and Human Development. We thank Bennett Bertenthal and Steven Kramer for making their computer programs available to us. Thanks are also given to Elizabeth Graham and Michael Carlin for their assistance in the testing of subjects. Wc extend our appreciation to Randy Blake, Al Baumeister, and Norman Bray for their many constructive comments on an earlier version of this chapter and to Dona Tapp and Linda Maggart for their assistance in its preparation. Finally, we are grateful to the staff and clients or students of the Metropolitan Public Schools System of Davidson County, Rochelle Training and Rehabilitation Center, New Horizons, Inc., and Prospect, Inc. for their generous assistance.

REFERENCES Accardo, P. J., & Capute, A. J. (1991). Mental retardation. In A. J. Capute & P. J. Accardo (Eds.), es in infancy and childhood (pp. 431-439). Baltimore, MD: Paul H. Brookes. Baker, C. L., & Braddick, 0. J. (1982). The basis of area and dot number effects in random dot motion perception. Vision Research, 22, 1253- 1260. Bruno, N . , & Cutting, J. E. (1988). Minimodularity and the perception of layout. Journal of Experimental Psychology: General. I 17, I6 1- 170. Cutting, J. E., & Bruno, N. (1988). Additivity, subadditivity, and the use of visual information: A reply to Massaro. Journal of Experimental Psychology: General, I 1 7 , 422-424. DeYoe, E. A,, & Van Essen, D. C. (1988). Concurrent processing streams in monkey visual cortex. Trends in Neuroscience, I I , 219-226. Fodor, J. A. (1983). The modulariry ofmind. Cambridge, MA: MIT Press. Fox. R. ( 1981). Stereopsis in animals and human infants: A review of behavioral investigations. In R. N. A s h , J. R. Alberts, & M. R. Petersen (Eds.), Development of Perception: Vol. 2. The visual system (pp. 335-381). New York: Academic Press. Fox, R.. A s h , R. N., Shea, S . L., & Dumais, S. T. (1980). Stereopsis in human infants. Scienre. 207, 323-324. Fox, R.. & Oross, S., 111. (1988). Deficits in stereoscopic depth perception by mildly mentally retarded adults. American Journal on Mental Retardation, 93, 232-244. Fox, R.. & Oross, S . , 111. (1990). Mental retardation and perception of global motion. Perception & Psyrhophysics, 48, 252-258. Fox, R., Patterson, R., & Langston, A. L. (1983).The effect of element density on global stereopsis. Investigative Ophthalmologv and Visual Science Supplement, 24, 96. Freedland, R. L., & Dannemiller, J. L. (1987). Detection of stimulus motion in 5-month-old infants. Journal of Experimental Psychdogy: Human Perception and Performance, 13, 566-576. Freyd, J. J. (1987). Dynamic mental representations. Psychological Review, 9 4 , 427-438. Frith, U . . Morton, J., & Leslie, A. M. (1991). The cognitive basis of a biological disorder: Autism. Trends in Neuroscience, 14. 433-438. Garner, W. R., Hake, H. W., & Eriksen, C. W. (1956). Operationalism and the concept of perception. Psychological Review, 63, 149- 159. Granrud, C. E., Haake, R. J., & Yonas, A. (1985). Infants' sensitivity to familiar size: The effect of memory on spatial position. Perception & Psyrhophysics, 37, 459-466. Granrud. C. E., & Yonas, A. (1984). Infants' perception of pictorially specified interposition. Journal of Experimental Child Psychology. 37, 500-5 11. Granrud, C. E.. Yonas. A., & Opland, E. A. (1985). Infants' sensitivity to the depth cue of shading. Perception & Psvchophysics, 37. 4 15-4 19.

26

Robert Fox and Stephen Oross 111

Crimson, W. E. L. (1981). From images to surfices. Cambridge, MA: MIT Press. Grossberg, S., & Mingolla, E. (1985). Neural dynamics of form perception: Boundary completion. illusory figures, and neon color spreading. Psychological Review, 9 2 , 173-21 I . Held, R., Birch, E. E., & Gwiazda, J. (1980). Stereoacuity of human infants. Proceedings of the National Academy of Science, 77, 5572-5574. Hinton, G. E., & Shallice, T. (1991). Lesioning an attractor network: Investigations of acquired dyslexia. Psychological Review, 98, 74-95. Julesz, B. (1960). Binocular depth perception of computer-generated patterns. Bell System Technical Journal, 30, 1125- 1162. Julesz, B. ( I97 I ) . Foundations of cyclopean perception. Chicago: University of Chicago Press. Julesz, B., & Frisby, P. (1975). Some new subjective contours in random-line stereograms. Perception, 4 , 187-192. Kaas, J. H . , & Krubitzer, L. A. (1991). The organization of extrastriate visual cortex. In B. Dreher & S. R . Robinson (Eds.), Neuroanatomy of the visualpathways and their retinotopic orgunizution (pp. 302-323). New York: Macmillan. Kosslyn, S . M., Flynn, R. A,, Amsterdam, J. B., & Wang, G. (1990). Compenents of high-level vision: A cognitive neuroscience analysis and accounts of neurological syndromes. Cognition. 34, 203-277. Kramer, S . J. ( 1986). Infants’ detection of motion in random-dot kinematograms. Paper presented at the International Conference on Infant Studies, Los Angela, CA. Kramer, S . J. (1988). Developmentul changes in the spatial properties of the short-range process during infancy. Unpublished doctoral dissertation, University of Virginia. Kramer, S . J., & Todd, 3. T. (1989.1.Spatio-temporal integration in short-range apparent motion. Investigative Ophthalmology and Visual Science Supplement, 31, 5 19. Linsker, R. (1990). Perceptual neural organization: Some approaches based on network models and information theory. Annual Review of Neuroscience, 13, 257-281. Livingstone, M., & Hubel, D. (1988). Segregation of form, color, movement, and depth: Anatomy, physiology, and perception. Science, 240, 740-750. Man, D. (1982). Vision. San Francisco: Freeman. Nakayama, K. (1985). Biological image motion processing: A review. Vision Research, 25, 625660. Newsome, W. T., Britten, K. H., & Movshon, J. A. (1989). Neuronal correlates of a perceptual decision. Nuture, 341, 52-54. Newsome, W. T., & Pare, E. B. (1988). A selective impairment of motion perception following lesions of the middle temporal visual area (MT). Journal of Neuroscience, 8 , 22012211. Ogle, K. N. (1959). Theory of stereoscopic vision. In S. Koch (Ed.), Psychology: A study of science (Vol. I , pp. 362-394). New York: McGraw-Hill. Oross, S., 111, Carlin, M. T., & Fox, R. (1990a). Motion perception deficits in mildly mentally retarded adults: An experimental attempt to ameliorate deficits via training. Gatlinburg Conference on Research and Theory in Mental Retardation and Developmental Disabilities Abstracts, 23, 66. Oross, S . , 111, Carlin. M. T., & Fox, R. (1990b). Short-range apparent motion: Investigations of detection and figural integrity. Investigative Ophthalmology and Visual Science Supplement, 31, 519. Oross, S . , 111, Francis, E., Mauk. D . , & Fox, R. (1987). The Ames window illusion: Perception of illusory motion by human infants. Journal of Experimental Psychology: Human Perception and Performance, 13, 609-613. Patterson, R., & Fox, R. (1990). Metacontrast masking between cyclopean and luminance stimuli. Vision Research, 30, 439-448.

PERCEPTUAL DEFICITS

27

Poggio, G. F., Motter, B. C., Squatrito, S., & Trotter, Y. (1985). Responses of neurons in visual cortex (V I and V2) of the alert macaque to dynamic random-dot stereograms. Vision Research. 25. 397-406. Protfitt, D. R., & Bertenthal, B. 1. (1990). Converging operations revisited: Assessing what infants perceive using discrimination measures. Perception & Psychophysics, 47, 1-1 I . Regan, D., & Hong, X. H. (1990). Visual acuity for optotypes made visible by relative motion. Optometry & Visual Science, 67. 49-55. Scott, K. G . , & Carran, D. T. (1987). The epidemiology and prevention of mental retardation. American Psychologist, 42, 801-804. Sejnowski, T. J., Koch, C., & Churchland, P. S. (1988). Computational neuroscience. Science, 241, 1299- 1306. Shapley, R. ( 1990). Visual sensitivity and parallel retinocortical channels. Annual Review of Psychology, 4 1 . 635-658. Shepard, R. N. (1984). Ecological constraints on internal representation: Resonant kinematics of perceiving, imagining, thinking, and dreaming. Psychological Review, 91, 417-447. Shepard, R. N., & Podgamy, P. (1978). Cognitive processes that resemble perceptual processes. In W. K. Estes (Ed.), Handbook of learning and cognitive processes (pp. 189-237). Hillsdale, NJ: Erlbaum. Snowden, R. 1.. & Braddick, 0. J. (1989). Extension of displacement limits in multiple exposure sequences of apparent motion. Vision Research, 29, 1777- 1787. Tyler, C. W. (1975). Stereoscopic tilt and size aftereffects. Perception, 4 , 187-192. Wheatstone, C. (1838). On some remarkable, and hitherto unobserved, phenomena of binocular vision. Philosophical Transactions of the Royal Society (London), 128, 371-394. Williams, D., & Phillips, G. (1987). Cooperative phenomena in the perception of motion direction. Journal of the Optical Society of America ( A ) , 4 , 878-885. Yonas, A.. Cleaves, W., & Pettersen, L. (1978). Development of sensitivity to pictorial depth. Science, 200, 77-79. Yonas, A., Granrud, C. E., Arterberry, M. E., & Hanson, B. L. (1986). Infants’ distance perception from linear perspective and texture gradients. Infant Behavior and Development, 9, 247-256. Yonas, A., Granrud, C. E., & Pettersen, L. (1985). Infants’ sensitivity to relative size information for distance. Developmental Psychology, 21, 161- 167.

This Page Intentionally Left Blank

Stimulus Organization and Relational Learning SAL A. SORACI, JR. DEPARTMENT OF PSYCHOLOGY THE UNIVERSITY OF ALABAMA TUSCALOOSA, ALABAMA 35487

MICHAEL T. CARLIN DEPARTMENT OF PSYCHOLOGY A N D HUMAN DEVELOPMENT VANDERBILT UNIVERSITY NASHVILLE, TENNESSEE 37240

I.

INTRODUCTION

The ability to detect interstimulus relationships such as similarity-dissimilarity is an important aspect of attention and is critical for successful learning involving visual stimuli. The ability to detect relational information may be particularly relevant to an understanding of attentional and learning deficits in mentally retarded individuals. Nettlebeck and Brewer (198 l), for instance, have suggested that “attentional” deficits in mentally retarded individuals may be related to control processes involving the orientation of the individual to different aspects of the learning task. We have suggested that mentally retarded children do not have a deficit in control processes as such, but rather may have a lower sensitivity to relational information (Soraci, Deckner, Baumeister, & Carlin, 1990). Our research has also demonstrated that procedures designed to enhance the salience of stimulus relations may be used to reduce the performance discrepancies between high- and low-functioning persons. Our perspective on the relationship between relational learning and attentional functioning is that, traditionally, both the cognitive/mediational and the behavior-analytic traditions have underemphasized the complexity and influence of INTERNATIONAL REVIEW OF RESEARCll IN MENTAL RETARDATION. Vol. I8

29

Copyright D 1992 hy Academic Press, Inc All rights of reproduction in any form reserved

30

Sal A. Soraci, Jr., and Michael T . Carlin

stimulus organization. Cognitive perspectives have tended to focus on higherorder mediational processes, far removed from “perceptual” factors (Campione & Brown, 1984; Zeaman & House, 1979). Behavior-analytic perspectives, on the other hand, have been based on a stimulus control approach. The research has been directed at examining the acquisition of arbitrary stimulus relations (e.g., Sidman & Tailby, 1982). Such stimulus relations are not constrained by structural aspects of the stimulus array; that is, they are by definition arbitrary relations. The present article begins by raising the question of the relationship between perception and cognition and the importance of this issue for understanding the nature of cognitive representation. We then focus on the delineation of interstimulus parameters in relational learning and how particular perceptual organizations structure cognitions. Research paradigms reviewed involve manipulations of the structure of interstimulus relations, as well as manipulations of the internal structure of individual stimuli. We then provide an overview of important recent advances in perceptual approaches toward information processing (Fox & Oross, 1990; Treisman, 1988), which we believe represent fruitful new directions in the field. These include the importance of factors involved in the detection of relational stimulus structure and recent findings in neurophysiology on the existence of loci that mediate relational discriminative processes (Livingstone & Hubel, 1988). Finally, we point out the commonalities between our approach and recent advances in connectionist modeling (Rummelhart & McClelland, 1986).

A.

Perception and Cognition

Historically, there has been an ambiguous relationship between perception and cognition (Gibson, 1979; Palmer, 1978; Pinker, 1985; Rock, 1985; Shepard & Podgorny, 1978). Pinker (1989, for example, claims that the classic approaches to visual representation do not specify “where perception ends and where cognition begins” ( p. 13). Pinker explores perception and recognition processes in a very precise manner. He notes that processing of perceptual variables (geometric properties of the visual array, interstimulus contrast, contextual factors, etc.) is a critical “front-end” variable requisite for subsequent mediational or conceptual utilization. As pointed out previously, cognitive psychologists have been concerned with delineating the nature of internal representations and positing mediational mechanisms requisite for effective, higher-order processing, with a minimal concern for perceptual or “front-end” processing. Cognitive strategies and/or mediational mechanisms are, however, constrained by relatively primitive informationprocessing operations (Detterman, 1986; Hunt, 1986). From our perspective, the relationship between perceptual variables, attentional functioning, mediational processes, and intelligence is an important and direct one (Soraci, Deckner, Baumeister, & Carlin, 1990). A delineation of the

DETECTION OF STIMULUS ORGANIZATION

31

structure of stimulus organization is, we believe, critical for specifying the constraints on (1) the acquisition of rule-based behaviors and (2) the detection of stimulus similarity and difference by individuals with and without mental retardation. Rescorla ( I 988) has recently characterized learning processes as entailing “the learning of relations among events so as to allow an organism to represent its environment” (p. 15 1). We concur with this statement, and would claim that key questions for research include how organisms learn to perceive relations among objects and what conditions enhance such acquisition. To answer these questions, our research has focused on stimulus organization, examining stimulus structure in visual arrays, in an attempt to understand processes of rule acquisition and discriminative responding in mentally retarded and nonretarded individuals (Soraci, Deckner, Baumeister, & Carlin, 1990). Our approach is consonant with the Gibsonian emphasis on stimulus information and, as will be discussed subsequently, with recent advances in computational models of visual processing (Marr, 1982). We believe that our approach has relevance for an understanding of issues typically considered to be in the cognitive/developmental domain, as well as the potential for facilitating understanding of processes involved in the acquisition of rule-based skills requisite for effective problem solving (see Siegler, I99 1).

B.

Relational Learning

Gibson’s (1960, 1979) basic message for researchers in psychology was to delineate the nature of the stimulus. For those of us oriented toward cognitive/mediational approaches, Gibson’s challenge was a relevant one, that is, to examine more precisely the informational basis of perceptual, cognitive, and behavioral processes. The present investigators’ approach has not been instructional in a traditional sense (cf. Belmont & Butterfield, 1977), that is, involving a focus on inducing sophisticated cognitive/mediational and/or metacognitive/ mediational strategies. Instead, we have attempted to enhance effective problem solving and discriminative responding through structural manipulations of visual and auditory arrays that facilitate the detection of relevant stimulus relations. Relational learning, from this perspective, involves the learning of interstimulus parameters (Soraci, Deckner, Baumeister, & Carlin, 1990). Early investigators interested in relational learning (Kohler, 1929; Kuenne, 1946; Lawson, 1960) emphasized the importance of discriminating interstimulus properties. Whether the paradigms involved two-choice discriminations or transposition-type generalization studies, and whether the investigators studied infrahumans (Lawson, 1960) or humans (Berman & Graham, 1964), the acquisition of relational responding appeared to be a very basic ability shared by a wide range of organisms. The fact that most discriminations require relational re-

32

Sal A. Soraci, Jr., and Michael T . Carlin

sponding, whether they involve novelty-familiarity, similarity-difference, or presence-absence, for example, Terrace’s ( 1963) “feature positive” effect, indicates the importance of delineating the interstimulus structure of visual and auditory arrays (Soraci, Deckner, Baumeister, Bryant, Mackay, Stoddard, & McIlvane, 1991). In a recent article, Brown, Collins, and Duguid (1989) espoused the notion of studying “situated cognition” as a radical approach toward the development of instructional formats. As Brown et al. (1989) state, “A theory of situated cognition suggests that activity and perception are importantly and epistemologically prior-at a nonconceptual level-to conceptualization in that it is on them that more attention needs to be focused” (p. 41). One commonality between the situated cognition approach of Brown et al. and that of the present investigators is a strong emphasis on the importance of exploring and examining how a particular learning situation structures cognitions. This article reviews particular experimental manipulations that enhance the perceptual relations among stimuli. The present investigators have used such manipulations in facilitating relatively sophisticated conceptually based skills such as oddity and match-to-sample. Shepard and Podgorny ( 1978) have argued convincingly that internal processes tend to be “more perceptual in character than linguistic or verbal” (p. 224). Their basic premise is that internally generated representations should bear a certain verisimilitude or symmetry with respect to external structure. From this standpoint, a critical aspect of an understanding of the mapping between external and internal relations and how they relate to intelligence-related differences must be predicated on a careful delineation of interstimulus relations in visual and auditory arrays.

11.

STIMULUS ORGANIZATION

An early attempt to delineate the role of organizational deficits in intelligencerelated differences was undertaken by Spitz (1963, 1966). Spitz believed that many of the differences exhihited between persons with and without mental retardation on learning and memory tasks could be accounted for by a lesser ability in mentally retarded individuals to efficiently organize the stimulus materials. The focus of this research, therefore, was on identifying the relative contributions of higher-order cognitive processes and the lesser organizational abilities of individuals with mental retardation. This approach was driven primarily by two theoretical approaches to organization, Gestalt psychology and information theory. Gestalt theory places heavy emphasis on the importance of the interrelationships between stimuli and parts of stimuli. Information theory, on the other hand, is concerned with quantifying the amount of information available and the organism’s ability to efficiently process

DETECTlON OF STlMULUS ORGANlZATION

33

the available information. The marriage of these two approaches, along with the general finding that persons with mental retardation have a lesser ability to organize stimulus information, led Spitz (1966) to study the role of input organization in the processing abilities of people with mental retardation. Spitz (1966) reviews a number of studies focusing on identifying types of materials that could be efficiently processed by lower-functioning individuals. It is possible that such organizational parameters could be used to facilitate the learning of persons with mental retardation. One type of organization that seems to be processed efficiently by persons with mental retardation is symmetry (e.g., Hoats, Miller, & Spitz, 1963; House, 1966). Symmetric stimuli have been found to facilitate the performance of persons with mental retardation on numerous tasks, including visual search (Spitz & Borland, 197I), discrimination learning (House, 1966), and match-to-sample (Carlin & Soraci, 1991). Spitz ( 1966) went further in developing methods for preorganizing stimulus materials for individuals with mental retardation. On such tasks as digit recall and paired-associate learning, it was found that organizational parameters such as contiguity and amount of redundancy could be used to enhance the performance levels of persons with mental retardation to near that of their mental age (MA)-matched and chronological age (CA)-matched peers. Thus, Spitz was successful in devising means for organizing materials for efficient processing in individuals who do not recognize information-reducing aspects of stimuli or who fail to spontaneously organize such stimuli. The approach taken by Spitz, though similar in many ways to that of the present investigators, does differ in some important ways. The main concern of Spitz was increasing the amount of information entering the processing system and enhancing its subsequent retrieval from memory, that is, increasing the amount of information available to the subject at acquisition and retrieval. Our approach, on the other hand, focuses primarily on stimulus organization manipulations that serve to increase attention to relevant aspects of stimulus arrays (i.e., global organization) or the internal structure of individual stimuli (i.e., stimulus-specific organization). A.

Global Stimulus Organization

Stimulus organization influences a range of perceptual processes. Uttal(l983) has emphasized the importance of stimulus attributes in detection processes involving visual forms. From the perspective of relational learning, one can construe organization in both a global (i.e., interstimulus) and a local (i.e., stimulus-specific) sense (Soraci, Carlin, & Baumeister, in press). Structural manipulations of global visual arrays entail changes in the characteristics or properties shared by stimuli in visual arrays (see Fig. 1). An example would be the oddity task, in which the subject must choose the “odd” stimulus from a group

34

Sal A . Soraci. Jr., and Michael T . Carlin

of three stimuli, two of which are identical. On the other hand, stimulus-specific parameters involve organizational manipulations of the characteristics of a single stimulus. An example of the latter would be a target stimulus consisting of a particular type of symmetry that distinguishes it from nontarget stimuli. 1. THE ODDITY PARADIGM

The oddity paradigm is an example of a task that is contingent on the presence of multiple stimuli and, hence, dependent on the detection of relational information. Oddity learning requires the subject to learn to choose the distinct stimulus from a display that also contains a set of stimuli that do not differ from one another. For effective performance, the subject must attend to and compare a minimum of three stimuli. As such, the oddity test requires attention to the relevant dimension and the relations among stimuli (Stella & Etzel, 1986), especially if effective transfer and generalization are to occur (Soraci et a]., 1991). The selection of an odd stimulus has therefore been posited to involve a range of skills, including attentional (Scott & House, 1978), conceptual (Gollin & Schadler, 1972), and perceptual (Soraci, Deckner, Haenlein, Baumeister, Murata-Soraci, & Blanton, 1987) processes. It is the emphasis on perceptual variables that we pursue in this article. Most discriminations, whether they involve novelty-familiarity (as in studies of language acquisition using the exclusion paradigm), same-different judgments (as in the oddity paradigm), or transposition judgments (as in repeatedly choosing the larger of two triangles with new sets), require the use of relational information. In the studies that follow, the evidence suggests that it is a differential sensitivity to relational information per se, rather than an orientation to discrete stimulus information or deficient mediational processing, that may best characterize individual differences in relational learning. Previous studies have shown that children with and without mental retardation with MAS of less than 5 years (e.g., Ellis & Sloan, 1959) typically are unable to succeed at a variety of oddity learning and transfer tests (see reviews by Greenfield, 1985; House, Brown, & Scott, 1974). Older children, both with and without mental retardation, are usually able to learn the oddity concept when the task is designed to focus attention on the relevant dimension (e.g., Hall, 1971) and/or the oddity relation (e.g., Brown, 1970). Brown (1970), in a study of mentally retarded and nonretarded children with MAS greater than 6 years, found that the mentally retarded subjects were slower to learn the oddity concept. These intelligence-related differences were attributed to attentional limitations in the subjects with mental retardation. To remediate these deficits, Brown introduced a number of “attention-engineering” manipulations that were intended to increase attention to the relevant dimension, the oddity relation, or both. Significant enhancements in oddity performance resulted from these manipulations; the number of subjects with mental retardation

DETECTION OF STIMULUS ORGANIZATION

35

achieving criterion performance rose to a level comparable to that of MAmatched children. This was, perhaps, the first systematic study of the effects of specific task manipulations on the oddity performance of children with mental retardation. Kemler (1983) suggested that it is a lack of differentiation per se that characterizes perception in children with mental retardation. Supporting this claim, the task manipulation in the Brown (1970) study that resulted in the greatest performance enhancement was one that increased attention to the oddity relation. This manipulation involved increasing the number of nonodd, homogenous distractors in the visual array. This enhances perceptual salience through improvement in the “figure-ground” contrast of the target and nontarget stimuli. Although investigators such as Gollin, Saravo, and Salten (1967), Gollin and Schadler (1972), and Zentall, Hogan, Edwards, and Hearst ( I 980) increased the number of nonodd distractors in their studies, subjects were not always able to transfer effectively from arrays with many nonodd stimuli to arrays with few nonodd stimuli. In addition, transfer was a function of age, with younger children (MA < 5 years) less able to transfer to other stimulus contexts or dimensions (Gollin & Schadler, 1972). Thus, younger children were believed to acquire a restricted, context-specific “rule,” as opposed to the decontextualized, transferable rule acquired by the older subjects. Soraci, Alpher, Deckner, and Blanton (1983) and Soraci et al. (1987) postulated that this difficulty was perhaps due to an inadequate structuring of the stimulus presentation format. Specifically, certain previous studies used transfer test formats with horizontal, extended arrays (e.g., Gollin & Schadler, 1972). In such a stimulus format, ( I ) individual stimuli must be scanned separately, and (2) temporal parameters, and hence memory demands, are necessarily involved. The memory demands induced by the scanning of individual stimuli increase task difficulty, minimize interstimulus contrast, and hence decrease the opportunity to use relational information. To eliminate these problems, in the several oddity studies we have conducted, we have used a matrix on a CRT screen in which all stimuli, both odd and nonodd, can be viewed relatively holistically. This task arrangement minimizes the need for scanning, decreases memory demands, and maximizes interstimulus contrast. This type of array therefore increases the subject’s opportunity to use the relational information available in the task. Several studies in our laboratory have demonstrated the efficacy of perceptual salience manipulations in inducing rapid oddity learning and transfer in young children with and without mental retardation (Soraci et al., 1983, 1987). In these studies, the number of nonodd stimuli was increased from 2 during baseline to 8 in an intervention phase. The assumption underlying this intervention was that the perceptual salience of the odd stimulus, specifically its relationship to the nonodd stimuli, would be enhanced. Accurate oddity responding was shown to be rapidly acquired with the change from baseline intervention. Following ex-

36

Sul A. Soraci, Jr., and Michael T. Curlin

Baseline

Intervention

Return to Baseline (Transfer)

FIG. I. An intervention involving an increase in the number of nonodd stimuli in an array has been shown to induce learning of the oddity rule and to result in transfer back to the original threeelement oddity task.

posure to the nine-element arrays during intervention, the children also demonstrated successful transfer when presented again with three-element arrays, even though they had been unsuccessful with this minimal array throughout large numbers of trials during the initial baseline phase (Fig. 1). It should be noted that this transfer performance, as assessed by the stringent reversal trial assessment, was achieved in subjects at a developmental stage wherein such oddity responding has been regarded as atypical (i.e., MAS < 5 years). In addition, the Soraci et al. (1 987) study demonstrated transfer of oddity responding across stimulus types (i.e., novel geometric forms and arabic numerals), and maintenance of the oddity learning for a minimum of 6 weeks. 2. STIMULUS CONTIGUITY An issue not addressed in the previous oddity studies conducted by Soraci and colleagues was that the facilitation obtained with the nine-element oddity arrays could be attributed to structural factors of the visual array other than an increase in the number of nonodd elements per se. Such factors include an increase in stimulus contiguity. The manipulation involving a change from a three-element array to a nine-element array eliminated blank spaces and, concomitantly, increased interstimulus contiguity. Enhanced stimulus contiguity has been shown to be related to stimulus detection in several studies (Caruso & Detterman, 1983; Royer & Weitzel, 1977). In a study involving stimulus encoding, Soraci, Carlin, Deckner, and Baumeister ( 1990) demonstrated that subjects benefited from high levels of stimulus structure as defined by symmetry and number of adjacencies (i.e., contiguity) on a match-to-sample encoding task. To address the unresolved issue of the role of stimulus contiguity in oddity learning, Bryant and Soraci (1991) explored the relationship between number of array elements and spatial contiguity.

37

DETECTION OF STIMULUS ORGANIZATION

Array Size

4 x

U

. ‘ I

1 bD

9

Low

.H U

u d

cd

.rl U

High

m FIG. 2. Depiction of the four types of arrays used in the oddity study focusing on the interaction of the variables of array size and spatial contiguity.

Preschool children were presented with four- or nine-element stimulus arrays that were either high or low in spatial contiguity (Fig. 2). Oddity performance in the high-contiguity conditions was facilitated regardless of the number of array elements. In addition, analysis of error patterns as a function of reaction time supported the notion that significant visual scanning of individual stimuli improved response accuracy in the noncontiguous array configurations, but was not critical for enhanced performance in the spatially contiguous arrays. These findings are consistent with a paralleUseria1 visual scanning trade-off model (Treisman, 1988). In such contiguous arrays the odd stimulus can be considered as “popping out” and detection is relatively immediate. These results strongly suggest that in the oddity task, interstimulus spatial contiguity is a critical variable in inducing effective stimulus detection. Our studies share a common emphasis on the induction of a relatively sophisticated rule-based skill (i.e., oddity) via alterations in the interstimulus organization of particular visual arrays. Such manipulations have resulted in the rapid induction of oddity responding in children heretofore thought incapable of demonstrating such a “conceptual” skill. Attesting to the fact that the manipulation does not simply result in a response style predicated on the presence of perceptual saliency, effective transfer, generalization, and maintenance have also been shown to occur subsequent to rule induction (Soraci et al., 1991). Given past failures in inducing oddity performance both in young nonretarded children and in children with mental retardation (Greenfield, 1985), the previous results are encouraging in several ways. They indicate that discriminations such as those involved in rule induction can be achieved via structural manipulations of stimulus arrays. This finding has theoretical and applied significance. With

38

Sal A . Soraci, Jr., and Michael T . Carlin

respect to theory development, it suggests the possibility that individuals differing in intellectual functioning can be characterized as having a differential sensitivity to relational information (Soraci et a]. , in press), as opposed to a mediational/structural deficiency. In terms of practical import, it is clear from the previous work that such manipulations also effect successful transfer, generalization, and maintenance of a newly acquired, rule-based skill. 3. PREDISPOSED SALIENCE AND ODDITY

As pointed out previously, successful performance on the oddity task requires observation of both the relevant vehicle dimension (e.g., form, color, or size) and the oddity relation among the cues on the vehicle dimension (Brown, 1970). A predisposition to make either of these observing responses should enhance the learning of the oddity concept (Brown, 1970; Martin & Qrrell, 1971). The approaches to facilitating the learning of the oddity concept discussed thus far have focused on increasing attention to the oddity relation (e.g, Gollin et al., 1967; Soraci et a]., 1987, 1991). Attention is now given to methods for increasing observation of the vehicle dimension on which the oddity relation is based. Observation of the vehicle dimension is influenced primarily by three factors: ( 1 ) the predisposed salience of the dimension for the individual child, (2) the magnitude of the physical differences between cues, and (3) the differential habit strengths of the specific cues on the vehicle dimension (Campione & Wentworth, 1969). For the purposes of the present paper, only studies focusing on predisposed salience are discussed. Little research has explored the role of the latter two factors in relation to the oddity task. Studies focusing on the effect of the predisposed salience of the vehicle dimension on oddity performance have either assessed the children preexperimentally for their individual dimensional preferences or trained the children to attend to a certain dimension prior to the oddity phase of the experiment. The majority of studies assessing the dimensional preferences of the children have replicated the finding of a developmental progression from color dominance in young children (e.g., 3- and 4-year-olds) to form dominance in children over 5 years of age (e.g., Brian & Goodenough, 1929; Suchrnan & Trabasso, 1966). Wright and Vlietstra ( 1 975) depicted this developmental shift in dimensional preference as a change in selective attention. Older children tend to focus on form and not color, perhaps based on their assumption or estimate that form is a more informative dimension. Aschkenasy and Odom ( 1982) discussed this developmental progression as the result of the discovery with age of more dimensions in which to divide visual scenes. These dimensions then become differentially salient to the child through experience. In contrast to preexperimentally assessing the salience of a given dimension, dimensional training artificially boosts observation of a particular dimension. The training usually consists of a two-choice match-to-sample task in which one

DETECTION OF STIMULUS ORGANIZATION

39

of the matching stimuli is identical to the sample on one dimension (e.g., color) but not the other dimension (e.g., form), and the opposing matching stimulus is identical to the sample on the latter dimension and different on the former dimension. In this way, one matching stimulus may be identical in form but of a different color than the sample, and the second matching stimulus may be identical in color to the sample but different in form. The child is reinforced for matching on one dimension and simultaneously taught not to select on the basis of the unreinforced dimension. Training is then carried out to some predefined criterion or total number of trials. Both methods of increasing attention to the vehicle dimension have resulted in positive findings. That is, nonretarded children (e.g., Brown, 1970) and children with mental retardation (Brown, 1970; Hall, 1971; House, 1989) perform much more efficiently on an oddity task when the relevant dimension is of high salience than when the relevant dimension is of low salience. An important limitation of these studies, however, is that only older children (MA > 5 years) have been studied. It has yet to be determined whether younger children who typically cannot perform above chance on the minimal three-choice oddity task will benefit from dimensional training as much as they do from procedures designed to enhance observation of the oddity relation. As previously discussed, important differences between the oddity learning of subjects with and without mental retardation have been documented. House ( 1989) demonstrated that persons with mental retardation maintain a longer period of chance performance prior to the initiation of learning. It takes these subjects longer to learn to attend to the relevant aspects of the task; however, once learning begins, the rate of learning is identical for the two groups (Brown, 1970). One reason for this group difference is the perseveration of negative response strategies that interfere with the learning of the oddity concept. On each oddity trial, the position of the odd form, the specific odd stimulus value, and the oddity relation are reinforced (Penn, Sindberg, & Wohlhueter, 1969). Thus, it is likely that younger children and children with mental retardation, who are typically more cognitively rigid in their strategy usage or automatic modes of response (cf. Ellis & Dulaney, 1991), may persist in the use of these negative response types. The rigidity of the negative strategies therefore interferes with the development of responses to oddity. The likelihood of these negative response strategies being eliminated can be enhanced in many instances through the use of relatively simple task manipulations (Croll, 1970; Gollin & Schadler, 1972; Scott & House, 1978; Soraci et al., 1991). Croll (1970) studied several types of inappropriate responding that can occur in the context of an oddity task. These negative response strategies included (1) repeated choice of a particular stimulus (stimulus perseveration) or position (position perseveration) and (2) alternating choice between particular stimulus

40

Sal A. Soraci, J r . , and Michael T . Carlin

values (stimulus alternation) and positions (position alternation). Croll ( 1970) found that the strategy most commonly used by young children on his oddity task was stimulus perseveration, with use of the other strategies relatively infrequent. The high frequency of stimulus perseverative responses was due most likely to the use of a limited stimulus set. The stimuli used included only two forms, two colors, and two sizes. Thus, a high degree of stimulus repetition was necessary across oddity trials. Children could therefore attain a relatively high rate of reinforcement for use of such a strategy. Scott and House (1978) confirmed that stimulus repetition interferes with oddity performance. Children were assigned to either a repeating set condition or a nonrepeating set condition. Not only did a greater proportion of children in the nonrepeating set condition learn the oddity rule, but they also learned the rule more quickly. A similar task manipulation was employed by Collin and Schadler (1972) to decrease the probability of reinforcement for positional perseveration responding. The number of possible positions in which the three oddity test stimuli could appear was increased from 3 to 9. This simple manipulation eliminated positional responding and increased the likelihood of the oddity rule being learned. The contextual factors related to task format investigated by Croll (1970), Scott and House ( 1 978), and Gollin and Schadler ( 1972) do not involve direct manipulations of stimulus organization; however, their findings indicate the importance of task-specific factors in oddity learning. Future research should assess whether young children with and without mental retardation will benefit from procedures designed to increase attention to the relevant vehicle dimension, and if so, whether these procedures are more efficient than those meant to enhance observation of the oddity relation (Brown, 1970). Ideally, an oddity task arrangement may be found that maximizes efficiency of learning through task manipulations focusing attention on the relevant vehicle dimension, the oddity relation, or both. Our emphasis on the importance of directing the child’s attention to relevant dimensions of difference is consistent with the learning theories of both Gibson (1969) and Zeaman and House (1963, 1979). The relational learning exhibited in the context of oddity consists of the growth of the ability to extract information from the learning environment that is relevant for task solution. Relational learning in this context is construed as being dependent on attentional/perceptual processes and not symbolic mediation (House et al., 1974). 4. BIMODAL RELATIONAL INFORMATION We have suggested that mentally retarded individuals may have a lower sensitivity to relational information and that this lower sensitivity is amenable to interventions that make the relationships among stimuli more salient. Most of the studies cited involve manipulations of visual stimuli only. A general question is

DETECTION OF STIMULUS ORGANIZATION

41

whether this lower sensitivity is domain specific or reflects a more general phenomenon. Kinney and Kagan (1976) demonstrated that nonretarded infants’ cardiac response to alterations in the pattern of six auditory tones was characterized by a curvilinear function. Cardiac deceleration was greatest at moderate levels of discrepancy and least at both minimal and high levels of discrepancy. Chang and Trehub (1977) extended the work of Kinney and Kagan not only by reordering the tonal patterns, but by transposing the pattern, thus preserving the pattern of stimulus relations while changing all the individual tonal elements. The results indicated that infants demonstrated a cardiac deceleration response to a simple reordering of the initial tones, but did not demonstrate deceleration when the specific tones of the pattern were transposed. Soraci, Barlean, Haenlein, and Baumeister (l986), in a variation of the Chang and Trehub (1977) study, compared the sensitivity of mentally retarded and nonretarded adults with alterations in tonal relationships. It was found that cardiac deceleration was indeed curvilinearly related to the magnitude of discrepancy between tonal patterns. More importantly, the curvilinear function was of a lower magnitude in the mentally retarded as compared with the nonretarded subjects. The results suggested a lower sensitivity in mentally retarded individuals to alterations in relational tonal structure. These findings also support the hypothesis that the lower sensitivity of mentally retarded individuals to relational information is cross-modal and may thus constitute a general phenomenon. This view is consistent with Gibson’s (1979) position that the detection of invariance is independent of modality. In a recent study, Soraci et al. (1991) demonstrated the facilitation of oddity acquisition in preschool children via a bimodal training procedure. Instead of manipulating the three-element visual array by increasing the number of nonodd stimuli, they presented auditory tones concurrently with visual stimuli. The manipulation was designed to provide redundant relational information in two modalities simultaneously. Put another way, the odd stimulus not only “looked different” but also “sounded different.” All subjects acquired visual oddity responding as a function of the auditory intervention on both reversal and novel set trials. In addition, the subjects demonstrated generalization to a visual-only condition on return to baseline. The auditory intervention was interpreted as a method of increasing the discrepancy (and hence the observing of differences) between the odd and nonodd stimuli by perceptual enhancement of the target in two modalities concurrently. The aforementioned results with auditory manipulations have implications for theories of attentional deficits in the mentally retarded. The results suggest that mentally retarded individuals may not have, as several investigators have suggested (Bradley-Johnson & Travers, 1979; Krupski, 1975; Porges & Humphrey, 1977). a gross attentional deficit mediated by a weak orienting response. An

42

Sal A. Soraci, Jr., and Michael T . Carlin

alternative explanation based on the studies described above suggests that mentally retarded individuals may be less able than nonretarded individuals to detect the relational characteristics among stimuli. Basic interstimulus relationships such as similarity-dissimilarity are embedded in both visual and auditory arrays. Thus, a differential sensitivity to relational information in mentally retarded and nonretarded individuals may be an essential factor in accounting for performance differences on experimental tasks regardless of modality (Soraci et al., in press). 6.

Stimulus-Specific Organization

Stimulus-specific organization refers to the structure of individual visual stimuli. Research to date has focused on the effect on encoding processes of varying the structure of such stimuli. The first to investigate the phenomenon in persons with mental retardation were Caruso and Detterman (1983). The stimuli used by these investigators were 16-cell checkerboard-type matrix patterns (Fig. 3). These stimuli are particularly informative in that three structural variables may be manipulated within each stimulus; symmetry, number of adjacencies, and number of cells filled. Caruso and Detterman (1983) concluded from their study that persons with mental retardation and their nonretarded, CA-matched peers processed the stimuli in a similar manner; however, inspection times were unlimited, thus minimizing memory demands, and a four-choice match-to-sample format was used that did not allow for direct examination of target-distractor disparity effects. Using similar checkerboard-type stimuli, Soraci, Carlin, Deckner, and Baumeister ( 1990) assessed the relation between stimulus organization and encoding in persons with and without mental retardation, with CAs between 15 and 18. Three important methodological differences existed between this study and that of Caruso and Detterman (1983). First, stimulus organization was manipulated across high, moderate, and low levels of structure. This allowed for a direct comparison of the encodability of forms varying in structure for each of the groups. Second, a two-choice match-to-sample format was used to permit direct study of target-distractor disparity effects. Finally, memory demands were increased by limiting target stimulus exposures to less than 150 milliseconds and instituting a 1-second interval between sample offset and presentation of the two comparison stimuli. Results demonstrated that persons with and without mental retardation more effectively detected stimuli when there was either high structural organization (e.g., symmetry) of the target or distractor stimulus or increased target-distractor structural disparity (e.g., highly structured target stimulus and low-structure distractor stimulus). Increased disparity has also been shown to enhance discrimination learning (Dinsmoor, 1985) and proficiency of visual search (Spitz & Borland, 1971) in persons with and without mental retardation.

43

DETECTION OF STIMULUS ORGANIZATION

ADJACENCIES

High

Low

FIG. 3. Examples of checkerboard-type stimuli defined by the interaction of the variables of symmetry and number of adjacencies.

An important intelligence-related difference was identified. Under conditions of reduced interstimulus disparity (e.g., with a moderate-structure distractor), subjects with mental retardation performed much more poorly than did the nonretarded subjects (Fig. 4). The individuals with mental retardation thus exhibited a lower sensitivity to the available relational information. This information was specified by conjunctions of symmetry and number of adjacencies within the individual stimuli. These findings suggest that the differential detection of moderate levels of stimulus organization could be responsible for intelligence-related differences on a wide range of tasks, as most experimental and “real-world” stimuli are composed of mixed organizational parameters (Soraci, Carlin, Deckner, & Baumeister, 1990). A further study (Carlin & Soraci, 1991) was designed to determine whether similar intelligence-related differences would be exhibited with formlike stimuli varying solely with respect to symmetry; that is, stimuli were either vertically symmetric, horizontally symmetric, doubly symmetric (i.e., both vertical and horizontal), or asymmetric (Fig. 5). Methods used were identical to those in the Soraci, Carlin, Deckner, & Baumeister (1990) study. Vertically symmetric stimuli were processed more efficiently than the other types of symmetric stimuli. Interestingly, no intelligence-related differences were found. The lack of intelligence-related differences indicated that, at least for manipulations involving symmetry, the groups were equally sensitive to the relevant dimension of difference and the relational information embedded in the task. These results are at odds with previous findings (Soraci, Deckner, Baumeister, & Carlin, 1990) that indicated a lower sensitivity to relational information in persons with mental retardation on an encoding task. This discrepancy may have resulted from the fact that the perception and discrimination of symmetries per se, in the absence of internal structure (i.e., contiguity), are more basic processing abilities.

Sal A . Soruci, J r . , and Michael T . Curlin

44 100

Nonretarded

90

I

80 +

0

t?

b

70 60

0

+

50

0

:100 r L

MODERATE

HIGH

LOW

= 0

Target High

Moderate

m Low

Mentally Retarded

C

m

2!

901

1

I

80

70 60 50

I

MODERATE LOW Distractor Structure

HIGH

FIG. 4. Depiction of the Target X Distractor structure interactions for nonretarded and mentally retarded subjects. Note the depressed performance of the mentally retarded subjects in the rnoderatestructure distractor condition.

The results of these studies indicate that persons with mental retardation are able to perform fairly well on a speeded encoding task when stimuli vary along only a single dimension (e.g., symmetry). When stimuli are defined by a conjunction of symmetry and number of adjacencies, however, subjects with mental retardation perform more poorly than CA-matched peers under conditions of reduced interstimulus disparity. This group difference makes salient the locus of an encoding deficiency, which involve central scanning mechanisms, in persons with mental retardation.

Ill.

APPROACHES EMPHASIZING EARLY STAGES OF PROCESSING

The discussions of the effects of structural manipulations on performance in the contexts of oddity and match-to-sample (i.e., checkerboard organization

45

DETECTION OF STIMULUS ORGANIZATION

A

B

C

D

FIG. 5 . Examples of the polygon stimuli used in the study: (A) vertical and horizontal symmetry,

(B)vertical symmetry, (C)horizontal symmetry, (D)asymmetry.

studies) have focused on attentional or encoding processes. The effects of structural manipulations on more primitive or “preattentive” stages of information processing have also been investigated (e.g., Fox & Oross, 1988, 1990). The work of Fox and Oross (1988, 1990), which is described in more detail by Fox and Oross in this volume, has focused on the ability of persons with mental retardation to process random element stereograms or kinematograms. In an early study of the processing of random element stereograms by persons with and without mental retardation, Webb (1972) concluded that the binocular processing of the subjects with mental ,retardation did not differ from that of the nonretarded subjects; however, the stereograms used by Webb were such that they placed minimal demands on the computational resources of the participants. Fox and Oross (1988, 1990), on the other hand, have manipulated the structure of the stereograms and kinematograms used in their studies by reducing dot density and/or temporal correlation, thereby increasing the demands placed on the subject’s computational resources (e.g., interpolation processes). With such reductions in the coherence of the forms, the performance of the subjects with mental retardation decreased dramatically, whereas that of the oonretarded subjects remained at or near ceiling. Additionally, these deficits have been shown to be resistant to improvement through training (Oross, Carlin, & Fox, 1990). This pattern of results may imply a deficit in an early stage of perceptual processing in persons with mental retardation when excessive demands are placed on the processing system. This research also identifies another circumstance in which intelligence-related processing differences are associated with alterations of the internal structure of individual stimuli.

46

Sal A . Soraci, Jr., and Michael T . Carlin

Another methodology for studying preattentive processes has been used by Treisman and her colleagues (e.g., Treisman & Gelade, 1980; Treisman & Paterson, 1984; Treisman & Souther, 1985). They have focused on identifying visual features that are automatically coded or extracted at a preattentional processing stage. Typically, a visual search task is used in which a pair of structurally simple stimuli with contrasting values (e.g., long versus short lines) on a specific dimension (e.g., length) are compared. Each stimulus value characterizes the target in one condition and the distractor(s) in the other condition. A feature is believed to be coded early in visual processing, or to “pop out,” if it is detected with little or no increase in latency as display size increases. The assumption is that a parallel search occurs in such a display. The identification of such emergent features provides critical information about the way in which the perceptual world is initially analyzed. Treisman and Paterson ( 1984) identified three defining characteristics of emergent features: (1) a flat function relating detection latency to number of distractors (i.e., a “pop-out’’ effect), (2) an ability to induce texture segregation, and (3) an inclination for forming illusory conjunctions when attentional capacity is overloaded. To date, the list of emergent features is limited to separable features within a single dimension (e.g., color). It should be noted that the identification of emergent features is complicated by a number of factors. First, search asymmetries have been the norm in that only one member of a pair of cues will exhibit the “pop-out,’’ parallel-search effect (Gurnsey & Browse, 1987; Treisman & Souther, 1985). As a result, search asymmetry itself has become one of the defining traits of the pop-out phenomenon. Generally, the member of a stimulus pair that exhibits pop-out is the member that would require more activity in a hypothetical population of feature detectors (Treisman & Gormican, 1988). A second complicating factor is that the search tasks used in studies of the pop-out effect are susceptible to the influences of higher-order cognitive processes. Specifically, the predisposed salience, or familiarity, of the vehicle dimension and task procedures has been shown to influence task performance (Gurnsey & Browse, 1987; Kinchla, 1974; Schneider & Schiffrin, 1977). Additionally, the ability to use information from instructions (e.g., percentage trials on which the target is present) can potentially alter the response criteria employed by subjects. Finally, the specification of the individual target stimulus may create an internal perceptual structure or set that also may affect response criteria (Kinchla, 1974; Williams, 1966). For example, subjects may be biased to decide in favor of target presence with relatively little evidence for such a decision (Duncan, 1980). To combat such difficulties, the majority of studies of this phenomenon have used highly practiced subjects. Finally, the question remains whether the information-processing system is coding stimulus relations (e.g., “longer than”) or specific distinctive perceptual features (e.g., “symmetry”). Central to this problem is defining the role of the distractor stimuli in the search task. Future research should explore whether pop-

DETECTION OF STIMULUS ORGANIZATION

47

out is facilitated by target-surround relations or an internal comparison of the target item with an internal representation of the target stimulus. The specification of emergent features through use of a simple guided-search task is a promising area for comparative research. It would be of interest whether the emergent features identified thus far with nonretarded subjects are similarly emergent for persons with mental retardation. If not, the locus of the deficit may lie in the early stages of visual processing.

IV.

NEUROPHYSIOLOGICAL CORRELATES

As we have seen with respect to the structural manipulations discussed above, variables involving relational information and stimulus structure have been operationalized in relatively precise ways. Important differences in the detection of stimulus organization between mentally retarded and nonretarded individuals have been demonstrated (Soraci, Carlin, Deckner, & Baumeister, 1990). Presumably, there is a neurophysiological substrate involved in the processing of relational information and other aspects of stimulus organization. In recent years, neuroscientists have examined in great detail the layered structures that compose the lateral geniculate nucleus (LGN) in the primate visual system (Desimone & Ungerleider, 1989; Hubel & Livingstone, 1987; Shapley, 1990). The LGN has an architecture comprising six layers: four dorsal, small-celled (or parvocellular) layers and two ventral, large-celled (or magnocellular) layers. Livingstone and Hubel (1988) have pointed out the functional distinctions between the magnocellular and parvocellular layers in recent experimentation involving stimuli that vary with respect to variables such as spatial frequency, spectral sensitivity, relative motion, acuity, and other perceptually related factors. The basic consensus is that the magnocellular layers are responsible for the processing of depth and movement cues. Low- to moderate-contrast stimuli and lower-acuity stimuli are the types of information processed by these layers. According to Livingstone and Hubel (1988) and other researchers, magnocellular layers are responsible for the detection of relative movement, that is, spatial displacement of one stimulus relative to a reference point. The magnocellular layering is also involved in processing general organizational variables of visual stimulation. In addition, figure-ground discriminations are assumed to be mediated by the magnocellular layers (Livingstone & Hubel, 1988). In contrast, the parvocellular layers are responsible for processing high-resolution information and information related to specific or discrete, and not relational, stimulus features. For example, the identification of surface properties and independent stimulus attributes such as color and angular characteristics is assumed to be mediated by the parvocellular system. A potential theoretical linkage with our work on relational information is

48

Sal A . Soraci, J r . , and Michael T . Carlin

possible if we assume that the discrimination of depth, relative movement, and figure-ground distinctions essentially involves detection of interstimulus parameters such as spatial distance, position changes, and similarity-dissimilarity. A critical question for future research, therefore, is whether the magnocellular layers are an important locus of intelligence-related differences.

V.

THE CONNECTIONIST PERSPECTIVE AND RELATIONAL LEARNING

Contemporary connectionist models have emerged in cognitive psychology (McClelland & Rummelhart, 1986; Rummelhart & McClelland, 1986) that address relatively complex learning processes via a nonmediational network model. Because they are nonmediational, they are consistent with our basic premise that relatively “higher-order’’ associative learning is dependent on and, in some sense, derivative of basic interstimulus relations and particular properties they instantiate. These models assume that computations are performed via the interaction of an extensive network of nodes or units. Simulations of complex behaviors such as rudimentary speech perception and category learning have demonstrated that an extremely sophisticated output can be simply the result of the interaction of nodes in the network. Differential “weighting” of values in the nodes and, typically, a high frequency of iterations of particular learning trials characterize the “learning” resulting in the acquired behavior. The classic notion of computer simulation involving higher thought processes assumed the utilization of rules, inferences, and hypotheses as integral parts of symbolic, computational systems. From the connectionist perspective, however, there is no need to posit the existence of an internal symbolic medium. As Hatfield ( 1986) has stated, the functional algorithms involved in such connectionist models are nonsymbolic and involve lawful relationships between stimulus input and output. Such algorithms are instantiated by the neural net. What is particularly interesting from the perspective of the present article is the notion of information in the connectionist perspective. According to Hatfield, “Information is defined in terms of the relation between real world events” (p. 10). The relations Hatfield (1986) is discussing, in a similar way as Rescorla (1988), involve patterning of light at the retina, geometric properties of the visual array, and other aspects of interstimulus relations. We would claim that experimental manipulations involving modifications of stimulus array structure that make salient the critical dimensions of difference do not necessitate mediational or symbolically based rule induction. Connectionist models that involve learning paradigms resulting in complex outputs have a commonality with our perspective on learning as the acquisition or detection of particular interstimulus relations. One important area for future research will be

DETECTION OF STIMULUS ORGANIZATION

49

the simulation of individual differences in relational learning using connectionist models.

VI.

CONCLUSION

The ability to detect basic interstimulus relationships such as similarity-dissimilarity is critical for the successful discrimination of complex sources of stimulation. There is substantial evidence that children of low mental ages, both mentally retarded children and nonretarded children, exhibit markedly deficient performance on tasks that demand relational responding (Greenfield, 1985; Soraci, Carlin, Deckner, & Baumeister, 1990; Soraci, Deckner, Baumeister, & Carlin, 1990; Soraci et al., 1987, 1991). Converging evidence from a number of studies conducted by the present investigators and others involving manipulations of stimulus organization suggests that an essential factor in the performance discrepancies between mentally retarded and nonretarded individuals is a differential sensitivity to relational information.

ACKNOWLEDGMENTS This research was supported by National Institute of Child Health and Human Development Research Grants HD-18754; HD-27336, awarded to the John F. Kennedy Center for Research on Education and Human Development at Vanderbilt University; and IK04 HD-00921-01, awarded to The University of Alabama. We appreciate the helpful comments provided by Norm Ellis and Theodore Hoehn on an earlier version of this article.

REFERENCES Aschkenasy, J. R., & Odom, R. D. (1982). Classification and perceptual development: Exploring issues about integrality and differential sensitivity. Journal of Experimental Child Psychology, 34, 435-448. Belmont, J. M., & Butterfield, E. C. (1977). The instructional approach to developmental cognitive research. In R. V. Kail & J. W. Hagen, Jr. (Eds.), Perspectives on the development ofmemory and cognition (pp. 437-481 ). Hillsdale, NJ: Erlbaum. Berman, P. W., & Graham, F. K. (1964). Children's response to relative, absolute, and position cues in a two-trial size discrimination. Journal of Compararive and Physiological Psychology, 57, 393-397. Bradley-Johnson, S . , & Travers, R. M. W. (1979). Cardiac change of retarded and nonretarded infants to an auditory signal. American Journal of Menral Deficiency, 83, 631-636. Brian, C . R., & Goodenough, F. L. (1929). Relative potency of color and form perception at various ages. Journal of Experimenral Psychology, 12, 197-2 13. Brown, A. L. (1970). Subject and experimental variables in the oddity learning of normal and retarded children. American Journal of Mental Deficiency, 75, 142-151.

50

Sal A. Soraci, Jr., and Michael T. Carlin

Brown, J. S., Collins, A,, & Duguid. P. (1989). Situated cognition and the culture of learning. Educational Researcher, 18, 32-42. Bryant, J. T., & Soraci, S. A.. Jr. (1991). Oddity facilitation as a function of array size and spatial contiguity. Manuscript submitted for publication. Campione, J. C., & Brown, A. L. (1984). Learning ability and transfer propensity as sources of individual differences in intelligence. In P. H. Brooks, R. Sperber, & C. McCauley (Eds.), Learning and cognition in the mentally retarded (pp. 265-293). Hillsdale, NJ: Erlbaum. Campione, J. C., & Wentworth. P. C. (1969). Differential cue habit strength as a determinant of attention. Journal of Experimental Psychology, 82, 527-53 I . Carlin, M. T., & Soraci, S. A., Jr. (1991). Similarities in the detection of stimulus symmetry by persons with and without mental retardation. Manuscript submitted for publication. Caruso, D. R., & Detterman, D. K. ( 1983). Stimulus encoding by mentally retarded and nonretarded adults. American Journal of Mental Deficiency, 87, 649-655. Chang, H. W., & Trehub, S. E. (1977). Auditory processing of relational information by young infants. Journal of Experimental Child Psychology, 24, 324-331. Croll, W. L. (1970). Response strategies in the oddity discrimination of preschool children. Journal of Experimental Child Psychology, 9 , 187-192. Desimone, R., & Ungerleider, L. A. (1989). Neural mechanisms of visual processing in monkeys. In F. Boiler and J. Gratman (Eds.), Handbook of neuropsychology (Vol. 2, pp. 267-299). AmsterdamlNew York: Elsevier Science. Detterman, D. K. (1986). Human intelligence is a complex system of separate processes. In R. 1. Sternberg & D. K. Detterman (Eds.), Whar is intelligence? (pp. 57-61). Nonvood, NJ: Ablex. Dinsmoor, J. A. (1985). The role of observing and attention in establishing stimulus control. Journal of the Experimental Analysis qffehavior, 4 3 , 365-381. Duncan, J. (1980). The demonstration of capacity limitation. Cognitive Psychology, 12, 75-96. Ellis, N. R.. & Dulaney, C. L. (1991). Further evidence of cognitive inertia of persons with mental retardation. American Journal on Mental Retardation, 95, 613-622. Ellis, N. R., & Sloan, W. (1959). Oddity learning as a function of mental age. Journal of Comparative and Physiological Pswhology, 52, 228-230. Fox. R., & Oross, S., 111. (1988). Deficits in stereoscopic depth perception by mildly mentally retarded adults. Americun Journal on Mentul Retardation, 93, 232-238. Fox, R., & Oross, S . , 111. (1990). Mental retardation and the perception of global motion. Perception & Psvchophysics, 4 8 , 252-258. Gibson, E. J. (1969). Principles of perceptual learning and development. New York: AppletonCentury-Crofts. Gibson, J. J. (1960). The concept of the stimulus in psychology. American Psychologist, 15. 697703. Gibson, J. J. (1979). The ecological approach to visual perception, Boston: Houghton-Mifflin. Collin, E. S., Saravo, A,, & Salten, C. (1967). Perceptual distinctiveness and oddity-problem solving in children. Journal of Experimental Child Psychology, 5, 586-596. Collin, E. W., & Schadler, M. (1972). Relational learning and transfer in young children. Journal of Experimental Child Psychology. 14, 219-232. Greenfield, D. B. ( I 985). Facilitating mentally retarded children's relational learning through novelty-familiarity training. American Journal of Mental Deficiency, 90, 343-348. Gurnsey, R.,& Browse, R. A. (1987). Micropattem properties and presentation conditions influencing visual texture segregation. Perception & Psychophysics, 4 1 , 239-252. Hall, J. E. (1971). Effect of response bias of mental retardates upon oddity learning. American Journal of Mental Deficiency, 75. 579-585. Hatfield, G. ( 1986). Information and representation in noncognitive explanations of perception. Baltimore, MD: Cognitive Neuropsychology Laboratory, Johns Hopkins University.

DETECTION OF STIMULUS ORGANIZATION

51

Hoats, D. L., Miller, M. B., & Spitz, H. H. (1963). Experiments on perceptual curiosity in mental retardates and normals. American Journal of Mental Deficiency, 68, 386-395. House, B. J. (1966). Discriminations of symmetrical and asymmetrical dot patterns by retardates. Journal of’Experimental Child Psychology. 3 , 377-389. House, B. J. (1989). Some current issues in children’s selective attention. In H. W. Reese (Ed.), Advances in childdevelopment and behavior (Vol. 22, pp. 91- 120). New York: Academic Press. House, B. J., Brown, A. L., & Scott, M. S. (1974). Children’s discrimination learning based on identity or difference. In H. W. Reese (Ed.), Advances in child development and behavior (Vol. 9, pp. 1-45). New York: Academic Press. Hubel, D. H., & Livingstone, M. S. (1987). Segregation of form, color, and stereopsis in primate area 18. Journal of Neuroscience, 7 , 3378-3415. Hunt, E. (1986). The heffalump of intelligence. In R. J. Sternberg & D. K.Detterman (Eds.), What is intelligence? (pp. 101- 107). Nonvood, NJ: Ablex. Kemler, (1983). Holistic and analytic modes in perceptual and cognitive development. In T. Tighe & B. E. Shepp (Eds.), Perception, cognition, and development: Interactional analyses (pp. 71102). Hillsdale, NJ: Erlbaum. Kinchla, R. A. (1974). Detecting target elements in multielement arrays: A confusability model. Perception & Psychophysics, 15, 149- 158. Kinney, D. K., & Kagan, J. (1976). Infant attention to auditory discrepancy. ChildDeveIopmenr, 4 7 , 155-164.

Kohler, W. (1929). Gestalt psychology. New York: Liveright. Krupski, H. (1975). Hcart rate changes during a fixed reaction time task in normal and retarded adult males. Psychophysiology, 12, 262-267. Kuenne, M. R. (1946). Experimental investigation of the relation of language to transposition behavior in young children. Journal of Experimental Psychology, 36, 47 1-490. Lawson, R. (1960). Learning and behavior. New York: MacMillan. Livingstone, M., & Hubel, D. (1988). Segregation of form, color, movement, and depth: Anatomy, physiology, and perception. Science, 240, 740-749. Marr, D. (1982). Vision. San Francisco: Freeman. Martin, A. S., & Qrrell, D. J. (1971). Oddity learning following object-discrimination learning in mentally retarded children. American Journal of Mental Deficiency, 75, 504-509. McClelland. J. L . , & Rummelhart, D. E. (1986). Parallel distributed processing (Vol. 2). Cambridge. MA: MIT Press. Nettlebeck, T., & Brewer, N. (1981). Studies of mild mental retardation and timed performance. In N. R. Ellis (Ed.), International review of research in mental retardation (Vol. 10, pp. 62-106). New York: Academic Press. Oross, S., 111, Carlin, M. T., & Fox, R. (1990, April). Motion perception deficits in mildly mentally retarded adults: An experimental attempt to ameliorate deficits via training. Paper presented at the Gatlinburg Conference on Research and Theory in Mental Retardation and Developmental Palmer, S. E. (1978). Fundamental aspects of cognitive representation. In E. Rosch & E. R. Lloyd (Eds.), Cognition und categorization ( pp. 259-303). Hillsdale, NJ: Erlbaum. Penn, N. E., Sindberg, R. M., & Wohlhueter, M. J. (1969). The oddity concept in severely retarded children. Child Development, 40, 154-161. Pinker, S. (1985). Visual cognition: An introduction. In S. Pinker (Ed.), Visual cognition (pp. 1-63). Cambridge, MA: MIT Press. Porges, S. W., & Humphrey, M. H. (1977). Cardiac and respiratory responses during visual search in nonretarded children and retarded adolescents. American Journal of Menral deficiency, 82, 162- 169. Rescorla. R. A . (1988). Pavlovian conditioning: It’s not what you think it is. American Psychologist, 43, 1.5-160.

52

Sal A . Soraci, Jr., and Michael T . Carlin

Rock. 1. (1985). Perception and knowledge. Acra Ps.vchologica. 59, 3-22. Royer, F. L., & Weitzel, K. (1977). Pemeptual cohesiveness of pattern in the block design task. Perception & Psychophysics, 2 1 , 39-46. Rummelhart, D. E., & McClelland, J. L. (1986). Parallel distributed processing (Vol. I). Cambridge, MA: MIT Press. Schneider. W., & Shiffrin, R. M. (1977). Controlled and automatic human information processing: 1. Detection, search and attention. Psychological Review, 8 4 , 1-66, Scott, M. S., & House, B. J. (1978). Repetition of cues in children's oddity learning and transfer. Journal of Experimental Child Psychology, 25, 58-70. Shapley, R. (1990). Visual sensitivity and parallel retinovertical channels. Annual Review ofPsycho1ogy, 41, 635-658. Shepard, R. N.. & Podgorny, P. (1978). Cognitive processes that resemble perceptual processes. In W. K. Estes (Ed.), Handbook of learning and cognitive processes (Vol. 5 , pp. 189-236). Hillsdale, NJ: Erlbaum. Sidman, M . , & Tailby, W. (1982). Conditional discrimination vs. matching-to-sample: An expansion of the testing paradigm. Journal of the Experimental Analysis of Behavior, 37, 5-22. Siegler, R. S. (1991). Children's thinking. Englewood Cliffs, NJ: Prentice-Hall. Soraci, S. A,, Jr., Alpher, V. S., Deckner, C. W., & Blanton, R. L. (1983). Oddity performance and the perception of relational information. Psychologia, 26, 175- 184. Soraci, S. A,, Jr., Barlean. J. L., Haenlein, M., & Baumeister, A. A. (1986). Lower sensitivity to alterations of auditory relational information in mentally retarded than in nonretarded adults. Physiological Psychology, 14, 146- 149. Soraci, S. A,, Jr., Carlin, M. T., & Baumeister. A. A. (in press). Stimulus organization and stimulus detection: Intelligence-related differences. In D. K. Detterman (Ed.), Current ropics in human intelligence (Vol. 3). Nonvood, NJ: Ablex. Soraci, S. A., Jr., Carlin, M. T., Deckner, C. W., & Baumeister, A. A. (1990). Detection of stimulus organization: Evidence of intelligence-related differences. Intelligence, 4 , 435-447. Soraci, S . A,, Jr., Deckner, C. W., Baumeister, A. A , , Bryant, J. T., Mackay, H. A., Stoddard, L. T., & Mcllvane, W. J. (1991). Generalized oddity performance in preschool children: A bimodal training procedure. Journal of Experimental Child Psychology, 51, 280-295. Soraci, S. A., Jr., Deckner, C. W., Baumeister, A. A , , & Carlin, M. T. (1990). Attentional functioning and relational learning. American Journal on Mental Retardation, 95, 304-3 15. Soraci. S. A . , Jr., Deckner, C. W., Haenlein, M., Baumeister, A. A , , Murata-Soraci, K., & Blanton, R. L. (1987). Oddity performance in preschool children at risk for mental retardation: Transfer and maintenance. Research in Developmental Disabilities, 8 , 137- 15 I . Spitz, H. H. (1963). Field theory in mental deficiency. In N. R. Ellis (Ed.), Handbook of menral deficiency: Psychological rheory and research (pp. I 1-40). New York: McGraw-Hill. Spitz, H. H. (1966). The role of input organization in the learning and memory of mental retardates. In N. R. Ellis (Ed.),Internationd review of research in mental retardation (Vol. 2, pp. 29-56). New York: Academic Press. Spitz, H. H., & Borland, M. D. (197 I ) . Effects of stimulus complexity on visual search performance of normals and educable retardates. American Journal of Mental Deficiency. 75, 724-728. Stella, M. E., & Etzel, B. C. (1986). Stimulus control of eye orientations: Shaping S + only versus shaping S- only. Analysis and Intervention in Developmental Disabilities, 6 , 137- 153. Suchman, R. G., & Trabasso, T. (1966). Color and form preference in young children. Journal of Experimental Child Psychology, 3, 177- 187. Terrace, H. S. (1963). Discrimination learning with and without errors. Journal of the Experimental Analysis of Behavior, 6 , 1-27. Treisman, A. M. (1988). Features and objects: The fourteenth Barlett memorial lecture. The Quarterly Journal of Experimental Psychology, 2 , 201-237.

DETECTION OF STIMULUS ORGANIZATION

53

Treismdn, A . M.. & Gelade, G. (1980). A feature-integration theory of attention. Cognitive Psychology, 12. 97-136. Treisman, A., & Gormican, S. (1988). Feature analysis in early vision: Evidence from search asymmetries. Psychological Review, 95, 15-48. Treisman, A,, & Paterson, R. (1984). Emergent features, attention, and object perception. Journal of Experimental Ps.vchologv: Human Perception and Performance, 10, 12-3 I . Treisman, A , , & Souther, J. (1985). Search asymmetry: A diagnostic for preattentive processing of separable features. Journal of Experimental Psychology: General, 114, 285-3 10. Uttal, W. R . (1983). Visual form detecrion in 3-dimensional space. Hillsdale, NJ: Erlbaum. Webb, T. E. (1972). Stereoscopic contour perception in mental retardation. American Journal of Mental Deficiency, 76, 699-702. Williams, L. G. (1966). The effect of target specification on objects fixated during visual search. Perception & Psychophysics, I , 3 15-3 18. Wright, J. C., & Vlietstra, A. G. (1975). The development of selective attention: From perceptual exploration to logical search. In H. W. Reese (Ed.), Advances in child development and behovior (Vol. 10, pp. 195-239). New York: Academic Press. Zeaman, D., & House, B . J. (1963). The role of attention in retardate discrimination learning. In N. R. Ellis (Ed.), Handbook of mental deficiency (pp. 159-223). New York: McGraw-Hill. Zeaman, 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. 63-120). Hillsdale, NJ: Erlbaum. Zentall, T. R., Hogan, D. E., Edwards, C. A,, & Hearst, E. (1980). Oddity learning in the pigeon as a function of the number of incorrect alternatives. Journal of Experimental Psychology: Animal Behavior Processes, 6 , 278-299.

This Page Intentionally Left Blank

Stimulus Control Analysis and Nonverbal Instructional Methods for People with Intellectual Disabilities WILLIAM J. MclLVANE BEHAVIORAL SCIENCES DIVISION E. K. SHRIVER CENTER WALTHAM, MASSACHUSETTS 02254 AND NORTHEASTERN UNIVERSITY BOSTON, MASSACHUSETTS 02 1 15

I.

BACKGROUND

People with intellectual disabilities present a myriad of challenges to those who hope to educate them. Among the challenges are language deficits that render many individuals difficult or impossible to teach via the verbal methods common to regular and special education. Even students who seem to have reasonably well-developed verbal skills may respond unreliably to verbal instruction. The teacher’s ability to teach is compromised when the student benefits little or unreliably from verbal explanations, verbal prompts, and other forms of verbal instruction. In such cases, both teacher and student can benefit substantially from teaching methodology that does not depend on developmentally normal verbal competency. Research conducted over the past several decades has sought to develop a scientific foundation on which to base efficient, effective nonverbal teaching methodology. Several noteworthy examples have been described in this series (e.g., Bijou, 1968; Conners, Caruso, & Detterman, 1986; Laniconi & Smeets, 1986; Sidman & Stoddard, 1966; Stoddard, 1982). The present contribution builds on these and other efforts. The main goal is to summarize and discuss findings of basic stimulus control research that are relevant to selected problems in mental retardation. A second goal is to review and propose applications of these findings to develop effective methods for teaching people who do not learn readily via verbal instruction. INTERNATIONAL REVIEW OF RESEARCH IN MENTAL RETARDATION. Vfll 18

55

Copyright 0 1992 by Academic Press, Inc. All rights of repduction in any form reserved.

56 A.

William J . Mcllvane

Stimulus Control Research and Intellectual Disabilities

When behavior is shown to be highly probable under certain stimulus conditions and not so under other conditions, the behavior is said to be under stimulus control. Many of the behavioral deficits characteristic of mental retardation reflect atypical and maladaptive stimulus control development. We are likely, for example, to identify a behavioral deficit in an adult whose street crossing is not influenced by red and green lights. Laboratory learning studies that seek to understand such a deficit may show, for example, that even protracted discrimination training with explicit differential reinforcement may not teach that individual to respond to green stimuli and not to respond to red ones (Zeaman & House, 1963). When such learning problems are pervasive, we are likely to term the individual intellectually disabled. Viewed broadly, there have been two main objectives of stimulus control research with people who have intellectual disabilities. The first has been to determine how these individuals differ from people of normal intellectual capability on measures of stimulus control acquisition, maintenance, and retention. The goal is to define better the phenomenon of mental retardation and to try to infer learning processes that may be defective (e.g., Soraci & Carlin, this volume; Borkowski & Cavanaugh, 1979). The second objective has been to develop methods that bypass or overcome characteristic stimulus control problems. The ultimate goal is a technology of teaching that can minimize the incidence and severity of mental retardation. This article focuses principally on stimulus control research having the latter objective. The nature of this work, however, is such that contributions consistent with the former objective are also inevitable.

B.

Framework for Developing Stimulus Control Teaching Technology

Research on methodology for teaching people with moderate and severe intellectual disabilities tends to adopt a fundamentally empirical, inductive approach. One begins by identifying a behavioral deficit or class of deficits that one wants to remediate. Then, one sets out to devise an appropriate remedial strategy. If available, techniques that have proven useful in solving related problems are applied. If the existing techniques do not prove adequate, as is frequently the case, they are modified or discarded in favor of other approaches. Where effective techniques cannot be found, a basic research problem is identified. An empirical, inductive approach presents certain challenges to developing a comprehensive science of teaching. For example, how does one organize and integrate an ever-expanding body of facts to foster the development of general scientific principles that have wide applicability and enduring value? In research conducted to test a particular theory, the theory provides an organizing frame-

STIMULUS CONTROL ANALYSIS

57

work for developing experiments, integrating data, and communicating the results. The empirical, inductive approach clearly requires a different strategy. The strategy that has emerged, often only implicitly, is to organize efforts around the comprehensive experimental analysis of complex, highly abstract, and obviously important scientific phenomena. A good example is the now highly advanced empirical analysis of how organisms choose among simultaneously available reinforcing alternatives (Williams, 1988). Researchers interested in intellectual disabilities typically organize their efforts around understanding and responding more effectively to problems imposed by those disabilities. Research described in subsequent sections addresses the problem of teaching highly complex and varied discrimination skill repertoires. The goal is not merely to teach large numbers of discriminations. Beyond that, the goal is to make progressive improvements in the discrimination learning abilities of individuals with learning disabilities. To illustrate the general approach, this article features research from a program that is aimed at developing methods for establishing prerequisites for symbol use in people with moderate and severe intellectual disabilities. The term symbol use is used here to refer to certain environmentbehavior relations that are established and maintained by one’s social environment for purposes of communication. Two examples help to define the scope and purposes of the program. For many individuals with severe intellectual and/or physical disabilities, communication cannot be readily accomplished via the typical response modes of speech or, for those with hearing impairments, sign. Successful communication with such individuals may depend on establishing a so-called “selection-based” system. Such systems present an array (or arrays) of stimuli (symbols), to which the individual makes some simple motor response (e.g., touching). Each of the symbols has meaning in that response to it serves to alter the environment in some characteristic way. Response to one symbol, for example, might produce a favored food, response to another might indicate the need to use the bathroom, and so on. Examples of selection-based communication systems include those that make use of Blissymbols (e.g., Harris-Vanderheiden, Brown, MacKensie, Reinen, & Scheibel, 1975), lexigrams (e.g., Romski, Sevcik, & Rumbaugh, 1985), and pictorial representations of events (e.g., Mirenda, 1985). For individuals whose disabilities are less severe, a more conventional form of symbolic communication might be appropriate. One might seek to establish a rudimentary reading repertoire, for example. Although fully developed and functional reading may not be a realistic goal for many individuals with disabilities, certain related skills could prove highly useful. Comprehension even of a relatively small number of words and word sequences could make a large difference in an individual’s ability to function independently-an ability all the more important given the current emphasis on deinstitutionalization. Consider, for example, the expanded opportunities and reduced supervision requirements en-

58

William J . Mcllvane

tailed by the ability to follow a set of simple printed instructions (e.g., recipes, travel directions). Symbolic communication systems have a number of inherent features that help one to organize a program of empirical research on discrimination learning. The nature of the subject matter is such that one can begin with simple stimuli, procedures, and discrimination tasks. As these are understood, it is natural to introduce further stimulus, procedure, and task complexity. The additional complexity, in turn, leads one to study progressively more advanced behavior and behavioral processes. Using the example of selection-based communication, one might initially be concerned merely with teaching the first instances of symbol use, even if an extensive program of direct teaching is required (e.g., Romski & Sevcik, 1989). As training progresses and the learner becomes more capable, however, the teaching goal is likely to change. One might become concerned with establishing more rapid and flexible learning, fostering the emergence of performances that have not been taught directly, and so forth. What follows describes ongoing efforts to develop the scientific foundation for a comprehensive stimulus control technology for teaching the discriminative prerequisites for symbol use in people with moderate and severe intellectual disabilities. A previous article addressed specific concerns that were important for establishing selection-based communication in people with severe disabilities (Mcllvane, Dube, Green, & Serna, in press). The present article is focused on more general issues, both conceptual and practical, that must be addressed in a scientific analysis and teaching of other behavioral repertoires that require welldeveloped discrimination skills. Examples of studies of discriminative prerequisites for rudimentary reading, spelling, and related performances are emphasized. The emphasis has been made in part because Singh and Singh’s (1986) contribution to this series pointed out the lack of available information on the topic with respect to the needs and capabilities of individuals with intellectual disabilities.

II.

OVERVIEW OF ANALYTICAL FRAMEWORK

This and subsequent sections further develop and elaborate an evolving framework for describing and analyzing stimulus control that results when subjects are exposed to various discrimination training regimens (e.g., Sidman & Tailby, 1982; Mackay & Sidman, 1984; Sidman, 1986a, 1986b; Stoddard & McIlvane, 1986; Dube, McIlvane, Mackay, & Stoddard, 1987; McIlvane, Dube, Green, & Serna, in press; Sidman, 1990). Behavior is analyzed explicitly or implicitly as two-, three-, four-, and sometimes higher-term contingency relations involving environmental and behavioral events. The two-term (response-stimulus) relation and the three-term (stimulus-response-stimulus) relation are those described in

STIMULUS CONTROL ANALYSIS

59

basic learning texts (e.g., Catania, 1984). In both, the term following the response is a reinforcing stimulus. The term prior to the response in the three-term relation is a discriminative stimulus that sets the occasion for the response to be reinforced. The four-term relation, a stimulus-stimulus-response-stimulus relation, and the more complex relations are useful in developing a coherent and fully detailed account of how some novel behavior emerges without direct training (Sidman, 1986).

A.

Two-Term and Three-Term Relations

1 . TWO-TERM RELATIONS

Some research problems require only looking at frequency of occurrence of a given response or response class, perhaps a simple touch to some manipulandum. If touching produces reinforcers and all other responses do not, then a twoterm (response-reinforcer) contingency relation is defined. Two-term relations suffice to analyze behavior that occurs when reinforcer delivery is not correlated with exteroceptive stimulus change (e.g., some reinforcement schedules). Often, however, it is critical to define and study the stimulus conditions under which behavior is emitted. 2. THREE-TERM RELATIONS When reinforcement depends on the presence versus absence of particular stimuli, a third term must be added to the analytical relation. Figure 1A shows an example, a simultaneous simple discrimination involving two visual stimuli. Pictures sometimes appear on the outer two of three response keys. Here, the picture of a cap is designated S + and the other, sun, is S - . Hereafter, capitalized words that refer to pictures are italicized, those that refer to print are roman, and spoken words are enclosed in quotation marks. Responses to CAP are reinforced. Responses to SUN or to the keys if no forms are present are not reinforced. Perfect visual discrimination is shown when the subject responds only to CAP and never to SUN or a blank key. Figure 1B illustrates another type of three-term relation with stimuli from a different modality. The task diagrammed here requires successive simple discrimination between two dictated words, “CAP” ( S + ) and “SUN” (S-); the response is to touch a key. Responses are reinforced only when “CAP” is spoken. When “SUN” is spoken or when no word is spoken, responses are never reinforced. Perfect auditory discrimination is shown when the subject responds only in the presence of “CAP” and never in its absence. Discussion of two additional examples of three-term behavioral relations illustrates an important point: what matters for analytical purposes are not specific structural characteristics of stimuli and responses, but rather the functional relations among environmental and behavioral events. Figure 1C diagrams a simul-

C

A

SR+

CAP +TOUCH

-0 OTHER -W

CAP +OTHER

+TOUCH

x -P x -+ x

-+ SR+ + X

PAC +TOUCH

-b

-b TOUCH -W

a a

-b OTHER

+ OTHER +

PAC

B

"CAP"+

TOUCH

X

D

+ SR+

"CAP" -W OTHER -b

X

"SUN" -b TOUCH 4

x

"SUN"

x

+ OTHER +

-b SR+

+"CAP"

-0 OTHER -b

a+ a+

X

X

"SUN" -b SR+ OTHER

+

X

c

"CAP"+

v -b TOUCH +SR+

"SUN-

"CAP"+

r 0 OTHER -0 X

"SUN"+

+ a + TOUCH+

"CAP" "CAP"

+ a +OTHER -0

X

"SUN"+

X

"SUN"+

TOUCH+ x'

v-0OTHER -0

a+ a

X

TOUCH +SR+

-& OTHER -& X

FIG. I . Three simple discrimination procedures and corresponding contingency diagrams. (A) Simultaneous visual discrimination procedure that requires a touch to one and not the other of two forms. (B) Successive auditory discrimination procedure that requires a touch to the center of three keys in the presence of one but not another auditory stimulus. (C) Simultaneous visual discrimination procedure in which both stimulus elements and order must be discriminated. (D)Stimulus displays and contingency diagrams for a naming test. (E) Stimulus displays and contingency diagrams for an arbitrary auditory-visual matching-to-sample procedure.

STIMULUS CONTROL ANALYSIS

61

taneous discrimination procedure that requires discrimination of other visual discriminative stimuli, printed words. The positive stimulus, CAP, is contrasted with several negative stimuli, some of which are made up of the same letters. Consistent selections of the positive stimulus require the subject to discriminate not only the letter elements but also their order. Figure 1D diagrams examples of contingencies arranged in an elementary naming test. CAP and SUN are successively displayed stimuli; the response requirements are met by speaking either “CAP” or “SUN ,” as appropriate. Unlike the preceding discrimination procedure examples, reinforcers are available for making two different, more complex responses occasioned by different stimuli. Thus, even though stimuli and responses become more complex, the critical functional relationships continue to be expressed in the same three-term analytical unit. 6.

Conditional Stimulus Control: Four-Term and Higher-Term Relations

1. FOUR-TERM RELATIONS Figure 1E shows a conditional discrimination procedure in which the picture to be touched changes from trial to trial, depending on the presence of another stimulus. When the spoken word “SUN” is dictated, for example, the subject must respond to SUN by touching it to produce a reinforcer. When “CAP” is dictated, however, CAP becomes the positive stimulus and SUN the negative. This procedure is technically termed an arbitrary audiovisual matching-to-sample procedure, because the discriminative (comparison) stimuli and the conditional (sample) stimuli are not physically similar. As Figs. 1A and 1B together suggest, the procedure shown in Fig. 1E combines some of the requirements of the simple successive and simultaneous discrimination procedures. Conditional discrimination requires successive discriminations between the dictated words and simultaneous discriminations between the pictures (cf. Saunders & Spradlin, 1989). As the conditional discrimination task is described here, there are four elements to the analytical contingencies: ( 1) conditional stimuli (samples), (2) discriminative stimuli (comparison stimuli), (3) responses, and (4) reinforcing stimuli (cf. Sidman, 1986). For this reason, four-term contingency analysis is sometimes perceived as a departure from classic three-term experimental analysis of behavior. The departure is more apparent than real, however, and chiefly a matter of definition and emphasis. Three-term contingency analysis recognizes that the important variables in behavior are (1) the situation in which behavior is emitted, (2) the behavior itself, and (3) the consequences of behavior. But the “situation” is always very complex, even in simplified experimental settings. As Skinner (1957, p. 2 1 ) has noted, capturing the complexities of situational control requires flexibility in defining analytical units to describe critical stimulus functions and relations.

62

William J . Mcllvane

2. HIGHER-TERM RELATIONS The simple and conditional discrimination procedures considered so far only begin to capture the complexity of stimulus control relations that make up a fully functional repertoire of adaptive behavior. Further complexity can be introduced by defining ever more elaborate contingency relationships. Several laboratories, for example, have sought to demonstrate five-term relations (Bush, Sidman, & de Rose, 1989; Catch & Osborne, 1989; Lynch &Green, 1991; cf. Weigl, 1941), and even higher-term contingency relationships are currently being studied. A five-term contingency could be arranged by expanding the procedure shown in Fig. 1E to include the printed words MATCH and MISMATCH as contextual (i.e., second-order) stimuli. When MATCH is present, the subject is to select C A P and SUN in the presence of “CAP” and “SUN,” respectively. When MISMATCH is present, however, the subject is to select C A P in the presence of “SUN” and SUN in the presence of “CAP.” Subsequent discussions merely touch on these relatively more complex conditional discrimination procedures. As yet, formal experimental study of higherorder contingency relationships has made little contact with subjects who are young and/or intellectually disabled. The priorities with these populations continue to be solving the many problems in analyzing and teaching the simpler forms of conditional discrimination. Notably, however, the current framework does not impose any obvious limits on the complexity of stimulus control relations that may be described and analyzed experimentally.

Ill.

MAJOR RESEARCH ISSUES AND FINDINGS

Among the most important aims in the scientific study of stimulus control are to ( 1 ) define stimuli appropriately, (2) identify the stimuli that are currently in control of behavior, and (3) describe major features of controlling environmentbehavior relations. Research seeking a stimulus control teaching technology has these aims and at least two more: (4) to describe comprehensively an effective methodology for establishing specified environment-behavior relations and (5) to provide a compelling logical and empirical basis for deciding that any given performance is not within the learner’s current capabilities. This section and the following one present and discuss selected examples of research issues and activities consistent with these aims. The examples have been selected primarily to illustrate an approach to developing rudimentary reading and related skills in people who have severe intellectual disabilities. Even within this relatively narrow focus, however, it is necessary to restrict the topics that can be considered. What follows will endeavor to review research issues and findings that have received only limited coverage in this series and other major compendiums of research in mental retardation (e.g., Ellis, 1979; Matson & Mulick, 1991).

STIMULUS CONTROL ANALYSIS

A.

63

Stimulus Definition and Stimulus Control Topographies

I . STIMULUS DEFINITION There are at least two aspects to the problem of stimulus definition. One is to specify structural properties of the stimuli presented (see Soraci & Carlin, this volume). Stimuli can be characterized, for example, with respect to distinctive features (Gibson & Levin, 1976) and/or values on various stimulus dimensions (e.g., size, color, time of onset; Zeaman & House, 1979). Structural analysis, however, does not answer questions about which of these characteristics actually control the learner’s behavior. The point can be illustrated by considering the examples in Fig. 1B. Previously, those stimuli were specified as the dictated English words “CAP” (S+) and “SUN” (S-). Stimulus control was shown after discrimination training engendered frequent responding to the former and infrequent responding to the latter. For some purposes, no further analysis may be necessary. Suppose, however, that the experimentedteacher is interested in assessing the degree to which the individual discriminates all of the properties that are implicit in the original stimulus definitions. Is his or her behavior under the control of all critical features of the dictated words or perhaps only some of them? There is wide agreement on the method for assessing control by stimulus properties. First, one establishes a stimulus control baseline, preferably one of high accuracy and stability. Then, one varies the properties of interest systematically and measures the frequency of response to the variation. If variation alters response frequency, then prior control by the varied property is inferred. If frequency is unaltered, however, the typical inference is that the varied property did not control behavior and that discrimination was based on some other S + / S - stimulus difference (Lashley & Wade, 1946).’ With respect to the performances shown in Fig. l B , one might apply the method of stimulus variation to ask about the basis for discrimination. One might ask, for example, whether discrimination depends on the specific dictated words used in training. This could be done by substituting a novel word, “BIT” perhaps, for either “CAP” or “SUN” and assessing the impact on performance. Suppose “BIT” were indeed substituted for “SUN.” If the learner responds to “CAP” and not to “BIT,” this outcome would be evidence that responding is controlled by one or more properties of the former. If discrimination were disrupted (i.e., the individual responded to both words or to neither reliably), the outcome would suggest that the previous discrimination was controlled by some other aspect of the S + /S- difference. Conceivably, for example, the individual might have learned to respond in the presence of the word that does not sound ‘One could also interpret this outcome as complete generalization of control along the dimension of variation, an interpretation that entails additional assumptions and seemingly circular reasoning; however, this position has advocates who do offer well-reasoned arguments (e.g.,Honig & Urcuioli, 1981; Macintosh, 1977).

64

Willium J . Mcllvane

like “SUN” (cf. Sidman, 1979, 1987). If so, responses might be occasioned by both “BIT” and “CAP.” To develop the illustration further, suppose that testing revealed that responses were controlled by “CAP” and not by “BIT.” Suppose also that the individual did not respond when several other test words were presented. One might be led to ask about the properties of “CAP” that did control. One might ask, for example, whether the controlling stimuli were the initial or final consonant sounds, the central vowel, or some combination-the whole word perhaps. To assess these possibilities, further stimulus variations (e.g., “COD,” “TOP,” “LAG”) would be presented and the response frequencies to each would be measured. To complete the example, suppose that testing revealed that the individual responded only in the presence of “CAP” and not when any other word was spoken. One might additionally go on to assess whether discrimination would be maintained when other characteristics of the dictated words were changed (e.g., from a female voice to a male voice). Provided that the original discrimination baseline (e.g., “CAP” versus “SUN”) remains intact, the method of stimulus variation can be used to define precisely the environmental conditions under which responding occurs. Familiarity with this basic approach may also prove helpful in establishing goals for teaching. Stimulus definition in functional as well as structural terms may render the teacher more sensitive to stimulus control possibilities and perhaps less content, therefore, with procedures that allow potentially wide variations in controlling stimuli. When learners with severe mental disabilities are involved, for example, restricted control by isolated stimulus features (i.e., stimulus overselectivity) is a likely outcome and the need for explicit procedures to broaden control must be anticipated (Allen & Fuqua, 1985; Schreibman, Charlop, & Koegel, 1982).

2. STIMULUS CONTROL TOPOGRAPHIES The issue of stimulus definitions leads directly to a second issue, characterizing the composition of a stimulus control baseline. When a subject is exposed to three-term or four-term contingencies, what behavioral relations are observed and in what frequency in the baseline? Does the same stimulus or stimulus dimension control all responses or are there variations in stimulus control across discrimination trials? If variation exists, how can it be identified and measured? In considering these questions, it may be helpful to review the generic nature of the concepts of stimulus and response in the analysis of behavior (Skinner, 1935). When examining response-reinforcer relations, across-trial variation is typically clearly observable. Some responses may be forceful and others less so, some responses may be made with the left hand and others with the right, and so on. Such variation is a major reason why the response is defined generically as a class of behavioral events that all have the same effect on the environment (e.g.,

STIMULUS CONTROL ANALYSIS

65

a key switch closure). For many studies of two-term relations (e.g., analyses of reinforcement schedules), topography variations go unmeasured without ill effect. This does not mean that response variation is totally ignored, however. Indeed, response variation may be studied explicitly as an important dependent variable in its own right (e.g., Notterman, 1959). Skinner also recognized the inherent variation in the stimuli that set the occasion for responding, if only because the subject often approached the response manipulandum from different directions. Hence, the controlling stimulus was also defined generically as a class of environmental events that all set the occasion for the same response. Paralleling studies of the two-term relation, many studies of three-term and four-term contingency relations have contributed valuable data without addressing the issue of across-trial variations in stimulus control. In fact, little experimental work has focused on this topic, probably for largely technical reasons. Whereas variations in response topography may be easy to detect, variations in the stimulus control of behavior- “stimulus control topographies” (McIlvane & Dube, in press; Ray, 1969)-are not directly observable (Sidman, 1979). Instead, they must be inferred on the basis of outcomes of tests that vary stimuli and discrimination procedures. This point can be made by describing illustrative studies of simple and conditional discrimination learning. a . Removal and Restoration of Stimulus Control. Early in this series, an article described programmed methods for (1) teaching people with severe intellectual disabilities to discriminate a circle (S+) from seven ellipses (S-) and (2) assessing individual circle-ellipse discrimination thresholds, that is, how rounded an ellipse could be discriminated from a circle (Sidman & Stoddard, 1966). Examples of the stimuli are shown Fig. 2A. The positive stimulus was always the circle; the negative stimulus during training was an ellipse having a minor-tomajor axis ratio of .53.During the threshold series, ellipses with progressively larger axis ratios were presented in the circle-ellipse discrimination task. Discrimination threshold was defined as the point within that range at which the circle-ellipse discrimination broke down. Using their methods, Sidman and Stoddard were able to demonstrate remarkably high discrimination thresholds in very young normally developing children, children and adults with severe intellectual disabilities, and nonhuman primates. Traditional interpretations of Sidman and Stoddard’s findings would suggest that the training procedures had established control by the dimension of circleellipse difference or one perfectly correlated with it (e.g., Spence, 1956). The threshold series would be viewed as defining the limits of capacity to discriminate within that dimension. An important but little-known study by Stoddard and Sidman (1 97 1 a) brought the traditional interpretation into question by demonstrating previously unsuspected variation in stimulus control in the discriminations that made up the threshold series. Stoddard and Sidman’s subjects were four rhesus monkeys who had reliably

William J . Mcllvane

66

A

1 .o

.53

0 .70

0

Test Series

-74

.77

.80

0 00 0 .86

.89

91

6 000 B IMMEDIATE HISTORY

TEST PROBE

IMMEDIATE HISTORY

TEST PROBE

lololollololol 10101ol FlG. 2. Experimental stimuli used in circle-ellipse studies. (A) Stimuli presented in Sidman and Stoddard’s ( 1966) fading and discrimination threshold programs. (B) Stimulus displays presented in Stoddard and Sidman’s (1971a) study of the removal and restoration of stimulus control.

discriminated stimulus differences ranging from circle versus .53 ellipse to circle versus .9 1 ellipse on the threshold series and subsequent stimulus generalization tests (Stoddard & Sidman, 1971b). A series of further tests were conducted to determine whether ( 1) presenting an “impossible” discrimination task would

STIMULUS CONTROL ANALYSIS

67

disrupt control by a given circle-ellipse difference and (2) stimulus control could be restored without resorting to repetition of the threshold training series. Figure 2B illustrates the test methodology. The impossible discrimination task presented eight circles, one of which was arbitrarily designated as the positive stimulus; ten such trials were presented as an “immediate history” prior to a “test probe” trial that evaluated control by one of the circle-ellipse differences. Such tests were replicated numerous times and covered the range of differences from circle versus .77 ellipse through circle versus .91 ellipse. If control by a given difference was not evident on a probe trial, the monkey was then exposed to ten further training trials that presented either ( 1) a circle-versus-blank key discrimination task that was an early step of the program used originally to establish the circle-versus-.53 ellipse discrimination, or (2) the circle-versus-.53 ellipse discrimination itself. The question was whether either of these immediate histories would suffice to restore control by a given circle-ellipse difference. Individual data for one monkey (R34) are used to summarize the major findings of the study: (1) the impossible discrimination history did disrupt stimulus control by all circle-versus-.80 and larger ellipse differences; (2) circle-versusblank key training restored control by the circle-versus-.80 difference but not the others; and (3) circle-versus-.53 ellipse training restored control by the circleversus-.83 and -.86 difference, but not the circle-versus-.89 and -.91 differences. Thus, there was a striking lack of uniformity in the effects of the various immediate histories on different circle-versus-ellipsediscriminations. Greater uniformity would be expected if the animal’s discrimination baseline performance was controlled by the continuous dimension of circle-ellipse difference. Instead, test results suggested that the baseline in fact was composed of several different stimulus control topographies that were differentially affected by the various immediate histories. Stoddard and Sidman’s probe data thus demonstrated stimulus control discontinuities in which might be otherwise interpreted as a baseline of continuous dimensional control. Until the threshold value was reached, the different stimulus control topographies that occurred during threshold testing were perfectly correlated with the circle-ellipse differences that the experimenters specified as relevant. High accuracy was maintained throughout the test, and thus provided no indication of changes that were occurring in the stimuli that controlled circle selections. Stoddard and Sidman’s study also provides, perhaps, the classic example of the dynamic and flexible character of stimulus control. Stimulus differences that controlled behavior reliably at one moment were made completely ineffective and then fully effective once again by providing the appropriate experimental history. A reasonable question is whether results obtained from monkeys are relevant to the concerns of those interested in analyzing and remediating learning problems of humans with intellectual disabilities. Notably, similarly dynamic,

68

William J . Mcllvane

FIG. 3. Three different controlling relations that are possible in an auditory-visual matching-tosample procedure and test procedures to assess those relations. Stimuli that are not critical to stimulus control of behavior are shaded in the displays on the left and replaced by other stimuli in those on the right. (A,B) Sample-S+ relations. (C,D) Sample-S- relations. (E,F) Mixed sample-S+ and sampleS - relations.

flexible stimulus control baselines have been demonstrated with individuals with a variety of behavioral limitations, including mental retardation and neurological disease (e.g., McIlvane, Kledaras, Dube, & Stoddard, 1989; Sidman, 1969). b. Controlling Relations in Marching to Sample. The coexistence of different stimulus control topographies within the same baseline can also be shown in more complex performances like matching to sample. Figure 3 shows several different ways in which an individual might perform the matching task shown in Fig. 1E. Shading is used to suggest that certain stimuli may not be involved in the control of matching selections. In Fig. 3A, negative (S-) comparison stimuli are shaded to suggest that the sample and positive (S+) comparison stimuli might be the critical controlling elements. The implication is that other stimuli could be substituted as negative comparison stimuli without disrupting consistent positive comparison selections (Fig. 3B). Hence, the controlling relations are termed samples+ relations. One often sees the apparent assumption that all controlling relations in matching to sample are of the sample-S+ variety, particularly when behavioral relations are conceptualized as “links” or “associations” between stimuli (e.g.,

STIMULUS CONTROL ANALYSIS

69

O’Mara, 199I). Other controlling relations are logically possible, however. Figure 3C shows an example in which the controlling elements are the sample and the negative comparison stimuli, sample-S - relations. When behavior is determined by a sample-S- relation, other stimuli can be substituted for the posirive comparison stimuli without disrupting conditional discrimination (Fig. 3D); the subject selects the substituted positive comparisons by continuing to reject the negative one. Figure 3E shows a third possibility. Here, conditional discrimination is the product of one sample-S+ and one sample-S- relation. When “CAP” is the sample, the subject selects CAP. When “SUN” is the sample, however, the subject rejects CAP and selects the remaining picture. In this example, the other comparison SUN is superfluous and can be replaced by other stimuli without disrupting conditionality (Fig. 3F). It is important to note that the behavior shown in Figs. 3A, 3C, and 3E is the same. What differs are the stimuli in control, and the differences are not detectable merely by observing the behavior. Studies of normally developing children and individuals with a variety of mental disabilities have examined the possibility of sample-S - control of matching-to-sample performance (Dixon & Dixon, 1978; Dixon, Dixon, & Spradlin, 1983; McIlvane, Withstandley, & Stoddard, 1984; Stromer & Osborne, 1982). Results of these studies have been consistently positive. Moreover, some studies have simultaneously tested for both sample-S+ and sample-S- relations in the same discriminative baseline, again with positive results (McIlvane, Withstandley, & Stoddard, 1984; Stromer & Osborne, 1982). It seems likely therefore that humans’ matching-to-sample baselines may sometimes reflect a mix of sample-S+ and sample-S- control. What is not known, however, is whether this is always the case and whether different stimulus control topographies occur with equal frequency in a given baseline. Unfortunately, extant test methodologies merely allow one to infer the presence versus absence of different stimulus control topographies; yet to be developed are “on-line” measurement techniques that would permit moment-by-moment inferences about the stimuli that are currently in control of behavior. c. Interpreting Intermediate Accuracy Scores. During discrimination acquisition on standard differential reinforcement procedures, the frequent errors made by the individual clearly demonstrate control by stimuli not correlated with the stimulus differences that the experimenterheacher specifies as relevant. When discrimination accuracy scores are persistently intermediate and variable over a number of training sessions, a problem of interpretation is encountered. This problem was well exemplified in data reported by House, Hanley, and Magid ( 1979). They sought to teach 19 adolescents with mild and moderate retardation a generalized oddity performance (given stimulus arrays displaying two identical and one different item, to select the latter). Results showed that 6 subjects learned generalized oddity, ultimately performing at virtually 100% accuracy levels. Seven subjects’ accuracy scores never rose significantly above chance levels

70

William J. Mcllvane

(33% correct). The remaining 6 subjects, however, achieved and maintained mean asymptotic accuracy scores of 78% correct. Attention theories of learning (e.g., Zeaman & House, 1979) suggest that the sufficient conditions for learning are attention to the relevant stimulus dimension ( “difference” in the oddity problem) and effective consequences for establishing and maintaining the proper motoric response (i.e.. touching S+). In House and colleagues’ study, attention to the relevant dimension was indicated by the 78% accuracy score; effective consequences were verified by other procedures. According to attention theory (and common sense, for that matter), therefore, learning should have continued until subjects performed virtually without error. But several subjects’ accuracy scores remained intermediate and variable-a common problem when subjects are very young and/or mentally disabled. One way to account for persistent intermediate accuracy scores is to suggest that there is something inherently wrong (or underdeveloped) about the discrimination abilities of individuals with intellectual limitations. House and colleagues suggested defective memory processes to account for their data, for example. Another interpretation is that the individual does not attend to the same stimuli or stimulus dimensions on every trial and that the discrimination baseline is composed of a mixture off stimulus control topographies. One or perhaps more may coincide with the stimulus differences specified as relevant by the teacher or experimenter. Each time that topography occurs, the response is reinforced. Other stimulus control topographies may also be present, however, and by chance they too often result in reinforced behavior, particularly when the number of choice stimuli is small. If the standard two-choice discrimination procedure is employed, for instance, responses controlled by irrelevant stimuli like position will be reinforced every other time on average (a variable-ratio 2 schedule) and the irrelevant topographies may persist. d. Experimental Separation of Stimulus Control Topographies. Mcllvane, Kledaras, Dube, and Stoddard (1989) recently reported a study that demonstrates experimental separation of relevant and competing irrelevant stimulus control topographies. Their study was conducted in the course of a larger effort that sought effective automated programming techniques for teaching individuals with profound mental disabilities (summarized in this series by Stoddard, 1982). One goal was to develop a way to prompt selection of one of two food items that were displayed in a modified Wisconsin General Test Apparatus (WGTA). The final performance was to be as follows: The subject first made a trial initiation (TI) response. When he or she did so, a screen was retracted to display two sideby-side compartments. Each contained a food item. A bank of red lights was lit in the back of one of the compartments to tell the subject which food was to be chosen. The methodology and general features of the data are presented in Fig. 4. In Fig. 4A, the leftmost portion, labeled Phase I, diagrams the simultaneous dis-

71

STIMULUS CONTROL ANALYSIS

A

+

Trial Initiation

Trial Initiation Ma1 State 1

I

I

+

WAIT

+

I

Trial State 2

+

+

FOOD

NOFOOD

+

FOOD

Phase 1

B 100

-

s

s -

3 F Y

8

E a

+

NOFOOD

Phase 2 Phase 1

Phase 2

Correct Red Light r n 2 ) Selections

-

50-

-

-

0-

TS1 Responses Sessions

FIG. 4. Procedures (A) and abstracted results (B)of the study by Mcllvane, Kledaras, Dube, and Stoddard (1989) of visual discrimination in people with severe mental retardation. TSI. trial state I ; TS2. trial state 2.

crimination procedure. Following a TI response, a red light was turned on immediately in one compartment, and the subject obtained food by choosing it. Under Phase 1 conditions, red light selection accuracy was often well above the 50% “chance” level on a two-choice task. Scores were highly variable, however. Further, a line fitted to the data would often be virtually horizontal, indicating

72

William J . Methane

little or no improvement over the course of a (sometimes very large) number of training sessions. Such results are suggested by the horizontal line drawn at the 70% correct level in the Phase 1 portion of the graph in Fig. 4B. The persistently unreliable discrimination accuracy scores became reliable when the Phase 2 procedure was implemented. In this “delayed S + ” procedure, each TI response was followed by a brief period in which S- stimuli were displayed in both compartments, Trial State 1 (TS1). If the subject did not select one of the compartments for a brief period (e.g., 3 to 5 seconds), Trial State 2 (TS2) commenced. One compartment was lit red (S +) and its selection could be reinforced by giving the subject access to a food item. If the subject did respond prior to the onset of the red light, that response delayed the onset of TS2 by the programmed length of TS 1. The subjects’ above-chance Phase 1 accuracy scores suggested the possibility of mixed control by the target stimulus (the red light) and by other competing stimulus control topographies. Because of the two-choice format of the discrimination task, even responses controlled by irrelevant stimuli would be reinforced on an average of every other trial, potentially maintaining the competing control. By programming Phase 1 and Phase 2 conditions, the delayed S + procedure arranged circumstances whereby all selections that occurred during the first phase (by definition not under the control of the red light) were never reinforced, thus selectively decreasing the probability of those competing stimulus-response relations through extinction. Perhaps this is why control by the target relation was demonstrated virtually always when the subject entered TS2; competing stimulus-response relations had an immediate extinction history and the target relation did not. Further, the TSl slope differences between higher- and lowerfunctioning subjects suggested more rapid extinction of competing stimulusresponse relations in the former. McIlvane and colleagues’ delayed S + studies produced at least five noteworthy findings. Two of them are illustrated in the right portion of Fig. 4B. The first finding was the immediate increase in red light discrimination (TS2) accuracy to virtually 100% correct. The accuracy increase was truly instantaneous; highly accurate responding was observed on even the initial Phase 2 trials. The second finding was the initially high levels of TSl responding, which declined with continued exposure to the delayed S + procedure. The third finding was the remarkable consistency of results across a low-functioning but otherwise heterogeneous subject population. Subjects tested on several variants of the basic procedure all yielded data of the general character shown in Fig. 4. The fourth finding was merely a suggested relationship between level of functioning and an aspect of Phase 2 performance. What seemed to differentiate higher- from lowerfunctioning subjects was the slope of the function plotting decreases in TSl responding. Although small subject numbers and procedure variations limit firm conclusions, higher-functioning subjects tended to have considerably steeper functions. The fifth and possibly most important finding was that continued

STIMULUS CONTROL ANALYSIS

13

exposure to the delayed S + procedure ultimately resulted in performances that were almost entirely error free in a number of very low functioning subjects; previously unreliable discrimination baselines became highly reliable, thus approximating the accuracy levels more commonly associated with individuals who do not have mental disabilities. One is led to ask about the generality of McIlvane and colleagues’ findings and their possible applicability to conventional discrimination training situations. This question was answered in two follow-up studies with subjects who were more capable. One study (McIlvane, Kledaras, Gilbert, & Stoddard, 1992) conducted a systematic replication using abstract two-dimensional stimuli presented on the screen of a microcomputer. Results were virtually identical for subjects whose baselines suggested a mixture of relevant and irrelevant stimulus control topographies. McIlvane, Kledaras, Stoddard, and Dube (1990) also demonstrated how a similar procedure might enhance matching-to-sample performances. 3. SUMMARY

Some general points can be made to summarize the discussion of stimulus definition and stimulus control topographies. First, stimulus definition in functional terms is potentially a very complex matter that may require elaborate test procedures to ascertain which properties of the stimulus are in control of behavior. Second, when the concept of stimulus control topography is incorporated, the complexity increases substantially. One must be concerned not only with which stimuli control but possibly also with when or how often different stimuli control. On its face, the complexity may seem to approach unmanageability. It is important to note, however, that analytical studies and techniques described so far have succeeded in rendering stimulus control measurable, interpretable, and to some degree manipulable. There seems to be a clear basis for optimism that a developing science of behavior will provide ever more powerful and effective approaches to managing this complex subject matter.

B.

Stimulus Functions and Relations

Appropriate stimulus definition and stimulus control topography specification do not complete the analysis of stimulus control. One must be concerned also with how stimuli function in the control of behavior and the nature of the controlling relations. These issues are taken up in this section. 1 . DEFINITION AND ASSESSMENT OF STIMULUS CONTROL FUNCTIONS

The problem of stimulus definition has a parallel in defining the way in which stimuli function to control responding. Discriminative (S+/S-) functions are defined empirically as different response rates resulting from exposure to con-

74

Willium J . Mdlvune

tingencies of reinforcement. The problem comes about, however, when one endeavors to define stimulus functions in conditional discrimination procedures. There are at least two ways of conceptualizing stimulus functions engendered by conditional discrimination procedures. The first is consistent with the fourterm contingency analysis described earlier. Stimuli are classified separately on the basis of apparent differences in stimulus function (e.g., Cumming & Berryman, 1965; Lashley, 1938; Sidman, 1986). In Fig. lE, the sample stimuli “CAP” and “SUN” are seen as serving “conditional” or “instructional” functions that determine the discriminative functions of the comparison stimuli CAP and SUN.2 Alternately, conditional discrimination has been thought of in terms of control by compound stimuli. With respect to a preceding example, the suggestion is that “CAP” and CAP together occasion a touch of the latter; the specific incorrect choice(s) (SUN in the example) can also be part of the controlling stimulus compound (cf. Carter & Werner, 1978). No separate stimulus functions are entailed. Rather, the compound is seen to function as a unitary stimulus, and analysis in terms of a three-term contingency relation is deemed appropriate (Thomas & Schmidt, 1989).3 The principal issue is whether ( I ) logically simpler discriminative compound stimulus control will suffice to explain the important features of conditional discrimination or (2) additional stimulus functions must be specified. The burden of proof is clearly on proponents of the latter position. At least two major questions must be answered. First, what behavioral phenomena compel one to specify stimulus functions beyond simple discriminative control? Second, how are these putative additional functions to be characterized? The instructionalversus-discriminative stimulus functional distinction seems to imply a hierarchical relationship among the stimuli. But is there any compelling empirical basis for establishing a stimulus hierarchy? Here again one encounters problems of measurement. Just as one cannot determine which stimulus properties are in control of behavior merely by observing any given instance, one also cannot determine how those stimuli are functioning. Inferences of stimulus function must be also be based on the results of ’Cumming and Berryman (1965) described these different stimulus functions as follows: “In simple successive or simultaneous discriminations, the discriminative stimuli have an invariant relation to reinforcement and extinction-thus, the presentation of S f is an occasion for reinforcement of the specified response, while the presentation of s- is an occasion for some other behavior. Here, the function of the discriminative stimuli is the control of specific responses. However, in more complex types of discriminative situations, a stimulus may function as a selector of discriminations, rather than individual response” (p. 285). 3A similar stimulus function definition issue is also encountered with higher-term contingencies. With respect to the example discussed earlier, do MATCH and MISMATCH ( I ) exert higher-order conditional control over the conditional functions of “CAP” and “SUN” or (2) function merely as elements of compound stimuli in three-term or possibly four-term contingency relations (see Bush, Sidman, & de Rose, 1989, and Lynch & Green, 1991, for further discussion of this issue)?

STIMULUS CONTROL ANALYSIS

75

varying stimuli and procedures. Nominal specifications of procedures as requiring conditional or simple (i.e., nonconditional) discrimination may be misleading. Two examples help to make this point explicit. Consider the procedure used in Lashley’s ( 1938)classic study of “conditional reactions” in the rat and subsequently replicated by others studying normally developing children (e.g., Gollin & Liss, 1962; Schilmoeller, Schilmoeller, Etzel, & LeBlanc, 1979). Traditional analyses of this procedure were consistent with four-term contingency analysis. The forms (upright versus inverted triangles) were viewed as the discriminative stimuli and the backgrounds (solid versus striated) as conditional stimuli that determined the S+/S- functions of the triangles. Inspection of the forms, however, shows that the problem can in fact be solved by making a simple simultaneous discrimination-responding to (or perhaps away from) invariant stimulus features that differentiate the positive and negative stimuli. Viewed in this way, Lashley’s task is only nominally a conditional discrimination problem; task mastery demonstrates neither true conditional discrimination (i.e., one based on a relation between stimulus elements) nor any conditional stimulus function. Lashley’s studies recognized the basic difficulty. Over a series of subsequent discrimination problems, he replicated the preceding procedure, replacing the triangles with a variety of other forms. These experiments sought unsuccessfully to show that changing backgrounds would result in immediate reversal of discriminative functions of forms presented on them. Such reversals would have demonstrated a stimulus function that was clearly different from simple discriminative control. He described that control as a generalized “rule” that might be verbalized as “any stimulus that is correct in Situation A is incorrect in Situation B.” (see Section III.B.2.c). A contrasting example can be found in the “ambiguous cue” problem (Thompson, 1954). Two trial types are intermixed. On one type, forms A and B are S + and S- , respectively; on the other, forms B and C are are S + and S - , respectively. Thus, the discriminative functions of form B change (i.e., are ambiguous) depending on the other form in the display. Although each trial type is nominally a simple simultaneous discrimination, the stimulus control requirements resemble those of a conditional discrimination. Minimally, compound stimulus control is required. Alternately, the problem may also be analyzed as a four-term contingency relationship wherein both forms together serve as the instructional stimuli determining the discriminative functions of each form individually. 2. DEFINITION AND ASSESSMENT OF STIMULUS RELATIONS Whether conditional control is conceptualized in terms of three- or four-term relations, there is an empirical necessity to address the issue of relational stimulus control. The facts of behavior are clear. Many of the stimulus control

76

William J . Mcllvane

phenomena seen in the laboratory and the classroom do not seem to be explainable by the presence of stimuli in naturally occurring or explicitly arranged contingency situations. Moreover, such phenomena are also not explainable by primary stimulus generalization (control by stimuli with physical features in common with stimuli present during direct conditioning). These points can be made by considering a few illustrative examples. a . Identiry Relations. Among the most important and well-studied topics in stimulus control research is whether or not a given discrimination performance reflects the concept of “sameness” or “identity.” The question is often asked by means of the identity match-to-sample p r o c e d ~ r eWhen .~ an individual selects a comparison stimulus that is physically identical to the sample stimulus, one might conclude that he or she has learned to select the comparison that is physically the same as the sample (i.e., generalized identity relations). If so, then the learner should be immediately capable of the new identity-matching tasks with stimuli that are physically dissimilar to those that appeared in training. Suppose, for example, that the individual is taught to match the printed words CAP and SUN. Generalized identity matching would be shown if he or she proved immediately capable of matching new printed words and perhaps other stimuli (pictures, forms, colors, etc.). Emergence of untrained identity-matching performances appears to be rhe critical test of whether or not matching has been based on identity relations. When positive test outcomes are lacking, even mastery of large numbers of identity-matching performances via direct training can still be accounted for in terms of more basic behavioral processes such as the acquisition of simple response chains (Carter & Werner, 1978; Cumming & Berryman, 1965). With respect to the issue of differentiating the three-term account of conditional discrimination from the four-term account, one might ask whether demonstrations of generalized identity-matching performances are helpful. One might argue that one cannot account for emergent identity matching in terms of mere discriminative control by a stimulus chain or compound. A higher-level (i.e., relational or rule-based) process must be involved. It is possible to make counterarguments, however. Dube, McIlvane, and Green (l992), for example, have suggested that certain emergent identity-matching performances might result from relatively simple discriminative processes, specifically very rapid withintrial acquisition of discriminative control by the sample stimulus. Their account, however, applies only to the special case of identity matching to sample. The following section returns to the more general issue raised when discussing emergent matching of physically dissimilar stimuli. b. Stimulus Equivalence. A classic problem in the analysis of behavior is to 4Studies of generalized oddity have the same goal. See Soraci and Carlin (this volume) for further details.

STIMULUS CONTROL ANALYSIS

77

provide an account of how physically dissimilar stimuli can have similar and apparently interrelated effects in control of behavior. The problem is illustrated in Goldiamond’s (1966) example of three physically dissimilar stimuli that are all functionally equivalent in controlling the behaviors involved in stopping one’s car (a red light, a stop sign, and a policeman’s upraised hand). An important aspect of the example is the membership of these stimuli in a functional equivalence class: A variable applied to one member of the class should have similar effects on other members (cf. Goldiamond, 1962). For example, a situation that might lead one to ignore a policeman’s upraised hand (e.g., a medical emergency) would have similar effects with respect to a red light and a stop sign. Study of the problem of stimulus equivalence has a long and venerable history in experimental psychology (e.g., Jenkins, 1963; Peters, 1935), culminating in a very active, sustained effort over the past two decades. Recent work was inspired by a series of studies by Sidman and his colleagues in the early 1970s (Sidman, 1971; Sidman & Cresson, 1973; Sidman, Cresson, & Willson-Morris, 1974). In the first study, Sidman taught an adolescent boy who was severely mentally retarded to select each of 20 pictures (from Set B) conditionally on its dictated English name (from Set A) (AB matching); the boy was also taught to select each of 20 three-letter words (from Set C) conditionally on the same dictated words (AC matching). The boy was also taught identity matching with the pictures and printed words. Stimulus equivalence was demonstrated when the boy proved immediately capable of matching each printed word to its corresponding picture (BC and CB matching); entirely visual arbitrary matching-to-sample performances emerged without further training. Thus, the pictures and their corresponding dictated and printed words were shown to be mutually substitutable for one another and functionally equivalent in the match-to-sample context. In Sidman’s early studies, “equivalence” was used in its traditional descriptive sense; however, Sidman and Tailby (1982) offered a formal specification of the equivalence relation. For a relation to be termed one of equivalence, they argued it had to have three properties: reflexivity, symmetry, and transitivity (cf. Stevens, 1951). In that same article, Sidman and Tailby also described tests that could determine whether stimuli were equivalent according to their rigorous specification. Reflexivity is evaluated by means of tests for generalized identity matching. That is, the subject must match each stimulus with itself without explicit trair~ing.~ Further details of Sidman’s procedural approach and the tests for symmetry and transitivity are illustrated by example in Figs. 5A and 5B. Figure 5A schematically presents conditional discrimination procedures that teach the subject to select CAP and SUN (Set B) conditionally on the spoken 3aunders and Green ( 1992) have recently argued for an additional requirement, specifically that the reflexivity test follow and not precede the conditional discrimination training so as to increase the likelihood that the test will actually reflect the properties of the baseline conditional relations.

78

William J . Mcllvane

A A

r &7 I ?

I.

D

I

I

I

I I I

I I I

4

CAP SUN

C

B Al

G Name dictated to subject UCAPU WSUNM

B

: 4 @

h

i

‘

I

b

I

I

‘

h

‘

I

!

I I

4:

by subject “CAP” “SUN”

4 I

C FIG. 5. Diagrams of directly trained (solid arrows) and emergent (dashed arrows) conditional stimulus-stimulus and stimulus-response relations in a study of stimulus equivalence.

words “CAP” and “SUN” (Set A), respectively. Also shown are trials that teach the subject to select the printed words CAP and SUN (Set C) conditionally on CAP and S U N , respectively. These directly taught AB and BC conditional relations are indicated by solid arrows, which point from sample to comparison. Dashed arrows indicate two new trial types with which the subject has no prior experience. In one type (AC), the samples are dictated words, “CAP” and

STIMULUS CONTROL ANALYSIS

79

“SUN,” and the comparisons are printed words, CAP and SUN. These trials ask whether the subject will prove capable of matching the dictated words with their corresponding pictures. If the subject does match the dictated and printed words appropriately, we have evidence for two transitive stimulus relations: (1) If “CAP” r CAP and CAP r CAP, then “CAP” r CAP, and (2) if “SUN” r SUN and SUN r SUN, then “SUN” r SUN. (Here, r means “bears a relation to.”) The second trial type presents another task with which the subject has no prior direct experience. The printed words are samples and the pictures are comparisons (CB), reversing the training relations. If the subject proves capable of the new matching performances, demonstrating sample-comparison reversibility, then the word-picture relations are symmetric: ( I ) If CAP r CAP, then CAP r CAP, and (2) if SUN r SUN, then SUN r SUN. Figure 5B shows another set of teaching and test procedures with the same stimuli. Samples on both types of teaching trials are dictated words (Set A). The test trials determine whether all-visual matching performances (BC and CB) emerge without explicit training. This procedure permits a combined evaluation of symmetry and transitivity (Sidman & Tailby, 1982). If the relations among “CAP,” CAP, and C A P , and among “SUN,” SUN, and SUN are symmetric and transitive, and if the subject has demonstrated generalized identity matching (reflexivity), then each set of stimuli (i.e., the dictated word, the picture, and the printed word) is a member of a class of equivalent stimuli. With subjects with intellectual disabilities, more than 20 published studies have demonstrated positive equivalence-test outcomes. Degree of disability has ranged from severe (e.g., Sidman & Cresson, 1973; Sidman, Willson-Morris, & Kirk, 1986) through moderate (Dube et al., 1987; McDonagh, McIlvane, & Stoddard, 1984). to mild (e.g., Spradlin & Dixon, 1976; Spradlin & Saunders, 1986; Stromer & Osborne, 1982). Equivalence classes have also been shown in studies with young normally developing children (Lazar, Davis-Lang, & Sanchez, 1984; Sidman & Tailby, 1982). The phenomena demonstrated in stimulus equivalence research, particularly positive outcomes on transitivity tests, provide perhaps the most compelling support for the four-term account of conditional stimulus control. Strictly speaking, the emergent matching that occurs on transitivity tests is not predictable from the three-term compound stimulus analysis. Recall that compound stimuli are conceptualized as unitary entities in which constituent elements together exert control over behavior; conceptualized in this way, the elements should not be shown to function independently (Sidman, 1986). That is, compound AB is one stimulus and compound BC is another independent stimulus; nothing is entailed with respect to the potential controlling properties of AC. To bypass this conceptual problem, one might first concede that stimulus compounding strictly defined cannot handle the transitivity phenomenon. One might, however, propose an “expanded” three-term account of conditional dis-

80

William J . Mcllvane

crimination in which the concept of the stimulus compound is reformulated. The issue is not whether the sample and comparison stimuli are separate elements; the phenomenon of transitivity shows clearly that they do function independently in some way. What is in question, however, is whether the elements truly function differently. What justification is there, other than the experimenter’s nominal specification, for terming one stimulus the sample (or instructional) srimulus and the other the comparison (discriminative) stimulus and employing a four-term unit of analysis (Thomas & Schmidt, 1989)? One answer might be found in the fact that the sample stimulus may serve additional functions that the comparison stimuli do not and cannot serve. A previous section, for example, reviewed the phenomena of sample-S+ and sample-S- relations in which the same sample stimulus set the occasion for selecting one comparison stimulus and rejecting another. The comparison stimuli, however, obviously cannot function in the same way. When sample-S+ and samples- control and transitivity are demonstrable in the same behavioral baseline (e.g., McIlvane, Kledaras, Munson, King, de Rose, & Stoddard, 1987; Stromer & Osborne, 1982), the requisite requirement of different sample and comparison functions seems to be met. c . Generalized Arbitrary Matching to Sample. On the face of it, generalized learning in arbitrary-matching procedures would not seem possible; lacking a basis in shared physical features, arbitrary stimulus-stimulus relations would seem to require explicit training. This fact notwithstanding, aspects of arbitrary matching performance do seem to generalize. Perhaps the best example is the “exclusion” phenomenon that has been studied in subjects with mental retardation ranging from mild (Dixon, 1977) to profound (McIlvane, Bass, O’Brien, Gerovac, & Stoddard, 1984) and also in young normally developing children (McIlvane, Munson, & Stoddard, 1988; cf. Vincent-Smith, Bricker, & Bricker, 1974). The phenomenon can be illustrated by the following example, by considering a subject who has mastered the conditional discriminations shown in Fig. IE. Suppose that one comparison stimulus, CAP perhaps, was replaced with a novel picture, that SUN remained, and that the dictated sample was also novel, “ZET.” That word has no explicit experimental history, and the subject may never have heard it before. Nonetheless, human subjects virtually always respond to the novel word by immediately and consistently selecting the novel picture. Logically, there are at least two different bases on which a subject might select a novel comparison stimulus in this situation (Dixon et al., 1983). On the one hand, the subject might explicitly reject SUN in the presence of “ZET,” a dictated name not previously related to SUN or any other stimulus in the same class (McIlvane, Munson, & Stoddard, 1988). Alternately, the subject might relate the novel dictated name to the novel picture on the basis of their shared property of novelty (cf. Mervis & Bertrand, in press). Recent evidence from

STIMULUS CONTROL ANALYSIS

81

studies of individuals with severe retardation suggests that both bases may in fact be operable, at least in some subjects (McIlvane, Kledaras, Lowry, & Stoddard, in press). The “arbitrary assignment” phenomenon reported by Stromer (1986) and Saunders, Saunders, Kirby, and Spradlin (1988) also merits discussion in this context. These studies showed that mentally retarded subjects with extensive arbitrary-matching histories spontaneously exhibited conditional discrimination on problems that presented entirely new stimuli. Given an entirely novel matching-to-sample display that presented comparisons X and Y and Sample A, for example, the subject might choose Y. If so, given the same comparison display and B as a sample, the subject would choose X. The arbitrary-assignment phenomenon seems to indicate that subjects learn that each comparison stimulus of a given comparison display is to be related to one and only one sample stimulus. The relationship to the exclusion phenomenon has yet to be determined. d. Transposition and Sequence Classes. Other stimulus relations have their basis in their relative position on continuous physical dimensions such as size, brightness, and spatial contiguity. Perhaps the classic example is the transposition problem (Reese, 1968) in which a subject is taught to select, for example, Form B and to reject a relatively smaller Form A. The question is asked whether training has established stimulus control by the absolute physical properties of Form B or the relational properties-the size disparity. Test trials present Form B and a new Form C, which is larger still. Subjects often select Form C, suggesting that the prior training had established control by the relational properties. Another area of current research interest is the potential relations among stimuli that are responded to in sequence. Historically, sequential responding has been analyzed in terms of chains of stimuli and responses (e.g., Skinner, 1938). More recent research, however, has raised questions about the adequacy of such analyses, even when subjects are nonhumans (e.g.. D’Amato & Columbo, 1989; Terrace, 1986). Research with children and adults of normal capabilities has shown clearly that sequence training results in stimulus classes based on ordinal position within the sequence (Lazar, 1977; Lazar & Kotlarchyk, 1986; Sigurdardottir, Green, & Saunders, 1990; Wulfert & Hayes, 1988). Sequence elements that have served as first, second, third, or fourth in a sequence prove substitutable in other independently trained sequences on the basis of previously established ordinal position. There are a number of reasons to study sequence learning in people who have intellectual disabilities. On the practical side, serial order is important in many functional skill areas, including reading (Elkind, 1975) and arithmetic (e.g., Baroody, 1988). On the basic side, fundamental questions are unanswered about sequence learning. One question, for example, is whether sequence learning in subjects with intellectual disabilities results in true ordinal relations or merely rote performances. McManis’ (1969) study suggests that ordinal concepts may be

82

William J. Mcllvane

lacking in those whose mental age (MA) is less than 10 or 1 1 years. Lutkus and Trabasso’s (1974) data suggest that ordinal concepts may be taught to individuals with MAS of 5 to 6 years (cf. Bryant & Trabasso, 1971). Stromer, Mackay, Cohen, and Stoddard (in press) have recently reported data that demonstrate sequence-class learning resembling that reported in the previously cited studies with normally capable individuals. Moreover, their data suggest the possibility of ordinal learning in individuals with MAS below 3 years. 3. SUMMARY

This section additionally emphasizes the importance of determining which of many possible stimuli, stimulus properties, and stimulus relations are in control of a subject’s behavior at any given moment. It becomes even more apparent that one cannot do this by merely observing the subject simply doing what he or she has been taught to do. What is necessary, in addition, is to vary the stimuli and procedures to which the subject is exposed to gather evidence to support one’s inferences about possible controlling stimulus relations. Moreover, as the subject’s behavior grows more complex, so too does the complexity of procedures that must be used to analyze stimulus control and the theoretical framework necessary to encompass a growing and ever richer data base. With respect to stimulus equivalence, for example, a comprehensive behavioral analysis must specify not only the important controlling aspects of individual stimuli but also the nature of the controlling relations among them in terms of behavioral process. The next section describes a few recent efforts to do so.

C. Theoretical Issues in the Analysis of Stimulus-Stimulus Relations This section considers a few of the more important theoretical issues in analyzing stimulus relations. Emphasis is placed on those that seem most relevant to problems of intellectual disability. 1. STIMULUS EQUIVALENCE: FUNDAMENTAL OR DERIVED PROCESS In analyzing the phenomenon of stimulus equivalence, it is natural to speculate about its origin. One possibility, recently suggested by Sidman (1990), is that positive outcomes on equivalence tests may reflect a fundamental behavioral process that may not be explainable in terms of other processes (cf. Catania, 1984). In support of his suggestion, Sidman cites the apparent failures of nonhumans to pass a1 of the necessary tests (e.g., D’Amato, Salmon, Loukas, & Tomie, 1985). He goes on to suggest that only humans may be capable of stimulus equivalence, and that this capability may underlie language and other behavioral repertoires with generative properties.

STIMULUS CONTROL ANALYSIS

83

Hayes (1990) offers another account, “relational frame theory.” He views positive equivalence test outcomes as only one instance of a general tendency of humans to display “arbitrarily applicable relational responding.” In Hayes’ theory, “stimulus equivalence” is merely one of a larger number of relations that could logically hold among stimuli (relational frames). In Hayes’ view, relational frames do not require new processes. Rather, they emerge as a product of one’s social experience, particularly that related to language. Thus, whereas Sidman suggests that equivalence classes may enable language development, Hayes proposes that language development may enable equivalence classes. In support of Hayes’ theory is also the fact that equivalence classes have not yet been demonstrated in nonhumans who also do not spontaneously develop language skills. Also offered in support is a widely cited study by Devany, Hayes, and Nelson (1986) conducted with young children with mental retardation. Children who displayed some evidence of language skills in a brief assessment also had positive equivalence-test outcomes. Children who did not display such skills also did not display equivalence classes. On its face, this evidence seems to indicate that stimulus equivalence is at least highly correlated with and possibly derived from linguistic experience. The results must be viewed with caution, however. Devany and colleagues’ study has a number of serious methodological limitations which must be acknowledged. Subject matching was questionable (e.g., the verbal children apparently had higher IQs, nonverbal children made many more training errors). In addition, there were several departures from typical equivalence-test procedures. Baselines were not maintained during testing (cf. Hamson & Green, 1990), and circumscribed testing did not provide the opportunity for classes to emerge gradually, a phenomenon seen routinely even in subjects who are much more capable (e.g., Spradlin, Cotter, & Baxley, 1973). Moreover, in a recent effort at systematic replication with verbal and nonverbal deaf children, positive equivalence-class outcomes were reported with one nonverbal deaf child and the possible beginnings of gradual emergence were seen in the other (Barnes, McCullagh, & Keenan, 1990). At the present time, therefore, one cannot make a data-based decision as to whether stimulus equivalence is a fundamental or a derived process. 2. STIMULUS-STIMULUS OR STIMULUS-RESPONSE RELATIONS IN EMERGENT BEHAVIOR Related to the preceding discussion is the issue of whether a given behavior should be analyzed in terms of stimulus-stimulus or stimulus-response relations. In the past century, this general issue has had many incarnations in the analysis of both human and nonhuman behavior (e.g, Trapold & Overmier, 1972). Figures 5A and 5B help to illustrate why it has arisen once again in relation to stimulus equivalence and related phenomena. Not mentioned previously were the DB and DC performances, which involve names spoken by rather than to the subject.

84

Willium J . McIlvane

In past research, the solid DB arrow indicates that subjects have often been able to name the pictures accurately on naming tests conducted at the start of the study. The DC arrow is dashed, however, because those same subjects do not name the printed words. It is frequently observed, however, that oral naming of those printed words does emerge after equivalence classes have been demonstrated. Emergent naming raises questions about the nature of equivalence relations (Dugdale & Lowe, 1990; Mackay & Sidman, 1984; Sidman & Tailby, 1982; Sidman et al., 1986). With respect to Fig. 5B, if the subject becomes capable of ( 1 ) matching the printed names and the pictures to the same dictated names and also of (2) producing those names when shown either type of visual stimulus, then the latter capability raises the possibility that the emergent BC and CB matching results from active or implicit naming (Sidman et al., 1974). The clearest exposition of this possibility suggests that the subject makes naming responses, covert sample naming perhaps, which then serve as the actual controlling stimuli for BC and CB matching. The account has much in common with stimulus-response “coding” analyses of matching to sample (Cumming & Berryman, 1965; Eckerman, 1970; Lowenkron, 1984). Much evidence contradicts this “coding” account, however. For example, classes composed of entirely visual stimuli can be formed (e.g., Stromer & Osborne, 1982). In this instance, if class formation depends on the application of common names, then the names would have to be supplied by the subject. If names were in fact supplied by the subject, one can reasonably argue that the subject should then produce those names if given the opportunity to do so on a naming test. Several studies have investigated this question and none has supported the suggested importance of common names in class formation (Lazar et al., 1984; Sidman & Tailby, 1982; Sidman et al., 1986). There is some controversy, however, whether the naming tests used in those studies were sufficiently sensitive (e.g., Stoddard & McIlvane, 1986). The controversy continues in part because of three as yet unexplained research findings: (1) classes that include dictated word samples apparently form more readily than those composed of all visual stimuli (Green, 1990; Sidman et al., 1986); (2) class formation and emergent matching are apparently facilitated when subjects are explicitly taught to apply names to the stimuli (Dugdale & Lowe, 1990), and the names need not be common to all members of the class; and (3) the findings of Devany et al. (1986). Taken together, these findings suggest a possibly important role for dictated auditory stimuli and their spoken counterparts in class formation. The possibility may be especially interesting in light of the apparent deficiencies shown by some subjects with mental retardation to show equivalence classes (MacDonald, Dixon, & LeBlanc, 1986; Sidman et al., 1986; Spradlin & Saunders, 1986). Perhaps it is not coincidental that members of this population also display production deficiencies (Flavell, Beach, & Chinsky, 1966; Kendler, 1972; cf. Bray, 1979), for example, failing to use spoken names

STIMULUS CONTROL ANALYSIS

85

to mediate delays in tasks that require the subject to remember stimuli that were recently present (Constantine & Sidman, 1975). The important issue, however, is not whether naming or other responses could mediate emergent matching. Research with rhesus monkey subjects (McIntire, Cleary, & Thompson, 1987) has shown clearly that positive equivalence-test outcomes can be simulated when mediating chains are explicitly developed. The issue is instead whether responses must mediate emergent behavior. Historically, questions of this type have proven extremely difficult to answer definitively (cf. Trapold & Overmier, 1972). Although many experiments are done, the findings can often be explained in other ways. Dugdale and Lowe’s findings provide an excellent example of data that on the surface seem compelling, but prove less so on closer examination. They studied a number of normally capable toddler-aged children, and none of them initially showed positive equivalence-test outcomes. Then, the investigators taught the children to name each of the experimental stimuli with a different spoken name. Thereafter, equivalence-test outcomes were immediately positive for most children. Dugdale and Lowe speculated that giving the children names for the stimuli encouraged covert verbal behavior, which then served in some way to mediate the emergent matching-to-sample performances. There are other ways to interpret Dugdale and Lowe’s data, however. Teaching the children to name the stimuli may have had effects other than creating stimuli to play a mediating role, as follows. Positive equivalence-class-test outcomes will not occur if the child performs the matching-to-sample tasks on the basis of mixed sample-S+ and sample-S- relations (cf. Saunders & Green, 1992) or simple compound stimulus control. When children are required to name the stimuli that appear on matching-to-sample trials, they must discriminate each element as separate from the others (cf. Saunders & Spradlin, 1989). Perhaps this successive discrimination requirement helped to encourage the development of the sample-S relations that were required for positive equivalence-class-test outcomes. The intent of this brief review was not merely to point out the limitations of specific studies. Rather, it was to illustrate the types of complexities that come up in trying to argue that ( 1 ) mediating responses were (or were not) involved in the stimulus control of behavior and (2) what the effects of the putative responses were. Perhaps all that may ever be known with certainty is whether empirically some procedures make equivalence classes more likely than others. Fortunately, if one’s concern is only effective teaching of people with intellectual disabilities, the empirical relationships may be sufficient.

+

3. STIMULUS EQUIVALENCE AND TRANSFER OF STIMULUS FUNCTIONS Among the most important features of stimulus equivalence is that it allows for transfer of stimulus functions without explicit teaching (cf. Hayes, 1990;

86

William J . Mcllvane

Spradlin & Saunders, 1984). For example, suppose a student with mental retardation learns that the price of a bus ride home is 65$ and, also, to pay the fare by selecting two quarters, a dime, and a nickel from her coin purse. Suppose further that the student is subsequently taught that several additional coin combinations are equivalent to 65$ (Stoddard, Brown, Hurlbert, Manoli, & McIlvane, 1989). Will she subsequently show that members of the equivalence classes are also functionally equivalent? Will she come to select appropriate alternate coin combinations to pay her bus fare? Questions of this general type-transfer or “generalization’’ of learned skills and strategies-are commonplace in mental retardation research, given the many failures of transfer that have been reported (e.g., Bray, 1979). A research issue is whether function transfers are characteristic when equivalence classes are established or whether classes are merely needed prerequisites. Given findings discussed earlier, the answer to the question might seem obvious. Naming responses occasioned by one member of an equivalence class may not transfer to other members without explicit training. Those transfer failures are difficult to interpret, however. One might also account for them, for example, in terms of variables such as competition from preexisting responses occasioned by those stimuli (Bucher & Keller, 1981). de Rose, McIlvane, Dube, Galpin, and Stoddard (1988) reported a simplified procedure that can be used to study transfer of stimulus functions in subjects with mental retardation. Stimuli A1 and A2 (nonrepresentative forms) were first established as the positive and negative stimuli in a simple simultaneous discrimination procedure. Then, A1 and A2 were presented as samples controlling selections of physically dissimilar forms B2 and B2, respectively, in an arbitrary matching-to-sample procedure. On subsequent probe trials, B I and B2 were presented in the simultaneous discrimination format to determine whether the S and S- functions of the samples transferred to the comparisons. With this simplified procedure, studies have shown that function transfers can be highly reliable in subjects with both moderate mental retardation and normal capabilities (de Rose, McIlvane, Dube, Galpin, & Stoddard, 1988; Smeets, Lancioni, & Striefel, in press). In experiments that ask whether function transfers will occur to all members of an equivalence class, however, transfer has not been reliable (de Rose, McIlvane, Dube, McDonald, & McIlvane, 1991). Even with this simplified preparation, however, apparent function transfer failure may be judged inconclusive. One might argue, for example, that the original stimulus function definition was inaccurate, perhaps that the simple discrimination reflected control by stimulus configurations and that separate S and S- functions had never been established (cf. Smeets et al., in press). These possibilities seem unlikely, however, given that positive equivalence-class-test outcomes are not consistent with configuration control. It seems, therefore, that although equivalence classes may be frequently accompanied by stimulus function transfers, such transfers are not entailed by such classes.

+

+

STIMULUS CONTROL ANALYSIS

87

4. SUMMARY One of the most striking aspects in reviewing recent research on stimulus equivalence is how many seemingly fundamental issues remain to be resolved. Given the very long history of research in this area, it seems surprising that there has yet to develop a consensus on some very basic issues. Perhaps the current state of affairs results in part because the problem does have a long history and has been tackled by investigators with highly diverse theoretical positions and experimental methods. The issue of whether or not stimulus equivalence is mediated by responses, for example, clearly has its roots in theoretical debates that flourished almost 50 years ago. What may be different about today’s research is the intensity and persistence of effort, which appear much greater than at any time in the past. At least some of these issues may be resolved in the next decade. One factor that may continue the current momentum is a new trend toward application of the basic research findings. It may be important in sustaining the area in which much of the research is conducted, implicitly or explicitly, in the context of efforts to improve methods for teaching people with developmental limitations. As practical teaching problems are solved, so too may be more basic theoretical ones. The next section provides an illustrative example of one such application effort.

D.

Techniques for Developing Stimulus Control

The seminal work of Sidman and Stoddard (1966) stands as a model of programmed instruction for individuals with intellectual disabilities. Their efforts and the basic laboratory studies on which they were based (e.g.. Terrace, 1963) have inspired a number of follow-up efforts (reviewed several years ago in this series by Lancioni and Smeets, 1986). The main rationale for these efforts is that programmed instruction appears substantially more effective than mere differential reinforcement (i .e., so-called “trial-and-error’’ procedures) in teaching people with developmental limitations (e.g., Richmond & Bell, 1986; Sidman & Stoddard, 1967). This general rule has exceptions (e.g., Koegel & Rincover, 1976), however, and an accumulating body of research permits one to develop a somewhat more informed perspective on this topic. Continuing a major theme of this article, the focus of the current section will be the development of programmed instructional methods for teaching discrimination skills to people with intellectual disabilities. Sidman and Stoddard’s research and follow-up efforts were typically concerned with teaching a single discrimination or a small number of discriminations rapidly with a low error rate. When the discriminative stimuli were letters or words (e.g., Domy & Zeaman, 1975; Wolfe & Cuvo, 1978), only a handful of stimuli were employed. Left unanswered was whether similar programmed methods could be applied to teach the much larger and more varied repertoire of discrimination skills like those required in rudimentary reading. This question

88

William J . Mcllvune

has been taken up only recently in a research program conducted by the author and his colleagues at the Shriver Center. Although several initial reports have appeared, the overall program has not yet been described. Description in the present context will help to illustrate further some points made in preceding sections. Also presented are brief descriptions of studies that have contributed to program evolution but have not yet appeared in the literature. Our research program is developing programming methods that will bring individuals with limited or essentially no relevant discrimination skills to the point that they can discriminate increasingly complex stimuli like printed words from one another. The entry requirements are minimal. Four response areas (keys) are presented on the screen of a microcomputer-based teaching apparatus, and the individual must learn merely to select the one key that displays a plus sign and to reject the other keys, which are blank (a form-versus-no form discrimination) (Dube & McIlvane, 1989; cf. Sidman & Stoddard, 1966). Thereafter, the discrimination requirements increase progressively. An initial assessment is provided to determine whether the subject has or can quickly acquire relevant match-to-sample skills. If no such skills are evident, they are developed in the following manner. I . ONE-TRIAL DISCRIMINATION LEARNING

The first major milestone is to teach the individual to discriminate line drawings and letterlike forms from one another (Dube, lennaco, Rocco, Kledaras, & Mcllvane, 1992; Mcllvane, Dube, Kledaras, Iennaco, & Stoddard, 1990). The teaching approach is illustrated in Fig. 6. Initially the individual is taught several simple simultaneous “form-versus-form’’ discriminations using relatively lengthy stimulus control shaping programs; shaping methods include fading size

FIG. 6 . Procedures for prompting simple simultaneous form discriminations and systematic withdrawal of prompts leading to one-trial discrimination learning.

STIMULUS CONTROL ANALYSIS

89

or intensity differences and delayed prompting. Representative steps of an intensity fading program are shown in the leftmost column of Fig. 6. When the individual learns discriminations reliably and with a low error rate via these programs, the teaching goal changes. It becomes to establish form-versus-form discriminations with progressively less stimulus control shaping (“fading out the fading,” cf. Herman & Arbeit, 1973). This approach is suggested by the center three columns in Fig. 6, which indicate the more abrupt withdrawal of intensity prompts. The ultimate goal is to establish form discrimination after a single reinforced selection of a form displayed by itself-a form-versus-no form or “demonstration” trial (House & Zeaman, 1958; Moss & Harlow, 1947)-without the need for any shaping at all (Fig. 6, right column). Thus, our use of stimulus control shaping is a marked departure from preceding research. Rather than using the shaping methods to teach individual discrimination, we use them to improve the individuals’ ability to learn discriminations more generally.

2. GENERALIZED IDENTITY MATCHING TO SAMPLE OF SINGLE ELEMENTS The second major milestone is to establish generalized identity matching with the forms. At the end of the OTDL phase, the individual is rapidly learning simple simultaneous discriminations. A program of gradual stimulus changes is then used to transform the demonstration stimulus into a sample stimulus that is present throughout the trial. The procedure becomes nonconditional identity matching to sample, so called because the discriminative functions of the comparison stimuli do not change from trial to trial (Dube, McIlvane, & Green, 1992). The stimulus function changes that characterize conditional identity matching procedures are introduced gradually over several sessions. In early sessions, there are relatively few function reversals and they are separated by a large number of trials that present other stimuli. As training progresses, the number of reversals increases and the number of stimuli and trials intervening between discrimination reversals decreases. Ultimately, the subject displays accurate generalized identity matching even when discriminative function reversals are programmed on irregularly alternating trials (McIlvane, Dube, Kledaras, Iennaco, & Stoddard, 1990). The seemingly elaborate programming to introduce reversals is provided in part because considerable research has shown us that many subjects with mental retardation may have great difficulty with discrimination reversals initially (cf. House & Zeaman, 1963). In one study, for example, we sought to develop a stimulus control shaping method to teach an ambiguous cue discrimination (McIlvane, Gilbert, & McDonald, 1992). Although we were able to teach two independent discriminations [i.e., A/B and C/D(S+/S-)] reliably with few or no errors, the same stimulus control shaping methods failed repeatedly when a discriminative function reversal was required (X/Y and Y/Z).

90

William J. Mcllvane

Another study with a related goal provided similar data. In earlier research (Mcllvane, Kledaras, Dube, Iennaco, & Stoddard, 1988), we sought to develop identity matching through discrimination reversals involving only two forms, for convenience, A and B. Lengthy stimulus control shaping programs were used to establish and reverse simultaneous discriminations in separate training sessions (i.e., AIB on one day and BIA on the following day). Next, we endeavored to establish and then reverse the discrimination within one training session. Almost without exception, the stimulus shaping methods established the initial discrimination (e.g., BIA) without error; those same shaping methods, however, did not suffice to establish the discrimination reversal (e.g., AIB) reliably. This result persisted even when the number of shaping trials to establish the initial discrimination was reduced to a handful and the number of shaping trials for the reversal was increased. Apparently, an immediate history of discrimination learning rendered the stimulus shaping methods much less effective for teaching a new discrimination with the same stimuli (i.e., the discrimination reversal). Perhaps the most striking finding of this kind was obtained in an experiment that sought to teach and reverse two different discriminations in the same session (i.e., Al/A2 and Bl/B2). The experiment was conducted with subjects with severe mental retardation. Here also, protracted and elaborate programming efforts failed to overcome discrimination reversal problems. Therefore, we decided to approach the problem in a different manner. To ask whether there was any potentially competing history that would not interfere with the stimulus control shaping procedure, we examined the effects of a single trial. An 18-trial stimulus control shaping program was used to establish AllA2 and Bl/B2 discriminations in some sessions and A2lA1 and B2/B1 in others. When the former discriminations were taught, the first trial of each session displayed either A2 or B2 alone and its selection was reinforced; when the latter were taught, the first trial displayed A 1 or B1. Thus, subjects began each session with a single reinforced selection of a stimulus whose function was then reversed in the discrimination training that followed immediately. With a BAB design, the results of such sessions (A) were compared with other sessions that were identical in all respects, except that the first trial was omitted (B). Each condition lasted five sessions. Results with one severely retarded adolescent boy, D.B.H., were remarkable. Accuracy scores in both B conditions were intermediate and variable, averaging about 70% correct. D.B.H.’s accuracy scores in the A condition, however, were always 100%. Thus, even a single reinforced selection of a given stimulus by itself interfered with the stimulus control shaping procedure; after the subject had been reinforced for selecting a stimulus, shaping could not establish its S- function errorlessly. Similar results were obtained with other subjects. Results of these studies so far suggest to us that programmed discrimination reversal may be unusually difficult for individuals with moderate and severe intellectual disabilities. D.B.H.’s results in particular make it difficult to see how

STIMULUS CONTROL ANALYSIS

91

discrimination reversal programming could be made more gradual. Results of these studies beg the question of why the gradual introduction of discrimination reversals in the identity-matching program does not lead to problems of a similar type. As yet, the answer is not clear. It seems possible, however, that the program may help to establish the sample as an instructional stimulus (see above), and that an additional function of such a stimulus is to minimize the effects of discrimination reversals. 3. ARBITRARY MATCHING TO SAMPLE

The next phase of the program is to establish the arbitrary matching-to-sample baselines required to teach matching relations involving pictures and their corresponding dictated and printed English words. Baseline development typically proves very easy with dictated sample stimuli and line drawing comparison stimuli. Most subjects who could reasonably acquire and benefit from a rudimentary reading repertoire will obtain at least a basal score on the Peabody Picture Vocabulary Test (PPVT-R; Dunn & Dunn, 1981), essentially a four-choice nonconditional arbitrary auditory-visual matching procedure. We have found that most subjects who are testable with the PPVT-R readily perform other similar auditory-visual matching tasks, probably because their past formal and informal training has included many instances of similar behavior. Difficulty is frequently encountered, however, when one endeavors to establish visual-visual arbitrary matching relations (e.g., between printed words and pictures). Although it is not uncommon to find individuals with intellectual disabilities who have at least some relevant skills, one frequently encounters subjects who have few to none. Hence, the visual-visual baseline may have to be developed de novo. Standard differential reinforcement techniques will be successful with a few subjects, but such training is often excruciatingly slow (e.g.. Sidman et al., 1974). Stimulus control shaping methods (intensity fading, delayed prompting, etc.) may be used in an effort to reach more individuals and to hasten acquisition (e.g., Dube, McIlvane, Maguire, Mackay, & Stoddard, 1989), but unfortunately these simple shaping methods do not work reliably (e.g., McIlvane, Kledaras, Killory-Andersen, & Sheiber, 1989; Rosenberger, Stoddard, & Sidman, 1972). Our research is seeking more effective and reliable programming approaches. a . Sample Stimulus Control Shaping. We are now studying a promising approach to teaching arbitrary matching by transforming our subject’s already established identity-matching baselines. The essential feature of this approach is to transform the physical characteristics of sample stimuli until they no longer resemble the comparisons. Recent studies have demonstrated the efficacy of these sample stimulus shaping methods. Adults with mental retardation and normally capable preschoolers rapidly learned arbitrary matching of nonrepresentative forms after failing to learn with standard training procedures (Zygmont,

92

William J . Mcllvane

Lazar, Dube, & Mcllvane, 1992). An ongoing study has replicated these findings with discrimination tasks that resemble very elementary reading exercises. Subjects with moderate to severe intellectual disabilities are learning to match pictures with corresponding initial letters (A matched with APPLE, B with BOTTLE, S with SHIRT, and so on). Further studies of this general approach have begun to examine stimulus-control shaping methods for teaching relations between pictures and whole words (cf. Miller & Miller, 1968). Figure 7A shows examples of sample stimulus control shaping programs that are used for experimental control purposes. Unlike those just discussed, these arbitrary-matching relations could not have been subject to past teaching efforts. During Phase 1 of the program, the die sample is gradually transformed into the Greek letter lambda; identity-matching trials with the shamrock are interspersed as baseline trials to maintain conditional control by the samples. During Phase 2, the shamrock is transformed into sigma. Baseline trials that display the lambda sample are interspersed to maintain conditionality. Figure 7B presents individual data on the experimental control problem for K.L.V., a young woman with moderate mental retardation. Prior to these sessions, she had failed to learn the arbitrary-matching problem despite extensive differential reinforcement. Figure 7B presents cumulative records of performance on four-trial blocks of teaching trials in the initial two sessions of Phase 1 (upper records) and Phase 2 (lower records); each block contained two shaping trials and two baseline trials. After every errorless block, the shaping program advanced one step, indicated by upward steps in the record. Any error in the block resulted in a backup to the previous shaping step in the next block, indicated by downward steps. Overall, the sample shaping proceeded rapidly and with relatively few errors. Note that K.L.V. made no errors until the penultimate program Step 7 during dieto-lambda shaping and achieved a 5 / 5 score on final performance trials in that session. The program was repeated in the following session with Step 6a added between Steps 6 and 7 to forestall any further difficulties. Repetition was scheduled because other subjects with comparable intellectual functioning often have not retained the final performance from one day to the next. Repeating the program and then gradually more abbreviated versions of it over several days seems to overcome the retention problem. The first Phase 2 session began with a brief review of the final steps of the Phase 1 program. Thereafter, shamrock-tosigma shaping commenced and proceeded without error until Step 7 of that program. Again, a new Step 6a was added to the program, and K.L.V. completed the shaping series with only one error in her next session. Overall, the sample-shaping program was conducted over 10 sessions that presented a total of 525 trials, including shaping, baseline, and retention combined, and K.L. V. made only 21 errors (96% correct overall). Notably, subsequent positive symrne-

93

STIMULUS CONTROL ANALYSIS

A

8

PHASE1

PHASE 2

8

K.L.V.

7

6a 6

PHASE 1 Die-to-Lambda

3 2 1

a!6

5 4

3 2 1

5 4 3 2 1

_i

#

Y

11/12

PHASE 2 Shamrock-lo-Sigma

3 2 1

SUCCESSIVE FOUR-TRIAL BLOCKS

FIG. 7. Stimuli used in a sample stimulus-shaping program to teach arbitrary matching (A) and results of the first two shaping sessions in each program phase (B).

94

William J . Mcllvane

try test outcomes showed that K.L.V. was immediately capable of also matching the letter comparison to picture samples without the need for any sample stimulus shaping. b. Other Programming Methods. In our work so far, sample stimulus shaping has proven to be a straightforward and effective method for establishing the first instances of arbitrary matching. Recent studies have also reported two other methods for teaching arbitrary matching to subjects who do not learn via standard methods. One method, reported by McIlvane, Dube, Kledaras, Iennaco, and Stoddard (1990), establishes contingency classes (Sidman, Wynne, Maguire, & Barnes, 1989) via yoked repeated reversal of simple discriminations, and then systematically transforms this baseline into arbitrary matching. A second method, reported by Saunders and Spradlin (1989, 1990), requires the subject to make differential responses to the sample stimuli, thereby ensuring successive discriminations among them. This method also provides programmed introduction of discrimination reversals to establish and maintain simultaneous discrimination of the comparison stimuli. Both procedures have potentially troublesome features. First, both appear to require a relatively large number of training sessions. Second, the early introduction of discrimination reversals may lead to many errors, even when gradual programming is employed. Where appropriate, therefore, sample stimulus shaping seems to be the method of first choice, with the others available to fall back on, as necessary. c . Learning by Exclusion. Once established, one may be able to extend auditory-visual and visual-visual arbitrary matching baselines via the “learning by exclusion” method (McIlvane & Stoddard, 1981). This procedure takes advantage of the replicability and reliability of the exclusion phenomenon that was discussed earlier. In this method, selections on exclusion trials are merely reinforced, and subsequent test trials are then scheduled to assess learning. For example, one might give a naming test to determine whether the subject will produce the novel name when shown the novel picture (e.g., McIlvane, Bass, O’Brien, Gerovac, & Stoddard, 1984; McIlvane, in press). The exclusion method has several advantages. It is conceptually straightforward, relatively easy to implement, and often highly successful (e.g., McIlvane, Kledaras, Lowry, & Stoddard, in press; Stoddard et al., 1989). Unfortunately, however, the procedure has proven unreliable in certain applications. Subjects may merely display exclusion and repeatedly fail outcome tests (the “exclusiononly” outcome; McIlvane, Munson, & Stoddard, 1988), even with protracted exposure to the procedure (e.g., Dixon, 1977). It is not yet clear why the procedure sometimes produces extraordinarily rapid learning in some circumstances and apparently little or none in others. We are currently asking whether a training program can be developed that will teach subjects who do not initially learn via the exclusion method to do so.

STIMULUS CONTROL ANALYSIS

95

4. GENERALIZED IDENTITY MATCHING OF PRINTED WORDS

Given reliable identity- and arbitrary-matching baselines, the next step is to determine whether the subject will display generalized identity matching of two-, three-, four-, and five-letter printed words. Use of these more complex stimuli raises the issue of how to ensure that the subject’s selections are in fact controlled by the critical stimulus features-all of the letters and their serial position. To meet this requirement, we employ so-called “constructed-response” matchingto-sample (CRMTS) procedures (Mackay & Sidman, 1984; McDonagh et al., 1984). The CRMTS procedure is shown in Fig. 8A. On each CRMTS trial, subjects are presented with a sample stimulus and a “choice pool” consisting of ten individual letters. Subjects are taught to construct identical arrays by selecting letters from the choice pool that match the sample letters. Sample letters are matched in left-to-right order; for example, given the sample EKS and the choice pool D S V T E B K 0 I M, the subject is taught to select the E, then the K, and finally the S. As each letter is selected, it moves from the choice pool to a position immediately below the matching sample letter. Thus, when all letters have been selected, the subject has “constructed” an array of letters that matches the sample. A different sample stimulus is presented on every trial to verify sample discrimination. Until recently, studies of the CRMTS method included relatively few subjects. We have begun to assess their general effectiveness, and to develop stimuluscontrol-shaping methods that might hasten acquisition of accurate CRMTS. Eighteen subjects have participated in development efforts to date. The entry requirement was highly accurate generalized identity matching with all individual letters and numerals in a two-choice format. When subjects were given individual letters in the constructed-response context, all but one continued to match with high accuracy as the number of comparison stimuli was increased from 2 to 10. In an early stage of program development, four subjects were exposed to twoletter samples with a flashing prompt. The letter to be selected flashed “on and off.” Flashing controlled responding for three of four subjects, and these three continued to perform accurately when flashing was discontinued. The remaining 13 subjects were given pretests of CRMTS with two-letter samples before any training procedure was introduced. Seven of these subjects’ accuracy scores were 85% or better. Six subjects did not demonstrate generalized CRMTS with twoletter samples; almost all errors were due to matching the right sample letter first on approximately half of the trials. For example, given the sample SP, the initial response was to P. With one exception, the PPVT-R MA scores for all subjects

A

a

b

Fl Kl LJ d

C

C

------------------

I

names

C

FIG. 8. (A) Constructed-response identity-matching-to-sample procedure. (B)Displays a and b: Constructed-response procedures for teaching word:picture relations. word construction to identical word prompt in the presence of the picture. Displays c , d , and e: Procedures for fading out word prompts. Displays f and g: Word construction in the presence of the picture alone. (C) Network of directly trained (solid arrows) and potentially emergent (dashed arrows) stimulus-stimulus and stimulus-response relations established by constructed-response (anagram) procedures.

STIMULUS CONTROL ANALYSIS

97

who passed the pretest were 4:8 or higher; with one exception, MA scores for all who failed were 4:7 or lower (cf. McIlvane, Kledaras, Killory-Andersen, & Sheiber, 1989). Those who failed the pretest were exposed to one or more stimulus-controlshaping procedures that attempted to prompt CRMTS performance. Various shaping methods were employed, using prompting differences such as relative size, intensity, and/or position. Five of six subjects ultimately learned via one of these methods, but typically after one or more of the other methods had been tried unsuccessfully. The remaining subject did not learn via these methods, but considering her case will help to exemplify how stimulus control analytic methods are used to diagnose learning problems. a. An Illustrative Case. M.L.D. was a 15-year-old deaf young woman with moderate mental retardation. Her CRMTS performance was virtually errorless when samples were single letters, but fell to approximately 50% correct when two-letter samples were introduced. All errors were due to matching the letter on the right first. As noted, these problems continued despite the use of several stimulus-control-shaping procedures. To devise an appropriate remedial procedure, it was necessary to determine the variables determining M.L.D.'s defective performance. One possibility was that her initial CRMTS selection was controlled by only one sample letter, on some trials the left letter and on other trials the right letter. If so, then further training would be geared toward broadening the range of controlling stimuli (Schreibman et al., 1982). A second possibility was that M.L.D. observed both sample letters on every trials, but not their relative (left/right) positions. If so, continued training might investigate procedures like those for remediation of letter reversals (e.g., Nolan & Siegel, 1980; Stromer, 1975). A series of discrimination tests evaluated these possibilities. M.L.D. was given standard (i.e., not constructed-response) three-choice matching-to-sample tests with two-letter stimuli. On common-letter tests, each S- had one letter in common with S + (e.g., given sample AB, comparisons were AB, AC, and DB). If matching were controlled by only one sample letter per trial, then selection of all three comparisons would be equally likely. On reversed-letter tests, one S- was composed of the S + letters in reversed order and the other S- had one letter in common with S + (e.g., given sample AB, the comparisons were AB, BA, and AC). If matching were controlled by both letters but not their relative positions, then selection of S + and the reversed-order Swould be equally likely, but there would be no selections of the comparison with only one letter in common with S + . On common-letter tests, M.L.D.3 accuracy scores were always 100% correct. On reversed-letter tests, however, mean accuracy was 55% correct, and all errors were selections of the reversed-letter S- comparison. Subsequent testing went on to show that M.L.D. could not discriminate letter combinations from their reversals even in a simultaneous

William J . Mcllvane

discrimination procedure, which merely required her to select the same letter combination over and over again. Thus, stimulus and procedural variation successfully isolated the basis for M.L.D.’s discrimination problems. She can readily discriminate individual letters and certain letter combinations. What she cannot yet do is discriminate letter order, and further training is being provided in an effort to supply this missing prerequisite. b. More Complex Sample Stimuli. CRMTS with three-, four-, and five-letter samples has been tested with 13 subjects. Seven were highly accurate with differential reinforcement only, and four others attained high accuracy after brief exposure to one of the stimulus-control-shaping methods. Two remaining subjects showed decreasing accuracy as a function of number of sample letters when tested with differential reinforcement only: For example, one subjects’ scores with three-, four-, and five-letter samples were 92, 80, and 64%, respectively. 5. TEACHING W0RD:PICTURE RELATIONS VIA STIMULUS CONTROL SHAPING The CRMTS method is also used to establish relations between spoken and dictated words and corresponding pictures. Figure 8B provides an example of an “anagram” procedure in which the subject constructs the word DOG in response to DOG (Mackay, 1985). To do so, the subject first selects the D from a choice pool, and places it over the D adjacent to the picture. Likewise, in series, the subject selects the letters 0 and then G. All letter selections can be prompted initially by identical letters on the choice card. As training progresses, however, the prompts are gradually faded out: first, the last letter of the word, then the middle, and finally the first. At this point, the picture, DOG, now occasions construction of D-0-G. The next step is to determine the nature of the controlling relations that lead to construction of DOG. The anagram training may have led merely to the acquisition of an extended chained performance cued by the picture. Or, we might find that the training had established relations between the word as a compound (i.e., DOG) and its picture, as well as its dictated and spoken counterparts. The possibly emergent relations are illustrated in Fig. 8C. Assume that the subject began training with relations AB and BD intact and that relation BE was trained as described above. If the response of constructing DOG was merely a rote chain, then the subject would not be capable of the emergent relations indicated by the broken arrows (AE, BC, CB, and CD). If, however, the anagram training had established the printed word DOG as a compound stimulus related to DOG, the picture, then all of those relations could emerge. This outcome has been reported in studies by Mackay (1985, 1991), Dube, McDonald, McIlvane, and Mackay (19911, and others (see below). In addition, some subjects who could copy printed words proved immediately capable of writing the correct word in response to both pictures and dictated words.

STIMULUS CONTROL ANALYSIS

99

Because stimulus-control-shaping procedures likes those just described are not effective will all subjects, we are also investigating alternative programming methods. One highly promising method, termed the delayed consrrucrion merhod, was reported by Stromer and Mackay (in press-a, in press-b). The procedure presents complex sample stimuli composed of both a picture and a printed word; for example, a sample complex might involve DOG and the printed word DOG. A touch to the sample removes it from the display and immediately produces the comparison stimuli. On some trials the comparison stimuli are pictures, and the correct selection is the matching picture. On other trials, a letter pool appears and selecting the letters D, 0, and G in that order is the correct behavior. During sample presentation, the subject cannot predict which comparison array will be displayed. In studies conducted with children with and without mental disabilities, Stromer and Mackay asked whether the complex-sample procedure was SUEcient to relate pictures and printed words to one another. This could occur because the procedure encouraged looking at and jointly remembering both the picture and the printed word. This procedural requirement probably limits problems of “blocking” and restricted stimulus control like those recently reported by Singh and Solman (1990). Studies so far have produced highly promising results, including the routine emergence of a network of untrained performances like that shown in Fig. 8C. The exclusion method and variants of the CRMTS procedure have also been used with some success to teach new printed word-picture relations by students in the Shriver Center UAP training program (Clark, 1985; Holcomb, 1980). In these procedures, students with moderate retardation copied the letters of the comparison stimulus by typing or writing. It is possible that this complex differential response requirement may have prevented the exclusion-only outcome that is sometimes seen when subjects merely touch the comparison stimuli. Holcomb’s study was particularly impressive because it demonstrated one-trial acquisition of new printed word-picture relations and virtually instantaneous emergence of a network of untrained matching and word construction relations.

6. TEACHING MORE COMPLEX ASPECTS OF RUDIMENTARY READING The material covered so far has concerned primarily relations involving single words (dictated, spoken, printed, or “written”) and their corresponding pictures. A functional reading repertoire, however, has a number of additional requirements. For example, it is necessary to incorporate other and more complex visual stimuli to establish similar relations involving verbs, adjectives, and other word classes. In addition to these primitive semantic relations, there is the need to address sequential relations among words, and thereby to begin analyses of stimulus relations that may be prerequisites for rudimentary syntax. Little rele-

100

William J . Mcllvane

vant research has been conducted with mentally retarded subjects, although related matters are examined in research on so-called matrix training procedures (e.g., Bunce, Ruder, & Ruder, 1985; Goldstein, 1985). Research on the development of sequence classes would also seem to be relevant, but only a few studies have been conducted with subjects with mental retardation (Mackay, Stoddard, & Spencer, 1989; Stromer et a]., in press). Only one study (Corder, 1982) used printed words as stimuli. A subject with moderate mental retardation was taught to point sequentially to the three stimuli in each of several stimulus sets. For example, the subject was to point first to the printed word HER, next to FAT, and finally to DOG. He then learned similar performances with two additional stimulus sets. Subsequent tests provided evidence that these performances were not merely rote response chains. The subject proved capable, without further training, of pointing correctly to all possible novel combinations of the training words. Although further and more comprehensive studies of this type will obviously be necessary, progress so far suggests that there are no conceptual obstacles to applying stimulus control methods to develop more complex repertoires. Rather, it appears to be primarily a matter of continuing the present work, broadening its scope, and solving the technique-related problems that seem inevitable when new topics are addressed. 7. TECHNICAL SUPPORT FOR STIMULUS CONTROL TEACHING TECHNOLOGY A particularly important and timely feature of the stimulus control methods described in this section is that they have been implemented almost entirely on inexpensive microcomputer equipment. There is burgeoning general interest in computer-assisted instruction (CAI), and it has begun to influence research on topics relevant to special education (Comers & Detterman, 1987; Goldenberg, 1984; Lally, 1981; Tashjian, 1984). Stimulus control research like that reviewed here is likely to prove highly relevant to this enterprise because currently available CAI material is apparently primitive, representing little more than an effort to “deliver flash cards electronically” in many cases (Comers et al., 1986, p. 126). This state of affairs suggests that the sophistication of the teaching technology employed so far does not yet match the sophistication of the microcomputer. Well-developed and validated stimulus control procedures are likely to help correct this problem. They promise precise skill taxonomies (cf. Holland, Solomon, Doran, & Frezza, 1976) that are especially appropriate for implementation on computer. Computer-generated graphics and computer-controlled procedures are particularly useful given that many stimulus control methods require more-or-less elaborate stimulus preparation and presentation capabilities. In the past, even certain relatively simple stimulus control procedures were difficult to justify because of the time and expense required to prepare stimulus materials. These facts in part

STIMULUS CONTROL ANALYSIS

101

led Etzel and LeBlanc (1979) to put forth their well-known “parsimony” principle with respect to selecting instructional procedures. According to this principle, stimulus-control-shapingprocedures would be used only after simpler methods like “trial-and-error” had proven unsuccessful or grossly inefficient. With appropriate computer support, however, the parsimony principle may come to have different implications. Stimulus-control-shaping methods might be considered the procedures of choice earlier and possibly even first in certain applications. Perhaps even more important, computer technology has begun to make feasible the routine use of a range of stimulus control analytic techniques like those described in this article. By taking advantage of computer branching capabilities, instructional programs can be implemented. If programs are unsuccessful, stimuli and procedures can be varied to diagnose learning problems and to select remedial strategies. Although this potential application of computer technology has been appreciated for decades, there have been virtually no efforts to exploit it in special education. Perhaps this is because special educators are not by and large sophisticated computer users. With the growing availability of appropriate authoring systems, however, the potential may come to be realized to a greater degree in the future.

IV.

GENERAL SUMMARY AND CONCLUSION

Research conducted over the past several decades has established the scientific foundation for a fine-grain analysis of progressively more complex behavioral phenomena. Both conceptual and procedural developments now permit detailed examination not only of control by specific stimuli established via direct training but also of the stimulus-stimulus relations that underlie the emergence of novel behavior. These developments form the basis for an empirical science of teaching that can reach many people with moderate and severe intellectual disabilities, particularly those who do not respond reliably to verbal instruction. Moreover, principles and techniques of this science have been applied with great success to teach rudimentary preacademic and academic skill repertoires to such individuals. Very recent research has contributed a number of new techniques that are likely to quicken the pace of development and application, particularly given the growing support that newly available computer technology can provide. ACKNOWLEDGMENTS The author gratefully acknowledges the continuing support of the National Institute of Child Health and Human Development, reflected currently in Grants HD-25995 and HD-28141. Also acknowledged is support from the Department of Mental Retardation of the Commonwealth of Massachusetts (Contract I OOZ ZOO23SC).

102

William J . Mcllvane

Thanks to William Dube, Robert Stromer, and Joseph Spradlin for various helpful suggestions and contributions and also to Joanne Kledaras, Fay Gilbert, Steve McDonald, and Deanna Antonacci for assisting with data analysis and manuscript preparation.

REFERENCES Allen, K. D., & Fuqua, R. W. (1985). Eliminating selective stimulus control: A comparison of two procedures for teaching mentally retarded children to respond to compound stimuli. Journal of Experimental Child Psychology, 39, 55-71. Barnes, D., McCullagh, P. D., & Keenan, M. (1990). Equivalence class formation in non-hearing impaired children and hearing impaired children. Analysis of Verbal Behavior, 8 , 19-30. Baroody, A. J. (1988). Number-comparison learning by children classified as mentally retarded. American Journal on Mental Retardation, 92, 461-471. Bijou, S. W. (1968). Studies in the experimental development of left-right concepts in retarded children using fading techniques. In N. R. Ellis (Ed.), International review of research in mental retardation (Vol. 3, pp. 65-96). 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 menral deficiency, psychological theory and research (pp. 569-617). Hillsdale, NJ: Erlbaum. Bray, N. R. (1979). Strategy production in the retarded. In N. R. Ellis (Ed.), Handbook of mental deficiency. psychological theory and research (pp. 699-726). Hillsdale, NJ: Erlbaum. Bryant, P. E., & Trdbasso, T. (1971). Transitive inference and memory in young children. Nature, 232, 456-458. Bucher, B., & Keller, M. F. (1981). Transfer to productive labeling after training in comprehension: Effects of three training variables. Analysis and Intervention in Developmental Disabilities, I , 3 15-33 1. Bunce, B. H.,Ruder, K. F., & Ruder, C. C. (1985). Using the miniature linguistic system in teaching syntax: ’lbo case studies. Journal of Speech and Hearing Disorders, 50, 247-253. Bush, K. M., Sidman, M., & de Rose, T. (1989). Contextual control of emergent equivalence relations. Journal of the Experimental Analysis of Behavior, 51, 29-45. Carter, D. E., & Werner, T. J. (1978). Complex learning and information processing by pigeons: A critical analysis. Journal of the Experimental Analysis of Behavior, 29, 565-601. Catania, A. C. (1984). Learning (2nd ed.). Englewood Cliffs, NJ: Prentice-Hall. Clark, J. C. (1985). The effects of an overt differential response to comparison stimuli in matching to sample. Unpublished master’s thesis, Northeastern University, Boston, MA. Conners, F. A., Caruso, D. R.. & Detterman, D. K. (1986). Computer-assisted instruction for the mentally retarded. In N. R. Ellis & N. W. Bray (Eds.), International review of research in mental retardation (Vol. 14. pp. 105-134). New York: Academic Press. Conners, F. A., & Detterman, D. K. (1987). Information-processing correlates of computer-assisted word learning by mentally retarded students. American Journal of Mental Deficiency, 91, 606612. Constantine, B., & Sidman, M. (1975). The role of naming in delayed matching to sample. American Journal of Mental Deficiency, 79, 680-689. Corder, C. M. (1982). Extending stimulus class formation to sequence-class membership via an exclusion procedure. Unpublished master’s thesis, Northeastern University, Boston, MA. Cumming, W. W., & Benyman, R. (1965). The complex discriminated operant: Studies of matchingto-sample and related problems. In D. I. Mostofsky (Ed.), Stimulus generalization (pp. 284330). Stanford. CA: Stanford Lhiversity Press. D’Amato, M. R., & Columbo, M. (1989). Serial learning with wild card items by monkeys (Cebus

STIMULUS CONTROL ANALYSIS

103

apella): Implications for knowledge of ordinal position. Journal of Comparative Psychology, 103, 252-261. D’Amato, M. R . , Salmon, D., Loukas, E., & Tomie, A. (1985). Symmetry and transitivity of conditional relations in monkeys (Cebus apella) and pigeons (Columba livia). Journal of the Experimental Analysis of Behavior, 44, 35-47. de Rose, J. C., Mcllvane, W. J., Dube, W. V., Galpin, V., & Stoddard, L. T.(1988). Emergent simple discrimination established by indirect relation to differential consequences. Journal of the Experimental Analysis of Behavior, 50. 1-20. de Rose, J. C., Mcllvane, W. J., Dube, W. V., & Stoddard, L. T. (1988). Stimulus class formation and functional equivalence in moderately retarded individuals’ conditional discrimination. Behavioural Processes, 17, 167- 175. Devany, J. M., Hayes, S.C., & Nelson, R. 0. (1986). Equivalence class formation in language-able and language-disabled children. Journal of the Experimental Analysis of Behavior, 46, 243257. Dixon, L. S. (1977). The nature of control by spoken words over visual stimulus selection. Journal of the Experimental Analysis of Behavior, 27, 433-442. Dixon, M. H., & Dixon, L. S. (1978). The nature of standard control in children’s matching-tosample. Journal of the Experimental Analysis of Behavior, 30, 205-2 12. Dixon, M. H., Dixon, L.S., & Spradlin, J. E. (1983). Analysis of individual differences of stimulus control among developmentally disabled children. Advances in Learning and Behavioral Disabilities, 2 , 85-1 10. Dony, G. W., & Zeaman, D. (1975). Teaching a simple reading vocabulary to retarded children: Effectiveness of fading and non-fading procedures. American Journal of Mental Deficiency, 79, 711-716. Dube, W. V., lennaco, F. M., Rocco, F. I., Kledaras, J. B., & Mcllvane, W.I. (1992). Generalized identity matching to sample: An analysis and new programming techniques. Journal of Behavioral Education, 2 , 29-51. Dube, W. V., McDonald, S. J., & Mcllvane, W. J. (1991). A note on the relationship between equivalence classes. Experimental Analysis of Human Behavior Bulletin, 9, 7- 11. Dube, W. V., McDonald, S. J., McIlvane, W. J., & Mackay, H. A. (1991). Constructed-response matching to sample and spelling instruction. Journal of Applied Behavior Analysis, 24, 307317. Dube, W. V., & McIlvane, W.I. (1989). Adapting a microcomputer for behavioral evaluation of mentally retarded individuals. In J. A. Mulick & R. F. Antonak (Us.Transitions ), in mental retardation, (Vol. 4, pp. 104-127). Nowood, NJ: Ablex. Dube, W. V., McIlvane, W. J., & Green, G. (1992). An analysis of generalized identity matching-tosample test procedures. The Psychological Record, 42, 17-28. Dube. W. V., Mcllvane, W. J., Mackay, H. A,, & Stoddard, L. T. (1987). Stimulus class membership established via stimulus-reinforcer relations. Journal of the Experimental Analysis of Behavior, 47, 159-175. Dube, W.V., McIlvane, W. J., Maguire, R. A,, Mackay, H. A,. & Stoddard, L. T.(1989). Stimulus class formation and stimulus-reinforcer relations. Journal of the Experimental Analysis of Behavior, 51, 65-76. Dugdale, N., & Lowe, C. F. (1990). Naming and stimulus equivalence. In D. E. Blackman & H. Lejeune (Eds.), Behaviour analysis in theory and practice: Contributions and controversies (pp. 115-138). Hilkddie, NJ: Erlbaum. Dunn, L. M., & Dunn, L. M. (1981). Peabody Picture Vocabulary Test-Revised. Circle Pines, MN: American Guidance Service. Eckerman, D. A. (1970). Generalization and response mediation of a conditional discrimination. Journal of the Experimental Analysis of Behavior, 13, 301-316.

William J . Mctlvane

1 04

Elkind, D. (1975). Child development and education. Canadian Psychological Review, 16. 81-87. Ellis, N. R. (1979). Handbook of mental deficiency. psvchological theory and research. Hillsdale, NJ: Erlbaum. Etzel, B. C., & LeBlanc, J. M. (1979). The simplest treatment alternative: The law of parsimony applied to choosing appropriate instructional control and errorless-learning procedures for the difficult-to-teach child. Journal of Autism and Developmental Disorders, 9. 361-382. Flavell, J. H., Beach, D. R., & Chinsky, J. M. (1966). Spontaneous verbal rehearsal in a memory task as a function of age. Child Development, 3 7 , 283-299. Catch, M. B., & Osborne, J. G. (1989). Transfer of contextual stimulus function via equivalence class development. Journal of the Experimental Analysis of Behavior, 5 I , 369-378. Gibson, E. J., & Levin, H. (1976). The psychology of reading. Cambridge, MA: MIT Press. Goldenberg, E. P. (1984). Computers in the special education classroom: What do we need, and why don’t we have any? In 1. A. Mulick & B. L. Mallory (Eds.), Transitions in mental retardation (Vol. I , pp. 107-127). Nonvood, NJ: Ablex. Goldiamond, I. (1962). Perception. In A. I. Bachrach (Ed.), Experimental foundations of clinical psychology. New York: Basic Books. Goldiamond, 1. (1966). Perception, language, and conceptualization rules. In B. Kleinmuntz (Ed.), Problem solving (pp. 183-224). New York: Wiley. Goldstein, H. (1985). Enhancing language generalization using matrix and stimulus equivalence training. In S . F. Warren & A. K. Rogers-Warren (Eds.), Teaching functional language: Gunera/ization and maintenance of language skills ( pp. 225-249). Baltimore, MD: University Park Press. Collin, E. S . , & Liss, P. (1962). Conditional discrimination in children. Journal of Comparative and Physiological Psychology, 5 5 , 850-855.

Green, G. ( 1990). Differences in development of visual and auditory-visual equivalence relations. American Journal on Mental Retardation, 95, 260-270. Harrison, R. J., & Green, G. (1990). Development of conditional and equivalence relations without differential consequences. Journal of the Experimental Anal.vsis of Behavior, 54, 225-237. Harris-Vanderheiden, D., Brown, W. P., MacKensie, P., Reinen, S., & Scheibel, C. (1975). Symbol communication for the mentally handicapped. Mental Retardation, 13. 34-37. Hayes, S . C. (1990). A relational control theory of stimulus equivalence. In L. J. Hayes & P. N. Chase (Eds.), Dialogues on verbal behavior. Reno, NV: Context Press. Herman, L. M., & Arbeit, W. R. (1973). Stimulus control and auditory discrimination learning sets in the bottlenosed dolphin. Journal of the Experimental Analysis of Behavior, 19. 379-394. Holcomb, W. L. (1980). Acquisition of reading prerequisites via a two-choice discrimination and stimulus class formation. Unpublished master’s thesis, Northeastern University, Boston, MA. Holland, J. G., Solomon, C., Doran, I . , & Frezza, D. A. (1976). The ana/.ysis of behavior in planning instruction. Reading, MA: Addison- Wesley. Honig, W. K., & Urcuioli, P. J. (1981). The legacy of Guttman and Kalish (1956): Twenty-five years of research on stimulus generalization. Journal of the Experimental Anal.vsis of Behavior, 3 6 , 405-445. House, B. J., Hanky, M. I., & Magid, D. F. (1979). A limitation on the law of effect. American Journal of Mental Deficiency, 84, 132- 136. House, B. J., & Zeaman, D. (1958). Reward and nonreward in the discrimination learning of imbeciles. Journal of Comparative and Physiological Psychology, 5 1 , 614-619. House, B. J., & Zeaman, D. (1963). Learning sets from minimal stimuli in retardates. Journal of Comparative and Physiological Psychology, 5 6 , 614-619. Jenkins, J. J. (1963). Mediated associations: Paradigms and situations. In C. N. Cofer & B. S. Musgrave (Eds.), Verbal behavior and learning: Problems and processes ( pp. 210-245). New York: McGraw-Hill. Kendler, T. S. (1972). An ontogeny of mediational deficiency. Child Development, 43, I- 17.

STIMULUS CONTROL ANALYSIS

105

Koegel, R. L.. & Rincover, A. (1976). Some detrimental effects of using extra stimuli to guide learning in normal and autistic children. Journal of Abnormal Child Psychology, 4 . 59-71. Lally, M. ( I98 I ). Computer-assisted teaching of sight-word recognition for mentally retarded school children. American Journal of Mental Dejiciency, 85, 383-388. Lancioni, G. E., & Smeets, P. M. (1986). Procedures and parameters of errorless discrimination training with developmentally impaired individuals. In N. R. Ellis (Ed.), fnternational review of research in mental retardation (Vol. 14, pp. 135-164). New York: Academic Press. Lashley, K. S. (1938). Conditional reactions in the rat. Journal of Psychology, 6 , 31 1-324. Lashley, K. S., & Wade, M. (1946). The Pavlovian theory of generalization. Psychological Review, 53, 72-87. Lazar, R. (1977). Extending sequence-class membership with matching to sample. Journal of rhe Experimental Analysis of Behavior, 27, 381-392. Lazar, R., Davis-Lang, D., & Sanchez, L. (1984). The formation of visual stimulus equivalence in children. Journal of the Experimental Analysis of Behavior, 4 1 , 251-266. Lazar, R. M., & Kotlarchyk, B. I. (1986). Second-order control of sequence-class equivalences in children. Behavioural Processes, 13, 205-213. Lowenkron, B. (1984). Coding responses and the generalization of matching to sample in children. Journal of the Experimental Analysis of Behavior, 4 2 , 1-18. Lutkus, A , , & Trabasso, T. (1974). Transitive inferences by pre-operational retarded adolescents. American Journal of Mental Deficiency, 78, 599-606. Lynch, D. C., & Green, G. (1991). Development and crossmodal transfer of contextual control of emergent stimulus relations. Journal of the Experimental Analysis of Behavior, 56, 139- 154. MacDonald. R. P. F.. Dixon, L. S . , & LeBlanc, J. M. (1986). Stimulus class formation following observational learning. Ana1.vsis and fntervention in Developmental Disabilities, 6 , 73-87. Macintosh. N. J. (1977). Stimulus control: Attentional factors. In W. K. Honig & J. E. R. Staddon (Eds.), Handbook of operant behavior (pp. 481-513). Englewood Cliffs, NJ: Prentice-Hall. Mackay, H. A. (1985). Stimulus equivalence in rudimentary reading and spelling. Analysis and Intervention in Developmental Disabilities, 5 , 373-387. Mackay, H. A. (1991). Stimulus equivalence. In B. Remington (Ed.), The challenge of severe mental handicap: An applied behaviour analytic approach ( pp. 235-259). London: Wiley. Mackay, H. A,, & Sidman, M. (1984). Teaching new behavior via equivalence relations. In P. H. Brooks, R. Sperber, & C. MacCauley (Eds.), Learning and cognition in the mentally retarded (pp. 493-513). Hillsdale, NJ: Erlbaum. Mackay, H. A , , Stoddard, L. T., & Spencer, T. J. (1989). Symbols and meaning classes: Multiple sequence production and the emergence of ordinal stimulus classes. The Experimental Analysis of Human Behavior Bulletin, 7 , 16- 17. Matson, J. L.. & Mulick, J. A. (1991). Handbook of mental retardation. Elmsford, NY: Pergamon Press. McDonagh, E. C., Mcllvane, W. J., & Stoddard, L. T. (1984). Teaching coin equivalences via matching to sample. Applied Research in Mental Retardation, 5 , 177- 197. McIlvane, W. J . , Bass, R. W., O'Brien, J. M., Gerovac, B. J., & Stoddard, L. T. (1984). Spoken and signed naming of foods after receptive exclusion training in severe retardation. Applied Research in Mental Retardation. 5 , 1-27. Mcllvane, W. J., & Dube, W. V. (in press). Stimulus control shaping and stimulus control topographies. The Behavior Analyst. McIlvane, W. J., Dube, W. V., Green, G., & Serna, R. W. (in press). Programming conceptual and communication skill development: A methodological stimulus class analysis. In A. P. Kaiser & D. B . Gray (Eds.), Enhancing childrens' communication. Baltimore, MD: Paul H. Brookes. McIlvane, W. J.. Dube, W. V., Kledaras, J. B., Iennaco, F. M., & Stoddard, L. T. (1990). Teaching relational discrimination to mentally retarded individuals: Some problems and possible solutions. American Journal on Mental Retardation, 95, 283-296.

106

William J . Mcllvane

Mcllvane, W. J., Gilbert, F. M., & McDonald, S. J. (1992). A note on teaching discrimination reversals to people with intellectual disabilities. Manuscript in preparation. McIlvane, W. J., Kledaras, J. B., Dube, W. V., Iennaco, F. M., & Stoddard, L. T. (1988). Programmed reversal of differential consequence relations in moderately retarded individuals’ conditional discrimination. Proceedings of the 21st Annual Gatlinburg Conference on Research and Theory in Mental Retardation and Developmental Disab McIlvane, W. J., Kledaras, J. B., Dube, W. V.,& Stoddard, L. T. (1989). Automated instruction of severely and profoundly retarded individuals. In J. A. Mulick & R. F. Antonak (Eds.), Transitions in mental retardation (Vol. 4 , pp. 15-76). Norwood, NJ: Ablex. McIlvane, W.J., Kledaras, J. B., Gilbert, F. M.,& Stoddard, L. T. (1992). Frequency analyses of stimulus control in discrimination learning by people with mental retardation. Proceedings of the 25th Annual Gatlinburg Conference on Research and Theory in Mental Retardation and Developmental Disabilities, 140. Mcllvane, W. J., Kledaras, J. B.,Killory-Andersen, R., & Sheiber, F. (1989). Teaching with noncriterion-related prompts: A possible subject variable. The Psychological Record, 39, 131- 142. McIlvane, W. J., Kledaras, J. B., Lowry, M. W., & Stoddard, L. T. (in press). Studies of exclusion in individuals with severe mental retardation. Research in Developmental Disabilities. Mcllvane, W. J., Kledaras, J. B., Munson, L. C., King, K. A,, de Rose, J. C., & Stoddard. L. T. (1987). Controlling relations in conditional discrimination and matching by exclusion. Journal of the Experimental Analysis of Behavior, 48, 187-208. McIlvane, W. J., Kledaras, 1. B., Stoddard, L. T., & Dube, W. V. (1990).Delayed sample developmental limitations. Experimental Analysis of Human Behavior Bulletin, 8 , 3 1-33. McIlvane, W. I . , Munson, L. C., & Stoddard, L. T. (1988). Some observations on control by spoken words in childrens’ conditional discrimination and matching by exclusion. Journal of Experimental Child Psychology, 4 5 , 472-495. McIlvane, W. J., & Stoddard, L. T. (1981). Acquisition of matching-to-sample performances in severe mental retardation: Learning by exclusion. Journal of Mental Defcienc.v Research, 25, 33-48. Mcllvane, W. J., Withstandley, J. K., & Stoddard, L. T. (1984). Positive and negative stimulus relations in severely retarded persons’ conditional discriminations. Analysis and lntervention in Developmental Disabilities, 4 , 235-25 I . McIntire, K. D., Cleary, J., & Thompson, T. (1987). The operant conditioning of response variability: Free-operant versus discrete-response procedures. Journal of the Experimental Analysis of Behavior, 47, 279-297. McManis, D. L. (1969). Conservation and transitivity of weight and length by normals and retardates. Developmental Psychology. I , 373-382. Mervis, C. B., & Bertrand, J. (in press). Acquisition of early object labels: The role of operating principles. In A. P. Kaiser & D. B. Gray (Eds.), Enhancing childrens’ communication. Baltimore, MD: Paul H. Brookes. Miller, A , , & Miller, E. E. (1968). Symbol accentuation: The perceptual transfer of meaning from spoken words. American Journal of Mental Dejciency, 73, 200-208. Mirenda, P. (1985). Designing pictorial communication systems for physically able-bodied students with severe handicaps. Augmentative and Alternative Communication, I , 58-64. Moss, E., & Harlow, H. F. (1947). The role of reward in discrimination learning in monkeys. Journal of Comparative and Physiological Psychology, 40, 333-342. emediating letter reversals: Cost considerations. Journal of LearnNotterman, J. M. (1959). Force emission during bar pressing. Journal of Experimental Psychology, 58, 341-347. O’Mara, H. (1991). Quantitative and methodological aspects of stimulus equivalence. Journal of the Experimental Analysis of Behavior, 5 5 , 125- 132. Peters, H. N. (1935). Mediate association. Journal of Experimental Psychology, 18, 20-48.

STIMULUS CONTROL ANALYSIS

107

Ray, 9.A. (1969). Selective attention: The effects of combining stimuli which control incompatible behavior. Journal of the Experimental Analysis of Behavior, 12, 539-550. Reese, H. W. (1968). The perception of stimulus relarions: Discrimination learning and transposirion. New York: Academic Press. Richmond, G., & Bell, J. (1986). Comparison of trial-and-emr and graduated stimulus change procedures across tasks. Analysis and Intervention in Developmental Disabilities, 6 , 127- 136. Romski, M. A,, & Sevcik, R. A. (1989). An analysis of visual-graphic symbol meanings for two nonspeaking adults with severe mental retardation. Augmentative and Alternative Communication, 5 , 109-1 14. Romski, M . A., Sevcik, R. A,, & Rumbaugh, D. M. (1985). Retention of symbolic communication skills in five severely retarded persons. American Journal of Mental Deficiency, 89,441-444. Rosenberger, P. B., Stoddard, L. T., & Sidman, M. (1972). Sample-matching techniques in the study of children’s language. In R. L. Schiefelbusch (Ed.), Language of the mentally retarded (pp. 21 1-229). Baltimore, MD: University Park Press. Saunders, R. R., & Green, G. (1992). The nonequivalence of behavioral and mathematical equivalence. Journal of the Experimental Analysis of Behavior, 57, 227-241. Saunders, R. R., Saunders, K. J.. Kirby, K. C., & Spradlin, J. E. (1988). The merger and development of equivalence classes by unreinforced conditional selection of comparison stimuli. Journal of the Experimental Analysis of Behavior, 50, 145-162. Saunders, K . 1.. & Spradlin, J. E. (1989). Conditional discrimination in mentally retarded adults: The effects of training the component simple discriminations. Journal of the Experimental Analysis of Behavior, 52, 1-12. Saunders, K. J . , & Spradlin, J. E. (1990). Conditional discrimination in mentally retarded adults: The development of generalized skills. Journal of the Experimental Analysis of Behavior. 54,239250. Schilmoeller, G. L . , Schilmoeller, K. J., Etzel, 9. C., & LeBlanc, J. M. (1979). Conditional discrimination after errorless and trial-and-error training. Journal of the Experimental Analysis of Behavior, 31, 405-420. Schreibman, L., Charlop, M. H., & Koegel, R. L. (1982). Teaching autistic children to use extrastimulus prompts. Journal of Experimental Child Psychology, 33, 475-491. Sidman, M. (1969). Generalization gradients and stimulus control in delayed matching-to-sample. Journal of the Experimental Analysis of Behavior, 33, 285-289. Sidman, M. (1971). Reading and auditory-visual equivalences. Journal of Speech and Hearing Research, 14, 5-13. Sidman, M. (1979). Remarks. Behaviorism, 7, 123-126. Sidman, M. (1986a). Functional analysis of emergent verbal classes. In T. Thompson & M. D. Zeiler (Eds.), Anal.vsis and integration of behavioral units (pp. 213-245). Hillsdale, NJ: Erlbaum. Sidman, M. (l986b). The measurement of behavioral development. In N. A. Krasnegor, D. 9.Gray, & T. Thompson (Eds.), Advances in behavioral pharmacology: Vol. 5. Developmental behavioral pharmacology ( pp. 43-52). Hillsdale, NJ: Erlbaum. Sidman, M. (1987). Two choices are not enough. Behavior Analyst, 5 , 11-18. Sidman, M. (1990). Equivalence relations: Where do they come from? In D. E. Blackman & H. Lejeune (Eds.), Behaviour analysis in theory and practice: Contributions and controversies (pp. 93-1 14). Hillsdale, NJ: Erlbaum. Sidman, M., & Cresson, 0. (1973). Reading and crossmodal transfer of stimulus equivalences in severe mental retardation. American Journal of Mental Deficiency, 77, 515-523. Sidman, M., Cresson, O., Jr., & Willson-Morris, M. (1974). Acquisition of matching to sample via mediated transfer. Journal of the Experimental Analysis of Behavior, 22, 261-273. Sidman, M., & Stoddard, L. T. (1966). Programming perception and learning for retarded children. In N. R. Ellis (Ed.), International review of research in mental retardation (Vol. 2, pp. 151208). New York: Academic Press. Sidman, M., & Stoddard, L. T. (1967). The effectiveness of fading in programming a simultaneous

William J . Mcllvane

108

form discrimination for retarded children. Journal of the Experimental Analysis of Behavior, 10, 3-15. Sidman, M., & Tailby, W. (1982). Conditional discrimination vs. matching-to-sample: An expansion of the testing paradigm. Journal qf the Experimental Analysis uf Behavior, 37, 5-22. Sidman, M., Willson-Moms, M., & Kirk, B. (1986). Matching-to-sample procedures and the development of equivalence relations: The role of naming. Anal.vsis and Intervention in Developmental Disabilities, 6, 1-19. Sidman, M., Wynne, C. K., Maguire, R. W., & Barnes, T. (1989). Functional classes and equivalence relations. Journal of the Experimental Analysis of Behavior, 52, 261-274. Sigurdardottir, Z. G., Green, G . , & Saunders, R. R. (1990). Equivalence classes generated by sequence training. Journal of the Experimental Analysis of Behavior, 53, 47-63. Singh, N. N., & Singh, J. (1986). Reading acquisition and remediation in the mentally retarded. In N. R. Ellis & N. W. Bray (Eds.), International review of research in mental retardation (Vol. 14, pp. 165-199). New York: Academic Press. Singh, N. N., & Solman, R. T. (1990). A stimulus control analysis of the picture-word problem in children who are mentally retarded: The blocking effect. Juurnal of Applied Behavior Analysis, 23, 525-532. Skinner, B. F. (1935). The generic nature of the concepts of stimulus and response. Journal of General Psychology, 1 2 , 40-65. Skinner, B. F. (1938). The behavior oforganisms: An experimental analysis. New York: AppletonCentury. Skinner, B. F. (1957). Verbal behavior. New York: Appleton-Century-Crofts. Smeets, P. M., Lancioni, G . M., & Streifel, S. (in press). Emergent simple discrimination in children: Transfer of stimulus control under non-reinforced conditions. Quarterly Journal of E.rperimenta1 Psychology.

Spence, K. W. (1956). Behavior theory and conditioning. New Haven, CT: Yale University Press. Spradlin, J. E., Cotter, V. W., & Baxley, N. (1973). Establishing a conditional discrimination without direct training: A study of transfer with retarded adolescents. American Journal of Mental Deficiency, 77. 556-566. Spradlin, J. E . , & Dixon, M. H. (1976). Establishing conditional discriminations without direct training: Stimulus classes and labels. American Journal of Mental Deficiency, 80. 555-561. Spradlin, J. E., & Saunders, R. R. (1984). Behaving appropriately in new situations: A stimulus class analysis. American Journal of Mental Deficiency, 88, 574-579. Spradlin, J. E., & Saunders, R. R. (1986). The development of stimulus classes using match-tosample procedures: Sample classification vs. comparison classification. Analvsis and Intervention in Developmental Disabilities, 6, 41-58. Stevens, S. S. (1951). Mathematics, measurement, and psychophysics. In S. S. Stevens (Ed.), Handbook of experimental psychology (pp. 1-49). New York: Wiley. Stoddard, L. T. (1982). An investigation of automated methods for teaching severely retarded individuals. In N. R. Ellis (Ed.), International review of research in mental retardation (Vol. 11. pp. 163-207). New York: Academic Press. Stoddard, L. T., Brown, J., Hurlbert, B., Manoli, C., & McIlvane, W. J. (1989). Teaching money skills through stimulus class formation, exclusion, and component matching methods: Three case studies. Research in Developmental Disabilities, 10. 413-439. Stoddard. L. T., & Mcllvane, W. J. (1986). Stimulus control research and developmentally disabled individuals. Anal.vsis and Intervention in Developmental Disabilities, 6, 155- 178. Stoddard. L. T., & Sidman, M. (1971a). Removal and restoration of stimulus control. Journal of the E.rperimenta1 Anal-vsis of Behavior, 16, 143- 154. Stoddard, L. T., & Sidman, M. (1971b). Stimulus control after intradimensional discrimination training. Psychological Reports, 2 8 , 147-157.

STIMULUS CONTROL ANALYSIS

109

Stromer, R. (1975). Modifying letter and number reversals in elementary school children. Journal of Applied Behavior Analysis, 8 , 21 I , Stromer, R. ( 1986). Control by exclusion in arbitrary matching-to-sample. Analysis and Intervention in Developmental Disabilities, 6 , 59-72. Stromer, R.,& Mackay, H. A. (in press-a). Delayed identity matching to complex samples: Teaching students with mental retardation spelling and the prerequisites for equivalence classes. Research in Developmental Disabilities. Stromer, R . , & Mackay, H. A. (in press-b). Spelling and emergent picture-printed word relations established with delayed identity matching to complex samples. Journal of Applied Behavior Analysis. Stromer, R . , Mackay, H. A., Cohen, M., & Stoddard, L. T.(in press). Ordinal relations in the sequence learning of individuals with behavioral limitations. Journal of Intellectual Disability Research. Stromer, R . , & Osborne, J. G. (1982). Control of adolescents’ arbitrary matching-to-sample relations. Journal of the Experimental Analysis of Behavior. 37, 329-348. Tashjian. T. A. (1984). Computerized vocal response training. In J. A. Mulick & B. L. Mallory (Eds.), Transitions in mental retardation (Vol. I , pp. 174-204). Nowood, NJ: Ablex. Terrace, H. L. ( 1963). Discrimination learning with and without errors. Journal of the Experimental Analysis of Behavior, 6 , 1-27. Terrace, H. L. (1986). Positive transfer from sequence production to sequence discrimination in a nonverbal organism. Journal of Experimental Psychology: Animal Behavior Processes, 12, 215-234. Thomas, D. R . , & Schmidt, E. K. (1989). Does conditional discrimination learning by pigeons necessarily involve hierarchical relations. Journal of the Experimental Analysis of Behavior, 5 2 , 249-260. Thompson, R. ( 1954). Approach versus avoidance in an ambiguous-cue problem in chimpanzees. Journal of Comparative Physiological Psychology, 47, 133- 135. Trapold. M., & Overmicr. J. B. (1972). The second learning process in instrumental learning. In A. H. Black & W. F. Prokasy (Eds.). CIa.~sicalconditioning 11: Current research and theory (pp. 426-452). New York: Appleton-Century-Crofts. Vincent-Smith, L., Bricker, D., & Bricker, W. (1974). Acquisition of receptive vocabulary in the toddler-age child. Child Development, 45, 189- 193. Weigl, E. ( 1941). On the psychology of so-called processes of abstraction. Journal of Abnormal and Social Psvchology, 36, 3-33. Williams, B. A . (1988). Reinforcement, choice, and response strength. In R. C. Atkinson, R . J. Herrnstein, G. Lindzey, & R. D. Luce (Eds.), Steven’s handbook on experimental psychology (Vol. 2 , pp. 167-244). New York: Wiley. Wolfe. V. F.. & Cuvo. A. J. (1978). Effects of within-stimulus and extra-stimulus prompting on letter discrimination by mentally retarded persons. American Journal of Mental Deficiency, 83, 297303. Wulfert, E., & Hayes, S. C. (1988). Transfer of a conditional ordering response through conditional equivalence classes. Journal of the Experimental Analysis of Behavior, 50, 125- 144. Zeaman, D., & House, B. J. (1963). The role of attention in retardate discrimination learning. In N. R. Ellis (Ed.), Handbook of mental deficiency, psychological theory and research (pp. 159223). New York: McGraw-Hill. Zeaman, D., & House, B. J. (1979). A review of attention theory. In N. R. Ellis (Ed.), Handbook of mental deficiency, psychological theory and research ( pp. 63- 120). Hillsdale, NJ: Erlbaum. Zygmont, D. M., Lazar, R. M., Dube. W. V., & McIlvane, W. J. (1992). Teaching arbitrary matching via sample-stimulus shaping to young children and mentally retarded individuals: A methodological note. Journal of the Experimental Analysis of Behavior, 5 7 , 109- 117.

This Page Intentionally Left Blank

Sustained Attention in Mentally Retarded Individuals PHILLIP D. TOMPOROWSKI DEPARTMENT OF PSYCHOLOGY THE UNIVERSITY OF ALABAMA TUSCALOOSA, ALABAMA 35487

LISA D. HAGER DEPARTMENT OF PSYCHOLOGY UNIVERSITY OF DENVER DENVER, COLORADO 80208

I.

INTRODUCTION

The ability to maintain a constant level of alertness to specific sources of information plays an important role in many aspects of everyday life for mentally retarded and nonretarded individuals alike. Central to many educational and employment situations is a requirement that the individual attend to a source of information and that a level of attention be maintained. It may be relatively easy for one to be attentive for brief periods; however, it is often more difficult to sustain attention when the task is difficult and when information is presented in an infrequent or unpredictable fashion. Such tasks often lead to a deterioration in performance over time. The observation that vigilance performance often declines as a function of time on task has resulted in considerable research on the sustained attention of nonretarded adults. Relatively little research, however, has been conducted on the sustained attention of mentally retarded individuals. The first comprehensive evaluation of sustained attention in mentally retarded individuals was provided by Warm and Berch (1985). This seminal work provided an overview of the vigilance paradigm and research conducted on intelligence-related differences in sustained attention. Topics addressed by Warm and INTERNATIONAL REVIEW OF RESEARCH IN MENTAL RETARDATION. Vol. 18

Ill

Copyright 8 1992 by Academic Press, Inc. All rights of reproduction in any form r e ~ e ~ e d .

Phillip D. Tomporowski and Lisa D.Hager

112

Berch include psychophysical aspects of the vigilance task, data analysis and interpretation issues, and theories of vigilance. Their review of the literature led them to conclude that the central issues concerning the relation between level of intelligence and sustained attention were yet to be resolved. They noted that many of the studies conducted to assess the relation between level of intelligence and sustained attention neglected to address task variables that are known to affect vigilance behavior and that a systematic analysis of vigilance in retarded individuals was lacking. This article focuses on methodological issues that are central to the study of sustained attention in mentally retarded persons. The literature reviewed by Warm and Berch (1985) is then reevaluated in the light of several recently conducted vigilance experiments. These studies provide evidence that the sustained attention of mentally retarded adults differs from that of nonretarded adults under some, but not all, conditions. The role of such factors as task type, task complexity, and memory load on sustained attention is discussed. Several theories of vigilance are described to explain why specific task conditions differentially affect the sustained attention of individuals with and without mental retardation.

11. A.

METHODOLOGICAL ISSUES

Laboratory Tests of Sustained Attention

The first controlled laboratory studies of sustained attention were conducted by Mackworth (1948) at the request of the British Royal Air Force. His studies were prompted by very practical considerations. During World War 11, military personnel were required to perform many tasks that demanded continued periods of watchkeeping, such as monitoring radar scopes for long periods. In his attempts to simulate the elements of these tasks, Mackworth developed a test requiring subjects to view the movements of a pointer along the circumference of a blank-faced clock. The pointer moved 0.3 in. every second; however, on an infrequent basis, the pointer executed a double jump of 0.6 in. The double jump served as the target event; sub.jects were asked to press a response key when they detected a critical event. In one of Mackworth’s studies, cadets performed a 2hour test in which 12 target events occurred during 20-minute blocks of time. Detection performance was found to decrease rapidly during the initial 60 minutes of the test and then decrease at a slower rate for the remainder of the vigil (Mackworth, 1948). The decline in performance is typically referred to as the vigilance decrement and it has been the focus of investigation in over 1500 experiments conducted since Mackworth’s early work (Parasuraman, 1986). Since Mackworth’s initial studies, several variations of the vigilance task have been introduced. One that has direct relevance to the study of sustained attention

SUSTAINED ATTENTION

113

in mentally retarded persons is the cognitively based vigilance task. Cognitively based vigilance tasks require comparisons between targets and nontarget symbols (e.g., letters or digits) that are clearly discriminable from one another. This contrasts with typical sensory-based vigilance tasks that require comparisons between target and nontarget events that differ in physical properties (e.g., shape, size, and intensity) presented at near-threshold levels. Two commonly used cognitively based tasks are the Bakan test (Bakan, 1955) and the Continuous Performance Test (CPT) (Rosvold, Mirsky, Sarason, Bransome, & Beck, 1956). In the Bakan test, digits are presented individually and the target event is a specific sequence of digits. The CPT originally consisted of two 10-minute parts in which subjects viewed individual letters of the alphabet. In the first part, subjects responded to all occurrences of the letter X ; in the second part, they responded to the letter X only if it was immediately preceded by the letter A . There was a 2-minute rest period between the two parts of the test. The CPT was introduced by Rosvold et al. as a clinical test for the diagnosis of brain damage. It has been widely used for assessment of sustained attention in attention-deficit/hyperactivity-disordered (ADHD) and learning-disabled (LD) children and, less often, in mentally retarded persons. Although the CPT provides measures of sustained attention, there are subtle methodological differences between the CPT and the more traditional vigilance test that make comparisons of the two types of tests difficult. The CPT differs from the standard, Mackworth-type vigilance test in a number of ways: (1) it is a brief test, (2) stimuli are often presented for short durations in rapid succession, (3) target events occur frequently, and (4) the distinct X and A-X conditions make it difficult to assess the vigilance decrement (Berch & Kanter, 1984; Krafft & Dietrich, 1985).

B. Measurement Issues Several dependent measures are used to evaluate sustained attention. Target detections, or hits, represent the number of target events correctly identified. False alarms reflect the number of responses made to nontarget events. Typically, hits and false alarms are expressed as proportions. Response latencies to target events, which tend to increase as a direct function of the length of the vigil (Buck, 19661, are also used. Signal detection theory (SDT) (Green & Swets, 1966) has also been applied to the analysis of sustained attention. This statistical method provides independent measures of detectability (d’) and decision criterion (p) on the basis of a mathematical formula that combines the probability of hits with the probability of false alarms. The d‘ value represents the estimated difference between the means of a distribution of target events and a distribution of nontarget events, and it is taken to reflect observers’ ability to detect differences between targets and nontargets.

114

Phillip D . Tomporowski and Lisa D . Hager

Decision criterion (p) measures are believed to reflect subjects’ level of caution in reporting the presence of a target event. The likelihood that subjects will make responses during a vigil is believed to be influenced by their evaluation of the trade-off between making correct target detections and avoiding false alarms. The application of SDT to vigilance has been seen as an important contribution to the study of sustained attention. As discussed by Davies and Parasuraman (1982, Chap. 3), attempts to explain sustained attention solely on the basis of target detection without also evaluating false alarms may provide a distorted view of vigilance behavior. Their view is that declines in vigilance performances are due to observers’ independent or joint shifts in their ability to detect differences between targets and nontargets (d’) and their decision criterion (p). Indeed, Davies and Parasuraman (1982) have developed a task taxonomy based on SDT that provides a method for isolating factors responsible for observers’ overall level of vigilance as well as changes in their performance over time. The task taxonomy is based on studies that have found d‘ and p to be differentially sensitive to specific task characteristics. Factors that have been found to have substantial effects on d‘ but little, if any, effects on p include stimulus presentation rate (Loeb & Binford, 1968; Parasuraman, 1979) and memory load (Parasuraman, 1979; Parasuraman & Mouloua, 1987). These factors place what Norman and Bobrow (1975) describe as data limitations on optimal detection. Fast event rates produce signal-data limitations on performance; they require the detection of weak signals in a noisy environment. Successive discrimination tasks produce memory-data limitations on performance; they require that information be remembered during a delay. Factors known to have effects on p but little, if any, effects on d’ include signal probability (Baddeley & Colquhoun, 1969; Williges, 1969, 1971, 1973), instructions (Colquhoun, 1967; Williges, 1973), previous training (Colquhoun & Baddeley, 1964), and feedback (Davies & Tune, 1970). These factors affect inferences that are made about the nature of the task and cause modifications in response decisions and strategies. Davies and Parasuraman’s (1982) taxonomic approach to the study of sustained attention is important for two reasons. First, it shows clearly that sustained attention is not a unitary phenomenon; it is related to specific task demands. Second, their taxonomy serves as an excellent model for evaluating research conducted on intelligence-related differences in sustained attention and for structuring a systematic approach to the study of sustained attention in mentally retarded individuals. It has been proposed that the performance of mentally retarded individuals is poorer than that of nonretarded individuals on a variety of tasks because of such factors as stimulus encoding (Merrill, 1990; Merrill, Sperber, McCauley, Littlefield, Rider, & Shapiro, 1987; Nettelbeck, 1985), shortterm memory (Detterman, 1975; Ellis, 1978), and response strategy (Bray, 1979; Brown, 1974). As outlined by the taxonomy, the vigilance performance of mentally retarded observers can be predicted to differ from that of nonretarded

SUSTAINED ATTENTION

115

observers under specific test conditions. Differences in detection performance ( d ’ ) should be most noticeable during tests with encoding or memory demands, whereas differences in response criteria (p) should be most apparent during tests that vary expectancy of targets. Studies that have assessed sustained attention in mentally retarded individuals are reviewed in the following section. Specific emphasis is placed on task and subject variables that are related to vigilance performance.

Ill. VIGILANCE IN MENTALLY RETARDED INDIVIDUALS Relatively few laboratory studies of sustained attention in mentally retarded persons have been published. Although the CFT has been used with mentally retarded individuals very infrequently, Rosvold et al. (1956) included mentally retarded individuals in their initial study using the CPT. They evaluated the performance of brain-damaged and non-brain-damaged retarded adults, nonretarded adults, and children. The target detection performance of braindamaged (mean IQ [MIQ] = 63.3, mean chronological age [MCA] = 27.1 years) and non-brain-damaged (MIQ = 66.8, MCA = 25.1 years) mentally retarded adults was lower than that of nonretarded brain-damaged and non-braindamaged children and adults. Regardless of level of intelligence, those with brain injury performed more poorly on the A-X part of the CPT than on the A part of the test. Rosvold et al. discussed the usefulness of the CPT as a test for diagnosing brain damage; relatively little was made of the differences between the mentally retarded and nonretarded groups. The CPT was also used by Crosby (1972) to determine whether mentally retarded adolescents were more distractible than nonretarded children with similar mental ages or adolescents with similar chronological ages. The study also assessed the role of brain injury and the effects of institutionalization on the performance of mentally retarded subjects. Observers performed the CFT under conditions of auditory distraction, visual distraction, and a combination of the two. Mentally retarded adolescents performed more poorly than CA-matched adolescents of average intelligence; however, the distractor manipulations did not differentially affect performance. Crosby interpreted the poorer performance of the mentally retarded group as indicative of an attentional deficit. He also found that institutionalized brain-damaged (MIQ = 63.7, MCA = 15.3 years) and nonbrain-damaged (MIQ = 63.9, MCA = 15.2 years) mentally retarded adolescents performed more poorly than noninstitutionalized mentally retarded adolescents (MIQ = 66.5, MCA = 14.8 years) and that the performance of the noninstitutionalized children did not differ from that of nonretarded children of similar mental age. The studies by Rosvold et al. (1956) and Crosby (1972) clearly indicate that

116

Phillip D. Tomporowski and Lisa D.Hager

mentally retarded individuals are less able to detect targets on the CPT than nonretarded individuals; however, these studies provide little information concerning retarded observers’ sustained attention because measures of changes in detection performance were not recorded as a function of time on task. Further, the differences in retarded and nonretarded observers’ performance on the CPT could be attributed to factors other than an attentional deficit. For example, one of the characteristics of the CPT is the use of stimuli that are presented rapidy for very brief durations. It is known that mentally retarded individuals encode and process information more slowly than nonretarded individuals (Maisto & Baumeister, 1984; Merrill, 1990); thus, the poorer overall performance of the retarded observers may have been the result of differences in encoding and speed of processing rather than attentional factors per se. As discussed by Baumeister (1984) and others (Krupski, 1980; Stanovich, 1978), the comparison of subjects’ performance under only one level of task demand is not adequate to isolate cognitive factors that contribute to IQ-related differences in performance. Several studies have assessed directly the change in vigilance over time in mentally retarded individuals using traditional Mackworth-type vigilance test methodologies. Ware, Baker, and Sipowicz ( 1962) assessed the vigilance performance of 14 mildly mentally retarded young adults (MIQ = 58, MCA = 17 years, 8 months) during a 3-hour vigil in which observers were asked to detect 300-millisecond interruptions in a light (0.45 footlambert [ftl]). Each subject was initially tested under a no-feedback condition and, a week later, under a feedback condition in which a cue was provided if the observer failed to detect a target event. A significant decrement in target detection occurred during both vigils. Knowledge of results improved target detection; however, it did not differentially influence the rate of the decline in performance. Performance in the nonfeedback condition was compared with that of 14 nonretarded adults randomly selected from a previous unpublished study conducted by Ware (1960). No differences were found in overall target detection between the two groups. Unfortunately, rate of decline in target detection of retarded and nonretarded observers was not presented. Semmel (1965) compared the sustained attention of mildly mentally retarded (MIQ = 68, MCA = 12 years, 4 months) and nonretarded (MIQ = 105, MCA = 12 years, 2 months) children on a 60-minute task in which brief interruptions (500 milliseconds) of a faintly illuminated light source served as the target event. The duration between target events was varied systematically and a short rest period was scheduled midway through the vigil. Both groups’ target detection declined significantly during the vigil, with mentally retarded children showing a more rapid decline than nonretarded children. The probability of target detection was also related to the duration between target events; 90-, 180-, and 240-second delays resulted in significantly fewer hits than 15- and 30-second delays. There was no difference in hit performance as a function of IQ level and delay interval,

SUSTAINED ATTENTION

117

however. Semmel concluded that sustained attention is related to observers’ general arousal level and that environmental events produce a more rapid dissipation of arousal in mentally retarded children than in nonretarded children. The arousal interpretation for the differences in the sustained attention of mentally retarded and nonretarded persons has been challenged, however. A group of researchers at the University of Adelaide in Australia have suggested that differences in sustained attention may be related to slower cognitive development of mentally retarded individuals. The view that cognitive abilities are developmentally delayed in mentally retarded individuals has been proposed by several researchers and it has been the central focus of theories of mental retardation (Zigler, 1982; Zigler & Balla, 1982). Further, studies of children with normal intelligence indicate that childrens’ ability to attend to tasks increases with age up to late adolescence (Anderson, Halcomb, Gordon, & Ozolins, 1974; Gale & Lynn, 1972; Mohan, 1982). The developmental-lag hypothesis is given support by a series of three experiments conducted at the University of Adelaide. The initial study was conducted by Kirby, Nettelbeck, and Bullock (1978). In their study, mentally retarded (MIQ = 70, MCA = 23 years) and nonretarded (MCA = 23 years) adults performed separate 50-minute visual and auditory vigilance tests. The visual test consisted of the presentation of a 3-cm circular light source for 500 milliseconds every 3 seconds. Light intensities of 4.8 ftl were nontarget events, and intensities of 10 ftl were target events. In the auditory test, white noise pulses of 500 milliseconds were presented every 3 seconds via headphones; 30-dB pulses were nontargets and 33-dB pulses were target events. Target events occurred on 5% of the stimulus presentations during successive 1O-minute periods in both tests. An evaluation of hit percentages in both tests revealed a significant decline in performance during the vigils; however, there was no difference in performance as a function of IQ level. An analysis of false alarms indicated that significantly more errors of commission were made during the visual detection task than during the auditory test, and that the number of false alarms decreased at the same rate for both IQ groups. Mentally retarded observers made significantly more false alarms than nonretarded observers on both tests. In the second study of the series, Kirby, Nettelbeck, and Thomas (1979) compared the vigilance performance of 9 mentally retarded children (MIQ = 68, MCA = 13 years) with the performance of 13 nonretarded children matched on chronological age (MCA = 12 years, 9 months) and 12 nonretarded children matched on mental age. The vigilance test was similar to the visual vigilance test used by Kirby et al. (1978). Stimulus presentation parameters were identical to those used in their earlier study; however, light intensities of 10 and 16 ftl served as nontarget and target events, respectively. An analysis of target detections of the three groups indicated that hits declined significantly during the 50-minute vigil; however, mentally retarded children and younger nonretarded children

118

Phillip D . Tomporowski and Lisa D . Hager

evidenced a more rapid decline in hit rate than older nonretarded children. The number of false alarms declined during the vigil at a constant rate for the groups, with mentally retarded children making significantly more errors of commission than younger nonretarded children who, in turn, made significantly more false alarms than older nonretarded children. The results of these two studies suggested to Kirby et al. (1979) that differences between the vigilance performance of mentally retarded and nonretarded individuals are due primarily to developmental factors. This conclusion may need to be qualified, however, as their interpretation was based on the premise that target detections alone reflect sustained attention. False alarms were interpreted to reflect observers' discriminative abilities. The significantly greater number of false alarms made by retarded observers in both experiments was thought to represent an intellectual deficit rather than retarded development. As discussed by Davies and Parasuraman (1982, Chap. 3), attempts to explain sustained attention solely on the basis of target detection without also evaluating false alarms may provide a distorted view of vigilance behavior. A recently conducted dissertation (Thomas, 1989) replicated the Kirby et al. (1978, 1979) studies. Mentally retarded (MIQ = 57, MCA = 14 years, 2 months) and nonretarded children and retarded (MIQ = 70, MCA = 18 years, 9 months) and nonretarded young adults participated in 50-minute auditory and visual vigilance tests. Target events in the visual task were light intensities of 20 ftl, and nontarget events were light intensities of 10 ftl; target events in the auditory task were 33-dB tones and nontarget events were 30-dB tones. The study provided additional support for the developmental-lag view of the sustained attention of mentally retarded persons. In both visual and auditory tests, the young mentally retarded subjects evidenced a more rapid decline in target detection than the older mentally retarded and nonretarded observers and the rate of decline was similar to that of MA-matched children. Further, the detection performance of mentally retarded and nonretarded young adults did not differ. As in the previous studies, mentally retarded observers made more false alarms than nonretarded observers on both tests. Thomas also performed a SDT analysis of subjects' hit and false alarm performance. Young mentally retarded subjects' perceptual sensitivity ( d ' ) declined during the vigil with little change in their response criteria (PI; all other groups showed little change in d' and an increase in P over time. Thomas suggested that the inferior vigilance of mentally retarded individuals is related to developmental factors and that, by the age of 17 years, their sustained attention is at the same level as that of nonretarded adults. Thomas' interpretation must be viewed cautiously, however. The difference in the level of intelligence between the mentally retarded children (MIQ = 57) and mentally retarded adults (MIQ = 70) may have contributed to the different vigilance decrements of the two groups. Das (1970) evaluated the sustained attention of mildly (MIQ = 66.1) and severely (MIQ = 47.8) mentally retarded

SUSTAINED ATTENTION

119

adolescents on two 30-minute cognitively based vigilance tests that differed in level of complexity. In the simple vigilance test, stimuli were ten common nouns, one of which served as the target cue. In the complex vigilance test, stimuli were six words and six numbers, and three of the words served as targets. The vigilance performance was similar for both IQ groups on the simple version of the test; however, mildly mentally retarded observers detected more targets than severely mentally retarded observers during the complex vigilance test. Although the studies conducted by Kirby et al. (1979) and Thomas (1989) suggest that sustained attention is not related to IQ level following late adolescence, this position has been challenged. Stankov (1983) pointed out that most vigilance studies employed paradigms that were quite simple and demanded only the detection of an unusual stimulus. He evaluated a number of psychometric studies and proposed that vigilance tasks that placed greater cognitive demands on observers would result in differential performance as a function of level of intelligence. On simple, noncognitively demanding vigilance tasks, Stankov predicts that lower-lQ observers would show less of a decrement in performance than higher-IQ observers; however, on complex vigilance tasks, lower-IQ observers were predicted to show a greater decrement in performance in the latter stages of a vigil. Warm and Berch (1985) likewise suggested that the studies conducted to assess sustained attention in mentally retarded individuals were simplistic and failed to recognize the role of several task factors that would be predicted to interact with IQ level. They recommended that research be conducted with mentally retarded and nonretarded individuals, and that such factors as ( I ) task complexity, (2) event rate, (3) memory load, and (4) target probability be manipulated in a systematic fashion. The views of Stankov (1983) and Warm and Berch (1985) prompted a series of experiments conducted in our laboratory at The University of Alabama. The initial study of the series addressed the issues of task complexity and its effects on sustained attention. Tomporowski and Allison (1988) compared the performance of mildly retarded (MIQ = 61.6, MCA = 19 years, 4 months) and nonretarded (MCA = 19 years, 5 months) young adults on three 50-minute visual vigilance tests. One was a letter-recognition task in which letters of the alphabet were presented for 150 milliseconds every second and the target event was the letter X. The second was a sensory-based task in which nontarget events were 3.0-cm squares and target events were 4.5-cm squares presented for 200 milliseconds every 2 seconds in the center of a computer monitor. The third was a cognitively based task in which individual digits were presented for 150 milliseconds every 2 seconds. The digits I to 9 were presented successively in a series and then repeated. The target event was a missing digit in a set (e.g., 1,2,3,5,6,7,8,9,1,2,. . .). In all three tests, target events were programmed to occur randomly on 5% of stimulus events during successive 10-minute periods. The letter-detection task was viewed as the least complex task because target

120

Phillip D . Tomporowski and Lisa D . Hager

identification required only simple letter recognition. The sensoj-type task was believed to be more complex than the letter-detection task because target detection was influenced by the observers’ ability to discriminate between targets and nontargets presented at near-threshold levels and by the memory load inherent in successive discrimination tasks. The missing-digit task was hypothesized to be the most complex of the three tasks because target detection depended on observers’ abilities to encode, reorganize, and retain constantly changing information. In the letter-detection task, the number of targets detected was quite high and the number of false alarms was extremely low for both IQ groups. Performance declined over time, and there were no differences between the groups in the vigilance decrement. During the sensory and missing-digit tasks, however, the overall vigilance performance of mentally retarded observers was poorer than that of nonretarded observers; mentally retarded observers detected significantly fewer targets and made more false alarms throughout the vigil. The rate of the vigilance decrement was the same for both groups, however. The results of the study suggest that mentally retarded adults are poorer watchkeepers on some, but not all, vigilance tests; overall detection of mentally retarded and nonretarded individuals did not differ on a simple task but did on two complex tasks. Although the absolute level of performance of mentally retarded adults may be inferior to that of nonretarded adults on some tasks, it appears that the ability of mentally retarded and nonretarded adults to sustain attention does not differ. Tomporowski and Allison (1988) hypothesized that the inferior performance of mentally retarded persons during the sensory and missing-digit tasks may have been influenced by specific task demands requiring well-developed encoding and short-term memory abilities. The differences between the sizes of target and nontarget events in the sensory task were designed to be minimal; the nearthreshold differences between targets and nontarget events presumably placed a demand on perceptual abilities. In both tasks, stimuli were presented successively and observers had to make decisions on the basis of an absolute judgment; that is, comparisons between stimulus events were made on the basis of memory for prior stimuli. Given that the encoding (Maisto & Baumeister, 1984; Merrill, 1990; Menill et al., 1987; Nettlebeck, 1985; Sperber & McCauley, 1984) and short-term memory (Carr, 1984; Detterman, 1975; Ellis, 1978) abilities of mentally retarded individuals are known to be less efficient than those of nonretarded persons, it is plausible that the poor performance of the mentally retarded subjects found by Tomporowski and Allison (1988) may be caused by the task demands related to these abilities. The Davies and Parasuraman (1982) taxonomy predicts that vigilance performance will be sensitive to factors that make it difficult to discriminate between target and nontarget events. Performance deteriorates when stimuli are presented rapidly (Krulewitz, Warm, & Wohl, 1975; Loeb & Binford, 1968; Parasuraman & Davies, 1976) or in a degraded fashion (Nuechterlein, 1983; Nuechterlein,

SUSTAINED ATTENTION

121

Parasuraman, & Jiang, 1983); this suggests that stimulus encoding plays a role in target detection. Performance will sometimes deteriorate when target and nontarget events are presented successively, suggesting that memory load affects vigilance performance. Parasuraman (1979) performed two studies that demonstrated that perceptual sensitivity (d’) changes only under conditions in which stimuli are presented successively and at a high rate. The combination of the memory load and the fast event rate was hypothesized to place an attentional demand on observers’ abilities to discriminate between target and nontarget events. Recent research tends to support this view (Lanzetta, Warm, Dember, & Berch, 1985; Parasuraman & Mouloua, 1987; Warm, Chin, Dittmar, & Dember, 1985). Davies and Parasuraman’s (1982) view that vigilance performance may be influenced by task type and by memory demands is given support by a study conducted recently by Koelega, Brinkman, Hendricks, and Verbaten (1989). They assessed the vigilance of nonretarded adults on two sensory-based and two cognitively based tasks, each 60 minutes in duration. One of the sensory tasks and one of the cognitive tasks required memory of prior stimuli. Koelega et al. ( 1989) reported that overall detection performance and the rate of the vigilance decrement were a function of both type of task and memory load, and suggested that observers’ specific abilities and those task demands interact. The study conducted by Tomporowski and Allison (1988) provided the initial step toward isolating specific factors that differentially affect the sustained attention of mentally retarded and nonretarded observers. Our next two studies evaluated the effects of one of these factors, memory demand, on vigilance. In the first study, Tomporowski, Hayden, and Applegate (1990) evaluated the effects of memory load on the sustained attention of mentally retarded (MIQ = 65, MCA = 28 years, 9 months) and nonretarded (MCA = 18 years, 9 months) adults. Observers performed missing-digit tests, similar to that used by Tomporowski and Allison ( I 988), in which the digits 1 to 9 were presented for 1 second with an interstimulus interval of either 1 second (30 events per minute) or 3 seconds (15 events per minute). Under these test conditions it was presumed that observers would be required to encode the value of a digit, hold the value in working memory until the next digit was presented, and then decide if the order of the digits was successive. Thus, the memory load of the tests was determined by the event rate. Each subject performed four 60-minute tests. In two tests, the event rates remained constant at either 30 or 15 per minute for the entire vigil. In the remaining tests the event rate was shifted without warning from one rate to the other (30 to 15 per minute or 15 to 30 per minute) 30 minutes into the vigil. There were 14 targets presented randomly during each 10-minute period of the four tests. The two event rates were 30 and 15 per minute; as such, they would be classified as fast and slow event rates according to Davies and Parasuraman’s (1982) taxonomic system. Mentally retarded observers detected fewer targets and made more false

122

Phillip D . Tomporowski and Lisa D . Hager

alarms than did nonretarded observers. Event rate did not differentially affect hits in either intelligence group; however, a slow event rate resulted in more false alarms for mentally retarded observers. An analysis of hits and false alarms was also conducted using measures from SDT; a nonparametric index of stimulus discriminability, P(A), was calculated (McNicol, 1972). The analysis revealed a ceiling effect in the performance of the nonretarded subjects and for this reason a direct comparison between the performance of the two groups of subjects could not be made. Nevertheless, the effects of event rate on the performance of mentally retarded observers could be assessed. It was found that P(A) was the same under both the fast and slow stimulus presentation conditions and that performance declined at the same rate. In the studies conducted by Tomporowski and Allison (1988) and Tomporowski et al. (1990), tasks that contained memory demands affected the overall level of target detection of mentally retarded adults more than they did nonretarded adults; however, the memory demands did not differentially affect observers’ sustained attention. Thus, the differences in performance between the intelligence groups may be attributable to the poorer short-term memory processes of retarded individuals and not to poorer attentional processes. As discussed by Berch and Kanter (1984) and Giambra and Quilter (l988), a veridical comparison of the sustained attention of individuals who are known to differ in cognitive abilities requires that subjects perform vigilance tasks that place similar information-processing demands on observers. In this way, changes in performance can be attributed to sustained attention alone and not to other cognitive processes. In an attempt to isolate the role of attentional processes during vigilance tests, a second experiment was conducted that evaluated the effects of memory demand on sustained attention. Tomporowski and Simpson (1990) developed a cognitive vigilance test in which memory demands were equated for individual subjects. The 60-minute test consisted of consecutive matching-digit sets; observers viewed a series of single digits presented successively (memory set) and, following a delay, compared the memory set with a series of digits presented simultaneously (test set). The target events were identical digit sets. The length of the digit set used for each subject was determined by the subject’s performance on a memory-span pretest. The study assessed the sustained attention of 16 mentally retarded (MIQ = 64.8 years, MCA = 23 years, 5 months) and nonretarded (MCA = 20 years) adults on two variations of the vigilance test. There was a 3second delay between the memory set and test set in one test; a 6-second delay was used in the other test. It was hypothesized that the two delay intervals would result in tests that differed in memory load. As seen in Fig. 1, the stimulus discriminability scores, P(A), of the two IQ groups was similar during the initial 30 minutes of the vigil; however, during the final 30 minutes, the performance of mentally retarded observers declined more rapidly than did that of nonretarded

123

SUSTAINED ATTENTION

0-0 0-0 A-A

A-A

NMR 3s IS1 NMR (6s lSl] MR 3s IS1 MR (6s lSl{

0.85

2 m

z

0.80

v

0.75

4

zE

v,

n

0.70 4

I

3 4 5 OBSERVATION TIME (1 0-MIN PERIODS)

1

2

6

FIG. I . Mean stimulus discriminability scores for retarded and nonretarded groups under two delay intervals as a function of time on watch. NMR, non-mentally retarded adults; MR, mentally retarded adults; ISI, interstimulus interval.

observers. The 3- and 6-second delays between memory and test sets did not influence either perceptual sensitivity or hits. False alarms, however, were found to be related to the delay interval; mentally retarded and nonretarded observers made more false alarms under the 6-second delay condition than under the 3second delay condition. The results of this study provide the first clear evidence to support the view that the sustained attention of mentally retarded adults is poorer than that of nonretarded adults. On tasks equated for memory demand, retarded adults’ vigilance performance declined at a more rapid rate than that of nonretarded adults. The poorer detection performance of mentally retarded individuals may result from differences in sustained attention; however, there are alternative possibilities. The number of digits in the memory sets in the Tomporowski and Simpson (1990) study was determined by a pretest measure of digit span. The average number of digits in the memory sets was significantly lower for mentally retarded observers (3.5digits) than for nonretarded observers (5.8 digits). Thus, the mentally retarded subjects performed more trials during the vigil than did nonretarded subjects. It is possible that the decline in performance was due to a greater buildup of proactive inhibition in retarded observers. Also, the number of targets presented during each period of the vigil was constant; as such, the probability of the occurrence of target events differed between the two groups of subjects. On the average, mentally retarded observers experienced lower target probabilities than did nonretarded observers. Target density has been shown to affect target detection in vigilance tasks (Craig, 1976, 1978; Parasuraman &

124

Phillip D . Tomporowski and Lisa D . Hager

Davies, 1976; Williges, 1969, 1971, 1973). In most cases, increasing target density results in increases in hits and false alarms (see Davies & Parasuraman, 1982; Parasuraman, 1986; and Warm, 1984 for reviews). False alarms were made more often during test conditions with longer intervals between events to be remembered. This pattern of false alarms was similar to that observed in the study conducted by Tomporowski et al. (1990), in which event rate was manipulated. The relation between the frequency of false alarms and delay interval in these two studies may be explained in several different ways. The change in false alarms could have resulted from differences in target density. When there is a high proportion of targets to nontargets, observers’ expectancies may influence the likelihood that they will make responses. The frequency of false alarms may have also been related to differences in the memory load present under the delay conditions. With longer delays, observers may have had more difficulty remembering the value of the digits in the memory set. Similar explanations have been offered for the increasingly poor performance exhibited by mentally retarded subjects (Baumeister, Smith, & Rose, 1965) and nonretarded subjects (Kinchla & Smyzer, 1967; Murdock, 1982) in delayed-recognition tasks as the interval between stimuli increases, and for increases in errors during repetitive counting tasks (Healy & Nairne, 1985). The false alarm responses may also reflect mentally retarded individuals’ tendency to perseverate on tasks (Balla & Zigler, 1979; Zigler & Balla, 1982); their responses to targets under the slow event rate (high target density) condition may have increased their general tendency to respond. Lastly, the difference in false alarms may simply reflect changes in the level of motivation to perform the task. During tests with long delays and infrequently presented stimuli, mentally retarded observers may tend to respond more because of boredom than do nonretarded observers. A recent study conducted by Tomporowski and Tinsley (1 99 1a) addressed these alternative explanations for the rapid decline in vigilance seen in the study conducted by Tomporowski and Simpson (1990). They modified the missingdigit task used by Tomporowski et al. (1990) to ensure that task demands were comparable for intelligence groups. Twenty mentally retarded observers (MIQ = 63, MCA = 18 years) performed tasks in which digits from 1 to 9 were presented successively and repetitively for 500 milliseconds every 1500 milliseconds; observers were to detect instances when two digits were not presented in sequence. Twenty nonretarded adults (MCA = 20 years, 1 month) performed tests in which three digits were presented simultaneously for 500 milliseconds every 1500 milliseconds. The digits began with 999 and declined by a value of 2 on each presentation (e.g., 999,997,995,993,991,989, . . .). Target events were digits presented out of sequence in either the second or third position of the three-digit series. The equivalency of the task demands of the two forms of the test was established in pilot studies. The subjects performed two 60-minute versions of the task; one version had a target density of 5% and the other test had a 30% target density.

-

‘2

1 .oo

-

4 0.95.rn W

5

0.90--

-0-0

0

rn

-

0.85--

f

0.80--

A m

2 a v

v,

0.75-

n

0.70 7

i

1 .oo T

12 I-

a a

0.90-

t 0 a

3

0.80--

0

I

P

0.70--

rn

z

0

$, 0.60W

a 0.50 4

I

1

2 3 4 5 OBSERVATION TIME (10-MIN PERIODS)

6

1 2 3 4 5 OBSERVATION TIME (1 0-MIN PERIODS)

6

1°00TC

600

4

4

FIG. 3. ( A ) Mean stimulus discriminability scores. (B) mean response-to-target ratios, and (C) iiicaii reaponhe latciicics for retarded and nonretardcd groups under two target density conditions ah a function of [imc on watch. N M R , non-mentally retarded adults; MR, mentally retarded adults.

126

Phillip D . Tomporowski and Lisa D . Huger

An analysis of the subjects’ stimulus discriminability scores, P(A), as seen in Fig. 2A, revealed that the performance of the nonretarded observers did not change during the vigil under the high target density condition. Under the low target density condition, however, performance declined during the initial 10minute period of the vigil and then remained unchanged for the remainder of the test. The performance of the mentally retarded observers was clearly different than that of the nonretarded observers. There were no significant differences between the groups during the initial 20 minutes of the test; however, after that point, the performance of the mentally retarded observers began to decline rapidly under both target density conditions. The relative change in performance between the intelligence groups over time was taken as evidence for poorer sustained attention in mentally retarded observers. These findings are similar to those obtained by Tomporowski and Simpson (1990) and support the view that, during complex vigilance tasks, mentally retarded individuals’ ability to sustain attention is poorer than that of nonretarded individuals. In Tomporowski and Tinsley’s (1991a) study, retarded and nonretarded observers were presented an equal number of stimulus events per unit of time; thus, differences in performance could not be attributed to differences in proactive inhibition or confounded by target density. An analysis of subjects’ response bias (p) was also conducted. A nonparametric analog measure of subjects’ response criteria was calculated following a method described by Craig (1978), in which the ratio of the number of responses (hits + false alarms) to the number of target events is calculated for each subject at each time period. The average response ratios for each of the six 10-minute periods of the vigil are seen in Fig. 2B. An analysis of response bias indicated that mentally retarded observers displayed a greater shift in their response criteria under the high target density condition between the second and third time periods. This change may represent a decreased willingness to respond to all events as targets. Thus, mentally retarded observers tend to respond less frequently late in a vigil, and responses that are made tend to be errors of commission. It might be suggested that the shift to a more conservative response bias could be the result of changes in motivation; however, an examination of subjects’ response times to target events fails to support a motivational explanation. As seen in Fig. 2C, there was no difference in response latencies for mentally retarded and nonretarded subjects in either the low or high density test conditions. It would be expected that if motivational factors differentially affected the tendency to key press, the response latencies of mentally retarded observers would increase with time. This was not the case, however. Like the study conducted by Tomporowski and Simpson (1990), this study indicates that sustained attention for mentally retarded individuals on complex, memory-demanding tasks differs in three ways from that of nonretarded adults;

SUSTAINED ATTENTION

127

mentally retarded individuals detect fewer targets, make more false alarms, and evidence a more rapid decline in vigilance. Further, on the basis of SDT, mentally retarded adults demonstrate more rapid decrements in stimulus detectability and, when targets are presented frequently, their response criteria become more conservative than those of nonretarded observers during the course of the vigil.

IV.

THEORIES OF VIGILANCE AND FUTURE RESEARCH

A general picture of the vigilance abilities of mentally retarded adults is beginning to emerge. First, mentally retarded adults perform more poorly than nonretarded adults on vigilance tasks that place demands on encoding processes. Their detection performance is poorer than that of nonretarded individuals on the CPT (Crosby, 1972; Rosvold et al., 1956) and on sensory-based tasks with target and nontarget stimuli that are difficult to discriminate (Tomporowski & Allison, 1988). Second, mentally retarded adults perform more poorly than nonretarded adults on vigilance tasks that tax memory capacities. They detect fewer targets and make more false alarms than nonretarded observers (Tomporowski & Allison, 1988; Tomporowski et al., 1990) and, under some conditions, show a more rapid vigilance decrement than nonretarded observers (Tomporowski & Simpson, 1990; Tomporowski & Tinsley, 1991a). Third, there are some tasks that mentally retarded and nonretarded adults perform in a similar fashion (Kirby et a]., 1978, 1979; Thomas, 1989; Ware et al., 1962). It may be that mentally retarded and nonretarded adults can be expected to perform at the same level on tasks that place limited demands on encoding abilities and memory abilities. Fourth, the attentional abilities of mentally retarded adults differ from those of nonretarded adults, and it is not simply a case of slower mental development as suggested by the studies of Kirby and his colleagues (Kirby et al., 1978, 1979; Thomas, 1989). Several studies conducted in our laboratory demonstrated that overall levels of vigilance and ability to sustain attention are poorer for mentally retarded than for nonretarded adults. At the very least, these results suggest that there is a developmental lag in sustained attention in mentally retarded children and adolescents that continues into early adulthood. These general findings provide a framework within which to evaluate theories of sustained attention that have been applied to mentally retarded persons. For instance, it has been suggested that vigilance performance is related to the level of subjects’ neurological arousal and it has been hypothesized that the poor vigilance evidenced by retarded individuals may be caused by lower levels of arousal and the habituation of arousal (Das & Bower, 1971; Rosvold et al., 1956; Semmel, 1965; Stankov, 1983). Mentally retarded individuals are believed to possess less intact neural systems and these systems are more likely to be affected by the type of repeated stimulation found in vigilance tests (Rosvold et al., 1956).

128

Phillip D . Tomporowski and Lisa D . Hager

Arousal theory, although intuitively appealing, is given little empirical support by the studies reviewed. Several studies reported that retarded observers made more false alarms as a function of time on task (Tomporowski and Allison, 1988; Tomporowski & Simpson, 1990; Tomporowski & Tinsley, 1991a); these results tend to run counter to predictions made by arousal theory. Distractibility theory has also been offered to explain differences in the attention of retarded and nonretarded persons (Crosby, 1972; Crosby & Blatt, 1968; Krupski, 1980). This theory suggests that mentally retarded individuals display more frequent lapses of attention and “off-task’’ gazing that affect their performance on vigilance tasks. It would be predicted that deficits in vigilance performance would be particularly noticeable when stimuli are presented to mentally retarded individuals for short durations, as in the CPT, or in a degraded fashion. In the studies reviewed, detection performance for mentally retarded adults was poorer than that for nonretarded adults on the CPT (Crosby, 1972; Rosvold et a]., 1956). The same pattern was found on one sensory-based task (Tomporowski & Allison, 1988); however, Crosby (1972) conducted his study specifically to assess the effects of distractors on the performance of retarded subjects, and he failed to provide support for the theory. At this time, there is insufficient evidence to favor the distractibility theory. Information-processing theories, however, appear to provide a compelling explanation for the results of many of the studies reviewed. These theories emphasize the role of stimulus selection mechanisms (Broadbent, 197I), expectancy (Craig, 1976, 1978; Jerison, 1970; Vickers & Leary, 1983), and cognitive processes (Fisk & Scerbo, 1987; Fisk & Schneider, 1981; Parasuraman, 1979; Posner, 1978) in attention. Taken as a group, these theories predict that intelligence-related differences in vigilance performance emerge as task complexity increases. Further, they predict that mentally retarded individuals’ lower target detection and higher false alarm rate on cognitively effortful tasks are attributable to poor memory abilities. In particular, their more rapid decline in vigilance is due to their inability to draw on available attentional resources to sustain information in working memory. These theories provide several directions for future research on the sustained attention of mentally retarded persons. It remains to be determined how encoding factors differentially affect the detection performance of mentally retarded and nonretarded observers. Methods of assessing attentional processes under perceptually demanding conditions have been developed by Nuechterlein ( 1983) and Nuechterlein et al. ( 1983); they found that perceptual sensitivity declined rapidly during the first few minutes of a CPT in which stimuli were presented for very short durations (40 to 50 milliseconds) and in a degraded fashion. Their methods seem well suited to the study of individual differences in attention. The role of memory processes and how they affect the sustained attention of mentally retarded individuals is another important area for future research. Sever-

SUSTAINED ATTENTION

129

a1 early theories of vigilance entertained the view that the role of memory processes is important to developing a comprehensive theory of sustained attention. Dornic (1967) proposed a “memory trace” theory in which the ability to detect a target depended on the interval between the presentation of target events and the number of nontarget events presented during the delay interval. The nontarget events were hypothesized to interfere with or weaken the memory trace of the target event. Johnston, Howell, and Williges (1969) developed a theory of vigilance for complex monitoring tasks that encompassed four processes: ( I ) attentional processes that determine an observer’s sensitivity to signals, (2) decision processes that control an observer’s willingness to respond, (3) scanning processes that affect an observer’s detection latency, and (4) memory processes that combine with attentional processes to determine an observer’s ability to detect target events. The role of each of the four processes was hypothesized to be fixed by task variables, for example, target density, event rate, and stimulus conspicuity. Unfortunately, the vigilance theories of Dornic and Johnston et al. and their emphasis on memory and cognitive processes have had relatively little impact on recent theories of vigilance (Davies & Parasuraman, 1982). For the most part, current theories of sustained attention have been proposed by researchers whose main interests are focused on the psychophysical aspects of the vigilance task rather than on its cognitive factors. Parasuraman, Warm, and Dember (1987) have voiced concern over the lack of vigilance theories that might help explain sustained attention during both complex monitoring tasks and cognitive vigilance tasks. A systematic analysis of factors that affect response bias and response strategies may also lead to a better understanding of the sustained attention of mentally retarded individuals. It is known that the performance of nonretarded observers is determined, in part, by their expectancy of target events (Colquhoun & Baddeley, 1964, 1967) and task instructions (Colquhoun, 1967; Williges, 1973). These variables tend to affect response bias (p) and not perceptual sensitivity ( d ’ ) ;observers tend to engage in active probability-matching strategies that produce shifts in their response patterns (Craig, 1978, 1981, 1987; Vickers, 1979; Williges, 1969, 1971, 1973). Given that mentally retarded individuals implement strategies differently from nonretarded individuals on a variety of tasks (Bray, 1979; Brown, 1974), the study of changes in retarded subjects’ response strategies during vigilance tasks as a function of such task parameters as pretraining, target location, payoff matrices, and target probability would appear particularly important. The data obtained by Tomporowski and Tinsley (1991a) were the first to suggest that mentally retarded individuals may fail to maintain strategies during the course of a vigil; when targets were presented frequently, retarded observers developed a more conservative response criterion and they became less willing to respond than nonretarded observers.

130

Phillip D . Tomporowski and Lisa D . Hager

It is also important, both theoretically and in the applied sense, to determine the conditions under which sustained attention is not related to level of intelligence. Many work settings demand that employees maintain high levels of performance on redundant tasks. An examination of task characteristics that differentially affect mentally retarded and nonretarded persons is important to education and job-training programs (Tomporowski & Hayden, 1990). There has been considerable discussion of the relation between laboratory methods of evaluating vigilance and work performance (Parasuraman, 1986), and little is known of the relation between intelligence level and the performance of specific jobs. It may be that subtle differences in the information-processing demands of some jobs will place retarded individuals at a disadvantage. Vigilance tests are often thought to be monotonous, and changes in performance over time have been suggested to simply reflect observers’ willingness to respond (Smith, 1966). Motivation theory would posit that many retarded individuals, because of their history of education and socialization, might approach the vigilance test situation differently. At this general level, a motivational explanation for differences in vigilance performance between mentally retarded and nonretarded observers might be plausible. Crosby ( 1972) found clear differences between the performance of institutionalized and noninstitutionalized retarded adults on the CPT. Zigler and his associates (Balla & Zigler, 1975; Zigler, 1961; Zigler & Balla, 1972) obtained data to suggest that mentally retarded children persist on repetitive tasks longer than do nonretarded children. The difference in responding was thought to reflect motivation for social interaction. Zigler and Balla (1982) suggest that mentally retarded individuals are affected by positive and negative reaction tendencies; they possess a heightened motivation to interact with others, but also have a reluctance to do so. Motivation theory can be interpreted to predict that the performance of mentally retarded individuals on simple cognitive tasks may actually be superior to that of nonretarded individuals. Although Zigler and Balla’s views are compelling, none of the studies reviewed provided support for their position. Indeed, one study demonstrated that reaction time for mentally retarded and nonretarded observers did not differ during a vigil, suggesting that subjects were equally motivated to perform the task (Tomporowski & Tinsley, I99 I a). Nevertheless, a systematic study of the effects of motivational factors on the sustained attention of retarded individuals would provide an important contribution to the field. Detection performance is influenced by several factors often thought to be motivational in nature. Knowledge of results has been found to improve the detection performance of mentally retarded (Perryman, Halcomb, & Landers, 1981; Semmel, 1965; Ware et al., 1962) and nonretarded observers (Dittmar, Warm, & Dember, 1985; Ware et al., 1962; Warm, Kanfer, Kuwada, & Clark, 1972; Warm, Epps, & Ferguson, 1974; Warm, Reichman, Grasha, &

131

SUSTAINED ATTENTION

Siebel, 1973; Wiener, 1968, 1974). Monetary rewards have also been found to improve nonretarded observers’ vigilance performance in some (Sipowicz, Ware, & Baker, 1962; Tomporowski, Simpson, & Hager, 1991), but not all (Bergurn & Lehr, 1964; Pollack & Knaff, 1958), cases. Other factors believed to affect vigilance performance include intrinsic motivation (Dember, Warm, Bowers, & Lanzetta, 1984; Noonan, Karwoski, & Loeb, 1985; Tomporowski & Tinsley, 1991b) and subjects’ perceptions of task boredom (O’Hanlon, 1981) and stress (Becker, Warm, & Dember, 1991; Sanders, 1983; Warm, Dember, & Parasuraman, 1991). The extent to which mentally retarded and nonretarded observers differ in their subjective responses to vigilance tests remains to be determined. In this article, several theories of attention were drawn on to help explain the vigilance performance of retarded individuals. The most compelling theories were those that focused on individual differences in information processing. Given the multifaceted characteristic of vigilance behavior, it is unlikely that a single, comprehensive theory of sustained attention will be forthcoming; however, it is suggested that continued theory-based evaluation of sustained attention in mentally retarded individuals will lead to a better understanding of both sustained attention and the nature of mental retardation.

REFERENCES Anderson, R. P., Halcornb, C. G . , Gordon, W., & Ozolins, D. A. (1974). Measurement of attention distractibility in LD children. Academic Therapy, 9, 261-266. Baddeley, A. D., & Colquhoun, W. P. (1969). Signal probability and vigilance: A reappraisal of the ‘signal rate’ effect. British Journal of Psychology, 60, 165-178. Bakan, R. (1955). Discrimination decrement as a function of time in a prolonged vigil. Journal of Experimental Psychology, 60, 165- 178. Balla, D., & Zigler, E. (1975). Preinstitutional social deprivation and responsiveness to social reinforcement in institutionalized retarded individuals: A six-year follow-up study. American Journal of Mental DeJciency, 80,228-230. Balla, D., & Zigler, E. (1979). Personality development in retarded persons. In N. R. Ellis (Ed.), Handbook of mental deficiency (2nd ed., pp. 143-186). Hillsdale, NJ: Erlbaum. 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 mentally retarded (pp. 1-38). Hillsdale, NJ: Erlbaum. Baumeister, A. A , , Smith, T., & Rose, J. D. (1965). The effects of stimulus complexity and retention interval upon short-term memory. American Journal on Mental Deficiency, 70, 129- 134. Becker, A. B.. Warm, J. S., & Dernber, W. N. (1991). Effectsof feedbackon perceived workload in vigilance performance. In Proceedings of the Human Factors Society 35th Annual Meeting (pp. 1491-1494). Berch, D. B., & Kanter, D. R. (1984). Individual differences. In J. S. Warm (Ed.), Sustained attention in human performance (pp. 143-178). New York: Wiley. Bergurn, B. O., & Lehr, D. J. (1964). Monetary incentives and vigilance. Journal of Experimental PSyChoIoRy, 67, 197- 198.

132

Phillip D. Tomporowski and Lisa D . Hager

Bray, N. W. (1979). Strategy production in the retarded. In N. R. Ellis (Ed.), Handbook of mental deficiency, psychological theory and research (2nd ed., pp. 699-726). Hillsdale, NJ: Erlbaum. Broadbent, D. E. (1971). Decision and stress. New York: Academic Press. Brown, A. L. (1974). The role of strategic behavior in retardate memory. In N. R. Ellis (Ed.), International review of research in mental retardation (Vol. 7, pp. 55-1 11). New York: Academic Press. Buck, L. (1966). Reaction time as a measure of perceptual vigilance. Psychological Bulletin, 65, 29 1-308.

Carr, T. H. (1984). Attention, skill, and intelligence: Some speculations on extreme individual difference in human performance. In P. H. Brooks, R. Sperber, & C. McCauley (Eds.), Learning and cognition in the mentally retarded (pp. 189-215). Hillsdale, NJ: Erlbaum. Colquhoun, W. P. (1967). Sonar target detection as a decision process. Journal of Applied Psychology, 51, 187-190. Colquhoun, W. P., & Baddeley, A. I). (1964). Role of pretest expectancy in vigilance decrement. Journal of Experimental Psychology, 68, 156- 160. Colquhoun, W. P., & Baddeley, A. D. (1967). Influence of signal probability during pretraining on vigilance decrement. Journal of Experimental Psychology, 73, 153- 155. Craig, A. (1976). Signal recognition and the probability-matching decision rule. Perception and Psychophysics, 20, 157-162. Craig, A. (1978). Is the vigilance decrement simply a response adjustment towards probability matching? Human Factors, 20, 441-446. Craig, A. (1981). Monitoring for one kind of signal in the presence of another: The effects of signal mix on detectability. Human Factors, 23, 191-197. Craig, A. (1987). Signal detection theory and probability matching apply to vigilance. Human Factors, 29, 645-652. Crosby, K. G. (1972). Attention and distractibility in mentally retarded and intellectually average children. American Journal of Mental Deficiency, 77, 46-53. Crosby, K. G., & Blatt, B. (1968). Attention and mental retardation. Journal of Education, 150, 67-81.

Das, I. P. (1970). Vigilance and verbal conditioning in the mildly and severely retarded. American Journal of Mental Deficiency, 75, 253-259. Das, J. P., & Bower, A. C. (1971). Orienting responses of mentally retarded and normal subjects to word-signals. British Journal of Psychology, 62, 89-96. Davies, D. R., & Parasuraman, R. (1982). The psychology ofvigilance. New York: Academic Press. Davies, D. R., & ' h e , G. S . (1970). Human vigilance performance. London: Staples. Dember, W. N., Warm, J. S . , Bowers, J. C., & Lanzetta, T. M. (1984). Intrinsic motivation and the vigilance decrement. In A. Mital (Ed.), Trends in ergonomicslhuman factors I (pp. 21-26). Amsterdam: Elsevier (North-Holland). Detterman, D. K. (1975). Memory in the mentally retarded. In N. R. Ellis (Ed.), Handbook of mental deficiency (2nd ed., pp. 727-760). Hillsdale, NJ: Erlbaum. Dittmar, M. L., Warm, J. S . , & Dember, W. N. (1985). Effects of knowledge of results on performance in successive and simultaneous vigilance tasks: A signal detection analysis. In R. E. Eberts & C. G. Eberts (Eds.), Trends in ergonomicslhuman factors I / (pp. 195-202). Amsterdam: Elsevier (North-Holland). Domic, S. (1967). Expectancy of signals and the memory trace. Studia Psychologia, 9, 87-91. Ellis, N. R. (1978). Do the mentally retarded have poor memory? Intelligence, 2, 41-54. Fisk, A. D., & Scerbo, M. W. (1987). Automatic and control processing approach to interpreting vigilance performance: A review and reevaluation. Human Factors, 29, 653-660. Fisk, A. P., & Schneider, W. (1981). Control and automatic processing during tasks requiring sustained attention: A new approach to vigilance. Human Factors, 23, 737-750.

SUSTAINED ATTENTION

133

Gale, A., & Lynn, R. (1972). A developmental study of attention. British Journal of Educational Psychology, 42, 260-266. Giambra. L. M.,& Quilter, R. E. (1988). Sustained attention in adulthood: A unique, large sample, longitudinal and multicohort analysis using the Mackworth clock-test. Psychology and Aging, 3,75-83. Green, D. M.. & Swets, 1. A. (1966). Signal detection theory andpsychophysics. New York: Wiley. Healy, A. F., & Nairne, J. S. (1985). Short-term memory processes in counting. CognirivePsychology. 17, 417-444. Jerison, H. J. (1970). Vigilance, discrimination, and attention. In D. 1. Mostosky (Ed.),Attention: Conremporury theory and analysis ( pp. 127- 147). New York: Appleton-Century-Crofts . Johnston, W. A., Howell, W. C., Williges, R. C. (1969). The components of complex monitoring. Organizational Behavior and Human Performance, 4 , 112-124. Kinchla, R., & Smyzer, F. (1967). A diffusion model of perceptual memory. Perception andPsychophysics, 2 , 219-229. Kirby, N. H., Nettelbeck, T., & Bullock, J. (1978). Vigilance performance of mildly mentally retarded adults. American Journal of Menral Dejiciency, 28, 394-397. Kirby. N. H., Nettelbeck, T., & Thomas, P. (1979). Vigilance performance of mildly mentally retarded children. American Journal of Mental Deficiency, 84, 184- 187. Koelega, H. S . . Brinkman, J., Hendricks, L., & Verbaten, M. N. (1989). Processing demands, effort, and individual differences in four different vigilance tasks. Human Factors, 31, 45-62. Krafft, K. M., & Dietrich, K. N. (1985). Methodological considerations in the assessment of vigilance in children. In R. E. Eberts & C. G. Eberts (Eds.), Trends in ergonomicslhuman factors I1 ( pp. 169- 179). Amsterdam: Elsevier (North-Holland). Krulewitz, J. E., Warm, J. S . , & Wohl, T. H. (1975). Effects of shifts in the rate of repetitive stimulation on sustained attention. Perception & Psychophysics, 18, 245-249. Krupski, A. (1980). Attention processes: Research, theory, and implications for special education. In B. K. Keogh (Ed.), Advances in special education: Vol. 1. Basic consrructs and theoretical orientations (pp. 101-140). Greenwich, CT: JAI Press. Lanzetta, T. M., Warm, J. S., Dember, W. N., & Berch, D. B. (1985). Information processing load and the event rate function in sustained attention. Proceedings of the 29rh Annual Meeting of the Human Factors Society, 2, 1084-1088. Loeb, M., & Binford, J. R. (1968). Variation in performance on auditory and visual monitoring tasks as a function of signal and stimulus frequencies. Perception and Psychophysics, 4 , 361-366. Mackworth, N. H.(1948). The breakdown of vigilance during prolonged visual search. Quarterly Journal of Experimental Psychology, I , 6-2 I . Maisto, A. A., & Baumeister, A. A. (1984). Dissection of component processes in rapid information processing tasks: Comparison of retarded and nonretarded people. In P. H. Brooks, R. Sperber, & C. McCauley (Eds.), Learning and cognition in the mentally retarded (pp. 165-188). Hillsdale, NJ: Erlbaum. McNicol. D. (1972). A primer of signal detection theory. London: Allen & Unwin. Merrill. E. C. (1990). Attentional resource allocation and mental retardation. In N. W. Bray (Ed.), International review of research in mental retardation (Vol. 16, pp. 51-88). New York: Academic Press. Merrill, E. C., Sperber, R. D., McCauley, C., Littlefield, J., Rider, E. A., & Shapiro, D. (1987). Picture encoding speed and mental retardation. Intelligence, 11, 169-191. Mohan, J. (1982). An experimental developmental study of performance, reminiscence and vigilance. Studia Psychologica, 24, 71-76. Murdock, B. B.. Jr. (1982). Recognition memory. In C. R. Puff (Ed.), Handbook of research methods in human memory and cognition (pp. 2-26). New York: Academic Press. Nettelbeck, T. (1985). Inspection time and mild mental retardation. In N. R. Ellis & N. W. Bray

I34

Phillip D. Tomporowski and Lisa D.Hager

(Eds.), lnrernational review of research in mental retardation (Vol. 13, pp. 109-143). New York: Academic Press. Noonan, T. K., Karwoski, W., & Loeb, M. (1985). Perception of challenge and difficulty in a cognitive vigilance task. In R. E. Eberts & C. G. Eberts (Eds.), Trends in ergonomicslhuman facrors I1 ( pp. 189- 193). Amsterdam: Elsevier (North-Holland). Norman, D. A., & Bobrow, D. G. (1975). On data-limited and resource-limited processes. Cognitive Psychology, 7 , 44-64. Nuechterlein, K. H. (1983). Signal detection in vigilance tests and behavioral attributes among offspring of schizophrenic mothers and among hyperactive children. Journal of Abnormal Psychology, 92, 4-28. Nuechterlein, K. H., Parasuraman, R., & Jiang, Q. (1983). Visual sustained attention: Image degradation produces rapid sensitivity decrement over time. Science, 220, 327-329. O’Hanlon, J. F. (1981). Boredom: Practical consequences and a theory. Acta Psychologica, 49, 53-82. Parasuraman, R. (1979). Memory load and event rate control sensitivity decrements in sustained attention. Science, 205, 924-927. Parasuraman, R. (1986). Vigilance. monitoring, and search. In K. R. Boff, L. Kaufman, & J. P. Thomas (Eds.), Handbook of perception and human performance: Vol. 11. Cognitive processes and performance ( pp. 43- I , 43-39). New York: Wiley. Parasuraman, R., & Davies, D. R. (1976). Decision theory analysis of response latencies in vigilance. Journal of Experimental Psychology: Human Perception and Performance, 2,587-590. Parasuraman, R., & Mouloua, M. (1987). Interaction of signal discriminability and task type in vigilance decrement. Perception & Psychophysics, 41, 17-22. Parasuraman, R., Warm, 1. S.. & Dember, W. N. (1987). Vigilance: Taxonomy and utility. In L. S. Mark, J. S . Warm, & R. L. Huston (Eds.), Ergonomics and human factors: Recent research (pp. 1 1-32). New York: Springer-Verlag. Penyman, R. E., Halcomb, C. R., & Landers, W. F. (1981). Operant conditioning of mental retardates’ visual monitoring. Perceptual and Motor Skills, 53, 507-5 12. Pollack, I., & Knaff, P. R. (1958). Maintenance of alertness by a loud auditory signal. Journal of the Acoustical Society of America, 30, 1013-1016. Posner, M. 1. (1978). Chronometric explorations of mind. Hillsdale. NJ: Erlbaum. Rosvold, H. E., Mirsky, A. F., Sarason, I., Bransome, E. D., & Beck, L. N. (1956). A continuous performance test of brain damage. Journal of Consulting Psychology, 20, 343-350. Sanders, A. F. (1983). Towards a model of stress and human performance. Actu Psychologica, 53, 61-97. Semmel, M. I. (1965). Arousal theory and vigilance behavior of educable mentally retarded and nonretarded children. American Journal of Mental Deficiency, 79, 44-5 1. Sipowicz, R. R.,Ware, J., & Baker, R. A. (1962). The effects of reward and knowledge of results on the performance of a simple vigilance task. Journal of Experimental Psychology. 64,58-61, Smith, R. L. (1966). Monotony and motivation: A theory of vigilance. Los Angeles: Dunlop and Associates, Inc. Sperber, R. D., & McCauley, C. (1984). Semantic processing efficiency in the mentally retarded. In P. H. Brooks, R. Sperber, & C. McCauley (Eds.), Learning and cognition in the mental1.v retarded (pp. 141-163). Hillsdale, NJ: Erlbaum. Stankov, L. (1983). Attention and intelligence. Journal of Educational Psychology, 75. 471-490. Stanovich, K. E. (1978). Information processing in mentally retarded individuals. In N . R. Ellis (Ed.), International review of research in mental retardation (Vol. 9, pp. 29-60). New York: Academic Press. Thomas, P. G . (1989). Vigilance performance of mildly mentally retarded children and adults. Unpublished doctoral dissertation, University of Adelaide.

SUSTAINED ATTENTION

135

Tomporowski, P. D., & Allison, P. (1988). Sustained attention of adults with mental retardation. American Journul on Mental Retardation, 92, 525-530. Tomporowski, P. D., & Hayden, A. M. (1990). Employment of individuals with developmental disabilities. In G . Reid (Ed.), Problems in movement control (pp. 133-158). Amsterdam: North-Holland. Tomporowski. P. D., Hayden, A. M., & Applegate, B. (1990). The effects of background event rate on the sustained attention of mentally retarded and nonretarded adults during cognitive vigilance tests. American Journal on Mental Retardation, 94,499-508. Tomporowski, P. D., & Simpson, R. G . (1990). Sustained attention and intelligence. Intelligence, 14, 27-38. Tomporowski, P. D., Simpson, R. G . , & Hager, L. (in press). Method of recruiting subjects and performance on cognitive tests. The American Journal of Psychology. Tomporowski, P. D., & Tinsley, V. (199 la). Effects of target density and event rate on the vigilance of mentully retarded and nonretarded adults. Manuscript submitted for publication. Tomporowski, P. D., & Tinsley, V. (1991b). Sustained attention in young and elderly observers during cognitive vigilance tests. Manuscript submitted for publication. Vickers, D. (1979). Decision process in visual perception. London: Academic Press. Vickers, D., & Leary, J. N. (1983). Criterion control in signal detection. Human Factors, 25, 283-296. Ware, 1. R. (1960). The effects of intelligence on visual and auditory vigilance tasks. Unpublished manuscript, University of Louisville. Ware, J. R., Baker, R. A., & Sipowicz, R . R. (1962). Performance of mental deficients on a simple vigilance task. American Journal of Mental Deficiency, 66, 647-650. Warm, J. S . (1984). An introduction to vigilance. In J. S. Warm (Ed.), Sustained attention in human perfvrmunce ( pp. I - 14). New York: Wiley. Warm, J. S., & Berch. D. B. (1985). Sustained attention in the mentally retarded: The vigilance paradigm. In N. R. Ellis & N. W. Bray (Eds.), lnternational review of research in mental retardation (Vol. 13, pp. 1-41). New York: Academic Press. Warm, 1. S . , Chin, K., Dittmar, M. L., & Dember, W. N. (1985). Effects of head restraint on signal detectability in simultaneous and successive vigilance tasks. Journal of General Psychology, I 14, 423-43 1. Warm, J. S., Dember, W. N., & Parasuraman, R. (1991). Effects of olfactory stimulation on performance and stress in a visual sustained attention task. Journal of the Society of Cosmetic Chemists, 4 2 , 199-210. Warm, J. S., Epps. B. D., & Ferguson, R. P. (1974). Effects of knowledge of results and signal regularity on vigilance performance. Bulletin of the Psychonomics Society, 4 , 272-274. Warm, J. S., Kanfer, F. H., Kuwada, S . , & Clark, J. L. (1972). Motivation in vigilance: Effects of self-evaluation and experimenter controlled feedback. Journal of Experimental Psychology, 92, 123- 127. Warm, I. S., Reichmann, W., Grasha, A. F., & Siebel, B. (1973). Motivation in vigilance: A test of the goal-setting hypothesis of the effectiveness of knowledge of results. Bulletin of the Psychonomic Society, I , 291-292. Wiener, E. L. (1968). Training for vigilance: Repeated sessions with knowledge of results. Ergonomirs, I I , 547-556. Wiener, E. L. (1974). An adaptive vigilance task with knowledge of results. Human Factors, 16, 333-338. Williges, R. C. (1969). Within-session criterion changes compared to an ideal observer criterion in a visual monitoring task. Journal of Experimental Psychology, 81, 61-66. Williges, R . C. (1971). The role of payoffs and signal ratios in criterion changes during a monitoring task. Humun Furtors, 13, 261-267.

136

Phillip D . Tomporowski and Lisa D . Hager

Williges, R. C. (1973). Manipulating the response criterion in visual monitoring. Human Factors, 15, 179-185. Zigler, E. (1961). Social deprivation and rigidity in the performance of feeble-minded children. Journal of Abnormal and Social Psychology, 62, 413-421. Zigler. E. (1982). Developmental versus difference theories of mental retardation and the problem of motivation. In E. Zigler & D. Balla (Eds.), Menzal retardation: The developmental-difference controversy (pp. 163- 188). Hillsdale, NJ: Erlbaum. Zigler, E., & Balla, D. (1972). Developmental course of responsiveness to social reinforcement in normal children and institutionalized retarded children. Developmental Psychology, 6, 66-73. Zigler, E., & Balla, D. (1982). Introduction: The developmental approach to mental retardation. In E. Zigler & D. Balla (Eds.), Mental retardation: The developmental-difference controversy (pp. 3-8). Hillsdale, NJ: Erlbaum.

How Modifiable Is the Human Life Path? ANN M. CLARKE SCHOOL OF EDUCATION THE UNIVERSITY OF HULL HULL HU6 7 R X , ENGLAND

ALAN D. 6. CLARKE DEPARTMENT OF PSYCHOLOGY THE UNIVERSITY OF HULL HULL HU6 7 R X , ENGLAND

I.

INTRODUCTION

The origins of the stream of research to be reviewed are to be found in the authors’ early studies of institutionalized mildly retarded adolescents and young adults (e.g., Clarke & Clarke, 1 9 5 3 , 1 9 5 4 , 1959; Clarke, Clarke, & Reiman, 1958). Subsequent work showed that the findings generalized to different samples of children who, to varying extents, suffered disablement (in some cases temporary) arising from adversity (e.g., Kadushin, 1970; Kagan & Klein, 1 9 7 3 ; Koluchova, 1972, 1976). The present thesis therefore goes well beyond the field of mental retardation, but includes it. It challenges the view that experiences in the first few years of life predetermine the life path. By life path we mean the biosocial trajectories of development from birth through childhood. adulthood, and senescence. These complex processes affect and are reciprocally affected by the experiences of individuals. As implied, the whole of the life span is perceived as important and, to varying extents, formative. Such a view contrasts with earlier (but still common) notions that regarded the first few years of life as critical for later development (cf. Bowlby, 1951, 1988). To provide a historical context, it may be helpful to indicate briefly some autobiographical details insofar as these illuminate the origin of the authors’ INTERNATIONAL REVIEW OF RESEARCH IN MENTAL RETARDATION, Vol. 18

I37

Copyright 0 1992 by Academic Press. Inc. All rights of reproduction in any form reserved.

138

Ann M . Clarke and Alan D.B . Clarke

current attempts to understand human development better. In 1951 we were asked to reopen a psychology department founded and then abandoned by several psychologists in a very low cost institution. This hospital contained some 1400 persons compulsorily detained under legal order, some three-quarters of whom were mildly retarded or, in some cases, merely backward. These were mainly adolescents or young adults who either had caused social problems or were regarded as at risk of perpetrating maladjusted or criminal behavior in the community. Two apparently unrelated streams of research followed, both of which gave results in the teeth of received wisdom. The more important, relating to this article, was the accidental discovery that some of our mildly retarded individuals in adolescence and early adulthood showed marked increments in IQ and social adjustment. A number of hypotheses were examined; these included the possibility that younger members would have made greater increments, that test practice might be responsible, that to be a recent admission to the hospital might result in underestimation because of emotional upset, and that those with lower IQs might show greater improvements. None of these appeared potent; however, it was established that a record of earlier severe, prolonged social adversity predicted later improvement. The rating of severe adversity depended on the presence in the individual case history of at least 2 of 12 specific criteria indicating an early record of cruelty and neglect that had led to official action to remove the child from the parents (Clarke & Clarke, 1954). Conversely, less severe adversity or its absence in the background was associated with a poorer prognosis for IQ increments, social adjustment, and discharge from care. After a pilot study, a “clean” series of studies were initiated, in which assessment was carried out independently by a colleague without knowledge of social history or earlier test scores, and another independent researcher without knowledge of the individual or test scores rated childhood history for the degree of adversity (severe or less severe). Altogether we undertook four studies, although our results were recorded in a number of articles in addition to the primary sources. The first (Clarke & Clarke, 1954) investigated the frequency of IQ changes over 24 years and their relationship to earlier severe adversity. The second, third, and fourth studies were all reported by Clarke et al. (1958). The second study employed a 4a-year test-retest time interval, confirming larger changes in relation to the longer time interval. The third increased the test-retest interval to 6 years. Some 78% of the severely deprived group showed increments of 15 or more IQ points, compared with some 25% of the less deprived group. The fourth study showed that changes were not related to differences in institutional programs, but represented a fading of the effects of earlier adversity-a self-righting tendency. Finally, we studied a control group of consecutive new admissions, reassessed after 3 months, enabling an estimate to be made of the greatest possible effects of regression toward the mean

MODIFIABILITY OF THE HUMAN LIFE PATH

139

and test practice. On average, these accounted for at the most only a quarter of the 6-year increments. The late or delayed development occurring in the most deprived mildly retarded persons seemed to represent one of several factors accounting for the marked decrease in administrative prevalence from adolescence onward, which is common in epidemiological research. For example, Richardson (1985, p. 374) reported some results from the Aberdeen study in which five birth-year cohorts were followed from ages 5 to 22. After school-leaving age (15-16 years) the prevalence of those receiving mental retardation services decreased by more than half, a drop accounted for by formerly designated mild cases, there being no change in the number of more severely retarded. Subsequently our four studies were replicated several times elsewhere. For example, Svendsen (1982, 1983) followed to adulthood a sample of educable mentally retarded and slow-learning children. Estimates of early family and personal problems showed that IQ changes were associated with the greater number of problems and with postschool education. Increments in the “problems” group averaged 18 IQ points, compared with 5 points in the group with one or no problems and without further education. Somewhat similar findings are recorded by Roswell Harris (1958) and Brown (1972), among others. This, then, was a problem that found the researchers rather than vice versa. It challenged the notion of a necessary developmental constancy, a view that has dominated developmental theory for most of this century (Brim & Kagan, 1980; Clarke & Clarke, 1984). It also challenged the idea, then recently revived (Bowlby, 1951), that the events of the first few years have a crucially formative influence. The early experiences of our “adverse” subjects could scarcely have been more damaging. Indeed, their compulsory incarceration provided testimony to their earlier disturbed behavior and significantly subaverage intellectual functioning. But these effects had tended to fade. As noted, improvements were much less common in those with backgrounds of less severe or no adversity. We were soon to be greatly encouraged by the identification of the catch-up phenomenon in physical growth after illness or mild malnutrition which seemed to rather closely parallel our behavioral results. It seemed possible that these findings might generalize (e.g. Clarke & Clarke, 1959, 1960). During the 1960s most of the published evidence appeared to go against our hypothesis and to support the existence of psychosocial critical periods very early in life. The work of the ethologists was very influential; Harlow (1963), Levine (1960), and Scott (1963, 1968) all published experimental results showing the long-term adverse effects of early deprivation in animals. While accepting the usefulness of animal studies at neural levels, we were very skeptical about the extrapolation of some of the results to social and educational problems in disadvantaged children. So far as Harlow’s work on the later maladjustment of formerly isolated

140

Ann M . Clarke and Alan D . B . Clarke

monkeys was concerned, we took this very seriously indeed, but lamented that an equal amount of ingenuity had not been spent constructing experiments that might reverse the damage. We argued that if, following isolation experiences, these monkeys had not been transferred to an improved, remedial situation, then a self-perpetuatingcycle of maladjusted behavior might be expected. There was no evidence that any such attempt had been made. Much later this problem was addressed by Suomi and Harlow (1972) and Novak and Harlow (1975) whose results were in keeping with our own hypothesis. There was also the problem of the many human studies that showed merely a correlation between early environmental circumstances and later behavioral status, without demonstrating a clear environmental change inbetween, strictly speaking the only valid test of the crucial effect of the early period. Finally, to our great surprise, Skeels (1966) entitled his important monograph “Adult Status of Children with Contrasting Early Life Experiences: A Follow-up Study,” [our italics], again following the zeitgeist of the 1960s. This research seemed clearly to b,” a study of children who had been shifted around a good deal during the “critical years” (Skeels & Dye, 1939) and then fortunately had been adopted (albeit late) and remained in stable, accepting homes. The outcome in adult life was thus almost certainly as much a product of their happy adoptive status as of their early experiences (see Clarke, 1982; Clarke, Clarke, & Berg, 1985, pp. 9, 10). By 1975 we believed that there was sufficient evidence to produce a book that suggested that early experiences, good or bad, did not in themselves set for the child an invariant life path. Contributions were included from such scholars as Urie Bronfenbrenner, Jerry Kagan, and Michael Rutter. It was stressed that conclusions should not be interpreted literally as a counterbalance, which might be an equal and opposite extreme, but rather as an attempt to achieve a balanced view. With one or two discordant voices, reviewers seemed to feel that in Early Experience: Myth and Evidence (Clarke & Clarke, 1976) a valid case had been made for a greater degree of potential openendedness in human development than had until then been accepted. Summarizing the main arguments, we indicated that there is no known adversity from which at least some children had not recovered if moved to something better and that the whole of the life path is important, including the early years. These are, of course, for most children, foundational-leading in most cases to confirmatory-influences, as age increases. Three additional points were also made. First, increasing age probably imposes constraints on potential responsiveness to environmental influences. This may be intrinsic to the aging process or may result from habit, forced or chosen life paths, and social pressures. For want of a better analogy we called this a “wedge” model. Second, one way in which early experience effects may be perpetuated, usually indirectly, is when one good or bad thing leads to another and a chain of

MODIFIABILITY OF THE HUMAN LIFE PATH

141

good or bad events then follows. For example, a maladjusted child in care may not be considered for fostering or adoption and, therefore, will remain in a less than satisfactory institution. The maladjustment may remain, justifying on later review no new intervention. Third, it seemed possible that early adversity, overcome by improved circumstances, might nevertheless leave the individual potentially more vulnerable to later stress. Experiences that affect the individual’s behavior in more than a transitory way must involve learning, broadly defined. Stress might reawaken earlier maladjusted behavior in the same way that unused skills are more easily relearned after the passage of time. Traditionally, researchers often sought powerful single main factors leading to single main effects; however, many variables may be involved in an outcome. These are not necessarily solely additive; some studies have shown potent multiplicative effects of two, three, or four variables (Rutter, Yule, Quinton, Rowlands, Yule, & Berger, 1975). Identifying six risk factors in childhood, Rutter et al. showed that the occurrence of one yielded no greater risk than was evident in control children. TWOresulted in a fourfold increase, and four factors produced a tenfold increase in emotional and conduct disorders. Genetic factors, constitutional factors, and micro- and macroenvironmental influences represent headlines for the complex processes involved in human development. One should also add transactions; there is empirical evidence that, to some extent, individuals create or choose their own environments, receiving reinforcing or modifying feedback (e.g., Bell, 1968; Scarr & McCartney, 1983; Thomas, Chess, & Birch, 1968). Finally, chance encounters or events sometimes potently alter the course of development. Even an armchair analysis of these overly simple headlines would suggest enormous complexity in human development, and that the probable nonlinearity of each variable, let alone in combination, would indicate some degree of unpredictability in the long term of individual (as opposed to group average) life paths. Indeed, this complexity suggests that the title of this article is overambitious, and that a choice of research data under just a few headings must represent an oversimplification. This criticism is both offered and accepted; however, it is hoped that some general principles will emerge. Hundreds of correlational studies of intelligence, temperament, or attainment over long periods indicate that there is a range of individual change or nonchange in ordinal position and relative level, with individual constancy at one extreme and marked change over time at the other. “How modifiable is the human life path?” We can already answer that, for many individuals, behavioral characteristics alter differentially in natural circumstances, in spite of social and educational pressures that tend to keep a person on a self-fulfilling prophetic track. And these characteristics, in turn, may affect the person’s life path for good or ill (Clarke & Clarke, 1986).

142

Ann M . Clarke and Alan D . B . Clarke

Contributors to Brim and Kagan’s (1980) Constancy and Change in Human Development offer a number of different ways of tracking development. Here perhaps the simplest has been chosen: investigations of changing or unchanging behavioral levels. Empirical evidence concerning children with actual or potential problems is outlined; here the question of modifiability of the life path is at its most pressing and its test the most exacting. We discuss the outcome for four groups: 1. children removed from very severe adversity; 2. children with prenatal or perinatal biomedical problems in different social classes; 3. children with adjustment difficulties; and 4. children escaping from disadvantage. Finally, an attempt is made to draw the threads together in an overview.

It.

REMOVAL FROM ISOLATION OR VERY SEVERE ADVERSITY

One test of human resilience is to be found in studies of children removed from isolation or very severe adversity. Legends about feral children date from Romulus and Remus; later 19th- and 20th-century cases suggest that these were recently abandoned autistic or retarded children. Davis (1947) provided the first modem studies concerning two children withdrawn from long isolation. With the help of a specialist, one made a remarkable recovery; without special attention, the other child did not. The latter may have been the victim of organic impairment. The best documented study is by the Czech psychologist Koluchova (1972, 1976), who has provided the authors with recent unpublished material. Her successive follow-ups concerned monozygotic twin boys, born in 1960, who lost their mother soon after birth, were cared for by a social agency for a year, and were then reared for a further 6 months by a maternal aunt. Their development was normal. The father, who may have been limited intellectually, remarried, but his new wife proved to be extremely cruel to the children, banishing them to a cellar for the next 54 years and beating them from time to time. On discovery they were dwarfed in stature, lacked speech, failed to recognize the meaning of pictures, and were severely malnourished and rachitic. They were legally removed from their parents and, after a medical rehabilitation program, were adopted by an exceptional lady and her sister. Scholastically, from a state of severe retardation they caught up with age peers and achieved intellectual and emotional normality. After basic education they went on to train as typewriter

MODIFIABILITY OF THE HUMAN LIFE PATH

143

mechanics but later decided to go on to secondary school, specializing in electrotechnology. After military service, one now works as an instructor in vocational training and the other as a computer technician. Both are married; one has a child, They are said to be entirely stable emotionally, lacking abnormalities and enjoying warm relationships. Note that they are twins, that they were together in appalling circumstances, and that soon after removal from their parents they were adopted into an unusually warm and supportive home. A second case, a girl adopted by the same lady at age 54, was more problematical. She had been isolated on her own and had a psychopathic natural mother; no information was available on the father. Cared for by a social agency during infancy, she appeared mentally retarded. Later, she was restored to her mother, who concealed her existence and neglected the child. She suffered repeated otitis and may have had meningitis. She was difficult for some time in her new family, showing anxiety and, at school, poor peer relations. But gradually she improved, becoming average scholastically; she now works in a factory making handbags (J. Koluchova, personal communication, 1991). Thompson (1986, and personal communication 1990) has published the case history of a little boy called Adam who was abandoned at the age of 4 months in Colombia and brought up in isolation in a reformatory for girls, where she first saw him. The conditions were appalling and the child was severely malnourished. He was removed to an orphanage at 16 months, at which time he weighed only 12 lb, 12 oz (i.e., less than 6000 g). Emotionally he was completely withdrawn. He could not sit, walk, or crawl. Adam was diagnosed as an extremely malnourished, mentally retarded spastic. Gradually he improved and, at 32 months, was adopted by a North American family, along with a little girl from the same orphanage. Initially, of course, there were problems. Adam was doubly incontinent and frequently bit his sister, but he has continued to improve in his adoptive family and, at age 8, his WISC-R was 113; his sister who was less severely deprived had an IQ of 102 at the same age. Both graduated from high school and are emotionally stable. A useful review of most known and properly recorded cases (nine in number) has been provided by Skuse (1984). Certain clinical features in such isolated children are, on discovery, virtually ubiquitous: motor retardation, absent or rudimentary language, grossly retarded perceptuomotor skills, paucity of emotional expression, lack of attachment behavior, and social withdrawal. If recovery is to occur, progress must be rapid at first. Skuse concludes that “in the absence of genetic or congenital anomalies . . . victims of such deprivation have an excellent prognosis. Some subtle deficits in social adjustment may persistmost human characteristics, with the possible exception of language are strongly “canalised” [in Scam-Salapatek’s ( 1976) conception] and hence virtually resistant to obliteration by even the most dire environments” (p. 567). If this is so, then the message is clear. Early behavioral development is not

144

Ann M . Clarke and Alan D . B . Clarke

critical in the same way that, for example, early prenatal central nervous system damage can be critical. Gross deficits can be made good and the self-righting tendency suggested by Waddington (1957, 1966) appears strong. Ill.

RECOVERY FROM PERINATAL PROBLEMS

Studies of some isolated children illustrate the gradual fading of physical and developmental problems granted special care in unusual circumstances. Is there evidence to support the idea that these ameliorative processes can take place without the removal of the child from the scene of his or her initial difficulties? The best known longitudinal study of children who suffered perinatal damage is, perhaps, Werner’s (Werner, 1985, 1989; Werner & Smith, 1982). She and her colleagues studied a total birth cohort of Oriental or Polynesian descent on the island of Kauai. The sample included 698 children, followed from birth until 2 years, 10 months, and in some cases until 18 and 32 years. Neonatal and infant mortality rates were very low, but a number of infants experienced perinatal stress involving a variety of congenital disorders. “At birth 9% of the cohort had some congenital defects, of which 3.7% were serious enough to require long-term, specialized care for either severe physical handicaps and/or severe mental retardation” (Werner, 1985, p. 337). Fourteen (or 2%) of the children surviving until 2 years had suffered severe perinatal stress; 69 (or 10%) had suffered moderate perinatal stress. At each of the follow-up stages intellectual outcome was a function of a significant interaction between characteristics of the caretaking environment and degree of perinatal stress. As early as 20 months, the children growing up in middle-class homes who had experienced the most severe perinatal complications had mean scores on the Cattell Infant Scale almost comparable to those of children with no perinatal stress who were living in disadvantaged homes. The most developmentally retarded (in physical as well as intellectual status) were those who had both experienced perinatal stress and been reared in the poorest homes. The impact of the caretaking environment appeared even more powerful at age 10 years. First, children with and without severe perinatal stress who had grown up in middle-class homes both achieved mean PMA (Primary Mental Abilities) IQ scores well above the average. Second, PMA IQ scores were seriously depressed in children from low SES [socioeconomic status] homes, particularly if they had experienced severe perinatal stress. Third, the family’s socioeconomic status showed significant associations with the rate of serious learning and behavior problems. By age 18 years, fen rimes as many youths wirh serious coping problems were living in poverry as had survived serious perinatal stress. (Werner, 1985, p. 341)

The latest report (Werner, 1989) focused on high-risk children, one third of whom had in adult life made satisfactory adjustments despite a combination of

MODIFIABILITY OF THE HUMAN LIFE PATH

145

moderate to severe perinatal stress, discordant or impoverished homes, and alcoholic or mentally disturbed parents. Many of the characteristics identifying the resilient children were already present at the age of 2. Pediatricians and psychologists noted their alertness, positive social orientation, and self-help skills. The author also drew attention to supportive social environmental factors that appeared to protect these young people at various points during their life span. A less well known but very elegant study was presented by Wilson (1985) as an offshoot of the famous Louisville Twin Study. This important prospective longitudinal research project included 494 pairs of monozygotic and dizygotic twins and their singleton siblings, demographically representative of the metropolitan borough of Louisville, Kentucky. These twins were tested at 3 months, 6 months, 12 months, 24 months, 3 years, and 6 years. A vast amount of personal and demographic information concerning the children and their families is available. The cohort yielded, naturally enough, babies who were small for gestational age and also twins falling below 1750 g birth weight. Would all these babies be at risk for slow mental development compared with the complete cohort of twins who as a group were somewhat premature, were somewhat smaller at birth, and showed delayed development? I t turned out that the small-for-gestational-age infants later exhibited only a modest deficit in 1Q scores (about 4 points) at age 6, reduced from a deficit of 16 points at age 3, when compared with the whole cohort. The statistical significance was greatly enhanced by the large numbers involved, outweighing their psychological importance. The author therefore concludes that small-forgestational-age babies are at no special risk for later retardation. The same, however, was not true of the very low birth weight babies, many of whom were also very premature. They showed a highly significant deficit at each test age throughout childhood, and although the initial deficit of 19 IQ points was eventually reduced to 9 points, there was no evidence of further recovery for this group. In contrast to the small-for-gestational-age twins, most, but not all, of these infants seemed to be at a long-term disadvantage. When the group was divided by social status, although the numbers were relatively small, 17 high versus I7 low, those in the high-SES group appeared to have recovered by age 6 with average IQs of 101, whereas those in the low-SES group were one standard deviation below them with average IQs of 86. Mother's education was significantly related to recovery from 24 months onwards, which suggests that maternal intelligence plays a prominent role in determining the level of recovery. When monozygotic twins of markedly unequal birth weight were compared, the twins who weighed less than 1,750 g attained the same level of IQ scores at 6 years as did their heavier cotwins. Among these genetic replicates, the initially powerful effects of low birth weight did not exert a long-term handicapping effect on mental development. The data argue for a high degree

146

Ann M . Clarke and Alan D . B . Clarke

of resilience in mental development in the face of prenatal stress and for a powerful effect of heritage and home environment in guiding the recovery from early deficit. ( pp. 69, 70)

Wilson (1985) concludes: Although the prior analyses have dwelt on the predictions of which infants will recover fully and which ones will not, perhaps the most heartening feature of these data is the pronounced weight gain shown by all risk infants up to school age. The effects of risk are not immediately overcome in the first 2 years, but even the low-SES infants showed a substantial upward gain between 3 years and 6 years of age. There are recuperative capabilities and qualities of resilience even among risk infants that steadily compensated for the burden ofprematurity. Such qualities bring into focus the fact that developmental processes are continuous and ongoing, and they possess intrinsic self-correcting tendencies. In fact, Waddington (1971) identified this as one of the cardinal properties of developmental processes (p. 84).

IV.

ADJUSTMENT DIFFICULTIES

We now turn to studies of children who showed adjustment difficulties and disorders that are usually defined as maladaptive reactions to psychosocial stressors leading to impairment of functioning, or symptoms that are in excess of normal reactions. These include severe anxiety, sleep disturbances, phobias, and depression, identified at different periods of childhood and adolescence. How do they fare in adult life? The answer, by contrast with many other psychiatric disorders, is well. As with many abnormal conditions, adjustment disorders may arise from a variety of different, and sometimes overlapping, causes. There are a number of obvious variables: the family context, especially the quality of parenting; the degree of match or mismatch between parents and between parents and child; the qualities of individual vulnerability and resilience during or following stress; the enhanced probability of disorder in those with organic brain dysfunction (e.g., Rutter & Sandberg, 1985); the doubled risk of emotional and conduct disorders in inner-city areas (London) compared with rural areas (Isle of Wight). The New York Longitudinal Study, initiated in 1956 by Chess and Thomas, provides a careful prospective longitudinal study of childhood disorders, uniquely with data gathered in all cases before onset (Chess & Thomas, 1984; Thomas et al., 1968). The sample consisted of 133 middle- to upper-middle-class subjects recruited through personal contact with parents during pregnancy or shortly after birth. An important incentive for joining the study and remaining in it was the promise of free high-quality medical care. There was no attrition among the 133 children whose families joined. Adjustment disorders represent a wide range of very common behavioral prob-

MODIFIABILITY OF THE HUMAN LIFE PATH

147

lems, which, as Chess and Thomas note, in many cases represent age-specific behaviors that, though troublesome, are not suggestive of pathological deviation. Sometimes the issue is simple, involving inappropriateness of the routines employed by parents. Suggesting alternative ways of handling the child can be effective. When problems are not resolved, psychiatric evaluation and sometimes treatment are necessary. As is to be expected in this sample there were virtually no conduct disorders. Of 45 clinical cases in childhood, 40 exhibited adjustment disorders, the majority classified as mild. By adolescence 24 had recovered and 2 had improved; by adulthood 29 had recovered and 5 had improved their original childhood diagnostic status, 2 were unchanged, 2 were moderately worse, and 2 were markedly worse. In addition, of 12 cases with onset during adolescence, 6 were completely recovered in early adult life, including 3 with severe problems (Chess & Thomas, 1984). Two recent prospective longitudinal studies, one from Sweden (Von Knorring, Anderson, & Magnusson, 1987) and one from Germany (Esser, Schmidt, & Woerner, 1990), confirm the fadeout of clinically diagnosed adjustment disorders with time. The von Knorring et al. study provided data on a large, representative sample, prospectively from 10 to 24 years and retrospectively from 0 to 9 years. In reviewing relevant literature these authors noted the higher incidence of childhood disorders in cities, compared with towns and rural areas, and the higher incidence in boys than girls, which, however, is reversed during adolescence. Conduct disorders may be dealt with other than clinically, but the authors underline the poor prognosis for these, with good outcomes for children with emotional disorders. The criteria for inclusion in the von Knorring et al. (1987) sample were different from those of the Chess and Thomas research, as von Knorring et al. depended on the rather tough criterion of psychiatric referral during childhood through early adulthood. The findings endorse other work, however; only 3 of 28 children exhibiting anxiety and emotional disorders before 9 years of age remained in psychiatric care in early adulthood, whereas a quarter of a large group with onset at 10 to 14 years and almost half of those with onset between 15 and 19 years were still in psychiatric care between ages 20 and 24. Males who had attended special classes were at particular risk for developing these disorders. Esser et al. followed a random sample of 356 eight-year-old children in West Germany until the age of 13. Although the overall prevalence rate of 16% for all psychiatric disorders remained constant across the two ages, the persistence of conduct disorders in individual cases was very much greater than that of emotional disorders. A related topic concerns the life path of children removed from problem families. Two recent studies reported by Quinton, Rutter, Dowdney, Liddle, Mrazek, and Skuse (1982) complement one another and illustrate an important principle. Prospective and retrospective studies on the same theme and with

148

Ann M . Clarke and Alan D . B . Clarke

similar populations normally yield different answers to the same question. A major reason is that an identified adult deviant group may enable a continuous deviant past history to be revealed in a retrospective investigation, whereas there is no way of knowing how many had similar early histories yet escaped unscathed. The research by Quinton et al. has a bearing on both intraindividual and intergenerational constancy and change. Both projects studied current parenting and associated psychiatric problems, relating these to past and present circumstances. The authors sought to assess the effects on child rearing of early emotional deprivation, as predicted by early learning theorists (Quinton et al., 1982). The first, a retrospective study, evaluated histories of parents whose children had been removed from home and placed in residential care. The second, a prospective study, examined the adult outcome of children originally assessed by the late Jack Tizard and colleagues in the 1960s when they were in residential care, having been themselves removed from home. In both investigations, appropriate carefully chosen comparison groups were used. Important differences emerged between the findings of these two studies: “our data . . . show that the picture of intergenerational continuities looking forwards is quite different from that looking backwards” (Quinton et al., 1982, p. 293). The retrospective sample of adults with children in care exhibited a strong link between their parenting breakdown and their own adverse childhood experiences; however, comparison group data showed that quite severe early adversities seldom led directly to such problems. Thus, an additional linking mechanism must be invoked in the former sample. Multiple personal and material circumstances seem to have been involved (see also Quinton & Rutter, 1984a, 1984b). The second, prospective sample of adults who as children had been assessed in care indicated that the experience of serious adversities in childhood or of seriously maladaptive and disruptive parenting does not always lead to parenting and other breakdown in the next generation. There is certainly an increased “risk” but this is not a “probability.” Many comparison children had also experienced severe adversity but had survived. Indeed, “intergenerational discontinuities outweigh continuities” (Quinton et al., 1982, p. 293). Nearly a third of the women, removed from their parents and reared in institutions showed good parenting, and a further quarter displayed only moderate difficulties of a type also shown by 40% of the comparison group (for an overview, see Rutter, Quinton, & Liddle, 1983). The general point here is that early adversity can set up a chain of consequences. The effects of early adversity can produce new environmental effects; there also appears to be an increased personal vulnerability to later adversity. Later events in adolescence or adulthood can be potent agents of amelioration or added difficulties. For example, indirect selective factors may lead to choosing a

MODlFlABlLlTY OF THE HUMAN LIFE PATH

149

stable and supportive marriage partner or an unstable, equally damaged spouse. Once again, the cumulative and transactional nature of development is implied by these important findings. One small subgroup showed up as particularly vulnerable (Rutter et al., 1983). Ten children had been admitted to institutional care before the age of 2 years and had remained there continuously until 16 years of age or older, inevitably experiencing the relatively detached, multiple-person caretaking that characterize institutions. Eight of the ten showed later problems, and although this subgroup difference fell short of statistical significance, it is possible that detached, multiple-person caretaking commencing in infancy may be as unsatisfactory in its consequences as rearing in seriously discordant multiproblem situations.

V.

ESCAPE FROM DISADVANTAGE

Our theme, that the life path is modifiable whether through internal or external factors, or both, may be further illustrated by reference to studies, often prospective surveys, that indicate that some individuals escape their adversities. This can occur without any special intervention. For example, a small British study (Tonge, Lunn, Greathead, McLaren, & Bosanko, 1983) indicated that at adulthood one-third of those reared in problem families achieved normal adjustment, one-third made marginal adjustment, and one-third remained clearly disadvantaged, like their parents. Many other studies show that there are always some children who escape the dire predictions suggested by their early circumstances. In 1972, Sir Keith Joseph initiated what became known as the Transmitted Deprivation Program. He had asked why it was that in spite of major social improvements since the end of the Second World War, “deprivation and problems of adjustment so conspicuously persist? . . . social attitudes and ways of life tend to recur in some families from one generation to the next. But what is more surprising (is that) this does not always happen.” In Cycles of Disadvantage (Rutter & Madge, 1976) and Despite the Welfare State (Brown & Madge, 1982) the field was reviewed at the beginning and end of the program, respectively. The 23 studies supported and the more than a dozen commissioned reviews of relevant world literature endorsed the view that cycles of deprivation do exist but do not inevitably recur between generations. These books offer a very comprehensive account of the problem under discussion. The Newcastle Study, one example of a late funded project, began well before the Transmitted Deprivation Program. In 1947 Sir James Spence and colleagues undertook what was initially envisaged as a 1 -year investigation into the ecology of infant illnesses in a total sample of 1142 infants born to Newcastle families in 2 months of that year. Successive follow-up studies, with much broader develop-

150

Ann M. Clarke and Alan D . B . Clarke

mental concerns, led to a final report by Kolvin, Miller, Scott, Gatzanis, and Fleeting (1990), who used baseline data from the 1952 assessments. The original sample had by then shrunk to 847, largely due as a result of movement out of the city, presumably of the less deprived. The final follow-up took place in 1979/ 1980; three randomly selected groups, comprising 264 families, were chosen from the 1952 cohort: (1) a nondeprived sample (14% of the original), (2) a deprived group (50% subsample) comprising those who met one or two of the criteria of deprivation, (3) a multiply deprived group (68% subsample) with three or more indices of deprivation. By determined efforts 96% of the sample was traced and 92% interviewed, more than 15 years (age 32) after the last contact. Among the more salient findings, it emerged that there were degrees of movement both into and out of deprivation. Indeed, as Kolvin et al. (1990, p. 167) indicate “change is the dominant feature.” Thus, of the sample of 62 children whose families did not show any deprivation in 1952, only 43.5% of the families of formation remained nondeprived in 1980, nearly 5% being multiply deprived. Escape from 1952 multiple deprivation was equally evident. Only 43.6% remained in the same situation; the remainder mostly shifted to lesser degrees of adversity, and almost 13% escaped completely. Kolvin et al. believe that deprivation increases personal vulnerability, modifiable, however, by such protective factors as intelligence, equable temperament, scholastic ability, and social skills. There is a degree of specific intergenerational transmission via marital disruption and poor quality of physical and emotional care of children. More general transmission, reflected in the sum of individual adverse factors, appears more potent. The findings of this study are, in general, congruent with the conclusions of other research. This work spans a period of immense social change perhaps somewhat attenuated in northeast England. It emphasizes what Rutter (1989) has termed chain effects. These can persist along the life path or, in certain circumstances, can be broken by environmental influences and personal qualities. The National Child Development Study (NCDS) has followed the development of all children born in the LJnited Kingdom during one week in March 1958. Various subsamples, as well as the whole cohort, have been the subject of separate analyses. Here the study of only one subsample is mentioned. In Born to Fail? Wedge and Prosser (1973) underlined data on 11-year-olds in the NCDS , the already robust findings concerning the physical and psychological correlates of socioeconomic status. They had a widespread impact, drawing public attention to “the massive accumulation of burdens afflicting disadvantaged children and their families, and which they are frequently expected somehow to overcome. Yet it should cause no surprise that so many of these children fail to ‘behave,’ fail to ‘learn’ and fail to succeed . . . one in sixteen, the disadvantaged group, suffered adversity after adversity, heaped upon them from

MODIFIABILITY OF THE HUMAN LIFE PATH

151

before birth; their health was poorer, their school attainment lower and their physical environment worse in almost every way than that of ordinary children” (p. 59). So the question for the future becomes, Did they all fail in later development? Fortunately, this study was continued by Essen and Wedge (1982), as part of the Transmitted Deprivation series. It concerned children in the NCDS who were identified as socially disadvantaged at 11, 16, or both 11 and 16. These data, together with earlier information, were designed to represent the ecology of these children at different stages of development. One of the main aims was to determine whether families experience long-term or transitory difficulties, and how these bear on outcome. Such families could at the same time be compared with ordinary families. The criteria of disadvantage were threefold: rearing in atypical families, poor housing, or low income. At ages 11 and 16, the percentages for the whole NCDS sample were, for the first criterion, 19.8 and 19.3; for the second, 14.7 and 10.2; and for the third, 13.1 and 11.9, respectively. Those who fulfilled all three criteria at age 11 constituted 4.5%, but by age 16 this had declined to 2.9%. Although the findings show some continuities between ages 11 and 16, marked discontinuities were also evident. Of all the disadvantaged at either age, about a quarter were disadvantaged at both ages, half at age 11 only, and a quarter at age 16 only. It is therefore clear that a larger group suffered disadvantage at some stage in childhood than is indicated by the cross-sectional findings at ages I1 and 16 (Essen & Wedge, 1982, p. 46). This underlines the fact that even an extreme ecology is not necessarily static. From the detailed findings, it became clear that those who at 11 or 16 years were multiply disadvantaged were found to show poorer achievement and behavior and to be shorter on average than ordinary children. Multiple disadvantages at either age were equally damaging to the children’s life chances. It is of interest that for outcome measures at age 16, those who were multiply disadvantaged at that age were at no greater disadvantage than those who had experienced such problems at age 1 1. “Neither of these ages appears to be particularly critical . . . [in] that multiple disadvantage of either, and therefore perhaps any, age is equally damaging to the children’s life chances” (p. 166). Recently this NCDS was followed up with interviews at age 26/27 of an educationally successful sample and a comparison group, both of whom had been disadvantaged during one or more periods in childhood. The former were somewhat less disadvantaged than the latter, and certain parental and adolescent aspirations for the future seemed relevant. It was already known that such qualities as easy temperament and good motivation were relevant to escape. This final study confirmed the importance of such factors, and indicated that the achievers had better relationships in their homes than did the comparison subjects. The strongest predictors related to parental interest in the child’s education,

152

Ann M . Clarke and Alan D . B . Clarke

and of course educational achievement is the key to a better adult life (Pilling, 1990).

VI.

DISCUSSION

It has been argued that the various behavioral characteristics may severally or in combination affect the direction of the individual life path. Of course, this concept of a life path, or what Rutter (1989) calls life pathways, is a broad one. It represents multiple correlations of dozens of contributing components, which may vary over time. This article has summarized some of the best recent evidence available. The implications of these studies could suggest a totally unrealistic proposal, namely, that to be absolutely safe so far as the environment is concerned, all the world’s children should from conception be cared for in well-educated, economically secure, stable and loving families that will support them in times of special need, discipline them, and engender self-discipline to meet life’s problems, ensuring that emerging talents are encouraged. Even so, it must be noted that some children reared in these ideal circumstances would still lead troubled lives. Conversely, some children with early problems and reared in dire circumstances will escape and achieve normal adult adjustment. Why do some succeed and others fail? Attempts will be made to answer this difficult question. Bear in mind, however, that no prospective study as yet conducted, no matter how carefully designed and or how many measures used, can give us detailed answers in individual cases. Nevertheless, there are some clues, culled from numerous studies, including those reviewed by Rutter et al., (1983) and Werner (l989), already discussed. Several factors seem to be protective: 1. Constitutional dispositions that will render some children attractive even to the most depraved parents and probably to other members of the family and to a wider community. These include sociability, problem-solving ability, and planning ability, leading gradually to an internal locus of control. Such children are likely to attract the positive attention of teachers in school; to acquire self-esteem and self-confidence; and to believe both in their own ability to adapt to changing circumstances and in their ability to change circumstances themselves. 2. Some network of affectional support, which may be absent even in the best institutions but may be present in very disadvantaged homes. 3. Schools in which children are valued and encouraged to learn. There are by now massive research data supporting the view that schools may differentially enhance both achievement arid adjustment. 4. A peer group, probably based in school, that is prosocial. This means that

MODIFIABILITY OF THE HUMAN LIFE PATH

153

individual children need to be sufficiently attractive in positive ways to be chosen as friends. 5. An ability to plan purposefully, which Rutter (1989) believes will make a big difference to choice of career andlor choice of continuing education and to delaying marriage and childbearing until an appropriate time. All these factors have been found to be protective of children at risk, in largescale studies based in ordinary communities. In addition, there are accounts of rare intervention programs that start in the preschool period and either maintain their hold over certain children (as in Ramey & Campbell, 1987) or initiate a sequence of positive, ongoing events, as in the well-publicized Perry Preschool Program (Bermeta-Clement, Schweinhart, Barnett, Epstein, & Weikart, 1984; Schweinhart & Weikart, 1980, 1981). Several factors mitigate against positive personal development: 1. Temperamental irritability and lack of sociability, often combined with some degree of mental retardation. Infants growing up in large, chaotic, discordant families may not have the opportunity to gain an understanding of social cause and effect, nor to develop planning ability or knowledge of the desirability of delayed gratification and an internal locus of control. 2. Lack of emotional security and strong affectional ties with any one person, be it a sibling, grandmother, neighbor, or other significant individual.

Children who lack these supports are likely to be rated troublesome by teachers and later by peers in school, to be unpopular with peers, and to seek attention and emotional support in socially undesirable ways. If they are unlucky they will find themselves in schools that lack the necessary tranquility and academic press to enable them to learn, returning to large, affectionless families that are unable to exercise kind but firm control. They may grow up in a social situation characterized by poverty and lack of hope, from which they lack the personal resources to escape. Problems of inner cities give rise to what Haggerty, Roughman, and Pless (1973, followed by Baumeister (Baumeister & Kupstas, 1990), term the new morbidity. This concept is not entirely new; it is the considerable increase in earlier well-known hazards that has prompted the evolution of the term. It includes behavioral and school problems among children and adolescents, environmental risks, drug and alcohol abuse, accidents, violence, adolescent pregnancy, psychiatric disorders, and family disruption (Baumeister & Kupstas, 1990, pp. 50-51). The theme addressed in this article is one that is beginning to attract attention from some unexpected quarters. The late John Bowlby, who in 1951 was primarily responsible for reviving the belief that the life path was predetermined in the very early years, wrote in 1988 that

Ann M . Clarke and Alan D . B . Clarke

154

the central task . . . is to study the endless interactions of internal and external (factors), and how the one is influencing the other not only during childhood but during adolescence and adult life as well. . . . Present knowledge requires that a theory of developmental pathways should replace theories that invoke specific phases of development in which it is postulated that a person may become fixated and/or to which he may regress (pp. 1-2).

It may be also worth recording an example of psychoanalytic commentary and surprise at findings like these, which in its original, orthodox view should not be possible. Moskovitz (1985) followed up a small number of child survivors of the Holocaust who had been brought to England at the end of the war. Noting that in the late 1970s the picture often presented suggested a grim outcome of irreparably damaged people “what (now) became apparent . . . was the wide range of adaptation when there was theoretically no reason to expect to see anything positive. . . . Many made adaptations that are not only impressive but inspiring” (p. 407).

VII.

CONCLUSIONS

Our thinking about the effects of early social experiences started in the 1950s in a large institution for the mentally retarded in which we were surprised to observe positive cognitive and social changes in adolescents and young adults who had previously experienced exceptional adversity. Explanations were sought and, at first, tentative hypotheses raised. The results of our early studies have been shown to generalize widely and to be relevant to other socially disadvantaged children in whom, for a variety of reasons, intellectual or emotional underfunctioning has been manifested. The results suggest that one cannot necessarily take at face value, and as of predictive accuracy, personal levels measured in the context of seriously adverse past or present circumstances. Rutter (1989) points out that “chain effects” during development are common. “Life transitions have to be considered both as end products of past processes and as instigators of future ones . . . as both independent and dependent variables” (p. 46). Similarly, we see development as a series of linkages in which characteristics in each period have a probability of linking with those in another particular period. But such probabilities are not certainties, and deflections, for good or ill, are possible, but always within limits imposed by genetic, constitutional, and social trajectories. Although increasing age appears to diminish the possibility of alterations in the life path, the point where no modification is possible may be extreme old age. ACKNOWLEDGMENTS This article is based on a joint lecture delivered in October 1990 at a meeting organized by the National Institute of Child Health and Human Development (NICHHD), held at Birmingham, Ala-

MODIFIABILITY OF THE HUMAN LIFE PATH

155

bama. The authors are very grateful to Dr. Phillip Nelson, Chief, Laboratory of Developmental Neurobiology, NICHHD, National Institutes of Health, Bethesda, Maryland, and to Dr. Craig Ramey, Co-director of the new Civilian Research Center, Birmingham, Alabama, for their generous invitations to attend and contribute to the conference on “Environmental Determinants on Nervous System Development .”

REFERENCES Baumeister, A,, & Kupstas, F. (1990). The new morbidity; Implications for prevention and amelioration. In P. L. C. Evans & A. D. B. Clarke (Eds.), Combatting mental handicap (pp. 46-72). Bicester: AB Academic. Bell, R. Q. (1968). A reinterpretation of effects in studies of socialization. Psychological Review, 75, 81-95.

Benueta-Clement, I. R., Schweinhart, L. J., Barnett, W. S., Epstein, A. S.,& Weikart, D. P. ( 1984). Changed lives: The effects of the Perry Preschool Program on youths through age I9 (Monographs of the HighlScope Educational Research Foundation). Ypsilanti, MI: HighlScope Press. Bowlby, J. (1951). Maternal care and mental health. Geneva: World Health Organization. Bowlby, J. (1988). Developmental psychiatry comes of age. American Journal of Psychiatry, 145, 1-10,

Brim, O . , & Kagan, J. (Eds.) (1980). Constancy and change in human development. Cambridge, MA: Harvard Educational Press. Brown, M . , & Madge, N. (1982). Despite rhe weyare state. London: Heinernann. Brown, R. I. (1972). Cognitive changes in the adolescent slow learner. Journal of Child Psychology and Psychiatry, 13, 183-193. Chess, S., & Thomas, A. (1984). Origins and evolution of behaviour disorders. New York: Bmnner/ Mazel. Clarke, A. D. B., & Clarke, A. M. (1953). How constant is the IQ? Lancet, 2 , 877-880. Clarke, A. D. B., & Clarke, A. M. (1954). Cognitive changes in the feebleminded. Brirish Journal of Psychology, 45, 173-179. Clarke, A. D. B., & Clarke, A. M. (1959). Recovery from the effects of deprivation. Acta Psychologica, 16, 137-144. Clarke, A. D. B., & Clarke, A. M. (1960). Some recent advances in the study of deprivation. Journal of Child Psychology and Psychiatry, I , 26-36. Clarke, A. D. B., & Clarke, A. M. (1984). Constancy and change in the growth of human characteristics. Journal of Child Psychology and Psychiatry, 25, 191-210. Clarke, A. D. B., Clarke, A. M.. & Reiman, S. (1958). Cognitive and social changes in the feebleminded: Three further studies. British Journal of Psychology, 49. 144- 157. Clarke, A. M. (1982). Developmental discontinuities: An approach to assessing their nahm. In L. A. Bond & J. M. Joffe (Eds.), Facilitating infant and early childhood development (pp. 58-79). Hanover, NH: University Press of New England. Clarke, A. M., & Clarke, A. D. B. (Eds.) (1976). Early experience: Myth and evidence. London: Open Books/New York: Free Press. Clarke, A. M., & Clarke, A. D. B. (1986). Thirty years of child psychology: A selective review. Journal of Child Psychology and Psychiatry, 27, 719-759. Clarke, A. M., Clarke, A. D. B., & Berg, J. M. (Eds.) (1985). Mental deficiency: The changing outlook (4th ed.). London: Methuen. Davis, K. (1947). Final note on a case of extreme isolation. American Journal of Sociology, 45. 554-565.

156

Ann M . Clarke and Alan D . B . Clarke

Essen, J., & Wedge, P. (1982). Continuities in childhood disadvantages. London: Heinemann. Esser, G., Schmidt, M. H., & Woemer. W. (1990). Epidemiology and course of psychiatric disorder in school-age children-Results of a longitudinal study. Journal of Child Psychology and Psychiatry, 31, 243-263. Haggerty, R. J., Roughman, K. J., & Pless, 1. V. (1975). Child health and community. New York: Wiley. Harlow, H. F. (1963). The maternal affectional system. In B. M. Foss (Ed.), Determinants ofinfant behavior (pp. 3-33). London: Methuen. Kadushin, A. (1970). Adopting older children. New York: Columbia University Press. Kagan, J., & Klein, R. E. (1973). Cross-cultural perspectives on early development. American Psychologist. 28, 947-961. Koluchova, I. (1972). Severe deprivation in twins: A case study. Journal of Child Psychology and Psychiatry, 13, 107-1 14. Koluchova, J. (1976). A report on the further development of twins after severe and prolonged deprivation. In A. M. Clarke & A. D. 9. Clarke (Eds.), Early experience; Myth and evidence (pp. 56-66). London: Open Books/New York: Free Press. Kolvin, I . , Miller, F. J. W., Scott, D. Mcl., Gatzanis, S.R. M., & Fleeting, M. (1990). Continuities of deprivation? The Newcastle lo00 Family Study. Aldershot: Gower House. Levine, S. (1960). Stimulation in infancy. Scientific American, 202, 80-86. Moskovitz, S . (1985). Longitudinal follow-up of child survivors of the Holocaust. Journal of the American Academy of Child Psychiatry. 24, 401-407. Novak, M. A , , & Harlow, H. F. (1975). Social recovery of monkeys isolated for the first year of life: I . Rehabilitation and therapy. Developmental Psychology, 1 1 , 453-465. Pilling, D. ( 1990). Escape from disadvantage. London: Falmer Press. Quinton, D., & Rutter, M. (1984a). Parents with children in care: I. Current circumstances and parenting. Journal of Child Psychology and Psychiatry. 25, 21 1-229. Quinton, D., & Rutter, M. (1984b). Parents with children in care: 11. Intergenerational continuities. Journal of Child Psychology and Psychiatry, 25, 231-250. Quinton, D., Rutter, M., Dowdney, L., Liddle, C., Mrazek, D., & Skuse, D. (1982). Childhood experience andporenting behavior. Final Report to the Social Science Research Council, U.K. Ramey, C. T., & Car.ipbel1, F. A. (1987). The Carolina Abecedarian Project: An educational experiment concerning human malleability. In J. J. Gallagher & C. T. Ramey (Eds.), The malleability of children. Baltimore, MD: Paul H. Brookes. Richardson, S. A. (1985). Epidemiology. In A. M. Clarke, A. D. B. Clarke, & J. M. Berg (Eds.), Mental deficiency: The changing outlook (4th ed.). London: Methuen. Roswell Harris, D. (1958). Some aspects of cognitive and personality test changes in a group of 100 feebleminded young men. Unpublished M.A. thesis, University of Reading. Rutter, M. (1989). Pathways from childhood to adult life. Journal of Child Psychology and Psychiutry, 30, 23-51. Rutter, M., & Madge, N. (1976). Cycles of disadvantage. London: Heinemann. Rutter, M., Quinton, D., & Liddle, C. (1983). Parenting in two generations: Looking backwards and looking forwards. In N. Madge (Ed.), Families at risk (pp. 60-98). London: Heinemann. Rutter, M., & Sandberg, S. (1985). Epidemiology of child psychiatric disorder: Methodological issues and some substantive findings. Child Psychiatry and Human Development. 15, 209-233. Rutter, M., Yule, B., Quinton, D., Rowlands, O., Yule, W., & Berger, M. (1975). Attainment and adjustment in two geographical areas: 111. Some factors accounting for area differences. British Journal of Psychiatry, 126, 520-533. Scam, S . , & McCartney, K. (1983). How people make their own environments: A theory of genotype-environment effects. Child Development, 5 4 , 424-435.

MODIFIABILITY OF THE HUMAN LIFE PATH

157

Scarr-Salapatek. S. (1976). An evolutionary perspective on infant intelligence: Species patterns and individual variations. In M. Lewis (Ed.), Origins of intelligence (pp. 165-197). New York: Plenum. Schweinhart, L. J., & Weikart, D. P. (1980). Young children grow up: The effects of the Perry Preschool Program on youths through age 15 (Monographs of the High/Scope Educational Research Foundation). Ypsilanti, MI: HighlScope Press. Schweinhart, L. J., & Weikart, D. P. (1981). Perry preschool effects nine years later: What do they mean? In M. J. Bejab, H. C. Haywood, & H. Garber (Eds.), Psychosocial influences on retarded development (Vol. 2, pp. 113-125). Baltimore, MD: University Park Press. Scott, J. P. (1963). The process of primaly socialization in canine and human infants. Monographs of the Society for Research in Child Development, 28, 1-47. Scott, J. P. (1968). Early experience and the organization of behavior. Belmont, CA: Brooks/Cole. Skeels, H. M. (1966). Adult status of children with contrasting early life histories. Monographs of the Society for Research in Child Development, 31(3), No. 105, 1-65. Skeels, H. M., & Dye, H. B. (1939). A study of the effects of differential stimulation on mentally retarded children. Proceedings of the American Association on Mental Deficiency, 44, 1 14- 136. Skuse, D. (1984). Extreme deprivation in early childhood: 11. Theoretical issues and a comparative review. Journal of Child Psychology and Psychiatry, 25, 543-572. Suomi, S . J . , & Harlow, H. F. (1972). Social rehabilitation of isolate-reared monkeys. Developmental Psychology. 6 , 487-496. Svendsen, D. (1982). Changes in IQ, environmental and individual factors: A follow-up of educable mentally retarded children. Journal of Child Psychology and Psychiatry. 23, 69-74. Svendsen, D. (1983). Factors related to changes in IQ: A follow-up study of former slow learners. Journal of Child Psychology and Psychiatry, 24, 405-413. Thomas, A., Chess, S., & Birch, H. G. (1968). Temperament and behavior disorders. New York: New York University Press. Thompson, A. (1986). Adam-a severely deprived Colombian orphan. Journal of Child Psychology & Psychiatry, 27, 689-695. Tonge, W. L., Lunn, J. E., Greathead, M., McLaren, S., & Bosanko, C. (1983). Generations of problem families in Shefield. In N. Madge (Ed.), Families at risk (pp. 37-59). London: Heinemann. Von Knomng, A.-L., Andersson, O., & Magnusson, D. (1987). Psychiatric care and course of psychiatric disorders from childhood to early adulthood in a representative sample. Journal of Child Psychology and Psychiatry, 28, 324-341. Waddington, C. H. (1957). The strategy of genes. London: Allen & Unwin. Waddington, C. H. (1966). Principles of development and differentiation. New York: Macmillan. Waddington, C. H. (1971). Concepts of development. In E. Tobach, L. R. Aronson, & E. Shaw (Eds.), The biopsychology of development (pp. 17-23). New York: Academic Press. Wedge, P., & Prosser, H. (1973). Born to fail? London: Arrow Books. Werner, E. E. (1985). Stress and protective factors in children's lives. In A. R. Nicol (Ed.), Studies in child psychology and psychiatry (pp. 335-355). London: Wiley. Werner, E. E. (1989). High-risk children in young adulthood: A longitudinal study from birth to 32 years. American Journal of Orthopsychiatry. 59, 72-81, Werner, E. E., & Smith, R . S . (1982). Vulnerable but invincible: A longitudinal study of resilient children and youth. New York: McGraw-Hill. Wilson, R . S. (1985). The Louisville Twin Study: Developmental synchronies in behavior. Child Development, 54, 298-316.

This Page Intentionally Left Blank

Unraveling the “New Morbidity”: Adolescent Parenting and Developmental Delays JOHN G. BORKOWSKI, THOMAS L. WHITMAN, ANNE WURTZ PASSINO, ELIZABETH A. RELLINGER, KRISTEN SOMMER, AND DEBORAH KEOGH DEPARTMENT OF PSYCHOLOGY UNIVERSITY OF NOTRE DAME NOTRE DAME, INDIANA 46556

KERl WEED DEPARTMENT OF PSYCHOLOGY UNIVERSITY OF SOUTH CAROLINA-AIKEN AIKEN, SOUTH CAROLINA 29801

I.

INTRODUCTION

Dramatic changes during the past two decades in employment patterns, family composition, and acceptable patterns of behavior have altered the character of American society. Most importantly, recent social and economic changes have magnified the extent and depth of poverty across America, increasing the distance between the “haves” and the “have nots.” As a consequence of insidious changes in the fabric of our society, the next generation of children will be placed at considerable risk as they progress through our educational systems. Already, there are frequent complaints from kindergarten teachers about student motivation, short attention spans, and diminished self-regulatory skills. Although their precise causes cannot be identified, these ominous signs were anticipated by Baumeister (1988) in his profound discussion of the “new morbidity.” The concept of “new morbidity” identifies complex, multivariate linkages between poverty and a host of related developmental delays. It emphasizes the INTERNATIONAL REVIEW OF RESEARCH IN MENTAL RETARDATION, Vol. I8

I59

Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

1 60

John G . Borkowski et al.

predisposing, catalytic, resource, proximal, and outcome factors associated with poverty, including drugs, alcohol, adolescent pregnancy, low birthweight, environmental risks, and accidents. Although these problems are not new in our society, they have multiplied in recent years and are occurring in more complex perturbations. Because of the complexity, multiple causality, and reciprocality underlying poverty, its causes, and its consequences, it is no small task to unravel the web of factors that surround each new morbidity phenomenon. The aim of this article is to analyze a specific instance of “new morbidity”-the case of adolescent parenting. Our goal is to develop a deeper understanding of the causes of developmental delays in children who have teen mothers and to link the social, emotional, and cognitive causes and correlates of mild mental retardation in this population within the framework of a single development model.

II.

ADOLESCENT PARENTING: HIGH RISKS FOR MOTHERS AND THEIR CHILDREN

The last two decades have witnessed the emergence of several longitudinal studies focusing on adolescent mothers and their children that give definitional clarity to the “new morbidity” concept (Baumeister, 1988). These projects have pinpointed some of the predisposing and catalytic variables present in today’s complex society that often interact to produce developmental disabilities (Furstenberg, Brooks-Gunn, & Morgan, 1987; Osofsky & Eberhart-Wright, 1988). Predisposing variables include the use of drugs, alcohol, and tobacco; poor nutrition; low birth weight; inadequate postnatal stimulation; low maternal intellectual ability; lack of knowledge about child development; unstable social supports; ill-defined personal goals; and tendencies toward personal instability. Because these factors occur in combination in the lives of teen mothers, they often produce significant developmental delays in their children. Recent data indicate that seven of every ten teenagers are sexually active and, among adolescent females, one of every ten becomes pregnant in the United States. This means that nearly one million adolescents become pregnant in the United States each year. Over 29% of all first births in this country occur to adolescents. These figures represent one of the highest fertility rates for this age group among the developed countries (Newberger, Melnicore, & Newberger, 1986). Of these, approximately 450,000will have abortions and nearly 500,000 will give birth. Of further importance is the fact that about 90% of live births to teen mothers will be raised by biological parent(s) or grandparents. Although the overall birth rate has decreased since 1980, the absolute number of adolescent parents remains extremely high. Two increasing trends are of particular concern: ( I ) the birth rates among young adolescents, ages 12 to 15,

UNRAVELING THE “NEW MORBIDITY”

161

and (2) the number of out-of-wedlock births. Children of very young mothers appear to be at greatest risk for a variety of biological, cognitive, and social outcomes (Ventura & Hendershot, 1984). Additionally, adolescent mothers typically confront the demanding task of parenting with little or unstable support from a partner, and single mothers often find inconsistent support from family or community professionals (Unger & Wandersman, 1985). Hence, it is not surprising that 25% of Caucasian infants and 80% of black infants are reported to experience significant stress as a consequence of being reared in single-parent homes (National Survey of Family Growth, 1982). For adolescent mothers themselves, the social consequences of parenthood include attenuated educational achievement (Russell, 1980) and related problems of underemployment and restricted lifetime earnings. As a consequence, many adolescent mothers are forced to rely on public aid. Currently, the United States invests over $16 billion annually to support adolescents and their infants. Other potential adverse consequences for the teenage mother, besides poverty, are marital disruption and the potential for inadequate or abusive parenting. A.

Adolescent Parenting and Childrearing

Because young mothers encounter their parenting roles simultaneously with other developmental tasks typical of adolescence (Bacon, 1974; Klein & Cordell, 1987; Levenson, Atkinson, Hale, & Hollier, 1978), they are likely faced with greater life challenges than their adult counterparts (see Elster, McAnarney, & & Lamb, 1983; Lancaster & Hamburg, 1986, for reviews). An unwanted early pregnancy causes high levels of stress, threatens the adolescent’s need for privacy, and delays her process of identity formation (Black & DeBlassie, 1985; Schaeffer & Pine, 1972). In addition, the struggle between her need for independence and her forced dependence on adults causes anxiety (Protinsky, Sporakowski, & Atkins, 1982; Schneider, 1982). These developmental conflicts are in direct opposition to the qualities required for parental competence: psychological stability, ego control, attentiveness, and stable relationships within established households. In short, the adolescent mother is likely to experience her offtime transition to parenting and the stressors associated with this transition at a time when she is developmentally unprepared to meet these demands and to provide optimal patterns of parenting for her child. The resulting heightened levels of stress may have a detrimental influence on her own development, as well as that of her child (Feldman & Feldman, 1975; Russell, 1980; Ventura, 1980).

It also appears that because of their age and related inexperience, adolescents are less well equipped to fulfill what is likely a more demanding parenting role than that of adult parents. Many adolescent mothers have a limited understanding of normal child development and a more restricted repertoire of parenting

162

John G . Borkowski et al.

skills than their adult counterparts (Field, 1979). Several researchers have described adolescents as being inclined toward the use of physical punishment with their infants, particularly when they fail to meet developmental expectations (cf. Field, 1980). All of these factors predispose children of teen mothers to elevated risk in their early development. It is the exposition of the potentially causal factors that can produce delays in children of adolescent mothers that represents the task of this article. We attempt to show how personal, social, and emotional factors, associated with teen parenting, interact to impede cognitive development in early childhood, thus setting the stage for a host of later development problems, such as mild mental retardation and general learning disabilities. 1. STRESS AND PROBLEMATIC PARENTING PRACTICE

To date, only a few studies have actually compared the stress levels of adolescent and adult mothers. Evidence that motherhood may be more stressful for adolescent mothers than for their adult counterparts has been presented by Brown, Adams, and Kellan (198 1). In this community-based longitudinal study, 828 women were interviewed when their children were 6 years old and, again, 10 years later. The teen mothers reported having more depressed feelings at the first interview in comparison with the older mothers, and also feeling more stress and distress in later years. In a more recent study, Wurtz (1989) compared the levels of stress (as reflected on the Parenting Stress Index) for adolescent and adult mothers when their infants were 6 months old. Results indicated that young mothers perceived their parenting roles as significantly more stressful than adult mothers. In addition to experiencing higher levels of parenting stress, adolescent mothers often display problematic parenting behaviors. A review of the literature suggests that in comparison to adult mothers, adolescent mothers are not as effective in their parent-infant interactions. They are less likely to interact verbally with their infants (Brooks-Gunn & Furstenberg, 1986; Epstein, 1980; McLaughlin, Sandler, Sherrod, Vietze, & O’Conner, 1979; Roosa, Fitzgerald, & Carlson, 1982), are often insensitive (i.e., less contingent, consistent, and appropriate) in their reactions to their infants’ cues, and are more likely to use physical rather than verbal communication (Elster et al., 1983). Schilmoeller and Baranowski ( 1985) found that adolescent mothers were significantly less responsible, less stimulating, more restrictive, and more punitive than older mothers. In a recent comparative study of adolescent and adult mothers, O’Dell (1990) found that adolescent mothers were less positive, flexible, and verbal with their infants than adult mothers. Younger mothers also gave less appropriate direction and encouragement to their infants, were more controlling in their interactions, and were less able to respond appropriately to their infants’ moods. Although several studies suggest that younger and older mothers do not parent

UNRAVELING THE “NEW MORBIDITY”

163

differently if factors such as race and socioeconomic status (SES) are taken into account (Philliber & Graham, 1981; Reis, Barbara-Stein, & Bennett, 1986), most research has found parenting differences between these groups even when these demographic variables are considered (Coll, Hoffman, & Oh, 1987; Jones, Green, & Krauss, 1980; Landy, 1981; McAnarney, Lawrence, Riccutti, Polley, & Szilagy, 1986; Stevens & Duffield, 1986; Teberg, Howell, & Wingert, 1983). These studies generally have found that adolescent mothers are less effective in their parenting behavior than older mothers. For example, results by Ragozin, Basham, Crnic, Greenberg, and Robinson (1982) indicated that when demographic (income, education, and parity) variables are controlled, older mothers exhibit more positive parenting behavior and report more satisfaction with, and greater time commitment to, their maternal roles. In another investigation of adolescent parenting, adult and adolescent low-income black mothers were compared with respect to parenting behavior (Stevens & Duffield, 1986). Results revealed that younger mothers provided lower levels of support for learning in the home environments, were less responsive, were more punitive, and were less able to promote the intellectual development of their infants. Researchers have also found that adolescent mothers exhibited less eye contact and smiling, as well as verbal and physical interactions (Teberg et al., 1983). They are often less understanding, cooperative, accessible, and sensitive in their interactions with their children (McAnamey et al., 1986). Additionally, parenting at a young age has been negatively associated with maternal praise and affection as well as positively associated with criticism and physical punishment (Conger, McCarty, Yang, Lahey, & Burgess, 1984).

2. IMPLICATIONS FOR DEVELOPMENT There is evidence that the problematic behaviors associated with adolescent parenting may, in turn, contribute to the developmental delays in cognitive, language, and social areas of infant development. For instance, verbal stimulation has been shown to be directly related to infants’ rates of vocalization and language acquisition and to children’s present and future levels of intelligence (Beckwith, 1976). Furthermore, parental sensitivity is considered by many to be a significant component of effective parenting, influencing children’s socialemotional development (Panzarine, 1988). Research has indicated that among adult mothers, parental sensitivity predicts secure infant-mother attachment (Ainsworth, Blehar, Waters, & Wall, 1978; Egeland & Farber, 1984), increased infant IQ scores (Beckwith, 1976), and present and future social competence (Lamb, 1982). On the basis of recent analyses of mothers from the Notre Dame Adolescent Parenting Project, and from informal contacts with other parenting projects, we estimate that the average IQ of teen mothers who elect to parent their children is around 85. Additionally, it is likely that the distribution is skewed such that

164

John G . Borkowski et al.

relatively few teen mothers are of average intelligence and many (perhaps 20%) have IQs below 70. Additionally, we suspect that many teen mothers, especially those who terminate schooling early or pursue a low-level degree program that does little to enhance functional, job-relevant skills, actually experience a decline in their intellectual abilities after the birth of their child. We hypothesize that these mothers, with borderline mental retardation, place their offspring at special risk for various developmental delays because they themselves may decline in intellectual ability and in other functional cognitive skills. Past theory and research suggest three factors that may have considerable impact on parenting behaviors and stress levels: the cognitive and emotional resources of the mother, the characteristics of her child, and her social support system. Whitman, Borkowski, Schellenbach, and Nath (1987) have proposed a multivariate model of adolescent parenting in which they hypothesize that the mother-infant interaction is directly influenced by each of these factors. It is hypothesized that the adolescent mother’s limited personal and social resources, in conjunction with the atypical characteristics of her young infant, can create parent-child interactional patterns that are problematic and may precipitate later developmental delays. In the next section, we review evidence for such delays and then outline a general model that provides a preliminary account of the determinants of adolescent parenting.

B.

Children of Adolescent Mothers

Children of adolescent mothers are at risk for a host of developmental problems. Without adequate prenatal care adolescent mothers often experience a large number of complications during delivery (Broman, 1981). Hence, their newborns are more likely to manifest developmental problems such as prematurity or low birth weight (Field, 1980). Even with adequate prenatal care and normal deliveries, children of teen mothers remain at risk for later development delays. For instance, data from the Perinatal Longitudinal Project of Broman, Nichols, Shaughnessy, and Kennedy ( 1987), which analyzed the outcomes of over 35,000 high-risk births, showed a higher percentage of mild retardation in the offspring of teen mothers versus adult mothers from similar SES and educational backgrounds. The actual percentage of retardation-attributable apparently to nonorganic causes--may be up to three times higher among teens. Similarly striking data have come from the Baltimore project of Brooks-Gunn and Furstenberg (1986). They found that up to one-half of the children of teen mothers showed behavioral and school-achievement problems, especially during their early teen years. These two important, longitudinal projects suggest that the social and psychological consequences of early pregnancy and teenage parenting are considerable, particularly for the children themselves and the society that helps guide their future development.

UNRAVELING THE “NEW MORBIDITY”

165

Belmont, Cohen, Dryfoos, Stein, and Sayac (1981) attempted to separate the effects of economic, social, and demographic disadvantages and maternal immaturity (mother’s age) on children’s cognitive development (WISC). The data were drawn from two large samples of the Health Examination Survey and from the Collaborative Perinatal Project, involving 12 medical centers and over 26,000 subjects. Children of mothers in the younger maternal age groups (519) had significantly lower IQ scores than children of mothers who were age 20 or older. When the effects of social variables such as family income, parental education, family size, and one- or two-parent families were controlled, maternal age continued to show a small, but significant, relationship with the children’s cognitive abilities. Mean IQ scores for several samples ranged from 91.36 to 94.41 for the children of young teens (14-17 years old), from 95.31 to 97.87 for children of older teens, and from 97.81 to 101.23 for children of adults. Dramatic developmental changes for children living in poverty have been described in more recent and extensive longitudinal studies. Both the North Carolina Abecedarian and the Milwaukee projects have found that cognitive development declines over time for the children of low-IQ adult mothers. For instance, children in the control group of the Milwaukee Project declined up to 20 IQ points by age 10, provided they did not receive early intervention. Children of very low IQ mothers in the Abecedarian Project showed a dramatic decline in IQ, up to 25 points, by age 10. In both projects, the intervention and control children were very similar on early measures of cognitive development, with most children appearing to be functioning within the normal range of development. In the Abecedarian Project, children in the control condition were 3.5 times as likely to have cognitive-educational deficits (i.e., IQs below 85). Older siblings of the target sample were also evaluated on the Wechsler scale (Ramey & Campbell, 1981); the mean IQ for older siblings was found to be 87.4. These findings support the general conclusion that IQs continue to decline for children in poverty who do not receive early intervention. We propose three reasons to explain why adolescent mothers tend to have children who experience cognitive-academic problems during early and middle childhood: (1) direct, genetic transmission of retardation; (2) poor quality of maternal teaching style, emotional instability, and impoverished home environments, factors sometimes independent of maternal IQ; and (3) genetic and environmental factors that interact to determine a mother’s parenting skills (Plomin, Loehlin, & DeFries, 1985). Thus, the development of cognition and intelligence in children of teenage mothers may be affected directly via genetic factors and indirectly through inadequate parental cognitive and emotional readiness, which combine to produce unstable and unstimulating childrearing environments. Some of these relationships, especially those dealing with environmental, social, and personal factors, are captured in a model of parenting and infant development, as described in the next section.

166

John G . Borkowski et al.

psychosocial Adjustment

FIG. I . .A model of adolescent parenting.

C. A Model of Adolescent Parenting Our multidimensional model of adolescent parenting (Whitman et al., 1987) contains some of the constructs we believe are needed to understand the causes and consequences of teenage childrearing (see Fig. 1). In addition to the parent, child, and social support constructs contained in Belsky’s (1984) general model of parenting, the interplay of three unique maternal constructs (cognitive readiness, learning ability, and maternal health) is highlighted in our model. Moreover, an enhanced role is assigned to the adolescent’s social support system. Thus, our model depicts relationships among maternal constructs, social support, and child characteristics, as well as illustrates how these constructs might conjointly affect parenting and subsequent child development. It is our view that a more complete understanding of mild mental retardation will flow from an analysis of these predisposing, catalytic factors associated with a single instance of new morbidity-adolescent parenting. Although the model sketched in Fig. 1 postulates numerous direct and indirect influences on the parenting behavior of adolescents during the first 6 months of life, these influences can be summarized as follows: To parent effectively, adolescent mothers must be cognitively and emotionally prepared. Their cognitive readiness rests on the formal and informal education they receive from their support systems and their ability to assimilate and use this information in specific

UNRAVELING THE “NEW MORBIDITY”

167

parenting situations. In addition, a mother’s ability to deal emotionally with the stresses associated with parenting depends on whether her child displays biological characteristics that complicate the parenting task, the type of physical and emotional assistance received from the social support system, and her own personal coping resources. By the end of the first year of life, the quality of infant cognitive (e.g., attentional flexibility) and emotional (e.g., attachment) development is likely related to the interaction of the factors represented in Fig. 1. Rather than focusing on the overall impact of age and socioeconomic status of the adolescent parent, the model suggests alternative, more specific hypotheses for explaining individual variations in parenting practices and their consequences for child development. These hypotheses, and their implications for developmental delays in children of teen mothers, are sketched in the next sections.

111.

INFLUENCE OF ENVIRONMENTAL AND PERSONAL FACTORS

In general, recent research has found that adolescent mothers exhibit less effective parenting behaviors than adult mothers. We believe that many adolescent mothers are cognitively and emotionally unprepared for the time-consuming and challenging task of parenting, have impaired learning ability and an unstable social support system, and are confronted by challenging infants. In this scenario, mothers are placed at elevated risk for problematic parenting and, in turn, place their children at risk for a variety of developmental delays, including mild retardation. Thus, the phenomenon of adolescent parenting provides an opportunity to create a complex, dynamic model of development that has the potential for explicating the concept of new morbidity in greater detail and with more precision. Unfortunately, little empirical work has attempted to identify how the psychological resources of the mother, her child’s temperament, and the social support factors in the environment place the adolescent at risk for poor parenting or to explain the wide range of individual differences in adolescent parenting (Panzarine, 1989). An increasing number of investigators, however, are recognizing the need for multivariate conceptual and methodological approaches to understanding this phenomenon (Phipps-Yonas, 1980; Reis & Hem, 1987; Whitman et al., 1987). We believe that the children of teen mothers are not only at significant risk because of the low IQs of their mothers but also because of a combination of complex biological, social, and psychological factors present in their lives. The next section describes the Notre Dame Adolescent Parenting Project, discusses evidence we have found in support of the contention that children of adolescent mothers display more developmental problems than children of adult mothers,

John G . Borkowski et al.

168

and outlines the various causes of these delays. In the remaining sections of the article, the effects of maternal intelligence, cognitive readiness, personal adjustment, child temperament, and social supports on adolescent parenting and early infant development are discussed in terms of our model. To explore later development we then introduce additional factors (such as infant attachment) that shape developmental trajectories during early childhood, especially as they interact with the major constructs in our model.

A.

Adolescents and Their Children: The Notre Dame Parenting Project

Adolescent mothers who are participating in the Notre Dame Parenting Project have been recruited over the past 4 years. Presently, 194 adolescents have been interviewed during pregnancy. Subsequently, 117 of this sample were seen when their children were 6 months, 54 at I year, and 32 at 3 years. Data are still being collected at these three age periods. In addition, 42 pregnant adults were interviewed during their pregnancy and 26 were seen again when their children were 6 months of age. Finally, 62 nonpregnant adolescents were interviewed within the same basic time frame as the pregnant adolescent sample. For all major variables (IQ, personality, SES, etc.), no significant differences have emerged between those subjects who have remained in the project and those who dropped out (approximately a 40% attrition rate at age 3). 1. MATERNAL CHARACTERISTICS The adolescent sample in the Notre Dame Parenting Project reflects many important characteristics of the population of teen mothers in the United States. Their average age at childbirth is 17.2 years. The sample is 38% white, 57% black, and 5% Hispanic. Only 8% were married at the time of childbirth. The teen mothers in our sample have completed an average of 10.6 years of school and have an average estimated IQ of 84. In a recent analysis of data collected when some of the children reached 18 months, only 25% of the teen mothers had completed high school; 37% percent were still attending school; 38% had dropped out. Only 12% were married or involved in a stable relationship. Thirtyseven percent were employed (most on a part-time basis), and 57% were receiving agency support. Finally, 37% had a second child, 6% had a miscarriage or stillbirth, and 14% were pregnant for the second time by the time their oldest child was 18 months of age. In short, adolescent mothers in the Notre Dame sample appear highly similar in terms of major personal-demographic variables to the teen mother population at large. 2. INFANT CHARACTERISTICS Infants in our adolescent sample do not differ from the infants of adult mothers. That is, very few children from either group were below 2500 g at birth,

UNRAVELING THE “NEW MORBIDITY”

169

there were no differences on APGAR scores at 1 or 5 minutes between groups, and there were no group differences on Brazelton scores. Similarly, at 6 months of age there were no significant differences in cognitive or motoric skills (as measured by the Bayley) between children of adolescent and adult mothers. The mean standardized score on the Bayley scale of Mental Development for children of adolescent mothers was 105. Motor development at 6 months was similarly unaffected by parental age, with the mean score on the Bayley scale of Motor Development being 108. These findings are consistent with Ramey and Ramey’s (1990) study of socioeconomically and biologically at risk children, which also did not find early developmental deficits in similar types of children. At 12 months of age, the Bayley Scales of Infant Development were again administered to the children of adolescent mothers. Scores on the Mental subtest were somewhat lower than they had been at 6 months, with an average score of 99; motor development scores remained high, with a mean scaled score of 1 1 1 . When children of adolescent mothers reached 3 years of age, the StanfordBinet test was administered. Children’s intellectual functioning showed a marked decline from 1 year, with the average score being 85. Moreover, they showed significant delays in receptive language skills, as assessed by the Peabody Picture Vocabulary Test-Revised (PPVT-R), with an average score of 74 (no child scored above 100). The Vineland Adaptive Behavior Scales were administered to teen mothers when their children were 3 years of age, to assess their children’s communication, daily living, socialization, and motor skills. Children of adolescent mothers had an average score of 100 in the communication domain, 102 on daily living skills, 89 on socialization, and 88 on the motor scale. The overall behavioral profile average score was 93, indicating that, as a whole, adolescent mothers did not report significant delays in their children’s acquisition of adaptive behaviors. It is important to note that many mothers may have overreported their child’s competencies. For example, in the communication domain the mothers rated their children as having normal communication skills, whereas interviewers frequently observed (and subsequent assessments confirmed) that these same children had marked receptive language deficits. In summary, as we examine the life course of children of adolescent mothers, we have observed that problems in the cognitive, verbal, social, and motor domains begin to emerge by the age of 3, even though there were no detectable problems at birth or during early infancy. That is, declines in mental functioning occur as these children develop, their receptive language skills are delayed, and they are reported to have developed fewer motor and social skills than would be expected in children of adult mothers. If this trend continues, as past research suggests it will, it is likely that children of adolescent mothers will experience significant school failures, in middle school or before, as a result of limited cognitive and social-interactional skills. We turn now to an analysis of the possible causal mechanisms that might give rise to early-appearing, developmental

John G. Borkowski et al.

170

delays in children of teen mothers, drawing from the Notre Dame Parenting Project as well as other data bases. B.

Maternal Intelligence

The Whitman et al. (1987) model of adolescent parenting and child development suggests three reasons why children of low-IQ teen mothers are themselves at risk for cognitive and intellectual delay: (1) direct genetic transmission of retardation, (2) inadequate stimulation and home environments that are relatively independent of heritability, and (3) genetic and environmental factors that interact to determine the quality of the caregiving environment. A strong genetic transmission model suggests that the high proportion of mild mental retardation for children of adolescent mothers results from the polygenic inheritance of intelligence. From this view, the importance of genes on intelligence increases across the life span; that is, genetic effects account for only approximately 15% of the variance in infant mental tests to around 50% of the variance in IQ tests in adolescence and adulthood (Plomin, 1990). As there is some evidence that individuals scoring in the borderline to low-average range of intelligence are overrepresented in the population of adolescent mothers (Keogh, 1991), it follows that mild mental retardation among their children may be directly passed from mother to child, to a large extent, via genetic transmission. In contrast, a strict environmental model proposes a “deprivation” view of development in accounting for the greater incidence of developmental delays in children of teen mothers with low IQs. A number of investigations have been undertaken to determine the environmental factors that enhance or inhibit cognitive development. Environmental measures such as warmth and affection toward the child, contingent responsivity, sensitivity to the child’s needs and interests, avoidance of restriction and punishment, and a well-organized physical environment have been shown repeatedly to be characteristics that promote intellectual development. From this perspective, developmental delays in children of young mothers may be a result of poor readiness for parenting and inadequate parenting skills which, in turn, produce impoverished childrearing environments. Although the importance of the child’s caregiving environment on cognitive and emotional development has been recognized, it seems clear that a simple environmental model of development (associated with low-IQ parents) fails to satisfactorily explain development. A more complex model, taking into account both environmental and genetic influences, appears necessary. The majority of studies that have investigated the relationship between caregiving environments and development have confounded the influence of genes and the environment (Plomin et al., 1985). On the basis of the Colorado Adoption Project, examining development in adoptive and nonadoptive families, Plomin and his colleagues suggested that, on the average, one-half of the en-

UNRAVELING THE “NEW MORBIDITY”

171

vironment-development relationship in biological families is genetically mediated. Parental characteristics also help to shape and create rearing environment; however, to the extent that parental characteristics are themselves genetically influenced, genes may also mediate the relationship between environmental measures and child development. In other words, intelligent parents create intellectually stimulating environments for their children: “When one observes that highly intelligent children have parents who read to them, discuss topics of interest and listen to their children’s opinions, these parental behaviors are not an arbitrary set of parenting skills applied willy-nilly to any child. Rather, they are a set of parenting skills used by intelligent parents toward their usually more intelligent children” (Scarr & Carter-Saltzman, 1983, p. 318). Data from the Notre Dame Parenting Project are generally in accord with hypotheses that relate maternal IQ to the quality of the caregiving environment and, in turn, associate both of these variables with the intelligence of the developing child. For instance, the correlation between prenatal maternal IQ and parenting at 6 months of age (as reflected in the teen mother’s knowledge of child development; her parenting style, e.g., reported use of punishment; and her attitudes about children and being a parent) was significant, r = .55, p C .001 (N = 74). Additionally, prenatal maternal IQs correlated with children’s Stanford-Binet IQs at 3 years of age (r = .52) as well as with receptive language as measured by the PPVT-R ( r = .49). These data are consistent with a direct linkage between maternal and child IQ and an indirect linkage between maternal IQ and child IQ via a mediating variable, “readiness” to be a successful parent. As this latter construct is unique to our model of parenting, we highlight its explanatory importance in the next section.

C.

Cognitive Readiness to Parent

Whitman et al. (1987) have suggested that it is the distinctive characteristics of adolescent mothers that place them at special risk for nonoptimal parenting and their children at risk for a variety of developmental delays. In contrast to more general models of parenting, such as that proposed by Belsky (1984), our model emphasizes the role of cognitive readiness to parent in predicting teenage parenting skills. We contend that to parent effectively, mothers should (1) be predisposed to becoming a parent; (2) understand how their children develop, especially in terms of the timing of developmental milestones; and (3) appreciate the importance of appropriate parenting practices. Because of their youth, adolescent mothers are less likely to accept their parenting roles. As a result of inexperience, their knowledge of children and of appropriate parenting practices is, in general, likely to be more limited than that of adult mothers. Whitman et al. (1987) pointed out that some teenage mothers may profit less from their life experiences

172

John G . Borkowski et al.

because of lower educational attainment, diminished intellectual ability, absence of social supports, and/or social-emotional immaturity, with the consequence that they are less cognitively ready to parent. More specifically, cognitive readiness to parent depends both on the information teen mothers receive from their formal and informal social support systems and on their ability to assimilate this information and subsequently use it in specific situations. To the extent that adolescent mothers are struggling with their own crises and searching for their personal identity, they are less likely to have interests in sustaining parenting or to have the time or capacity to consider fully their long-term parental responsibilities. We hypothesize that a deficit in cognitive readiness to parent predisposes teenage mothers to greater parenting stress as well as to less responsive parenting. From a research perspective, a review of the teenage parenting literature yields disparate conclusions concerning the preparedness of adolescent mothers and whether they are less equipped for parenting than adult mothers. One possible reason for seemingly discrepant conclusions is the fact that researchers define and evaluate “cognitive readiness for parenting” in narrow and often diverse ways. Whereas some studies have assessed maternal knowledge of child development (Vukelich & Kliman, 1985), others have examined attitudes toward the parenting role (de Lissovoy, 1973). Still others examine parenting knowledge and style (Parks & Smeriglio, 1983; Roosa et al., 1982). Another possible reason for the disparity in research findings relates to the samples employed, which have been idiosyncratic and generally small. Although a considerable amount of research has examined the cognitive characteristics of adolescent mothers, relatively little is known about whether these characteristics actually influence parenting behaviors. Several studies suggest a relationship between parental knowledge of child development and parenting style; however, the majority have employed only indirect measures of parenting, such as the HOME scale (Parks & Smeriglio, 1983). To gain a more complete understanding of adolescent parenting, the links between cognitive readiness and parenting behavior need to be examined in greater depth. This type of exploration can assist in the early identification of teen parents who are at special risk for maladaptive parenting as well as in the formulation of more effective parenting programs. In contrast to most previous research evaluating cognitive readiness, we employed a variety of cognitive readiness measures, including assessments of child development knowledge, parenting knowledge, and parenting attitudes; a more comprehensive assessment of parenting behavior; a larger sample size; and two comparison groups (adult and nonpregnant adolescent samples) (cf. Sommer and Rellinger, 1989). As described earlier, we selected the groups randomly, recognizing that each sample would likely have unique characteristics that reflected the population demographics of adult and adolescent mothers in the United

UNRAVELING THE “NEW MORBIDITY”

173

States. For example, demographic data indicate that adolescent mothers are more likely to be black and of lower socioeconomic status than adult mothers (National Survey of Family Growth Cycle 111, 1982). Our initial aim was to describe “naturally occurring differences” in cognitive readiness among representative groups of primiparous adolescent and adult mothers to infer population characteristics. A second goal was to assess the contribution of important sample characteristics, including SES, IQ, race, and education, to age-related differences in cognitive readiness. Finally, we identified relationships between cognitive readiness to parent and actual parenting in our adolescent sample, examining both the shared (with major demographic characteristics) and the unique contributions of cognitive readiness to parent (Sommer & Rellinger, 1989). All expectant mothers were initially seen during the last trimester of their pregnancy; the nonpregnant participants were seen at various intervals within the same time frame as the pregnant samples. For all participants, the three aforementioned measures of cognitive readiness for parenting were administered, as were multiple measures of other constructs in our model (see Fig. 1). Adolescent and adult mothers were seen again 6 months following the birth of their babies. At this time, the same three cognitive readiness measures were readministered to assess changes in parental knowledge and attitudes. In addition, a self-report measure of parenting stress was administered, and each mother and child was videotaped while interacting in a toy play situation. The videotape was subsequently rated via two assessment procedures, a global scale of maternal interactions (MIS) and a more molecular assessment of moment-to-moment reciprocal mother-child interactions (MBA). In accordance with our hypothesis, a series of analyses of prenatal and postnatal cognitive readiness indicated that the adolescent mothers were significantly less knowledgeable about child development, displayed a less desirable parenting style, and held more undesirable attitudes toward their parenting roles than the adult mothers. These findings are consistent with the existing literature, indicating that adolescent mothers possess less precise knowledge about development and more negative attitudes about parenting than adult mothers (Field, Widmayer, Stringer, & Ignatoff, 1980; Roosa & Vaughan, 1984). It was also predicted that age, serving as a proxy variable for educational background and social-emotional maturity, might differentially influence the ability of adolescent and adult mothers to benefit from their early childrearing experiences and, as a consequence, to enhance their readiness to parent. In contrast to expectations, neither adults nor adolescents showed significant changes in the three components of cognitive readiness from the prenatal to the postnatal assessment periods. As expected there were differences in the IQ, SES, race, and education of adult versus adolescent mothers. To determine whether these differences ac-

174

John G . Borkowski et al.

counted for differences in cognitive readiness between adolescent and adult mothers, the effects of personal-demographic variables were sequentially partialled out of all major analyses. Although age-related differences in maternal attitude, parenting style, and knowledge of child development were influenced by IQ and SES, age per se continued to be an explanatory factor, implying that age differences are not entirely attributable to important demographic factors; however, when the effects of all four personal-demographic variables (IQ, SES, education, and race) were controlled, age was no longer a significant explanatory factor in accounting for group differences in cognitive readiness. Adolescent mothers were under significantly greater stress than the adult mothers, with adolescents reporting more stress in both the Parent and Child Domains of the Parent Stress Index (244 versus 215 on the total PSI scores). On the Parenting Stress Inventory, scores above the 75th percentile are considered to be within the deviant range. Whereas 39 adolescent mothers were in this range, only 2 adult mothers scored this high. Adolescent mothers reported feeling more socially isolated, unsupported, and restricted by their parenting responsibilities. These results are consistent with past research and theory which suggest adolescent mothers experience greater stress than adult mothers (Brooks-Gunn & Furstenberg, 1986), and with the observation that adolescent motherhood represents an additional strenuous developmental challenge added onto the multiple demanding normative tasks routinely confronting adolescents (Ventura, 1980). That is, adolescent mothers must cope not only with the normal stressors of adolescence but also stressors associated with pregnancy and parenthood at a time when their personal and social resources are still developing. However, when IQ, SES, race, and education were controlled, adolescent and adult mothers did not differ significantly in their reported levels of stress. This finding suggests that age differences in maternal stress are mediated by demographic factors. We expected that adolescent mothers would also exhibit less optimal parenting skills than adult mothers, as assessed on micro-level behavioral frequency counts (the MBA scale) as well as on overall judgments of parenting competence (the MIS). The MBA assessed derived measures, such as maternal sensitivity and responsivity collected over brief (5-second) time intervals within a 15-minute toy play session. In contrast, the MIS, a 10-item Likert-type rating scale (measuring characteristics such as interactional orientation, flexibility, and rate of stimulation) was used by judges to assess overall maternal parenting style following their observation of the 15-minute tape of parent-child toy play. It is interesting that the more global rating of the quality of maternal interactions, the MIS, presented a different portrayal of adolescent-versus-adultmother’s parental functioning than the more molecular, behavioral instrument, the MBA. That is, consistent age-related differences in parenting competence were found only with the MIS.

UNRAVELING THE “NEW MORBIDITY”

175

Most importantly, the results supported the hypothesized link between an adolescent’s cognitive readiness and her parenting affect and behavior. Prenatal and postnatal maternal cognitive readiness measures were found to be significant predictors of parenting stress. When the effects of maternal IQ, SES, race, and education were statistically controlled, the results, although somewhat attenuated, still indicated the unique influence of cognitive readiness on maternal stress. A different and interesting pattern of results was found in examining the relationship between cognitive readiness and maternal interactional style. Although postnatal cognitive readiness predicted MIS scores even when SES and IQ were removed, the cognitive readiness-parenting style relationship was no longer significant when all demographic variables were removed. Overall these results are consistent with past theory (Whitman et al., 1987): Adolescent mothers who lack a rich and detailed knowledge about their children, who are less cognizant of positive appropriate parenting practices, and who espouse more negative attitudes about their parenting roles are more likely to experience stress in carrying out their parenting roles during the first 6 months of life and, relatedly, to display less competent parenting skills.

D.

Personal Adjustment and Parenting

Belsky’s (1984) model of adult parenting, as well as our own model of teen parenting, suggests the critical role of maternal adjustment in successful parenting. In general, research suggests that a significant percentage of adolescent mothers are psychologically maladjusted (Hamburg, 1986). Young women who become early childbearers are probably those adolescents least prepared in terms of personal adjustment and maturity. These findings are alarming given other research that indicates that an individual’s personal resources are directly related both to the amount of stress they experience and to their ability to cope with that stress. When early parenting occurs, the mother-child relationship may be affected by both the teen mother’s immaturity and her developmental conflicts (Koniak-Griffin, 1988). Although this consequence is not inevitable, deviations in social-emotional development often significantly affect a teen mother’s adjustment to parenting as well as her capacity to regulate parenting skills (Dix, 1991). Hamburg’s (1986) data suggest that adolescent parents often do not have the patience, judgment, or emotional resiliency necessary to cope with the behavior of their children and to make accurate decisions regarding their care. Because adolescent mothers are often affected by social isolation, interpersonal conflicts, and feelings of loneliness and depression (Barth & Schinke, 1984), they may reject or become overly dependent on their children. Either reaction constitutes a maladaptive style of exchange (Rickel, 1989). In an in-depth study of the effects of personal adjustment on adult parenting,

176

John G . Borkowski et al.

Leifer (1980) categorized the psychological functioning levels of adult women during early pregnancy. Women in a high-psychological-functioning group demonstrated a stable personality integration early in pregnancy. They approached the transition to pregnancy by eliciting help from others (e.g., they obtained information about childbirth and involved their husbands in the pregnancy). They were more confident about changes in their personal identity and body image and showed very little anxiety about themselves. Most importantly, during their pregnancy these women identified and initiated a “relationship” with the fetus. This bonding continued into the postpartum stage and facilitated a sense of continuity and emotional attachment between the mother and her child. In contrast, women in a low-psychological-functioning group demonstrated poor personal integration. As they were more easily threatened by the identity changes caused by pregnancy, these women frequently ignored or misunderstood the psychological processes occurring. They exhibited significantly lower levels of self-esteem during pregnancy, expressed more ambivalent and negative feelings toward the fetus, and experienced higher levels of personal anxiety. These women also expressed serious doubts about their abilities to be good mothers. Women in this group were passive, remained uninformed about the birth process, and were quite dependent on their physicians. They also used more passive coping mechanisms in dealing with childbirth, nursing, and other challenges of parenting. The emotional attachment shown by these women to their infants were significantly less than that shown by the other women in the study. On the birth of their child they were more frequently depressed and anxious, and continued to have low self-esteem. Depressed mothers have been found to be less involved in functional reciprocal interactions with their young children (Mills, Puckering, & Pound, 1985). Under the assumption that teen mothers are less psychologically stable, we believe that serious implications for insecure attachment are raised, producing additional negative consequences for infant development. Although research suggests that adult maternal adjustment influences maternal parenting behavior and stress, only a handful of studies have examined the relationship between adolescent personality and actual parenting behaviors. In a study of adolescent parenting, Unger and Wandersman (1985) indicated that selfesteem and feelings of mastery by adolescent parents predicted parenting skills, life satisfaction, and less parenting anxiety. In the Notre Dame sample, we found that pregnant adolescents were less socially adjusted than pregnant adults. Pregnant adolescents were found to report more behavior problems than adults on both Internalizing and Externalizing scales of the Achenbach Youth Self-Report Scale (1987). This finding suggests that the pregnant adolescents had more difficulty than the pregnant adults in appropriately expressing and releasing their feelings. The Internalizing and Externalizing scales respectively reflect fearful, inhibited, and overcontrolled behavior versus aggressive, antisocial, and undercontrolled behavior. Additionally, a comparison of problem-solving abilities of

UNRAVELING THE “NEW MORBIDITY”

177

the adolescent and adult samples on a social situations problem-solving test indicated that pregnant adolescents were significantly less effective than nonpregnant peers and pregnant adults at generating relevant solutions to social problems. This result squares with past research indicating that pregnant adolescents are generally poor problem solvers (Mechanic, 1974). As a consequence of these types of problems, it seemed likely that the adolescent’s transition into assuming her maternal role would likely be highly stressful. Further analysis indeed indicated that personal adjustment among teen mothers predicted parenting stress. Mothers who were less adjusted found their children less adaptable, less reinforcing, and more moody and demanding. More poorly adjusted mothers also experienced their maternal roles as quite restrictive, did not feel as competent, viewed themselves to be socially isolated and depressed, and reported having more health problems than did more adjusted mothers. These results suggest that young mothers in our sample who had adjustment problems also had a difficult time adjusting to the demands of parenthood and the needs of their children while still attempting to address their own personal needs. E.

Child Temperament and Adolescent Parenting

Our model implicates infant characteristics, such as temperament, in determining parenting behavior. Temperament, in a broad sense, is an important characteristic that defines the infant’s behavioral style. In the mother-infant relationship, the temperament of the young infant affects the type of exchange that occurs between mother and infant (Bates, 1980; Carey & McDevitt, 1978). For example, infants who are pleasant and responsive to their mothers’ behaviors may reinforce maternal engagement and encourage further positive interactions. In contrast, difficult children may inhibit positive interactions and make it more difficult for mothers to be responsive and display warmth and affection. An infant with a difficult temperament may present even greater challenges to the immature adolescent mother, with the consequence that she is more stressed and less responsive. These factors, in turn, inhibit responsive parenting. A child’s tendency to respond to his or her mother may be a primary determinant of the subsequent behaviors of the mother (Eagle & Brazelton, 1977). Crockenberg (1987) points out that mothers who have infants with difficult temperaments react to their irritable children in an insensitive and aloof fashion and use a power-oriented disciplinary approach, thereby creating tensions between mother and child. Similarly, Belsky (1986) believes that irritable and demanding children create higher levels of stress in their parents. Thus, a child’s temperament, or individual behavioral style, may be an influential component in the dyadic relationship and may be one of the primary determinants of maternal stress and the quality of mother-child interactions. In a review of the literature on influences of child temperament on caregiving,

178

John G . Borkowski et al.

Crockenberg (1986) found that nearly half of 16 studies suggested that mothers are more interactive with irritable or difficult babies. In contrast, the other half indicated that babies who are pleasant and easy to interact with receive more responsive maternal reactions. Crockenberg ( 1986) hypothesized that these discrepant results could be reconciled by understanding the interaction of infant temperament with the disposition of the caregiver as well as the actual caregiving environment. She suggested mothers with adaptive personalities and a positive life situation may perceive an irritable or difficult infant as a challenge, thus resulting in a positive correlation between difficult temperament and increased maternal behavior: Such mothers use their personal and environmental resources to meet the challenging demands of difficult children. Conversely, mothers with maladaptive personalities or unstable life situations are likely to experience difficult infants as stressors and avoid responsive-sensitive contact, resulting in a negative correlation between infant temperament and maternal interaction. Crockenberg further pointed out that in studies where difficult temperaments led to sensitive and responsive maternal behaviors, none of the mothers were identified as at risk for problematic parenting. The influence of child temperament on adolescent parenting is of special concern in our theory (see Fig. I), especially given that infants of young mothers are more likely to present unique problems to their parents. For instance, infants of adolescent mothers have been found to be hyperactive and less responsive to their mothers (Osofsky & Osofsky, 1970). They often are more distractible, less adaptable, and passive in interactions with their mothers (Field, 1981). In a comparative study of adolescent and adult mothers’ interactions with their infants (Teberg et al., 1983), 47%of the infants of teenage mothers demonstrated limited ability to cope with stress, compared with only 26% of the infants of adults. During infant-maternal interactions, the infants of the younger mothers appeared to have more difficulty with individual play activities and with separation from their mothers. As will be described in the final section of this article, immature interactional patterns may eventually lead to insecure attachment, thus providing an unstable foundation for subsequent attentional and metacognitive development.

F. Social Support and Adolescent Parenting Within the past decade, the influence of social support on the psychological well-being and behavior of parents has received increasing attention within the mental retardation research community. The emphasis on contextual sources of support can be traced to Bronfenbrenner’s ( 1 977) discussion of the importance of ecological variables in stable family functioning. Subsequently, Unger and Powell (1980) indicated that social support may buffer the effects of stressful events and serve to prevent problems in parent-child interactions. More recently, Brooks-Gunn and Furstenberg (1986) have emphasized the importance of an

UNRAVELING THE “NEW MORBIDITY”

179

ecological framework for understanding adolescent parenting. In their review of the literature on social support and adolescent parenting, Whitman et al. (1987) noted that a young mother is at increased risk for problematic parenting and her infant at risk for developmental delay if the social support given to her prenatally as well as following the birth of her baby is inadequate. The influence of social support on adult parenting behavior and stress has been studied extensively (Cmic, Greenberg, & Slough, 1986; Dunst, Trivette, & Cross, 1986a; 1986b). Researchers have found that social support affects both attitudes and behaviors. For example, women involved in emotionally close relationships generally report lower levels of stress and depression. With regard to family functioning, support for mothers has been related to positive motherchild interactions, child adjustment, and marital satisfaction (Grossman, Eichler, & Winickoff, 1980; Hetherington, Cox, & Cox, 1976; Wandersman, Wandersman, & Kahn, 1980). In families with disabled children, adequate social support has been found to buffer the stresses associated with parenting a challenging child. Parental behavior, marital satisfaction, personal well-being, and child outcome within families with disabled children have all been shown to be positively affected by social support (Dunst, Trivette, & Cross, 1986a, 1986b; Friedrich & Friedrich, 1981). Likewise, support given to mothers during a crisis has been correlated with greater nurturance, more positive interactions, and fewer negative responses to their children (Longfellow, Zelkowitz, Saunders, & Belle, 1970). Other studies have found that active social support facilitates sensitive, responsive maternal behavior and child development (Cmic, Greenberg, Ragozin, Robinson, & Basham, 1983; Weinraub & Wolf, 1983). Conversely, lack of support and social isolation have been connected with child abuse and neglect (Polansky, Chalmers, Buttenweiser, & Williams, 1979). Studies specifically investigating adolescent mothers have found similar relationships between social support and adolescent parenting. Results by Barth, Schinke, and Maxwell (1983) indicated that support systems that meet the needs of the young mother reduce tension and frequently contribute to positive feelings toward the parenting role. In addition, social support may reduce the negative impact of stress on adolescent parenting (Unger & Wandersman, 1985). In general, interpersonal relationships that are supportive serve as buffers to protect the individual from stress by enabling the parent to cope more effectively (Thompson, 1982). Supportive relationships generally offer information, financial and instrumental assistance, encouragement, and emotional understanding. As a result of their parenting status, adolescent mothers are often isolated from typical sources of support usually available to nonparenting adolescents and married mothers, (e.g., support from a peer group, individuals at school, spouse, and co-workers within a place of employment). In a comparative study of adolescent and adult mothers in an urban, low-income community, young mothers were found to be the heads of single-parent households more frequently than older

180

John G . Borkowski et al.

mothers. These young mothers received little assistance with childrearing, and participated minimally in community and church-related activities (Kellam, Adams, Brown, & Ensinger, 1980). The amount of total child care from others was significantly less for adolescent mothers than for their older counterparts in a study of lower- to middle-class mothers (Levine, Garcia-Coll, & Oh, 1985). Moreover, adolescent mothers were likely to receive only marginal levels of support from peers, as they tended to be socially isolated and less socially competent than their nonpregnant peers (Rogeness, Ritchey, Alex, Zuelzer, & Morris, 1981; Wurtz, 1989). In general, adolescent mothers lacked encouraging and consistent support from peers and community agencies, in terms of both emotional and instrumental support. Because adolescent mothers tend to remain single (Furstenberg et al., 1987), they often face the demanding task of parenting with limited support from the father of the baby or an intimate partner (Furstenberg & Crawford, 1978; Unger & Wandersman, 1985). When, however, a male partner does remain involved and provides support, adolescent mothers have been observed to interact with their children in a more positive and less punitive manner (Crockenberg, 1987). Similarly, Colletta (1981) found that both emotional support and the total amount of instrumental support received were inversely related to the adolescent mother’s rejection of her infant. In this study, family-of-origin support was found to be more important than support from a partner. Adolescent mothers who live with their parents are more likely to receive social support, such as financial assistance and child care, than are adolescent mothers who live independent from their families of origin (Furstenberg & Crawford, 1978). Those adolescents who live with family members find the transition to parenting much easier than their isolated peers (Panzarine, 1988). Family support has also been associated with more rapid development in children of adolescent mothers (Field et al., 1980) and with the heightened probability that the mother will complete high school, be employed, and eventually become economically independent (Furstenberg & Crawford, 1978). In an investigation of black, lower-SES, adolescent mothers, those who received support from relatives exhibited higher levels of parenting effectiveness, reported having more satisfaction with their parenting roles, and derived more pleasure from their infants. Finally, research has shown that young mother’s social connectedness to extended family members, especially their own mothers, can improve their parenting skills (Stevens, 1984). Data from the Notre Dame Project suggest that the character of the social network itself also influences parenting behaviors: For instance, teens with a more positive network orientation and less dense networks (i.e.. with a lower percentage of kin) performed better in their parenting tasks. This finding seems to support the contention of Unger and Wandersman (1985) that a teen mother’s overdependence on her family can, at times, have negative effects in developing

UNRAVELING THE “NEW MORBIDITY”

181

independence as a parent. Clearly, the processes through which the teen mother’s dependence on her family influence her adjustment to parenting tasks need to be explored in future research. Thus, both theory and research suggest that adolescent mothers benefit from receiving social support. When adequate levels of assistance are provided, adolescent mothers cope more effectively with the numerous stresses of early parenting and demonstrate more effective parenting behaviors. It appears that adequate social support can buffer deficiencies in personal adjustment and mediate the translation of “cognitive readiness” directly into parenting practices.

IV.

COMPONENTIAL ANALYSES: THE RELATIVE IMPORTANCE OF COGNITIVE READINESS

To assess the adequacy of our model of teen parenting (Fig. I), we employed a series of regression analyses to assess the contributions of the model’s major constructs to actual parenting behavior as well as to child development at 6 months of age. First, canonical correlations were used to decide which measures were the most appropriate indicators of each construct; then regression analyses were used to find the best predictors of maternal and child development at 6 months. The best measure of social support was a linear combination that equally weighted emotional support of mother, financial support from the family, and social network size. The measure of personal adjustment was social competence minus externalizing and internalizing behavior (from the Achenbach Youth SelfReport Scale). The cognitive readiness measure included assessments of knowledge about infant and child development, knowledge about development milestones, beliefs about punishment, attitudes about role reversal, and child centeredness. Vocabulary and block design subtests (WAIS-R) were used to estimate maternal intelligence. Next, measures of each construct were placed in regression equations, predicting parenting and child outcomes. The most interesting findings were the following: (1) Stress, as measured by the Parenting Stress Index, and child temperament (Carey Scale) were significantly related to personal adjustment during pregnancy and cognitive readiness to parent (25 and 19%, respectively, of the variance accounted for). (2) Maternal sensitivity (as derived from videotaped toy play episodes) was significantly related to cognitive readiness to parent. (3) Judgments about the overall quality of parenting (MIS)were related to the IQ of mothers as well as to cognitive readiness. (4) Bayley mental developmental scores were predicted by measures of prenatal social supports. These findings reveal the diverse and unique impact of each of the constructs in our model to parenting and child development. Furthermore, the central role played by cognitive readiness (relative to other potentially influential factors) in

182

John C . Borkowski et al.

determining early parenting skills and stress is highlighted. The results give credibility to the addition of the “readiness to parent” concept to Belsky’s (1984) general model of parenting, especially when it is applied to adolescent mothers. Cognitive readiness to parent appears to have unique predictive power in understanding unsuccessful adolescent parenting, at least during the infancy period. In summary, our longitudinal investigation to date has indicated the importance of considering individual differences when investigating adolescent parenting. More specifically, the major results show how an understanding of diversity within the population of first-time parents can lead to identification of the environmental and personal factors that place young mothers at risk for poor parenting and provide tentative information concerning how interventions might be structured so as to assist teen mothers in fulfilling their early childrearing responsibilities. Consistent with past theory and research, adolescent mothers were found to be less cognitively prepared for parenting than adult mothers. This deficiency in cognitive readiness is, even when detected during pregnancy, associated with subsequent parenting stress and insensitive maternal interactionspotential causes of subsequent developmental delays. With multiple factors tentatively identified as contributors to poor parenting and potential harbingers of developmental delays, we have begun to examine their longer-term impact on child development, especially in combination with additional factors that arise during early childhood and that can exacerbate various types of developmental delay. One such factor, known to be critical for social development but not generally considered highly related to cognitive-intellectual deficits, is attachment. In the final section of this article we make a case for how attachment might be related to elementary information processing and to early signs of metacognition, and thus potentially related to the eventual onset of mild mental retardation.

V.

INFLUENCE OF EARLY ATTACHMENT ON LATER COGNITIVE DEVELOPMENT

The attachment between infants and their caregivers has been studied extensively in both normal and at-risk groups (Colin, 1991). Considering the recent attention given to the attachment process in high-risk groups, it is somewhat surprising that research programs have not focused directly on its role as a correlated or causal factor in mild retardation and/or generalized learning disabilities. For instance, no articles on attachment appeared in either the American Journal of Mental Retardation or Mental Retardation during 1990, even though Cicchetti and Serafica (198 1) documented the importance of attachment in understanding development in children with Down syndrome over a decade ago. In this section, we provide a brief overview of the attachment literature, present an

UNRAVELING THE “NEW MORBIDITY”

183

information-processing account of the relationship between attachment and impaired cognitive development, provide preliminary data examining this relationship, and suggest theoretical implications for the subsequent emergence of mild to moderate mental retardation in insecurely attached children of teen mothers. A.

Attachment Defined

Attachment is defined as an enduring emotional bond between the infant and a specific person(s). The infant may have one or more attachment relationships (e.g., to mother, father, older sibling, or day-care provider). Critical episodes of attachment behaviors, aimed at seeking and maintaining closeness to the attached person, are evoked by stressful situations. Although the formation of attachment relationships is an important component in emotional and social development, its definition implies a hidden agenda: a reciprocal relationship with cognitive development. More specifically, to establish an enduring relationship to a specific person, the infant needs to recognize, remember, and believe in the ongoing presence of that person. For the average infant, evidence suggests these abilities come together around the age of 7 months. The consolidation of early-appearing cognitive processes allows the formation of enduring emotional bonds. Separation anxiety and stranger anxiety or wariness are thus outward manifestations of the infant’s more advanced cognitive development, and are suggestive of the ongoing formation of an attachment relationship. Hence, emotional or cognitive impediments that hinder the formation of enduring bonds may result in the formation of insecure attachment relationships. Relatedly, the formation of a secure attachment relationship provides a solid foundation for the development of social, emotional, attentional, and self-regulatory processes. The Strange Situation, developed by Ainsworth and Wittig (1969), has been the primary method used to assess the quality of infants’ attachment to their caregivers. During the Strange Situation test, the 12- to 18-month-old infant is exposed to moderate stress induced by systematic, brief separations from the primary caregiver and by contact with a friendly stranger. The infants’ behaviors during the separation and reunion periods are evaluated to the extent they seek and maintain closeness, resist the soothing efforts, and avoid the caregiver and the stranger. Ratings are then used to evaluate the overall quality of the attachment relationship. Classification of infants as securely or anxiously attached has been an evolving process (Ainsworth et al., 1978; Main & Solomon, 1986). The early classification system contained secure, avoidant, and ambivalent categories. In 1986, Main and Solomon (1986) suggested the addition of a disorganized/disoriented classification to the initial categories of anxious attachments, avoidant and ambivalent.

184

John G . Borkowski et al.

1. Secure. Infants with secure attachments to their caregivers are able to use their caregivers, when present, as a base from which to explore their environments. For example, secure infants show interest in the attractive, age-appropriate toys provided in the Strange Situation room. When the caregiver is not present, securely attached infants show distress, and the stranger is not able to substitute for the caregiver in soothing the infants. During reunion episodes, securely attached infants actively seek contact with the caregiver and are genuinely comforted by that contact. Studies with normal populations find that approximately 60 to 65% of infants are categorized as secure (Colin, 1991). 2. Avoidant. At first glance, avoidant infants appear independent. These infants seem to have little need for their caregivers. When the caregiver is not present, avoidant infants often continue their toy play or show minimal distress. During the reunion episodes, avoidant infants do not, however, seek contact with the caregiver, may avoid eye contact, or actively turn away from the caregiver. Their avoidance is specific to the caregiver. That is, the infant may even approach and interact with the stranger. Even though the infant may appear independent, there is evidence the infant is actually experiencing consider distress (Belsky & Braungart, 1991). Approximately 20% of infants from normal populations are categorized as avoidant. 3. Ambivalent. The ambivalent classification is relatively infrequent (about 10% of infants). An ambivalent infant often mixes anger and hostility with contact seeking. Distress is obvious when the caregiver is not present, but the return of the caregiver may not alleviate the distress. Ambivalent infants appear passive, and may cry or fuss rather than go to the caregiver for comfort. 4. Disorganizedldisoriented. The disorganized classification, proposed by Main and Solomon (1986), has only recently begun to appear in the attachment literature. Disorganized infants have no consistent strategy to cope with the stress produced by separation and exposure to the stranger. They may alternate between avoidance, hostile behaviors, and clinging to the caregiver. Infants are classified as disorganized if their behavior cannot be reliably established as secure, avoidant, or ambivalent and shows inconsistencies.

B.

Parenting and Attachment

Parenting behavior occupies a critical position in our model. We have discussed several qualities of effective parenting that distinguish adolescent and adult mothers, suggesting that these qualities can be explained by a mother’s cognitive readiness for parenting, psychosocial adjustment, social support system, as well as the characteristics of her infant. The effects of parenting behavior on child development may be straightforward (e.g., lack of verbal interactions predisposing the child to language delays) or may interact dynamically with other

UNRAVELING THE “NEW MORBIDITY”

185

variables to produce an attachment relationship that, in its own right, shapes the child’s developmental trajectory. In a review of the attachment literature, Colin (1991) concluded that “the mother’s success in responding sensitively to her baby’s needs and signals is a principal determinant of the baby’s attachment pattern” (p. 44).Other background factors in and of themselves do not seem to increase the likelihood of an insecure attachment. It is only when the background factors impact on a mother’s tendency to respond consistently and contingently, or when more than one risk factor is present, that the quality of the attachment relationship is affected. Maternal sensitivity may, therefore, have a complex impact on cognitive-social development, by focusing and sustaining the child’s attention, by providing reinforcement for individual actions, and by promoting a secure attachment relationship.

C.

Social and Cognitive Consequences of Attachment

A large body of evidence suggests that the attachment categories are relatively stable, provided the infant’s environment remains consistent. In addition, the attachment classification in infancy may have consequences for further social, behavioral, and cognitive development. Sroufe, Fox, and Pancake (1983) concluded that 4- to 5-year-old children who were classed as secure at 18 months were more socially competent, had more friends, and were more compliant and empathic than anxiously attached children. Other studies have found a high incidence of aggressive behaviors and other behavior problems in early childhood among anxiously attached infants (Lyons-Ruth, Alpem, & Pepachold, 1991). Cognitive consequences of early attachment are less well established. Although differences in intelligence or achievement have not been found between securely and anxiously attached infants, secure infants have been characterized as more persistent (Frodi, Bridges, & Grolnick, 1985), more ego-resilient, and curious (Arend, Gove, & Sroufe, 1979). Colin (1991) suggests that avoidant and ambivalent children may have less adaptive responses to frustrating and challenging problem-solving tasks. Relatedly the security of attachment influences the development of symbolic play (Belsky, Gardoque, & Hmcir, 1984; Matas, Arend, & Sroufe, 1978; Slade, 1987). For instance, Belsky et al. (1984) measured the quality of free play in 60 infants at approximately 1 year of age. Infants were rated for their extent of undifferentiated exploration (i.e., mouthing and simple manipulation of toys), transitional play (relating two or more objects in play and simple symbolic play), and pretend play. In addition, the difference between the quality of infants’ free play and their play when encouraged by a trained adult was evaluated. Securely

186

John G . Borkowski et al.

attached infants engaged in significantly more undifferentiated exploration and transitional play than insecure infants. Further, securely attached infants showed less of a discrepancy between their free play and elicited play, suggesting a tendency to routinely perform at their highest level of competence; insecure infants fail to show this consistency. Belsky et al. (1984) speculated that “infants and children who are inclined not to deploy their most sophisticated skills may deprive themselves of self-generated learning experiences that probably derive from functioning at one’s optimal level” (p. 416). Alternatively, they suggest a continuity in motivational factors affecting free play at 12 months of age as well as later problem-solving competence. Although research in this area is necessarily correlational, and the causal influence of early attachment on cognition has not been clearly demonstrated, we hypothesize that attachment is related to sustained attention and to the early appearance of self-regulatory, metacognitive processes.

D.

Attachment and Information Processing

Information-processingtheory suggests an alternative explanation for how the quality of attachment in infancy might relate to cognitive development in early childhood. The hypothesis is that an infant’s attempts to cope with an insecure attachment relationship require most of his or her attentional resources, leaving fewer resources for exploring and learning about new aspects of the environment (cf. Main, 1991). In a series of experimental and correlational studies, Dunham and colleagues (Dunham, Dunham, Hushman, & Alexander, 1989; Dunham & Dunham, 1990) examined the relationship between contingent and noncontingent social interactions and subsequent attention to a novel stimulus in 3-month-old infants. Results suggest that, as early as 3 months of age, aspects of the interaction between infants and their caregivers begin to influence the quality of infants’ attention to relevant aspects of the environment. Dunham and Dunham (1990) concluded that “the more time the mother-infant pair spent in the dyadic state of vocal turntaking the longer the infants fixated on the subsequent stimulus pattern and the shorter their interfixation intervals during the contingency task” (p. 789). Similarly, Tamis-LeMonda and Bornstein (1989) investigated the relationship between infants’ habituation at 5 months and their competence in a free-play setting at 13 months, with a sample of 37 middle- to upper-SES households. Results indicated that habituation (a measure of early attentional processing), but not the maternal encouragement of attention, predicted the quality of infants’ symbolic play. It is interesting to note that the same contingent interactions that we believe affect infants’ attentional processes are also related to the quality of attachment (Colin, 1991).

UNRAVELING THE “NEW MORBIDITY”

E.

187

Attachment and Attention: Precursors of Developmental Delays

Research with adolescent mothers in the Notre Dame Parenting Project provides further support for interesting relationships between sociallaffective development and the emergence of information-processingskills, and has implications for understanding the interactive causes of mild mental retardation from birth through the early childhood years. We have recently examined the concurrent relationship between the security of the infant’s attachment and attentional flexibility, as reflected in the type and quality of toy play, and then linked these variables with intellectual development at age 3. We expected that attentional differences between securely and anxiously attached infants would be particularly pronounced during the two reunion episodes. For secure infants, the caregiver’s return was anticipated and expected; once reunited, such infants should be free to return their attention to exploring their environment. In contrast, anxiously attached infants experienced stress during and following separations; hence, they were expected to experience cognitive confusion on the caregiver’s return. This confusion was expected to impede cognitive flexibility as the child explored his or her environment. This hypothesis was tested with a sample of 59 first-time adolescent mothers and their infants. Infants were videotaped in the Strange Situation when they were approximately 12 months of age. Twenty-two percent of the dyads were categorized as avoidant, 42% were categorized as secure, and 34% were disorganized. Only one mother-infant dyad was found to be ambivalent. In addition to scoring the videotapes according to standard attachment procedures, tapes were also coded to determine the flexibility of infants’ attention, as defined by their interaction with toys present in the room. Flexibility of attention to toys was rated on a 5-point scale: 0 = no interaction toys, 1 = nonfunctional interaction with one toy, 2 = nonfunctional interaction with more than one toy, 3 = functional play with one toy, 4 = functional play with more than one toy. As only two of the infants failed to engage in interaction with the toys during baseline or the first reunion, these infants were combined with those given a score of 1. Infant behaviors were rated during the baseline episode and during the first reunion. Coding began 10 seconds after physical contact between the mother and the child was terminated at the beginning of each episode, and continued for 60 seconds; interrater reliability was satisfactory. Although infants generally interacted with from one to six different toys during baseline, only 36% showed functional play. During the first reunion the number of toys interacted with did not increase over baseline levels; however, more infants engaged in functional play (47%).Importantly, the correlation between security of attachment and attentional flexibility during the first reunion episode

John G . Borkowski et al.

188

was significant ( r = .27). It should be noted that this outcome occurred in the presence of attenuated ranges for each variable. Additionally, the measure of attentional flexibility correlated significantly with Bayley mental development at 1 year ( r = .45), Stanford-Binet IQ at 3 years (r = .47), and receptive language (PPVT-R) at 3 years ( r = .48; for the baseline period only). Although these preliminary analyses are only suggestive of the early reciprocal linkages between the flexibility of attention and attachment, as well as their subsequent ties with later cognitive-intellectual impairments (i.e., the average IQ at age 3 for this sample was 85), they do provide an intriguing framework for broadening our views about the potential causes of mild retardation.

F. Attachment: Implications for Mental Retardation The abilities of infants to attend selectively to relevant aspects of their environments and to adapt flexibly to changing environments are critical components of cognitive development (cf. Columbo & Mitchell, 1990). Deficiencies in attentional processing slow the amount of new information absorbed by the infant and, over time, likely lead to other cognitive delays (Rose, Feldman, Wallace, & McCarton, 1991). From this perspective, we hypothesize that the quality of attachment may, in part, be related to the adequacy of early attentional processing. More importantly, insecure attachment may cause information-processing impairments. Specifically, in stressful situations (such as the Strange Situation paradigm), infants and children with anxious attachments may be compelled to direct their attention to emotional relationships rather than to exploring, and learning from, important aspects of their environments. Although yet to be substantiated, the implications of this hypothesis for understanding delayed cognitive development are straightforward: Contingent social interactions tend to direct an infant’s attention to critical aspects of environmental stimulation, whereas noncontingent interactions reduce attention to such events. Simultaneously, consistent and contingent responses to an infant’s needs and verbalizations also contribute to the quality of attachment relationships (Isabella, Belsky, & vonEye, 1989). Although contingent responsivity itself provides explanations for both anxious attachment and the regulation of attention, we believe the emergence of anxious attachment, especially disorganized attachment, further interferes with attentional flexibility, inhibiting the development of critical self-regulatory skills that are at the heart of metacognition (Borkowski, Day, Saenz, Dietmeyer, Estrada, & Groteluschen, in press). VI.

UNRAVELING THE “NEW MORBIDITY”

The story we are in the process of telling about the children of adolescent mothers begins with the cognitive, social, personal, and biological charac-

UNRAVELING THE “NEW MORBIDITY”

189

teristics of the mother that are manifested during pregnancy. These states and dispositions predispose the young mother to parent with differential responsivity, sensitivity, and warmth. Whereas some parents are quite responsive, contingent in their interactive behavior, and positive in style, others are detached, neglectful, and/or abusive. In turn, parenting-in part, determined by infant characteristics and the personal and social resources of the parent-gives rise to major emotional and cognitive developmental consequences, including differential infant attachment outcomes. These outcomes, such as insecure attachment, may further hinder the development of the attentional, emotional, and or selfregulatory processes necessary for realizing each child’s cognitive-intellectual potential.

A.

Complex Causality and New Morbidity

The story of adolescent parenting and child development is complex, interactive, and developmental. We believe that teen mothers are generally unprepared, emotionally (e.g., inadequate personal adjustment such as depression) and cognitively (e.g., limited and inaccurate knowledge about parenting practices and child development), to assume full responsibility for the heavy demands of parenting. Lack of emotional and instrumental support from significant others and problematic child characteristics (e.g., a difficult temperament) can further exacerbate existing parenting problems. The consequences for the child are widespread, encompassing cognitive, emotional, and social domains of development. Teen mothers-often borderline in intelligence themselves-manifest personal and social characteristics that lead to problematic, unsystematic parenting practices, which in turn hinder their infants’ cognitive and social development. For instance, problematic parenting styles, as well as more general maternal characteristics, impede both early attentional and emotional development in many infants of teen mothers. Habituation, scanning, and a preference for novelty in infants with teenage mothers may be adversely affected by a combination of genetic, prenatal, or postnatal risk factors. Similarly, insecure attachment, which appears to result from insensitive parenting, may predispose the child at I year of age to subsequent cognition and emotional developmental delays. Our perspectives on adolescent parenting are in the formative stages, as are all explanation for new morbidity phenomena. Many questions remain to be investigated. Research is needed to identify how risk factors change with time, differentially predicting developmental delays during the early school years: Do certain combinations of factors lead to mild mental retardation, others to general learning disabilities, and still others to conduct disturbances? It seems likely that factors causally related to early development may become subsumed under a later-developing process that itself becomes the causal proxy for a specific type

I90

John G . Borkowski et al.

of developmental delay. In particular, research needs to identify specific causal factors in the complex chain of events that should be targeted for intervention, pinpoint when intervention should optimally occur, and determine whether multiple processes should be focused on simultaneously to maximize treatment gains. We suspect that early intervention, introduced at critical points in our hypothesized chain of interactive events, holds the potential for forestalling the emergence of many types of learning impairments that first become evident to professionals during the school years.

B.

Defining Mental Retardation: A Broader Perspective

The implications for broadening our understanding of cultural-familial mental retardation and its remediation at an early age are considerable. Both existing and proposed definitions of mental retardation refer to “subaverage abilities in cognitive functioning, accompanied by deficits in adaptive skills” as its salient characteristics. Recent definitional perspectives focus more on personal competence, document deficits within the community environments typical of an individual’s age peers, broaden the domains of adaptive deficits, and focus on individual strengths as well as the potential for change. What continues to remain hidden and elusive in the definition of mental retardation are the linkages among cognitive and adaptive deficits, as well as their emotional underpinnings, and how these linkages change over time. The analysis of adolescent parenting-as an instance of new morbidityimplies a different and broader perspective on mental retardation. It suggests that emotional development in the mother (e.g., depression) and child (e.g., insecure attachment) should be more actively examined as a precursor of both cognitive and adaptive deficits (cf. Hubbs-Tart, Eberhart-Wright, Ware, Osofsky, Yockey, & Frisco, 1991). Perhaps, the field of mental retardation has far too long ignored the emotional underpinnings of mild forms of intellectual disabilities. ACKNOWLEDGMENTS The writing of this article was supported by NIH Grant HD-25456. Authors Rellinger and Sommer were supported by NIH Training Grant HD-07184.

REFERENCES Achenbach, T., & Edelbrock, C . (1987).Manualfor the youth selfreport and profile. Burlington, VT: University of Vermont. Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Attachment: A psychological srudy of the strange siruarion. Hillsdale, NJ: Erlbaum. Ainsworth, M. D. S., & Wittig, B . A. (1969). Attachment and exploratoly behavior of one-year-olds in a strange situation. In B. M. Foss (Ed.),Determinants ofinfant behavior IV (pp. 113- 136). London: Methuen.

UNRAVELING THE “NEW MORBIDITY”

191

Arend, R., Gove, F., & Sroufe, L. A. (1979). Continuity of individual adaptation from infancy to kindergarten: A predictive study of ego resiliency and curiosity in preschoolers. Child Development, 50, 950-959. Bacon, L. (1974). Early motherhood, accelerated role transition and social pathologies. Social Forces, 52, 333-341. Barth, R., & Schinke, S. P. (1984). Coping with daily strain among pregnant and parenting adolescents. Journal of Social Service Research, 7(2), 51-63. Barth, R. P., Schinke, S. P., & Maxwell, J. (1983). Adolescent parenting. Journal of Youth and Adolescence, 12, 471-485. Bates, J. E. (1980). The concept of difficult temperament. Merrill-Palmer Quarterly. 26. 299-319. Baumeister, A. A. (1988). The new morbidio and the prevention of mental retardation. Research progress report, John F. Kennedy Center, Vanderbilt University. Beckwith, L. (1976). Intervention strategies for high risk infants and young children (pp. 119-139). Baltimore, MD: University Park Press. Belmont, L., Cohen, P., Dryfoos, J., Stein, Z., & Sayac, A. (1981). Maternal age and children’s intelligence. In K. G. Scott, T. Field, E. G. Robertson (Eds.), Teenage parents and their o#Yppring (pp. 177-197). New York: Grune & Stratton. Belsky, J. (1984). The determinants of parenting: A process model. Child Development. 55, 83-96. Belsky, J. (1986). Transition to parenthood. Medical Aspects of Human Sexuality, 20, 56-59. Belsky, J., & Braungart, J. M. (1991). Are insecure-avoidant infants with extensive day care experience less stressed by and more independent in the strange situation? Child Development, 6 2 , 567-57 I . Belsky, J., Gardoque, L., & Hrncir, I. (1984). Assessing free play, elicited play, and executive capacity in infant exploration: Relationship to home environment and attachment. Developmental Psychology, 20, 406-417. Black, C., & DeBlassie, R. R. (1985). Adolescent pregnancy: Contributing factors, consequences, treatment, and plausible solutions. Adolescence, 20, 281-290. Borkowski, J. G., Day, J. D., Saenz, D. S., Dietmeyer, D., Estrada, T., & Groteluschen, A. (in press). Expanding the boundaries of cognitive interventions. In B. Wong (Ed.), Intervention research with students with learning disabilities. New York: Springer-Verlag. Broman, S. (1981). Longterm development of children born to teenagers. In K.G. Scott, T. Field, & E. G. Robertson (Eds.), Teenage parents and their offspring (pp. 195-225). New York: Grune & Stratton. Broman, S., Nichols, P. L., Shaughnessy, P., & Kennedy, W . (1987). Retardation in young children. Hillsdale, NJ: Lawrence Erlbaum Associates. Bronfenbrenner, U. (1977). Toward an experimental ecology of human development. American Psychologist, 32, 513-531. Brooks-Gunn, J., & Furstenberg, F. (1986). The children of adolescent mothers: Physical, academic, and psychological outcomes. Developmental Review, 6 , 224-25 1. Brown, H.,Adams, H., & Kellan, S. (1981). The longitudinal study of teenage motherhood and symptoms of distress: Research and community mental health (pp. 83-213). Greenwich, CT: JAI Press. Carey, E. B., & McDevitt, S. C. (1978). Revision of the Infant Temperament Questionnaire. Pediatrics, 61, 735-739. Cicchetti, D., & Serafica, F. (1981). The interplay among behavioral systems: Illustrations from the study of attachment, affiliation, and wariness in young Downs syndrome children. Developmental Psychology, 17, 36-49. Colin, V. L. (1991). Human attachment: What we know now. A report prepared under Contract HHS-100-90-00 for the Office of Planning and Evaluation, Department of Health and Human Services. Coll, C., Hoffman, J., & Oh, W. (1987). The social ecology and early parenting of Caucasian adolescent mothers. Child Development. 58, 955-963.

192

John G . Borkowski et al.

Colletta, N. D. (1981). Social support and the risk of maternal rejection by adolescent mothers. Journal of P sychology. 109, I9 I - 197. Columbo, J., & Mitchell, D. N. (1990). Individual differences in early visual attention: Fixation time and cognitive processes. In 1. Columbo & J. Fagan (Eds.), lndividual direrences in infancy: Reliabiliry. stability, prediction (pp. 193-228). Hillsdale, NJ: Erlbaum. Conger, R. D., McCarty, J. A., Yang, R. K., Lahey, B. B., & Burgess, R. L. (1984). Mother’s age as a predictor of observed maternal behavior in three independent samples of families. Journal of Marriage and Family, 46, 41 1-424. Crnic, K. A . , Greenberg, M. T., Ragozin, A. S., Robinson, N. M.. & Basham, R. B. (1983). Effects of stress and social support on mothers and premature and full-term infants. Child Development, 54, 209-217. Crnic, K. A., Greenberg, M. T., & Slough, N. M. (1986). Early stress and social support influences on mothers’ and high-risk infants’ functioning in late infancy. lnfant Mental Health Journal. 7(1), 19-33. Crockenberg, S. (1986). Are temperamental differences in babies associated with predictable differences in caregiving? In J. V. Lerner & R. M. Lerner (Eds.), Temperament and social interaction during infancy and childhood: New directions for child development (Val. 3 I). San Francisco: Jossey-Bass. Crockenberg, S. (1987). Predictors and correlates of anger toward and punitive control of toddlers by adolescent mothers. Child Development. 58, 964-975. de Lissovoy. V. (1973). Child care by adolescents. Children Today. 2, 22-25. Dix, T. (1991). The affective organization of parenting: Adaptive and maladaptive processes. Psychological Bulletin, 110. 3-25. Dunham, P., & Dunham, F. (1990). Effects of mother-infant social interactions on infants’ subscquent contingency task performance. Child Developmenr, 61, 785-793. Dunham, P., Dunham, F., Hushman, A., & Alexander, T. (1989). Social contingency effects on subsequent perceptual-cognitive lasks in young infants. Child Development. 60. 1486- 1496. Dunst, C. J., Trivette, C. M., & Cross, A. H. (1986a). Mediating influences of social support: Personal, family, and child outcomes. American Journal of Mental Deficiency. 90, 403-4 17. Dunst, C. J., Trivette, C. M., & Cross, A. H. (1986b). Roles and support networks of mothers of handicapped children. In R. Fewell & P. Vadasy (Eds.), Family of handicapped children: Needs and supports across the life span (pp. 167-192). Austin, TX: PRO-ED. Eagle, D. B., & Brazelton, T. B. (1977). The infant at risk: Assessment and implications for intervention. In M. F. McMillan & S. Henao (Eds.), Childpsychiutry (pp. 37-50). New York: Brunner/Mazel. Egeland, B., & Farber, E. A. (1984). Infant-mother attachment: Factors related to its development and changes over time. Child Development, 55. 755-771. Elster, A. B., McAnarney, E. R.. & Lamb, M. E. (1983). Parental behavior of adolescent mothers. Pediatrics. 71, 494-503. Epstein, A. (1980). Assessing the child development information needed by adolescenr purents with very young children. A report to the Department of Health, Education, and Welfare. Ypsilanti, MI: High/Scope Educational Research Foundation. Feldman, H., & Feldman, M. (1975). The family life cycle: Some suggestions for recycling. Journal of Marriage and the Family, 37(2), 277-284. Field. T. M. (1979). Interactional patterns of high-risk and normal infants. In T. Field (Ed.). Infants born at risk. New York: Spectrum. Field, T. M. (1980). Interactions of preterm and term infants with their lower- and middle-class teenage and adult mothers. In T. Field, S. Goldberg. D. Stern, & A. Sostek (Eds.), High-risk infants and children. New York: Academic Press. Field, T. M. (1981). Early development of the preterm offspring of teenage mothers. In K. Scott, T. Field, & E. G. Robertson (Eds.), Teenage parents and their offspring. New York: Gmne & Stratton.

U N R A V E L I N G THE “NEW MORBIDITY”

193

Field, T. M., Widmayer, S. M., Stringer, S., & Ignatoff, E. (1980). Teenage, lower class, black mothers and their preterm infants: An intervention and developmental follow-up. Child Development. 51. 426-436. Friedrich, W. N.. & Friedrich, W. L. (1981). Comparison of psychosocial assets of parents with a handicapped child and their normal controls. American Journal of Mental Deficiency, 85, 551553. Frodi. A , , Bridges, L.. & Grolnick, W. (1985). Correlates of mastery-related behavior: A short-term longitudinal study of infants in their second year. Child Development, 5 6 , 1291- 1298. Furstenberg, F., & Crawford, A. (1978). Family support: Helping teenager mothers to cope. Family Planing Perspectives. 10. 322-333. Furstenberg, F. F., Jr., Brooks-Gunn, J . , & Morgan, S. P. (1987). Adolescent mothers in later life. LondoniNew York: Cambridge University Press. Grossman, F., Eichler, L., & Winickoff, S . (1980). Pregnancy, birth. andparenthood: Adaptations of mothers, ,fathers, and infants. San Francisco: Jossey-Bass. Hamburg, B. A . (1986). Developmental issues in school-age pregnancy. In E. Purcell (Ed.), Aspects ofpswhiutric problems of childhood and adolescence ( pp. 299-325). New York: Josiah Macy, Jr., Foundation. Hetherington, E. M.. Cox, M., & Cox, R. (1976. September). The aftermath of divorce. Presented as an invited address at the meetings of the American Rychological Association, Washington, DC . Hubbs-Tart, L.. Eberhart-Wright, A., Ware, L., Osofsky, J., Yockey, W., & Frisco. J. (1991, April). Maternal depression and infant attuchment: Behavior problems at 54 months in children of udolescent mothers. Paper presented at the biennial meeting of the Society for Research in Child Development, Seattle, WA. Isabella, R . A., Belsky, J., & vonEye, A. (1989). Origins of infant-mother attachment: An examination of interactional synchrony during the infant’s first year. Developmental Psychology. 25. 12-21. Jones, F. A., Green, V.. & Krauss, D. R. (1980). Maternal responsiveness of primiparas mothers during the postpartum period: Age differences. Pediatrics, 65. 579-584. Kellam, S . G.. Adams. R. G.. Brown, C. H., & Ensinger, M. E. (1980). The longterm evolution of the familv structure of teenage and older mothers (NICHD Contract I-HD-728321). Washington, DC: U.S. Govt. Printing Office. Keogh. D. A. ( I 991. May). A descriptive look lit changes in dolescent mothers. Paper presented at the Gatlinburg Conference on Research and Theory in Mental Retardation and Developmental Disabilities, Miami, Florida. Klein, H. A,, & Cordell, A. S . (1987). The adolescent as mother: Early risk identification. Journal of Youth and Adolescence. 16( I ) , 47-58. Koniak-Griffin, D. (1988). The relationship between social support, self-esteem, and maternal-fetal attachment in adolescents. Research in Nursing and Health. 11. 269-278. Lamb, M. E. (1982). Parent-infant interaction, attachment, and socioemotional development in infancy. In R . E. Emde & R. R. Harmon (Eds.), Attachment and affliative systems: Neurf)biOkJgiCdand psychobiological aspects ( pp. 1-28), New York: Plenum Press. Lancaster, J. B., & Hamburg, B. A. (1986). School-age pregnancy and parenthood. Hawthorne, NY: Aldine de Gruyter. Landy. S. ( 198 I ). A n investigation of teenage mothers. their infants. and resulting mother-infant dvads. Unpublished doctoral dissertation, University of Regina, Regina, Saskatchewan. Leifer, M. (1980). Psychological effects of motherhood: A study of first pregnancy. New York: Praeger. Levenson, P., Atkinson, B., Hale, J., & Hollier, M. (1978). Adolescent parent education: A maturational model. Child Psychiatry and Human Development, 9, 104-1 18. Levine, L., Garcia-Coll, C . , & Oh, W. (1985). Determinants of mother-infant interaction in adolescent mothers. Pediarrics, 75. 23-29.

194

John G . Borkowski et al.

Longfellow, C., Zelkowitz, P., Saunders, E., & Belle, D. (1979, April). The role of support in moderating the effects of stress and depression. Presented at the Biennial Meeting of the Society for Research in Child Development, San Francisco. Lyons-Ruth, K.,Alpern, L., & Pepachold, B. (1991). Disorganized infant attachment classificurion and maternal psychiatric problems as predictors of hostile-aggressive behavior in the preschoo/ classroom. Manuscript submitted for publication. Main, M. ( 1991, April). Metacognition consequences qfattachment. Paper presented at the annual meeting of the Society for Research in Child Development, Seattle, WA. Main, M., & Solomon, J. (1986). Discovery of an insecure-disorganizedldisorientedattachment pattern: Procedures, findings and implications for the classification of behavior. In T. B. Brazehon & M. Yogman (Eds.), .4ffective development in injancy (pp, 95- 124). Norwood. NJ: Ablex. Matas, L., Arend, R. A,, & Sroufe, L. A. (1978). Continuity of adaptation in the second year: The relationship between quality of attachment and later competence. Child Development, 49, 541-556. McAnamey, E. R., Lawrence, R. A.. Riccutti, H. N., Polley, J.. & Szilagy, M. (1986). Interactions of adolescent mothers and their one-year old children. Pediatrics, 78, 585-590. McLaughlin, F. J., Sandler, H. M., Sheriod, K. S . , Vietze, P. M., & O’Conner, S . (1979). Socialpsychological characteristics of adolescent mothers and behavioral characteristics of their firstborn infants. Journal of Population, 2 , 69-73. Mechanic, D. (1974). Discussion of research programs on relations between stressful life events and episodes of physical illness. In B. S . Dohrenwend and B. P. Dohrenwend (Eds.), Stressful life events: The nature and effects. New York: John Wiley and Sons. Mills, M., Puckering, C . , & Pound, A. (1985). What is it about depressed mothers that influences their children’s functioning? In J. E. Stevenson (Ed.), Recent research in developmental psychopathology ( pp. I I - 17). New York: Pergamon Press. National Survey of Family Growth, Cycle 111. (1982). National Center for Health Statistics. Washington, D.C.: Department of Health and Human Services. Newberger, C. M., Melnicore, L. H., & Newberger, E. H. (1986). The American family in crisis: Implications for children (Vol. 16, No. 12). Chicago: Year Book Medical. O’Dell. P. C. ( 1990). Mother interaction: Assessment ofadolescent and adult behavior. Unpublished manuscript, University of Notre Dame. Osofsky, J. D., & Eberhart-Wright. A . (1988). Affective exchanges between high risk mothers and infants. International Journal of Psycho-Analysis, 69. 22 1-231. Osofsky, H. J., & Osofsky, J. D. (1970). Adolescents as mothers: Results of a program for lowincome pregnant teenagers with some emphasis upon infants’ development. American Journal of Orthopsychiatry, 40. 825-834. Panzarine, S . (1988). Teen mothering: Behaviors and interventions. Journal of Adolescent Health Care, 9. 443-448. Panzarine, S . (1989). Interpersonal problem solving and its relation to adolescent mothering behaviors. Journal of Adolescent Research, 4( I ) , 63-74. Parks, P. L . , & Smeriglio, V. L. (1983). Parenting knowledge among adolescent mothers. Journal of Adolescent Health Care. 4 . 163- 167. Parks, P. L., & Smeriglio, V. L. (1986). Relationships among parenting knowledge, quality of stimulation in the home, and infant development. Family Relations. 35. 41 1-416. Philliber, S., & Graham, E. (1981). The impact of age of mother on mother-child interaction patterns. Journal of Marriage and Family. 43(2). 109- 115. Phipps-Yonas, S. (1980). Teenage pregnancy and motherhood: A review of the literature. American Journal of Orthopsychiatry, 50, 403-43 I . Plomin, R. (1990). Nature and nurture: An introduction to human behavioral genetics. Pacific Grove, CA: Brooks/Cole.

UNRAVELING THE “NEW MORBIDITY”

195

Plomin, R., Loehlin, J. C., & DeFries, J. C. (1985). Genetic and environmental components of “environmental” influences. Developmental Psychology, 21, 391-402. Polansky, N . , Chalmers, M.. Buttenweiser. M., & Williams, D. (1979). The isolation of the neglected family. American Journal of Orthopsychiatry. 49, 149- 152. Protinsky, H., Sporakowski, M., & Atkins, P. (1982). Identity formation: Pregnant and non-pregnant adolescents. Adolescence, 17, 73-80. Ragozin, A. S . , Basham, R. B., Crnic, K. A., Greenberg, M. T., & Robinson, N. M. (1982). Effects of maternal age on parenting role. Developmental Psychology, 18, 627-634. Ramey, C. T., & Campbell, F. A. (1981). Educational intervention for children at risk for mild retardation: A longitudinal analysis. In P. Mittler (Ed.), Frontiers of knowledge in menrol retardation (Vol. I). Baltimore: University Park Press. Ramey, C. T., & Ramey, S . L. (1990, April). Prevention of intergenerational mental retardation. Paper presentcd at the Gatlinburg Conference on Research and Theory in Mental Retardation and Developmental Disabilities, Gatlinburg, Tennessee. Reis, J., Barbara-Stein, L., & Bennett, S. (1986). Ecological determinants of parenting. Family Relations, 35, 541-554. Reis, J . S., & H e n , E. J. (1987). Correlates of adolescent parenting. Adolescence, 22(87), 599-609. Rickel, A. U. (1989). Teen pregnancy and parenting. New York: Hemisphere Publishing Company. Rogeness, G. A., Ritchey, S., Alex, P. L., Zuelzer, M., & Moms, R. (1981). Family patterns and parenting attitudes in teenage parents. Journal of Community Psychology, 9, 239-245. Roosa, M. W., Fitzgerald, H. E., & Carlson, N. A. (1982). Teenage and older mothers and their infants: A descriptive comparison. Adolescence, 17, 1-17. Roosa, M . W., & Vaughn, L. (1984). A comparison of teenage and older mothers with preschool age children. Family Relations. 33, 259-265. Rose, S. A , , Feldman, J. F., Wallace, 1. F., & McCarton, C. (1991). Information processing at I year: Relation to birth status and developmental outcome during the first 5 years. Developmental Psychology. 27, 723-737. Russell, C. (1980). Unscheduled parenthood: Transition to “parent” for the teenager. Journal of Social Issues, 36, 45-63. Scarr, S., & Carter-Saltzman. L. (1983). Genetics and intelligence. In J. L. Fulles & E. C. Simmel (Eds.), Behavior genetics: Principles and applications ( pp. 217-335). Hillsdale, NJ: Erlbaum. Schaeffer, C., & Pine, R. (1972). Pregnancy, abortion and the developmental tasks of adolescence. Journal of American Academy of Child Psvchiatry, 1 1 , 51 1-536. Schilmoeller, G. L., & Baranowski, M. D. (1985). Childrearing of firstborns by adolescent and older mothers. Adolescence. 20, 805-822. Schneider, S. ( 1982). Helping adolescents deal with pregnancy: A psychiatry approach. Adolescence, 17, 285-292. Slade, A. (1987). Quality of attachment and early symbolic play. Developmental Psychology. 23, 78-85. Sommer, K . S., & Rellinger, E. (1989, March). Cognitive readiness and its influence on the parenting behavior of adolescenr and adult mothers. Paper presented at the Gatlinburg Conference on Research and Theory in Mental Retardation and Developmental Disabilities, Gatlinburg, TN. Sroufe, L. A,, Fox, N. E., & Pancake, V. R. (1983). Attachment and dependency in developmental perspective. Child Development, 54, 1615-1627. Steinlauf, B. (1979). Problem-solving skills, locus of control, and the contraceptive effectiveness of young women. Child Development. 50, 268-27 I . Stevens, J. H. (1984). Child development knowledge and parenting skills. Family Relations. 33, 237-244. Stevens, J. H., Jr., & Duffield, B. N. (1986). Age and parenting skill among black women in poverty. Early Childhood Research Quarterly. I , 221-235.

196

John G . Borkowski et ul.

Tamis-LeMonda, C. S., & Bornstein. M. H. (1989). Habituation and maternal encouragement of attention in infancy and predictors of toddler language, play, and representational competence. Child Development. 60, 738-75 I . Teberg, A. J . . Howell, V. V., & Wingert, W. A. (1983). Attachment interaction behavior between young teenage mothers and their infants. Journal of Aclulescenr Health C a r e , 4 , 61-66. Thompson, M. S. (1982). Adolescent morherhoodc Social supporr and parenring. Unpublished doctoral manuscript, University of Wisconsin-Madison. Unger, D. G., & Powell, D. R. (1980). Supporting families under stress: The role of social networks. Family Relations. 29. 566-574. Unger, D. G., & Wandersman, L. P. (1985). Social support and adolescent mothers: Action research contributions to theory and application. Journul of Social Issues, 4 / ( I ) , 29-45. Ventura. S. J. (1980). Trends and differentials in births to unmarried women: U.S. 1970- 1976. Vitul Health Statistics, Series 21, No. 36. Ventura, S . J., & Hendershot. G. E. (1984). Infant health consequences of childbearing by teenagers and older mothers. Public Health Reports. 99, 138. Vukelich, C., & Kliman, D. S. (1985). Mature and teenage mothers’ infant growth expectation and use of child development information sources. Family Relutions, 34, 189- 196. Wandersman, L., Wandersman, A , , & Kahn, S. (1980). Social support in the transition to parenthood. Joirrnal of Community Psychology. 8,332-342. Weinrdub, M . , & Wolf, 9. M. (1983). Effects of stress and social support on mother-child interactions in single- and two-parent families. Child Development. 54(5), 1297- I3 1 I . Whitman, T. L., Borkowski. J. G.. Schellenbach, C., & Nath, P. (1987). Predicting and understanding developmental delay of children of adolescent mothers: A multidimensional approach. American Journal of’ Mental Deficiency, 92, 40-56. Wurtz, A. C. ( 1989). Personality and its impuct on adolescent parenting. Unpublished master’s thesis, University of Notre Dame.

Longitudinal Research in Down Syndrome JANET CARR DEPARTMENT OF PSYCHOLOGY ST. GEORGE’S HOSPITAL MEDICAL SCHOOL UNIVERSITY OF LONDON LONDON sw17 ORE, ENGLAND

I.

INTRODUCTION

Down syndrome, discoverable at birth and the most common single cause of mental retardation, has been a prime target of longitudinal investigations. These have been of two main kinds: (1) those that aim at charting the patterns of natural development of the children; (2) those that document the influence of intervention programs on the children’s progress. Each is briefly reviewed and succeeded by an outline of the writer’s 21-year prospective study and a more detailed consideration of some of the longitudinal findings from it. The review refers to the following criteria for high-quality longitudinal research, and attempts to assess the extent to which the investigations discussed conform to them: 1. All available subjects in a given geographical area and time span should be included. 2. The research should continue over a matter of years rather than months. 3. All subjects should be examined on every occasion. 4. Subjects who drop out should be accounted for. 5. Measures should be taken on several, rather than on only two, successive occasions.

Where intervention is involved, five further points obtain: 6. The subjects should be carefully assessed at the outset of the experiment. 7. The subjects should be randomly assigned to experimental and contrast groups. INTERNATIONAL REVIEW OF RESEARCH IN MENTAL. RETARDATION. Vul. I X

I97

Copyright 0 1992 by Academic Press. Inc. All rights of reproduction in any lorn reserved.

198

Janet Carr

8. Subsequent assessment should be carried out by researchers who are “blind” to the subjects’ groups. 9. The regime for each group should be clearly described. 10. Follow-up beyond the termination of the intervention should be reported.

These are rigorous requirements and some at least may not always be attainable. Nevertheless, if they are transgressed the reasons for this and any effect it may have on longitudinal findings should be explicitly considered and discussed.

II.

NONINTERVENTION RESEARCH

Early papers provided consistent information on the decline of IQs in young children with Down syndrome, both hospitalized and home reared. Dameron (1963) followed 12 hospitalized infants from 3 to 18 months. Each was tested within 4 days of the birth day but the homogeneity of this sample is attenuated by numbers that fluctuate between I 1 and 5. Mean developmental quotients (DQs) were 59 at 12 months and 46 at 18 months, well below those of home-reared children (described later), findings that were repeated subsequently (Shipe & Shotwell, 1965; Shotwell & Shipe, 1964; Stedman & Eichorn, 1964). In homereared samples mean scores decreased from 70 in the first year, to 60 in the second, 50 in the third, and 45 in the fourth year. (Koch, Share, Webb, & Graliker, 1963; Share, Webb, & Koch, 1961). These means, however, were commonly derived from individual results gained at different ages and then grouped around chronological ages (CAs); the children were clinic attenders whose initial test results happened to be available, with no information given as to why or at what ages they had been tested. Numbers diminished with increasing age, no explanation being given (Fishler, Share, & Koch, 1964). Dicks-Mireaux (1972) appears to have maintained the numbers in her research but this goes only as far as 2 years of age. Melyn and White (1973) amassed findings from 532 children, from 1 to 14 years of age, who had attended an outpatient clinic and been given a wide variety of tests. “Most” are said to have been tested on a longitudinal basis, but after 7 years the number dropped to less than half the original (76), and at 12, 13, and 14 years, only 9, 6, and 4 children are represented. More recently, Berry, Gunn, and Andrews (1984) have provided information on 39 Australian children with Down syndrome, tested between 2$ and 48 months. “Most” children born in one area over a 4- to 5-year span were included. The children were not all assessed at the same intervals on the same number of occasions; each child is estimated to have had an average of nearly five assessments spread over 12 occasions up to 6 years of age. The data are presented only as scattergrams, making it impossible to compare them with any other data.

LONGITUDINAL RESEARCH IN DOWN SYNDROME

199

Thus, of the nonintervention projects reviewed, all continued over a matter of years and contained measures taken on several successive occasions. Most have not examined all subjects on every occasion (Berry et al., 1984; Dameron, 1963; Fishier et al., 1964; Melyn & White, 1973) nor given reasons for omissions. None included all available subjects in a geographical area and time span or accounted for missing subjects. The research gives some consistent information on the early development of children with Down syndrome, but very little is available after about age 7. The majority of this research is so insecurely based that, without access to the original data, few longitudinal conclusions can be drawn.

111.

INTERVENTION STUDIES

Excellent reviews of the remedial effectiveness of intervention projects already exist (Farran, 1990; Gibson & Harris, 1988). Here such projects are considered for their longitudinal validity, assessed on the 10 points outlined above. One of the earliest interventions was that of de Coriat, Theslenco, and Waksman (1967). Details are given of the 12-month stimulation program provided and the 6-monthly assessments made over a total of 5 years. The intervention group initially comprised 30 subjects, rising to 80 at 18 months and diminishing to 9 by 5 years. Numbers in the contrast group ranged from 182 to 15. No information is given on how the children were allocated to their groups nor on the identity of those carrying out the assessment. Subsequently, 16 children between 21 and 69 months of age were divided into two groups matched for age and sex and for mean mental age (Aronson & Fallstrom, 1977). One group received a half-hour of training twice a week for 17 months; the other received the normal preschool regime. All the children were tested on the Griffiths (1954) scale before, during, and after the training period and again 1 year later. Ludlow and Allen (1979), in a potentially valuable investigation, employed an intervention group and two contrast groups, one home and the other institution reared. The stimulation program continued for 2 to 3 years and the children were then followed until age 10. The groups were large, containing respectively 72, 79, and 33 children, but apart from one occasion the full groups are not represented at any age, numbers in the intervention group varying from 25 at 1 year, to 72 at 4 years to 28 at 10 years. Moreover, the children were tested on average only about once in 2 years and “interpolated quotients” (gained by connecting measured points for any child by a straight line and estimating intermediate figures from this) are included. These fictitious figures are not acceptable although the authors state the procedure was preferred by the statisticians consulted. A general description of the intervention regime, but not of the conditions for the other two samples, is given. It is not true, however, as Farran (1990)

200

Janet Carr

avers, that parents of the honie-reared contrast group “had actually refused to participate in the early intervention project” (p. 526); Ludlow and Allen (1979) make it clear that this applied to only four (5%) of these families (p. 30). Piper and Pless (1980) assigned 37 infants with Down syndrome to different groups according to time of entry into the program. Twenty-one infants available from July to December and 16 infants available from March to June were allocated to treatment and no treatment groups, respectively. The former had biweekly center-based 1-hour sessions and their mothers were given instructions to follow between visits. The children were tested “blind” on the Griffiths scale at the outset and at the termination of the project, 6 months later. Cunningham and his associates in Manchester have been engaged in a largescale ongoing study of many facets of intervention, including intensity, frequency, and the age at which the intervention was begun (Cunningham, 1987). The population consisted of about 90% of all births in the area and was judged a representative sample. Means on the Bayley and Binet tests are given from 1 to 5 years, the number of children involved varying from 59 to 112. In one substudy of the timing of visits (Cunningham, 1982) and in another of intensity of training (Cunningham, 1983) 40 and 15% respectively of assessments were done “blind” as to the child’s status, the latter comparison revealing no evidence of bias. The considerable body of data that emerge was subjected to meticulous longitudinal analysis. Children in a recent American program were estimated to constitute about onethird of those in the geographical area and “assumed” to be representative of its population (Pueschel, 1984). A program of motor and sensory stimulation was developed for each child and taught to the parents through demonstration and written guidelines. Between 75 and 88 children were tested biannually on the Bayley scales from 6 months to 3 years, DQs throughout being between 55 and 58 (Schnell, 1984). No controls were used or referred to, attention being given primarily to within-group variables such as cardiac involvement and parental “follow-through” (derived from ratings made at each visit of how well the parents were carrying out the program of Stimulation provided). The policy of enrolling all children with Down syndrome born in Jerusalem in a stimulation program offered an opportunity to study a true population sample (Sharav & Shlomo, 1986). E’ifty-one children were included. A nonspecific description of the intervention program is given, consisting of home visits by an occupational therapist at about 6 weeks, followed after 18 months by a kindergarten milieu with extra speech and physiotherapy. One hundred and eighty-nine assessments were made over 19 occasions from 3 months to 13 years, a maximum of 22 children appearing at 1 and 2 years and 4 or fewer children represented after the age of 6. The findings were compared with those from two previous studies, by Carr (1970) and by Dicks-Mireaux (1972); however, because in Jerusalem infants with severe cardiac and other malformations were

LONGITUDINAL RESEARCH IN DOWN SYNDROME

20 1

excluded, the conclusion that they were superior to Carr’s group, which included such children, is inappropriate. To summarize, all the intervention projects assessed the children at the outset; all except one (Piper & Pless, 1980) continued over some years and included measurements taken on several occasions. All except two (Piper & Pless, 1980; Schnell, 1984) followed up their sample at least one year beyond the termination of the intervention, although the small numbers reported by Sharav and Shlomo (1986) detract from the value of the latter part of their follow-up. In no case were subjects randomly assigned to experimental or contrast conditions, but in one study the groups were compiled geographically (Ludlow & Allen, 1979) and in another, temporally (Piper & Pless. 1980). implying a degree of nonselectivity in their composition. Otherwise matched controls (Aronson & Fallstrom, 1977), noncontemporary controls (Cunningham, 1982; Sharav & Shlomo, 1986), or no controls (Schnell, 1984) were used. Cunningham is, however, unique in employing specific controls for specific purposes. No study gave a clear description of the regime experienced by the controls; all gave some description of that for the intervention children, varying from minimal (Schnell, 1984; Sharav & Shlomo, 1986) to more detailed (Aronson & Fallstrom, 1977; Cunningham, 1987). Finally, in only one study were all (Piper & Pless, 1980) and in another a proportion (Cunningham, 1982, 1983) of the assessments carried out blind. None of the research reviewed fulfills all the criteria. The most satisfactory studies are those of Cunningham (1987) and Aronson and Fallstrom (1977); the most fallible, that of Sharav and Shlomo (1986). A judgment of the methodological strength of any piece of research can be made only on the basis of information supplied; researchers who give clear details of negative, as well as positive, aspects of their work may be penalized, whereas others who fail to provide such details may escape. It is nevertheless apparent that much of both intervention and nonintervention research suffers from serious methodological weaknesses, and that these limit the credence that can be given to their findings.

IV.

THE SURREY STUDY

The Surrey study is of a population-based cohort of Down syndrome subjects. All babies born in one year, December 1963 to November 1964, in part of southeast London (the boroughs of Camberwell and Lewisham, 23 infants) and in the county of Surrey (31 infants) were included. Camberwell and Lewisham are densely populated, city areas, whereas Surrey is predominantly suburban and rural. Because more children came from Surrey than from Camberwell and Lewisham, the project is referred to as the Surrey study. In the first instance the intention was to provide developmental data to complement neurological and chromosomal investigations in the first 10 months of life (Cowie, 1970). Subse-

202

Janet Carr

quently it was decided to extend the developmental study to 4 years, to include nondisabled controls for the home-based children, and to explore, through interviews, the effects on the families of either type of child. Fifty-four babies with Down syndrome were referred, 25 males and 29 females. Careful inquiry over the years has not revealed any missed cases. Fortyfive children were brought up in their own homes, at least until the 11-year assessment, and nine in various foster homes. The children were seen at home at 6 weeks, 6 months, 10 months, 15 months, and 2 , 3, 4, 11, and 21 years. Psychological tests were given on every occasion, and at 15 months, 4 years, 1 I years, and 21 years, the mother (or principal caretaker) was interviewed. Visits were scheduled to occur, at 6 weeks, within 3 days either way of the birth date and, at 6 months to 4 years, within 2 weeks either way. Over 80% were seen within the limits (Carr, 1975, p. 13). At I 1 and 21 years, all subjects were seen within a 3- to 6-month period. All the children were tested on every occasion apart from those who died and 2, 1 , and 3 children who refused parts of the 15month, 3-year, and 4-year tests (Carr, 1975, p. 14). By 21 years, 41 subjects remained in the study: all losses were due to death (5 by 2 years, 2 by 4 years, 1 by 1 I years, and 2 by 21 years) apart from one child whose family refused contact after the first visit (this child was not included in any analysis), one whose family emigrated before she was 2 years old, and one young man whose mother refused contact at 2 1 because of her anxieties about computerized data. Each home-reared child was matched for sex, age, and social class with a nondisabled child; this contrast group was tested, from 6 weeks to 3 years, and the families were interviewed on every occasion as were those with Down syndrome. This contrast group provided contemporary early developmental test data with which to compare those from the children with Down syndrome; in the family study they allowed identification of those problems that arise in rearing a child with Down syndrome as distinct from problems encountered by any family. Because they are not relevant to the topics under discussion in this paper the contrast group will not be referred to further. Formal support for the project was ended at 4 years of age; however, contact was maintained (through Christmas cards) with the Down syndrome families and, at 11 and 21 years, it was possible to follow up both groups. All testing, at every age, for both groups was done by the author. At 1 I years (Carr and Hewett, 1982), Dr. Sheila Hewett interviewed 37 of the 43 Down syndrome families; all other interviews were carried out by the author. The instruments used to assess the children’s abilities were, in the early years, the Bayley scales of Mental and Motor Development (Bayley, 1964); at 11 years, the Merrill-Palmer Scale of Mental Tests ( 1948) and the Reynell Developmental Language Scale (Reynell, 1969); and at 21 years, the Leiter International Performance Scale (Leiter, 1980) and a range of language and academic tests (see Carr, 1988a). The interview schedules used to collect data on the families were based

LONGITUDINAL RESEARCH IN DOWN SYNDROME

203

on, at 15 months, those devised by Newson and Newson (1963) and, at 4 years, those devised by Hewett (1970); at 11 and 21 years, items from the Health, Behavior and Skills (HBS) schedule (Wing, 1980) were included. These are properly devised and commonly used measures, enabling comparisons to be made with other investigations. The strengths of this study are that it concerns a population-based 1-year cohort of children with Down syndrome and their families, with a matched comparison group, who have been followed by a single investigator for a long period, virtually all attrition resulting from death. The weaknesses are, first, that the sample is a relatively small one and, second, that there are periods of several years in which no testing was done. After seven test occasions in the first 4 years, there was a 7-year interval to age 11 and a 10-year interval to age 2 1. These gaps occurred because continuous funding was not available for the project, the two later follow-ups being made possible only through the provision of brief periods of study leave to the author. Nevertheless, it is believed that no other study currently exists that includes both a Down syndrome group and a contrast group, is concerned with both children and families, and has followed these subjects from infancy to adulthood.

A.

Cognitive Aspects

1 . RESULTS

Figure 1 illustrates the results on developmental and intelligence tests for the whole cohort,' showing that mean ratio IQs fell from 80 at 6 months to 45 at 4 years and 37.2 at 11 years. The decline in scores was most rapid in the first 2 years, an average of 17.5 ratio IQ points per annum, whereas the average between 2 and 4 years was 5 points and that between 4 and l l years, 1.1 points per annum. At 21 years mean IQ rose to 41.9, an increase of 4.7 points from the mean at 11 years. Correlations between mental ages from 6 weeks to 4 years were insignificant for the early tests, but from 10 months onward were significant, reaching 0.92 between 3 and 4 years (Carr, 1975). Correlations between IQs at 11 and 21 years with those at all other ages were low until 10 months, between 0.10 and 0.43, confirming the failure of early infant tests to predict later development (Bayley, 1955). At 15 months the correlations rose and, from 2 years onward, were between 0.61 and 0.83. Thus, in this population mean IQs remained very stable over long periods, tests even in early childhood predicting with considerable accuracy scores of the same group in middle childhood and early adulthood, albeit on different tests. 'Figures for the 41 surviving to age 21 are very similar, being two points higher at 6 and 10 months, one point higher at 3 years, and identical at all other ages.

204

Janet Carr

'"1 -

0-0

6 highest at 21 whole cohort 6 lowest at 21

0-0

0-0

90

80-

7060-

a

50 -

40 30 -

20

~

10-

'

l U l +

4

11

OJ

1.5

6

10

months

15

2

3

21

years

FIG. I . Mean IQs and ranges, 6 months to 21 years.

Between 6 months and 4 years the scores of the non-home-reared children were significantly below those of the home-reared children, but neither at 1 I nor 21 years was any difference apparent. Mean scores of the females were consistently above those of the males, and by 21 years this difference was significant. Not only did females predominate in the upper quartile (Carr,1988a) but Fig. 2 shows that, over the whole range, females more frequently scored at the upper and males at the lower levels. Similar distributions are seen at other ages also (obtainable on request from the author). At no stage has there been a significant difference in IQ between children from the middle- and working-class family groups, although those from the workingclass families have consistently had a slight advantage. Academic skills (reading, writing, and arithmetic) at 21 years were significantly and positively associated with social class and were significantly correlated with IQ from 2 years onward; for example, correlations between 21-year reading level score and IQ at 2, 3, 4, 1 1 , and 21 were, respectively, 0.57, 0.57, 0.62, 0.71, and 0.65. This apparent paradox of significant relationships between academic skills and both IQ and social class when the latter two are not related may be explained by missing data;

205

LONGITUDINAL RESEARCH IN DOWN SYNDROME

60

Males 50

Females

40

0

B

E

30

a

20

10

I 61-70

C

21-30

31-40

41 -50

51-60

IQ FIG. 2. Distribution of 21-year IQs by sex (percentages).

all the subjects were involved in calculating the association between IQ and social class, but when calculating the association between academic skills and social class many subjects (e.g., 10-12 for reading) could not be included as they failed to score on the tests. The relationship of IQ to parental education was also explored, using two measures: (1) years of education reported by each parent and (2) parental educational level, a score of 1 representing education that did not go beyond secondary school, 2 any further training that did not include academic qualifications, and 3 training that included academic qualifications. Parental educational factors were significantly related to social class, but in this population neither showed a clear

206

Janet Carr

association with child ability. This agrees with the findings of Bennett, Sells, and Brand (1979) but conflicts with the positive association between child IQ and parental education reported by Cunningham (1987) and by Fraser and Sadovnick (1976). Multivariate analyses were undertaken to explore the relative importance of the effects of social class, place of rearing, birth weight, illness, health, hospitalization, vision, hearing and weight, parental age at birth, parental education, mother’s scores on the Malaise scale (Rutter, Tizard, & Whitmore, 1970), and “telling,” a score derived from mothers’ descriptions of how well or badly the news their babies’ condition was broken to them. None of these factors had any demonstrable effect on IQ scores except social class at 2 years, the advantage being to the working-class children. When previous IQ was added into the equation, this was shown to be by far the most important predictor of later IQ. The major factor predicting IQ from 15 months onward was previous 1Q and no other factor contributed to any great extent. The scores of two subgroups, those at the extremes at 21 years, were selected for comparison (mean IQs of each of these groups are included in Fig. 1). The six young people who had the lowest scores and were profoundly disabled by 21 years have been described elsewhere (Carr, 1988b), and to balance these an examination has been made of those six (15% of the whole population at 21 years) who gained the highest scores at 2 1 years. (The means for 6 weeks are not given as only two in the profound group and four in the high group were seen at that time.) The two sets of scores run on fairly even lines, the gap between them gradually widening from 6 points at 6 months to 24 points at 3 years. Thereafter, the scores of the profound group continued to decline but those in the high group rose minimally at 4 years, declined slightly to 11 and rose, by 1 I points, to a mean of 60 at 21 years. Means for the high group were consistently above the means of the cohort as a whole, and those for the profound group, consistently below. Nevertheless, the two groups overlapped somewhat, especially in the early tests. At 15 months the score of one subject in the profound group exceeded those of three subjects in the high group; there was no overlap at 2 and 3 years, but at 4 years, the score of this same girl exceeded (by 1 point) that of one subject in the high group. Thereafter no overlap was seen. About half the children with very high or very low early (10-month) scores continued in those positions. The majority of the remaining subjects had scores that varied somewhat, but one girl’s scores changed dramatically, moving from an early position in the high group to the profound group at age 2 1. This girl became phobic and withdrawn at about 9 years of age, but to date no other factor such as health or family circumstances can be adduced to account for this massive change. The reasons for the differences between the groups are hard to tease out. Of those in the high or low groups at age 21, females predominated in the high

LONGITUDINAL RESEARCH IN DOWN SYNDROME

207

group and males in the low group. Social class was similar in each, four being middle class, one working class, and one placed out of home in early life. Heart, vision, and hearing problems did not differ between the groups (the young woman with the highest score at age 21 was deaf and wore a hearing aid); however, the profoundly disabled group had more health problems; at age 21 they had had an average of three hospital appointments and two spells as an inpatient in the last year and 40 days as an inpatient since age 1 1, compared with average figures of 1.3,0.2, and 5 , respectively, in the high group. Three subjects in the low group had two or more severe medical problems (such as very poor vision or hearing, fits, and heart or kidney problems); one subject in this group had two, one had three, and one had four such problems compared with only one subject in the high group who had two of these severe problems. Although these problems may have contributed it seems unlikely that they would be sufficient to account for the low group’s more severe impairment. Graphs of IQ scores over time for groups of people with Down syndrome, such as those in Fig. 1, commonly show a relatively smooth, declining curve. This is composed of means derived from the scores of individuals, which may deviate considerably from this pattern. Although IQs are relatively stable in Down syndrome populations, with high correlations from one age to another, individuals do not necessarily make steady progress but may vary in their performance from one age to another. It is not possible to predict with certainty how any one will develop, although approximately half may be expected to remain at or near the positions they originally occupied relative to the group. 2. DISCUSSION

Five main points emerge from this part of the investigation. First, mean IQs declined steadily from infancy to childhood, increasing slightly at age 21. Second, this group trend did not necessarily mirror the curve of scores for individuals, some of whom demonstrated large changes. Third, the deleterious effect of out-of-home rearing was much less evident at ages 11 and 21 than it had been earlier. Fourth, females had consistently higher mean scores than did males. Fifth, neither social class nor parental education was significantly associated in this group with IQ. Turning to a more general discussion of these points, the curve of scores for the whole group seen in Fig. 1 is, in its early stages, similar to that shown in other nonintervention studies (Bayley, Rhodes, & Gooch, 1966; Dicks Mireaux, 1972; Fishler et al., 1964; Shipe & Shotwell, 1965; Shotwell & Shipe, 1964). This decline has been variously suggested as resulting from deterioration of cerebral function, artifacts of test construction or content, or specific neuromotor and sensory disabilities emerging with age, but no definitive conclusion has been reached (see Gibson, 1978, pp. 30-34). Intervention studies have shown a less marked decline in scores (Cunningham,

208

Janet Carr

1987; Schnell, 1984), this effect being commonly reported while the intervention was ongoing. Cunningham (1987) has documented the failure of such effects to survive the termination of the intervention, although children in the intervention group were found to be healthier and more likely to have attended a mixed/mainstream preschool; also, their mothers were more likely to be employed and willing to leave the child unsupervised for long periods. These findings, both those on child health and schooling and those directly concerned with the mother, reflect findings from other long-term intervention studies, in the light of which it has been postulated that a more lasting effect of intervention may be seen in parental attitudes and motivation than can be observed in the children themselves (Clarke & Clarke, 1981, p. 354, 1989, pp. 292-293). At the young adult level mean Leiter IQ is similar to that reported from a group with an average CA of 28 (Holmes & Cam, 1991). Berry, Groenweg, Gibson, and Brown (1984) present results from 28 adults, between 15 and 42 years of age, (mean CA = 21), who gained a mean mental age of 5 t on Raven’s Coloured Progressive Matrices. This corresponds exactly to the 21-year mean mental age in Surrey, when the six profoundly disabled young people who would not have been able to attempt Raven’s test are excluded. In the Berry et al. study mean mental age rose from 54 to 74 six years later, at a mean CA of 27 years, more than twice the increase seen from 11 to 21 years in Surrey. In the former study, however, this gain was demonstrated on identical tests. Although the possible remedial effects of the program in which the group was involved must be taken into account, it appears that a real gain has been demonstrated. In another study nearly half of a group of 39 adults with Down syndrome, reassessed after 6 i years, had increases in mental age, most of the increase coming from those in their twenties and thirties (Fenner, Hewitt, & Torpy, 1987). These findings demonstrate the importance of charting intelligence levels of adults with Down syndrome throughout their third and fourth decades of life and beyond, especially if in later years comparisons are to be drawn with previous levels to establish the process of aging in this population. If IQ can be shown reliably to increase beyond the second decade, this needs to be taken into account when documenting any later decline. Identifying differences between the scores of children brought up within their own homes and those reared elsewhere was a prime concern of some of the early studies (Bayley et al., 1966; Dameron, 1963; Shipe & Shotwell, 1965; Shotwell & Shipe, 1964; Stedman & Eichorn, 1964). In all these studies marked differences were seen between the groups, to the advantage of those brought up within their own homes, as indeed was found in the early stages in Surrey. By 11 years, however, and again at 21 years, there was no difference in nonverbal IQs, although there was a suggestion of some superiority in language for those brought up at home (Carr, 1988a). It should be remembered that the non-homereared Surrey children were brought up in small homes and foster homes, not in

LONGITUDINAL RESEARCH IN DOWN SYNDROME

209

large institutions. Nevertheless, the specific effect on IQ of in-home rearing may be less pronounced in the long term than it is in the early ages. The Surrey finding of higher scores by females is mirrored in other longitudinal (Cunningham, 1987; Schnell, 1984) and nonlongitudinal (Clements, Bates, & Hafer, 1976; Connolly, 1978; Gath & Gumley, 1984) studies. Gibson (1978) suggests that this difference is due to selective mortality, the more severely disabled females dying earlier (p. 109), but this conjecture could not be substantiated in the Surrey study (Carr, 1988a). Others have pointed to the greater linguistic facility of females as an explanation for these differences (Cunningham, 1987; Schnell, 1984). Cunningham suggests that this “may not be some inherent factor but related to interactional style within families” (p. 175), paralleling ideas put forward about the same phenomenon in nondisabled children (Newson & Newson, 1977, p. 186). This hypothesis deserves exploration. Nevertheless, in Surrey the differences between the scores of the sexes increased steadily over the years and was largest at age 21, when the test used required no language, either verbal or expressive; the females’ superiority on tests of language was not significant when allowance was made for IQ. No study, so far as is known, has shown males to be superior. Bearing in mind the distribution of scores seen in Fig. 2 and despite the need, in view of the small numbers, for this finding to be replicated, the possibility must be entertained that female sex confers some advantage to populations with Down syndrome. In Surrey, neither social class nor parental education was associated with IQ. In contrast, Cunningham (1987) found both factors to be significantly associated with child mental age, parents of higher social class and with higher educational levels having children with higher mental ages. In Surrey, the direction of the social class effect, although not significant, was the reverse of that seen by Cunningham, with a slight advantage to children and adults from working-class families. Despite measures of parental education being similar in the two studies, the significant effects on IQ seen in Manchester could not be seen at any age in Surrey. In seeking an explanation of these discordancies, we note that an obvious difference between the two studies lies in the proportion of those with profound disabilities: none in the Manchester study (Cunningham, personal communication) and 5% at 4 years rising to 15% at 2 1 in the Surrey study. The Surrey figures were recalculated omitting these children, resulting in a slight shift toward the positive end of the scale; for example, correlations of IQ with parental educational level went from between -0.27 and 0.025 to between -0.18 and 0.012, all remaining very small and nonsignificant. It therefore seems that the presence in the Surrey study of people with profound disabilities is not sufficient to explain these differences. Another difference is that of size, the Manchester group being roughly four times larger than its Surrey counterpart; however, the numbers in the latter are

210

Janet Carr

similar to the numbers in other studies in which positive relationships with parental abilities have been reported (Fraser and Sadovnick, 1976). Although larger numbers may have resulted in associations more like those found in Manchester, nothing in the figures as they stand suggests that this is likely to be the case. It may be that such findings will not be seen invariably and that populations of people with Down syndrome, like the individuals within such a population, may vary considerably from each other.

B.

Self-Help and Independence

The degree to which a child or adult can take responsibility for fulfilling his or her own personal needs can have a major effect on the burden of care for the family and on the person’s own sense of independence. In the Surrey study the skills were not observed directly, but questions were asked about them in the course of structured interviews carried out with the parent (usually the mother) or principal caregiver at 15 months and 4, 11, and 21 years. The data presented here focus on four areas-feeding, washing, dressing, and toiletingand on a combined score labeled total independence. 1. RESULTS Figure 3 indicates the proportion of people independent and those totally dependent in each area at 4, 1 1, and 21 years of age. Fairly steady progress was made over the years; most progress was made by age 11 although further improvement was still possible beyond that age. Nevertheless, by age 21 only 18 young people, 44% of the whole group, were fully independent in all four areas. Where full independence had not been achieved washing was the most frequent difficulty for nearly half (49%), problems with hair washing accounting for most of that. After 4 years those totally dependent were very severely and profoundly disabled individuals. Correlations of each self-help score with the others at ages 1 1 and 21 show that at 1 1 years dressing was significantly related to feeding and washing and these to each other ( r = 0.43 to 0.521, but toileting was not related to any other score. Total independence was highly significantly related to feeding, washing, and dressing ( r = 0.72 to 0.81); however, by age 21 all areas, including toileting, were strongly related to each other ( r = 0.70 to 0.86). IQs at 2, 3, 4, 11, and 21 years were each correlated with each self-help score at ages 11 and 2 1. At age 11 all previous IQs were significantly related to total independence, but only IQs at 3 and 4 years were related to washing and, at 3 years, to dressing. None of the previous IQs was related to feeding or toileting. All correlation coefficients, although significant, were small, between 0.43 and 0.56. At 21 years, IQs at all the previous ages were significantly related to all areas. Between 2 and 4 years, the coefficients were between 0.43 and 0.55; from 11 years onward, between 0.6 and 0.78.

21 1

LONGITUDINAL RESEARCH IN DOWN SYNDROME

Feeding (knife & fork) independent

Bathing

loo[ 80 60 40 20

0

-3

'O0

a,

80

C a,

6o

ea,

a

Dressing

4o 20

0

l o80 o[

60 40 20 0 loo

80

- All four areas -

60 40

-

20

-

0

1//,//1

FIG. 3. Independence and dependence in four self-help areas at 4, I I , and 21 years (percentages).

212

Junet Curr

If regression analysis is used to explain the scores at 4 years, IQ at age 3 has some effect on scores for feeding and nighttime continence, but even with sex and social class these account for less than 23% of the variance. IQ has no perceptible effect on total independence. At 1 1 years, total independence is quite well accounted for by 1Q at age 4 (42%), and this is hardly added to by the inclusion of sex and social class. Using 1 I-year IQ in place of that at age 4 results in a dramatic improvement, with 77% of the variance now accounted for; however, substituting the 4-year independence score gives a much reduced percentage, to less than 10%. So total independence at age 1 I can be largely accounted for by IQ, and of the measures available at 4 years, IQ has much greater predictive power than has independence at that time. At 21 years, total independence is well explained by 1 I-year IQ ( 6 5 % )and sex and social class add only another 4%. Using IQ at age 21 in place of that at age 1 1 increases the proportion explained by 9%, but substituting independence at I I increases this to 85%. Of the individual skills at age 21, IQ at age 1 1 made the greatest contribution apart from self-management at the toilet, which is more influenced by the same factor at age I I . With IQ allowed for, no effect could be seen of sex or social class but those young people brought up from infancy out of their own homes had a higher mean score for dressing and for total independence. In general, neither sex nor social class had any denionstrable effect on selfhelp skills, either singly or in combination, at any point. Total independence was difficult to explain at age 4, and was predominantly influenced by IQ at age I 1 and by the previous level of independence at age 21. It seems then that the timing of the acquisition of these skills may be quite variable in young children so that the level they achieve by the middle school years is little related to their achievement in the preschool years but more closely governed by their innate ability. By the middle school age, however, the pattern of children’s skills is more clearly defined and, although still strongly associated with ability level, predicts with considerable accuracy the level of independence they will demonstrate as young adults. 2. DISCUSSION The data from the Surrey study may be compared with those from a number of other studies that have included examination of self-help skills in children with Down syndrome, although few of these studies have been longitudinal. Buckley and Sacks (1987) examined the self-help skills of 90 children with Down syndrome: 40 girls and SO boys, aged 1 1 to 17, divided roughly equally into those over and under the age of 14. There are references to “improvement with age” ( pp. 24, 34) but the study is not a longitudinal one. In general, the authors are careful not to imply progression from one age to another, but discuss the levels of achievement of the “younger” and “older” children. The only significant dif-

LONGITUDINAL RESEARCH IN DOWN SYNDROME

213

ferences between the age groups concerned the preparation of snacks, and the summed self-help scores, both in favor of the older group. Neither sex, health, nor family position had a significant effect on skill development. Elsewhere, a significant increase in self-help with increasing CA has been shown in adolescents with Down syndrome, although not in those with autism (Loveland & Kelley, 1988). Two other British studies give some information on self-help skills. Gath (1985) had not assessed these at earlier ages but, in the 23 children remaining at 9 years, scores on the AAMD Adaptive Behavior scales (covering self-help as well as social function and communication) were related to IQ (r = 0.44) and to birth weight ( r = 0.56) but not to any other factor considered. Byme, Cunningham, and Sloper (1988) do not discuss feeding, washing, or dressing, and toileting and sleep are considered as behavior problems (raising the question of what should be regarded as a behavior problem). Sleep and toileting headed the list of such problems, shown by between 20 and 40% of the children, but longitudinal data are not given. In a later paper (’hmer, Sloper, Knussen, & Cunningham, 1991) an Index of Self-Sufficiency, includlng self-help, domestic, and community skills, was related to a variety of child and family factors. Overall scores for 127 children aged 6 to 14 were significantly correlated with MA (r = 0.71) and IQ (r = 0.52). Also positively related were a lower level of behavioral problems, more social contacts and activities for the child, and the mother’s use of practical coping methods. Cullen, Cronk, Pueschel, Schnell, and Reed (1984) used the Vineland Social Maturity Scale to evaluate self-help skills in a group of 70 to 80 Down syndrome children aged from 6 to 36 months, seen biannually. Mean social quotient declined, though not significantly, from 84 at 6 months to about 70 by 24 to 36 months. The presence of moderate and severe congenital heart disease was a powerful factor, associated with a 3-month deficit on the Vineland at 36 months. Excluding those with moderate and severe heart disease, those with good muscle tone, and those whose parents were able to “follow through” on the guidance given them did better than those with poor muscle tone and those whose parents did not “follow through” ( p. 62). Feeding milestones were particularly carefully assessed and were shown to be passed in the same developmental sequence, although delayed, as in normal children. Sex had no effect, cardiac status affected finger feeding and independent spoon use and, more particularly, chewing and weaning, but parents’ good “follow-through” affected positively all nine feeding variables. The authors considered the question of the effect of “follow-through,” pointing out that the group rated adequate in this respect contained more children with good muscle tone and without significant heart disease. Despite this, “followthrough” was shown to have an independent effect; however, IQ may also have had an effect. IQ is related to “follow-through” (pp. 218-222), and although no

214

Junet Carr

correlations between IQ on the one hand and Vineland and feeding scores on the other are given, the Vineland and feeding scores were all significantly correlated, being “frequently higher than those among the feeding milestones” themselves (p. 248). The implication is that good “follow-through’’ was instrumental in bringing about better self-help skills, but it is also possible that the reverse obtained and that parents were more willing to exert themselves and to “follow through” where they saw the child as responsive and as having a good measure of learning ability. This study demonstrates the extra difficulties imposed by cardiac and muscle weakness on the acquisition of self-help skills by a young child with Down syndrome, and the advantages that accrue when parents enthusiastically embrace the advice they are offered. These effects were demonstrated in the course of a continuing program, and there is no evidence as to how far they would endure beyond its discontinuation. In both Surrey and Manchester (Turner et al., 1991) strong associations were found between self-help and IQ. None of the studies discussed has shown an effect of sex on self-help. This seems somewhat unexpected. If “interactional style within families” is thought to play a part in the higher intellectual abilities of females it might be expected to do the same for self-care skills; however, no such effect is found. Perhaps family interactional style affects intellectual and self-help skills differently, or this finding may diminish somewhat the force of the argument for its effect on intellectual ability.

C. Behavior and Management Important as is the developmental level of a child with Down syndrome, personality, temperament, and behavior are of almost equal salience in determining how well he or she will fit into a family and be accepted by society. Mothers were asked about behavior difficulties-aggressiveness, tantrums, naughtiness, and destructiveness-and how they dealt with these. When their children were ages 1 1 and 21, mothers were also questioned about personality, cooperativeness, and manageability; about behavior problems such as rebelliousness, pestering, self-injury, and public disturbances; and about their disciplinary methods, use of rewards, and attitudes to discipline. 1. RESULTS

Looking at the consistency of behaviors over time, nearly half (45%) of the 1 1-year-olds who had tantrums and half of those aggressive at that time were similarly rated at 21 (though only two of those said to be aggressive at age 4 were said to be so at ages 1 1 and 21). Two-thirds (65%) of those rebellious at age I 1 were said to be so also at age 21, though now only occasionally. Nearly half (47%) of those who pestered for attention but only 19% of those who showed difficult behavior in public at age 11 were still doing so at 21. Few children who at age 1 1 neither pestered nor rebelled nor were difficult in public showed these

LONGITUDINAL RESEARCH IN DOWN SYNDROME

215

behaviors at age 2 1. Self-injury was infrequent in this population, affecting less than one-sixth at any one time; only five ever were head-bangers and in no case was this a serious problem. Four of the five, however, continued to show some degree of self-injury; in the case of two profoundly disabled boys, this led to severe self-scratching. Two who had shown other forms of self-injury at age 11 continued it in a minor form at age 21. Thus, 11 subjects (a quarter of the total) in this population showed self-injurious behavior at some time and this persisted to some degree in 6 subjects, being severe and frequent in 2 profoundly disabled young men. Social class appeared not to be a relevant factor in persistence into adulthood of behavior difficulties, around half being implicated in each group (middle class = 55%, working class = 53%). Rather more females showed persistent behavior problems than males (males = 55%, females = 76%), more females continued to pester (males = 1, females = 6 ) , and more males continued to be difficult in public (males = 3, females = 0). Some differences in IQ appear when sex is allowed for; of the 9 males in whom behavior difficulties persisted, 6 (67%) were below the mean male IQ, compared with 3 of the 12 females (25%) who were below their mean. None of these differences reaches significance. Five of the six profoundly disabled young people had persistent problems, two showing severe self-injury and one of these other problems as well. To sum up, in this population about half of those who were aggressive, who pestered for attention, or who had tantrums and somewhat more than half of those who were rebellious at age 1 1 continued to show these behaviors at age 21. Difficult behavior in public tended to die out. Neither sex nor social class had a significant effect and these behaviors were somewhat more likely to persist in the more severely disabled men and in the more able women. Data were collected at 15 months and 4, 11, and 21 years on how parents handled their offspring’s behavior and their attitude to and use of disciplinary methods (e.g., supervision, smacking, sending to bed early, depriving of sweets or TV, and use of threats and rewards; see Carr, 1992). These factors were individually related to cooperativeness and manageability at ages 11 and 21 and to frequency of behavior problems at age 21. Manageability, rated at 4, 11, and 21 years of age, and cooperativeness, rated at 11 and 21 years of age, were also correlated across ages. The expectation was that although there might or might not be some stability in behaviors across ages, a major effect would be seen in the relationships between the disciplinary approaches used by mothers and the outcome in terms of cooperativeness, manageability, and presence or absence of behavior problems in their children. Considering first the relationships between behaviors across ages, young people with Down syndrome were found easier to manage at age 21 if they had had few tantrums at 15 months, and were more cooperative if they had been easy to manage at age 4 and, at 11, had kept out of mischief and been cooperative.

216

Janet Carr

There was then a tendency for those who were not subject to tantrums, who kept out of mischief, and who were cooperative and easily managed at early stages to be more cooperative and easy to manage later on, indicating some consistency of behavior across ages. Turning to the relationships between the measures of parental discipline and later cooperativeness, manageability, and presence or absence of behavior problems, only one (marginally) significant result was obtained: fewer behavior problems were seen at age 21 where at 4 years there had been little physical punishment for any child in the family. Given the number of associations examined this may have been a chance finding. No other factor nor the combination of individual factors into disciplinary scales at each age could be shown to have any association with later behavior. Finally, associations were explored with a factor labeled mother’s attirude. This was derived from responses to a question asked at ages 11 and 21: “How would you describe N as a person now?” Replies were grouped into those consisting of only positive, both positive and negative, and only negative comments. Because there were so few negative comments, they were combined with the positive and negative group, and this bipolar factor at ages 11 and 21 was related to cooperativeness at each age. Mother’s attitude at age 11 was related to cooperativeness at both 11 and 21 years; mother’s attitude at age 11 was not related to her attitude at 2 1, nor was attitude at 21 related to cooperativeness at 2 1, nor was cooperativeness at 11 related to mother’s attitude at 2 1, suggesting that it was not the child’s cooperativeness at the earlier age that led to a more positive attitude on the part of the mother later. Further, although, as we saw earlier, cooperativeness at age 1 1 was associated with cooperativeness at age 2 1, the association of mother’s attitude at 1 I with cooperativeness at 21 is stronger. These data may be regarded as only indicative, but they suggest that a positive attitude by the mother toward her Down syndrome child may play some part in the child’s later cooperativeness. In summary, in this cohort some consistencies were seen in behavioral ratings from one age to another: almost no effect could be demonstrated for parental disciplinary practices, and cooperativeness at a later age may have been influenced by the mother’s attitude earlier. 2. DISCUSSION Few data exist from other studies of children with learning disabilities with which the currently discussed data may be compared. Richardson, Koller, and Katz (1985), following up a group of young people with mild learning disabilities at ages 16 and 22, did not consider the effect of parental discipline, but concluded that behavior disorder was most likely attributable to the stressful conditions in which the young people grew up. (Similar conclusions, relating child behavior disorders to family discord, are reported from studies of nondisabled

LONGITUDINAL RESEARCH IN DOWN SYNDROME

217

children; see Rutter, 1976.) Byrne et al. (1988), in their longitudinal study of children with Down syndrome in Manchester, found some consistency of behavior, in that 75% of mothers who were concerned about their children’s behavior between ages 2 and 10 were still concerned about their behavior 2 to 3 years later. Efforts to relate the mother’s disciplinary methods to outcome resulted in only two significant associations: less problem behavior where the mother felt happy about her handling of the child ( r = 0.23) and more problems where she had threatened to send the child away ( r = 0.2; Sloper, personal communication). Data from large populations of children without disabilities suggest that poor adjustment in early adult life is related to difficult temperament, behavior disorder, and adjustment of the child and with parental conflict at age 3 (Chess & Thomas, 1984, p. 99). Also, difficult 1 I-year-olds were likely to be difficult at age 16 and later, a child frequently physically punished (once a week or more) at age 1 1 was more likely to be difficult at age 16 and to go on to acquire a criminal record than those not so punished, and this association was maintained even when sex, family size, and social class were controlled (Newson & Newson, 1989). Comparing these data with those from the Surrey study, both the Manchester and, especially, the Newsons’ studies concur with the present one in finding some agreement between the occurrence of problem behaviors at different ages. In the Newsons’ studies, these behaviors were generally less frequent in girls; the same tendency was seen in Surrey apart from pestering, which was more frequent in females. Neither of the relationships seen in Manchester, fewer problems where the mother felt confident and more problems if she has threatened to send the child away, could be detected, but some similarity was seen to the Newson’s findings on the relationship with physical punishment. Twenty-one percent of 1 1year-olds with Down syndrome were physically punished once a week or more often; at 2 I , all relationships-with being difficult to manage, uncooperative and having two or more behavior problems-were in the same direction as in the Newsons’ study, of more problems in those physically punished more often, although the figures do not reach significance. It seems, first, that there is evidence of some consistency in the behavior of children, with or without disabilities and that easy children tend to be easygoing as they get older and those with behavioral difficulties at one age are likely to be difficult later (although there are also many exceptions to this). Second, child behavior problems are associated with instability of upbringing and family discord in both groups, although there was insufficient of either unstable upbringing or family discord in the Surrey study to bear this out. Third, no evidence can be found in the studies discussed to show that frequent physical punishment leads to improvement in the child’s behavior. The evidence, so far as it goes, points to the contrary. Many mothers of children with learning difficulties who use physical punishment do so because they say, “It’s the only thing he understands.” If,

218

Janet Carr

however, it is ineffective in the long run we may need to help the mothers to look for other methods. Besides methods based on learning principles (which have not been evaluated in large-scale longitudinal studies), it may be useful to help mothers see, enjoy, and respond to positive aspects of their children and to generate more positive aspects of the “interpersonal context” of childrearing (Caldwell, 1964), which may have a beneficial influence on later behavior. Of great interest would be the discovery of factors that militated against the ernergence of behavior problems, in the way that Cunningham (1987) suggests that higher parental education “can overcome the boys’ ‘disadvantage’ compared to girls” (p. 174). In their work with nondisabled children Newson and Newson ( 1989) indicate that greater paternal involvement, apart from that connected with discipline, reduced the risk of future behavior disorder; with limited data it was not possible to replicate this finding in the Surrey study.

V.

CONCLUSIONS

Like other groups of children with Down syndrome, those in Surrey had relatively high scores in infancy, the scores decreasing rapidly in early childhood although, in this instance, more slowly after the age of 10 months. As the decline to 1 1 years had been a steady one, the rise at age 21 was the more unexpected. The scores of individuals do not, by any means, follow a uniform downward trend, so an increase might occur by chance in small groups of subjects (Connolly, 1978; Kostrzewski, 1974). A substantial rise has been shown by Berry et al. (1984) in a group of adults, although it is not clear to what extent this was due to the remedial program in which the group was involved. Nevertheless, the downward curve of scores, seen in numerous studies in early childhood, may not inevitably continue. It has long been accepted that people with Down syndrome continue to be able to learn and to develop skills well after leaving school; it may be that, in the course of the learning that occurs in childhood and adolescence, they also acquire techniques, strategies, and insights that can be applied in psychological tests. If this were the case one might expect such learning to be more readily applicable in verbal rather than nonverbal tests; but in the Surrey study the gain was shown on a nonverbal test (the Leiter scale) following one with some verbal content (Merrill-Palmer Scale), and in the study by Berry et al. (1984) the gain on a nonverbal test (RCPM) was substantially greater than that on a verbal test (PPVT-R). Turning from a discussion of the data from the aggregated group to those of individuals within it, one of the most striking features of the study is the wide variation in ability shown by people with Down syndrome. Although recognizing that “IQ” scores do not have the same meaning at the extremities as they do over

LONGITUDINAL RESEARCH IN DOWN SYNDROME

219

the middle part of the range, the distribution of scores, of about 60 points, is of the same order in the Surrey group as for the majority of the more to the less able in a normal population (IQs 70-130). Certainly the difference experienced in a social encounter is at least as great in the former as in the latter case. So a point of particular interest has been to try to discover what determines an individual’s ability level. These efforts have been largely unsuccessful. The association with gender is well established but factors such as parental education and social class, seen as important in other studies, were not found to be so here. Neither has a diligent search among other measured factors produced any answers: although anecdotally the most profoundly disabled could be seen to have suffered from more adverse health factors than the most able, it was not possible to establish that these, or any other, factors had had a significant effect on the scores of the whole group. It is tempting to wonder whether there are differences within the condition itself, as yet unidentified, that may be at least partly responsible for the wide differences, some of which remain quite stable over time, between individuals with an apparently identical condition. Where the development of self-help skills is concerned, overall children with Down syndrome acquire these in the same order as nondisabled children, but more slowly, so that even as young adults nearly two-thirds were not able to look after their own self-care entirely. Most skills were acquired by age 11, after which further acquisition, although still possible, was less often seen. A major finding has been the importance of IQ, the relationship to skill learning being very similar to that relating to academic subjects (Carr, 1988a). Both kinds of skills were more easily acquired by able than by less able children with Down syndrome, although other factors, such as the opportunities and encouragement provided, were seen in individual cases to have been important. Cullen et al. ( 1984) have shown children with physical disabilities to have more cognitive difficulties in the early years; although this is eminently plausible it remains to be seen if these effects are maintained later in life. Although in Surrey the self-help scores of the females were slightly higher, no study has found sex to be a significant factor where these skills are concerned. It may be that the self-help areas under consideration are so basic that there can be no question of their being seen as more appropriate to one or the other sex; all need as much expertise as possible, both to enhance their own feelings of competence and independence and to maximize their capability of functioning in and being accepted by society. It is interesting to speculate how the skills may best be taught to people with Down syndrome. Cullen et al. (1984) point to the importance of the mother’s commitment; the indication in Surrey of the small advantage to the young people who grew up out of their own homes suggests that, if parents are not on hand to help, young people may be better able to develop their competence. Berger and Cunningham (1983) showed that infants’ vocal and

220

Janet Carr

social skills blossomed when their mothers allowed them more time to respond rather than rushing in with help and stimulation. It might be that a similar approach to self-help teaching, with parents and caregivers keeping a lower profile, would pay dividends in this area too. It should, however, be noted that these speculations were not investigated in the Surrey study. One of the most useful functions a longitudinal study like the present one might be expected to serve is a demonstration of the disciplinary methods that are most successful with this population. Many mothers felt uncertain as to how best they should handle their children and uncertain as to whether the methods they used with their other children would be appropriate and, if not, what they should do. Apart from some indication that with these young people, as with others, frequent physical punishment is ineffective and that a harmonious family and positive attitude on the part of the mother may be helpful, almost no confident advice may be proffered. It may, however, be of some comfort to mothers to know that problem behaviors were fewer in this population at age 21 than they had been at age 1 I . Many commented on how much easier their young people were as adults; this was due partly to the gains made in self-help (“he can d o so much more for himself now”) and partly to the decline in some behaviors, for example, being difficult in public, which at age 21 was shown by only three of the profoundly disabled young people. Finally, we may consider whether and how the three areas discussed-IQ, social competence, and behavior-are interrelated. At any one point in time subjects with higher 1Qs were also the more competent socially. Despite the decline in IQ, in this population, at least to age 1 1, social competence increased over time, as did mental age, both reflecting greater social maturity. Regarding behavior, the trend here also was toward improvement with age; this may similarly have been related to better social competence seen in increasing mental age and self-help skills, notwithstanding the tendency, seen also in other studies of normal and disabled populations, for certain personality traits to persist from early to adult life. Since this study began great changes have taken place for people with Down syndrome. Babies whose fanlilies feel unable to look after them are no longer placed in institutions, much progress has been made in the treatment of heart and respiratory conditions, the long-term future for any infant born with Down syndrome today is envisaged as in the community and not, as the Surrey mothers anticipated with horror, in the back ward of a subnormality hospital. All this places limitations on the usefulness of a single cohort study like the present one and the lessons that can be learned from it; circumstances have changed so much that conclusions drawn from the earlier group may not apply to a later one. Nevertheless, it is hoped that the project has something in it of interest to later generations; certainly, the intention is to continue following the present cohort into their thirties and, perhaps, beyond.

LONGITUDINAL RESEARCH IN DOWN SYNDROME

22 1

ACKNOWLEDGMENTS I am grateful to Professors Ann and Alan Clarke, Dr. Beate Hermelin, and Dr. Sheila Hewett for reading and commenting on an earlier draft of this article, to Sean McGuigan for statistical assistance, and to Hermione Holmes for typing repeated redrafts.

REFERENCES Aronson. M.. & Fallstrom, K. (1977). Immediate and long-term effects of developmental training in children with Down’s syndrome. Developmental Medicine & Child Neurology. 19. 489-494. Bayley, N. (1955). On the growth of intelligence. American Psychologisr, 10, 12-15. Bayley, N. (1964). Manual of directions for infant scales of development (Restricted Research Forms). Bethesda, MD: National Institute of Mental Health. Bayley, N.. Rhodes, L., & Gooch, B. (1966). A comparison of the development of institutionalized and home-reared mongoloids. California Mental Health Research Digesr. 4 , 104- 105. Bennett, F. C., Sells, C. J., & Brand, C. (1979). Influence on measured intelligence in Down’s syndrome. American Journal of Mental Deficiency. 133. 700-703. Berger, J., & Cunningham, C. (1983). Early social interactions between infants with Down’s syndrome and their parents. Health Visitor, 56, 58-60. Berry, P., Groenweg, G., Gibson, D., & Brown, R. I. (1984). Mental development of adults with Down syndrome. American Journal of Mental Deficiency, 89. 252-256. Berry, P., Gunn, V. P., & Andrews, R. J. (1984). Development of Down’s syndrome children from birth to five years. In J. M. Berg (Ed.), Perspectives & progress in menral rerardation: Vol. 1. Social. psychological & educational aspects (pp. 167- 177). Austin, TX: PRO-Ed. Buckley. S . , & Sacks, B. (1987). The adolescent with Down’s syndrome: Life for the teenager and f;w the fami1.y. Portsmouth Polytechnic, Portsmouth Down’s Syndrome Trust. Byrne, E. A,, Cunningham, C. C., & Sloper, P. (1988). Families and their children wirh Down’s svndrome: One feature in common. London: Routledge. Caldwell, B. M. (1964). The effects of infant care. In M. L. Hoffman & L. W. Hoffman (Eds.), Review ofchilddevelopment research (Vol. I , pp. 261-298). New York: Russell Sage Foundation

Carr, 1. (1975). Young children wirh Down’s syndrome. London: Butterworths. Carr, J. (1988a). Six weeks to twenty-one years old: A longitudinal study of children with Down’s syndrome and their families. Journal of Child Psychology & Psychiatry, 29. 407-43 I , Carr, J. (1988b). Down syndrome. In J. Hogg, J. Sebba, & L. Lambe (Eds.), Profound rerardation and multiple impairment (pp. 40-53). London: Chapman & Hall. Carr, J. (1992). Assets and deficits in the behaviour of people with Down’s syndrome-A longitudinal study. In B. Tizard (Ed.), Human resilience and vulnerability (pp. 55-69). London: Jessica Kingsley Publishers. Carr, J., & Hewett, S . (1982). Children with Down’s syndrome growing up. Association of Child Psychology & Psychiatry News, Spring, 10- 13. Chess, S., & Thomas, A. (1984). Origins and evolutions of behavior disorders: From infancy to adult life. New York: Brunner-Mazel. Clarke, A. D. B., & Clarke, A. M. (1981). “Sleeper effects” in development: Fact or artifact? Developmental Review, 1 . 344-360. Clarke, A. M., & Clarke, A. D. B. (1989). The later cognitive effects of early intervention. lnrelligence. 13, 289-297. Clements, P. R.. Bates, M. V., & Hafer. M. (1976). Variability within Down’s syndrome (trisomy 21): Empirically observed sex differences. Mental Rerardarion, 14. 30-31.

222

Janet Carr

Connolly, J. A. (1978). Intelligence levels of Down’s syndrome children. American Journal of Mental Deficiency. 83. 193- 196. de Coriat, L. F., Theslenco, L.. & Waksman, J. (1967). The effects of psychomotor stimulation on the IQ of young children with trisomy 21. In B. Richards (Ed.), Proceedings of the first congress of the International Association for the Scientific Study of Mental Deficiency (pp. 377-385). Reigate, U.K.: Michael Jackson. Cowie, V. (1970). A study of the early development of rnongols. Oxford: Pergamon Press. Cullen, S. M., Cronk, C. E., Pueschel, S. M., Schnell, R. R., & Reed, R. B. (1984). Social development and feeding milestones. In S. M. Pueschel (Ed.), The young child with Down syndrome (pp.227-252). New York: Human Sciences Press. Cunningham, C. C. (1982). Psychological and educational aspects of handicap. In F. Cockburn & R. Gitzelman (Eds.), Inborn errors of metabolism (pp. 237-253). Lancaster, England: MTP Press Ltd. Cunningham, C. C. (1983). Early development and its facilitation in infants with Down’s syndrome (Final Report to DHSS, Part 1). Cunningham, C. C. (1987). Early intervention in Down’s syndrome. In G. Hosking & G. Murphy (Eds.), Prevention of mental handicap: A world view (RSM Services International Congress & Symposium Series No. 112, pp. 169-182). London: Royal Society of Medicine Services Ltd. Dameron, L. E. (1963). Development of intelligence of infants with mongolism. Child Development, 34, 733-738. Dicks-Mireaux, M. J. (1972). Mental development of infants with Down’s syndrome. American Journal of Mental Deficiency, 77, 26-32. Farran, D. C. (1990). Effects of intervention with disabled and disadvantaged children. In S. J. Meisels & J. P. Shonkoff (Eds.), Handbook of early childhood intervention (pp. 501-539). Cambridge, U.K.: Cambridge University Press. Fenner, M. E., Hewitt, K. E., & Torpy, D. M. (1987). Down’s syndrome: Intellectual and behavioural functioning during adulthood. Journal of Mental Deficiency Research, 31, 241 -249. Fishler, K., Share, J., & Koch, R. (1964). Adaptation of Gesell developmental scales for evaluation of development in children with Down’s syndrome (mongolism). American Journal of Mental Deficiency, 68, 642-646. Fraser, F. C., & Sadovnick, A. D. (1976). Correlation of IQ in subjects with Down syndrome and their parents and sibs. Journal of Mental Deficiency Research, 20, 179-182. Gath, A. (1985). Down’s syndrome in the first nine years. In A. R. Nichol (Ed.), Longitudinal studies in child psychology and psychiatry ( pp. 203-2 19). LondonKhichester: Wiley. Gath, A,, & Gumley, D. (1984). Down’s syndrome and the family: Follow-up of children first seen in infancy. Developmental Medicine and Child Neurology, 27, 392-400. Gibson, D. (1978). Down’s syndrome: The psychology of mongolism. London: Cambridge University Press. Gibson, D., & Harris. A. (1988). Aggregated early intervention effects: Patterning and longevity of benefits. Journal of Mental Deficiency Research, 32, 1-17. Grifiths, R. (1954). The abilities ofbabies. A study in mental measurement. London: University of London Press. Hewett, S. (1970). The family and the handicapped child. London: Allen & Unwin. Holmes, N., & Cam. J. (1991). The pattern of care in families of adults with a mental handicap: A comparison of families of autistic adults and Down’s syndrome adults. Journal of Autism & Developmental Disorders, 2 I 159- 174. Koch, R., Share, J., Webb, A,, & Graliker, B. V. ( I 963). The predictability of Gesell developmental scales in mongolism. Journal of Pediatrics, 62, 93-97. Kostrzewski, J. (1974). The dynamics of intellectual development in individuals with complete and incomplete trisomy of chromosome group G in the karyotype. Polish Psychological Bullerin, 5 . 153- 158.

LONGITUDINAL RESEARCH I N DOWN SYNDROME

223

Leiter, R. G . (1980). Leiter International Performance Scale instruction manual. Chicago: Stoelting co. Loveland, K. A,, & Kelley, M. L. (1988). Development of adaptive behaviour in adolescents and young adults with autism and Down’s syndrome. American Journal on Mental Retardation, 93, 84-92. Ludlow, J. R., & Allen, L. M. (1979). The effect of early intervention and pre-school stimulus on the development of the Down’s syndrome child. Journal of Child Psychology & Psychiatry. 23, 29-44. Melyn, M. A., & White, D. T. (1973). Mental and developmental milestones of noninstitutionalised Down’s syndrome children. Pediatrics 52, 542-545. Merrill-Palmer Scale of Mental Tests (1948). New York: Harcourt Brace & World. Newson, J., & Newson, E. (1963). Infant cure in an urban community. London: Allen & Unwin. Newson, 1.. & Newson, E. (1977). Perspectives on school at seven years old. Oxford: George Allen & Unwin. Newson, J., & Newson, E. (1989). The extent of parental physical punishment in the U K . London: APPROACH (Association for the Protection of All Children). Piper, M. C . , & Pless, I. B. (1980). Early intervention for infants with Down syndrome: A controlled trial. Pediatrics, 65, 463-468. Pueschel, S. (1984). The young child with Down’s syndrome. New York: Human Sciences Press. Reynell, J. (1969). Reynell Developmental Language Scale. Slough, Bucks.: National Foundation for Educational Research. Richardson, S . A,, Koller, H.,& Katz, M. (1985). Relationship of upbringing to later behavior disturbance of mildly mentally retarded young people. American Journal of Mental Deficiency, 90, 1-8. Rutter, M. ( 1976). Helping troubled children. Harmondsworth: Penguin. Rutter, M., Tizard, I., & Whitmore, K. (1970). Education, health and behaviour. London: Longmans. Schnell, R. R. (1984). Rychomotor development. In S. M. Pueschel (Ed.), The young child with Down syndrome (pp. 207-226). New York: Human Sciences Press. Sharav, T., & Shlomo, L. (1986). Stimulation of infants with Down’s syndrome: Long-term effects. Mental Retardation, 24, 8 1-86, Share, J., Webb, A,. & Koch, R. (1961). A preliminary investigation of the early developmental status of mongoloid infants. American Journal of Mental Deficiency, 66, 238-241. Shipe, D., & Shotwell, A. M. (1965). Effect of out-of-home care on mongoloid children: A continuation study. American Journal of Mental Dejiciency, 69, 649-652. Shotwell, A. M., & Shipe, D. (1964). Effect of out-of-home care on the intellectual and social development of mongoloid children. American Journal of Mental Deficiency, 68, 693-699. Stedman, D. J., & Eichom, D. H. (1964). A comparison of the growth and development of institutionalized and home-reared mongoloids during infancy and early childhood. American Journal of Mental Deficiency. 69. 391-401. Turner, S . , Sloper, P., Knussen. C., & Cunningham, C. (1991). Factors relating to self-sufficiency in children with Down’s syndrome. Journal of Mental Deficiency Research, 35. 13-24. Wing, L. (1980). MRC Handicaps, Behaviour & Skills (HBS) Schedule in epidemiological research. Acta Psychiatrica Scandinavia, 62(Suppl. 285), 241-247.

This Page Intentionally Left Blank

Staff Training and Management for Intellectual Disability Services CHRIS CULLEN PSYCHOLOGICAL LABORATORY UNIVERSITY OF ST. ANDREWS ST. ANDREWS, FIFE K Y 1 6 9 J U , SCOTLAND

I.

INTRODUCTION

Services for people with intellectual disabilities are in transition in many countries. For several decades it has been acknowledged that service users are among the least powerful and most devalued groups in society, and there are many situations and occasions when they suffer abuse and a poor quality of life. Sometimes these are evidenced in the form of nationwide scandals that reveal instances of neglect, poor treatment, abuse, and the intimidation of those who “blow the whistle” (cf. Beardshaw, 1981; Blatt & Kaplan, 1 9 6 6 ) . Often these problems are seen in the findings of researchers-frequently buried in academic journals-that show that people with intellectual disabilities, especially those receiving residential services, have a relatively impoverished life-style. For example, Cullen, Burton, Watts, and Thomas (1983) observed staff and 10 residents of a medium-sized institution (approximately 120 residents) in Britain and found that the behavior of residents received no attention of any kind, from either staff or other residents, more than 90% of the time. Less than 10% of the residents’ behavior was judged to be socially appropriate but, on the average, a staff person responded to a resident’s behavior only once every 25 minutes, that interaction lasting less than 10 seconds. Hile and Walbran (1991) recently conducted a more extensive observational study in an American institution and obtained similar results. A common “solution” to the problems that such findings highlight has been to advocate for institutional closure and for community-based residential services. Bronfenbrenner (1979) describes this as the “social address” model of service provision; what is thought to be most important is where a person lives. We now, INTERNATIONAI. REVIEW OF RESEARCH IN MENTAL RETARDATION. Vul I 8

225

Copyright 0 lW2 hy Academic Prcv. lnc All righls of reproduclion in any form mewed.

226

Chris Cullen

however, have enough experience in community services to know that “there is more than a zip code to changes in services” (Bruininks, 1990, p. 13). Although moving from a large institution to a smaller community residence may be a necessary condition for an improved quality of life (and even this is open to debate, cf. Williams, 1991; Zigler, Hodapp, & Edison, 1990), it is rarely a sufficient one. Flynn (1989), for example, interviewed 88 people who had been discharged from an institution in the northwest of England and who were now living in their own homes. Professionals described almost 75% of the group as living a “successful” life, but 25% of the group were experiencing victimization by neighbors; their average weekly income was very low; most needed help with budgeting and medical problems; and most reported that their only friends were their social workers. Nevertheless, few of the 88 wished for a return to the institution. Data from a project of my own that is nearing completion support the view that the social address model is inadequate. Various measures of functioning and of quality of life have been taken for 50 people discharged from an institution and 50 controls who remained in hospital, matched for age, sex, IQ, and score on the Adaptive Behaviour Scale. None of the subjects were worse off after discharge compared with their controls, although initial gains in some aspects of functioning tended to level off after a few months rather than continuing to improve. For example, although the level of interaction between staff and residents remained at around 2% for the 18 months during which data were gathered for the controls, it increased to 7% for the subjects after 6 months in their smaller community dwellings. Interaction was 11% after 12 months, and remained at that level for the duration of the project. Interactions of any kind between residents remained very low-less than 2% of the time-for both groups throughout the project. It seems plausible to assume that these levels of interaction are probably lower than for people without disabilities living in their own homes, but we do not have normative data. During this closing decade of the 20th century, therefore, there still remains the problem of how to provide services for intellectually disabled people that will lead to more positive relationships and to higher-quality life-styles in general. It seems clear that some things should not be done. Large institutions are difficult environments in which to arrange for good services; however, the necessary conditions for good services, even in community residences, are not entirely clear. To many, at least part of the answer seems to lie with the behavior of those who are charged with the responsibility of caring for people with intellectual disability. Rice and Rosen ( 1991) write: Every program administrator acknowledges the critical role of direct-care workers in applying sound tenets of learning and development to the daily lives of residents with mental retardation. The most carefully conceived residential program applied in the most state-of-the-art architectural structure rests in the hands of direct-care workers. This truism cuts across the issue of large versus small facilities and institutional versus community-living settings. ( p. iii)

STAFF TRAINING AND MANAGEMENT

227

Although there is no doubt that relatively stable setting events such as architectural features do have an effect on the behavior (cf. Wahler & Fox, 1981), development, and quality of life of individuals with intellectual disabilities, the nature of interactions between people and their caregivers must be of considerable importance. Rice and Rosen (1991) suggest that “the plight, motivation, training, counselling, and material reinforcement of persons working in directcare positions be accepted as a priority” (p. iv). The crucial question that follows is how to address this challenge.

II. A MODEL OF STAFF TRAINING AND MANAGEMENT Traditionally, staff training and management models have drawn upon psychological theories that emphasize the individual characteristics, such as personality and “style,” of managers and staff. There is little evidence to support this as an effective model. A more radical approach is to emphasize the behaviors of staff and managers that are necessary to produce good service. Such an approach has been reviewed by Reid, Parsons, and Green (1989). They have represented the process of staff training and management as a generic model (Fig. 1). (The only variation from Reid, et al., 1989, is the replacement of their “provide negative consequences” with “implement corrective procedures”). Figure 1 identifies some of the important factors that need to be considered when asking how we are to get staff to interact with residents in ways likely to enhance quality of life. Each of these has been addressed by research during the past two or three decades, and it is now possible to identify some relevant factors and to see where further research is necessary. First, it is essential to identify precisely how staff are to behave in their day-today work and to do this in ways that make it possible to know when the specified staff behaviors are occurring. “Treat clients with dignity” is less helpful than “knock on bedroom doors before entering.” It is necessary then to implement procedures for monitoring these behaviors. If they meet preagreed criteria of acceptability, then they ought to be followed by positive consequences that (it is hoped) will act as reinforcers and will ensure that the behaviors maintain. If staff behavior is not acceptable it may be because the required performances are not in their repertoire-staff do not know how to do what is requiredwhereupon training may be necessary. If the behaviors are (potentially) in their repertoire but staff are not behaving as required, then corrective actions (which may or may not involve negative consequences, hence my variance from Reid et a]., 1989) will be required. Providing consequences is a part of staff management. There are thus two areas, staff training and staff management, that both interact with the issue of what should staff be doing in disability services. In the recent past it has been common to find each of these dealt with as if it were separate. It has now become clear that it is essential to give equal attention to each if high-quality service is desired.

228

Chris Cullen

0 start

define staff responsibilities

.1

monitor staff perform a nc e

I

L

train staff provide positive consequences

no

Yes

I

implement corrective procedures

FIG. I . Generic model of staff training and management process. Adapted with permission. from Reid, Parsons, and Green (1989), Smjf Munugemeiil in Humuii Services. Courtesy of Charles C Thomas, Publisher, Springfield, IL.

STAFF TRAINING A N D MANAGEMENT

229

In this article, I review a selective sample of experimental studies from each area and attempt to identify the major lessons to be learned. Most of the studies are examples of organizational behavior management (cf. Cullen, 1988) and draw on the field of applied behavior analysis, arguably the branch of psychology that has had the most practical impact on service in the field of intellectual disability. I first review the types of staff behavior that have been the focus of interest-the dependent variables-and then I examine what interventions have been used-the independent variables. In concluding I want to draw attention to a question that has become very important for the behavior analytic community: To what extent does choice of dependent and independent variables match the concerns and aspirations of the wider community of which intellectual disability is a part? This is the question of social validity.

A.

What Staff Performances Are Required?

I . ABSENTEEISM Researchers have investigated several aspects of staff behavior. For example, actually being present at work is important and many studies have addressed the question of how to reduce staff absenteeism. Zaharia and Baumeister (1978) drew attention to the high rates of absenteeism in human services, and different approaches to the problem have involved ( 1) punitive disciplinary systems (Ford, 1981); (2) group contingencies for decreased absenteeism, such as opportunities for staff groups to arrange their own work schedules (Reid, Schuh-Wear, & Brannon, 1978); (3) positive consequences such as extra leave time (Durand, 1983); and (4) multifaceted procedures. Briggs (1990) reports a reduction of 27% in absenteeism achieved by a combination of positive consequences (verbal acknowledgment, letters of commendation, and increasing promotion chances) and negative consequences (progressively more serious disciplinary actions). Shoemaker and Reid ( I 980) used a combination of feedback on attendance from supervisors, commendation letters for perfect attendance, and eligibility to participate in a lottery. (Later in this article there is a fuller discussion of multifaceted procedures.) 2. BURNOUT Getting staff into the workplace is clearly important, but once they are there other issues become relevant. Burnout, or emotional overload, is relatively common among staff who work with people who have intellectual disabilities (cf. Edwards & Miltenberger, 199l), and a few studies have addressed this aspect of staff behavior. For example, Keyes and Dean (1988) gave stress inoculation training to direct-contact staff, with some positive results; 71% of staff rated the training useful at 3-month follow-up, indicating some impact on staff burnout. Most interest, however, has been on staff behaviors thought to impinge more

230

Chris Cullen

directly on the quality of life of clients. These might be general repertoires, such as staff interacting with their clients, or more focused repertoires, such as staff actually training clients. 3. STAFF-CLIENT INTERACTIONS Doerner, Miltenberger, and Bakken ( 1989) investigated the differential effects of goal setting and self-management on the frequency of interactions between staff and residents in two group homes. Positive interactions involved (1) physical contacts such as hugs and handshakes; (2) verbalizations expressing approval; (3) engagement in activities such as playing games and helping with household chores; and (4)appropriate tone of voice (“requests made by staff that were spoken at normal voice volume expressed a desire for appropriate behaviour by the client(s) and did not involve any physical guidance by staff’ [p. 3171). Negative interactions were those that expressed disapproval, such as verbal reprimands, physical guidance, and inappropriate tone of voice (requests “made at louder than usual voice volume that were directed towards stopping inappropriate behaviour” [p. 3 171). The (implicit) rationale for identifying participation in interactions as an important set of staff behaviors is the reasonable assumption that positive interactions should lead to, or at least be associated with, a higher quality of life for clients. This notion has good face validity, although some authors have pointed out that adaptive behaviors in “normal” environments may sometimes be maintained, or established by negative consequences. For example, Tennant, Hattersley, and Cullen (1978) pointed out that much ‘‘normal” social behavior is maintained as avoidance responding. That is, behaving in certain ways is avoided because of the likely consequences that would ensue. In the same vein, Iwata ( 1987) described examples where negative reinforcement plays an important role in some therapeutic procedures, including toilet training and behavior replacement strategies such as teaching a person how to request help rather than throw a tantrum when faced with a difficult task (cf. Cam & Durand, 1985). Nevertheless, as mentioned earlier, there is good reason to be concerned about the quantity and quality of interactions between staff members and clients. Hile and Walbran (1991) observed staff for 20 consecutive days at various locations in a large residential facility, and found that the majority of staff time per hour was spent in supervision (32.5%), which was directing attention to residents with a specific interest, such as getting the person to engage or desist in an activity. Other interactions included leisure (20.7%), which was time spent in activities unreluted to clients, such as taking a smoking break; and socialization (19.6%), which was engaging in warm interpersonal interchanges with residents. Training residents (1.8%) was one of the lowest-frequency behaviors. Just over half (52.1%)of the staff member’s time, therefore, was spent in potentially interactive activities, but over half of that time was spent in supervision, the least interactive of the various options. Excluding supervision, the amount of time

STAFF TRAINING AND MANAGEMENT

23 1

staff spent in meaningful interactions drops to 19.6%. During an average hour, each client who was observed would receive only 0.21 minute of training, 0.67 minute of personal care, and 2.3 minutes of socialization. Only 11.7% of the time would a resident be engaged in potentially interactive activities with staff members. These data are consistent with previous findings (cf. Cullen et al., 1983; Dailey, Allen, Chinsky, & Wet, 1974). One of the strengths of the Hile and Walbran (1991) study is that they also showed that particular environmental variables were related to the behaviors observed. For example, two factors associated with training clients were the presence of a professional staff member and conduct of the observation sessions in classrooms. It is not always that people with intellectual disabilities are involved in poorquality interactions. Rusch, Hughes, Johnson, and Minch (1991), for example, examined the nature of interactions between workers without disabilities and 341 supported employees served by community rehabilitation facilities. They used the categories of associating, evaluating, training, advocating, collecting data, and befriending. In contrast to the majority of studies carried out in residential settings, Rusch et al. (1991) found an encouraging variety of interactions. For example, 87% of the supported employees associated with co-workers, 55% were trained by co-workers, and 23% were befriended by co-workers. With the exception of advocacy, supported employees with mild levels of disability experienced the highest levels of interaction. A major problem with the Rusch et al. (1991) study, however, is that data were not collected on the level of interactions. So, although the majority of supported employees associated with co-workers, it is not clear whether this was frequent or infrequent. Further, there was little information on the quality of the interaction. 4. CLIENT TRAINING Many staff training and management studies have focused on staff behaviors more specific than interacting, such as staff training clients. The important point about such specific staff behaviors is that they are related directly to specific aspects of client behavior. Early room management studies, for example, focused on interventions that led to increases in the engagement of clients in day-center activities (Cullen, 1985). Specifying specific roles for staff-of individual helper, supervisor, and room manager-has led in some cases to reported levels of engagement as high as 80% (Porterfield & Blunden, 1978). The precise behaviors involved in the roles of room manager, individual helper, and supervisor were carefully specified, and training was given by the researchers. It was realized quite early, however, that although there was a direct relation between staff engaging in these behaviors and dramatic improvements in client engagement, the staff did not continue in their new roles (cf. Woods & Cullen, 1983). It became important to introduce a management system to maintain staff behavior. Even more focused staff behaviors have been studied. Such behavioral pro-

232

Chris Cullen

cedures as prompting, shaping, fading, giving reinforcers, and recording behavior have been common targets (cf. Davis, McEachern, Christensen, & Voort, 1987; Harchik, Sherman, Hopkins, Strouse, & Sheldon, 1989). Highly specialized staff behaviors, such as lifting physically disabled clients (Alavosius & Sulzer-Azaroff, 1986) and using manual sign language with profoundly retarded persons (Faw, Reid, Schepis, Fitzgerald, & Welty, 1981), have also been studied.

5. PROCESS AND OUTCOME In recent years, however, it has become clear that there is an important distinction to be considered when determining what it is about staff behavior that is important. Should the focus be staff behavior per se (the process), or effects of staff behavior on clients (the outcome), or both? Arc0 and Birnbrauer (1990) point out that there has been some variation in findings from studies that have concentrated on either process or outcome. For example, Pomerleau, Bobrove, and Smith (1973) found that monetary rewards for psychiatric aides’ behavior not contingent on the behavior of patients had little effect on appropriate patient behavior compared with cash awards contingent on patient behavior. Greene, Willis, Levy, and Bailey (1978) also found very slight improvements in client behavior when there was immediate feedback (supervisor praise) to staff, but substantial improvements when clients’ performance was publicly displayed. In both of these studies it was important to identify and consequate the outcome of staff behavior rather than just the behavior (the process) alone. Woods and Cullen (1983) found that client behavior alone may not be enough to maintain staff behavior. In two toilet training programs behavior change was imperceptible from week to week. In one instance client behavior (appropriate urination) did improve very slowly over many months; in the other case it actually deteriorated, again very slowly, over many months. The behavior of care staff in carrying out the programs was, however, maintained, probably because of the long history of a particular supervisor in carrying out instructions, almost regardless of their effects. In a room management program Woods and Cullen (1983) found that staff behavior was not sustained, even though there was an obvious improvement in client behavior. And finally in a token economy program, there were gradual improvements in client self-care skills that were not noticed by staff, although staff behavior was maintained. The situation is clearly complex. Arc0 (1991) used a procedure including outcome performance feedback to staff in a small community residence for clients with severe/profound intellectual disabilities. He found that client gains were maintained for 4 months but did not stabilize. He concluded that staff did not become fluent (cf. Sulzer-Azaroff, 1990) in the use of therapeutic procedures, and perhaps would only have become so by the introduction of process feedback. It may be essential to consequate staff behavior directly and ensure that it links to changes in client behavior.

STAFF TRAINING A N D MANAGEMENT

233

One other matter ought to be mentioned. The assumption is usually made that competent staff behavior will have positive effects on clients, but that may not always be so. Reid, Phillips, and Green (1991) reviewed research on teaching persons with profound multiple disabilities and concluded that “meaningful and durable” behavior change is not often established in clients. In such cases these outcomes (or lack of them) may have potentially punishing effects on staff behavior, making it even more imperative that process feedback be particularly effective (i.e., being precise about what staff should be doing). As is clear from this brief review, the definition of “staff performance” has not been standardized. “Appropriate behavior” ranges from general injunctions to behave “well,” to highly specified skills. Some studies have produced desirable change for clients by specifying the outcomes of staff behavior alone, some by specifying the process alone, and some by specifying both. Which will be effective is not yet predictable in advance, although it is clear that a relatively high degree of specificity is necessary to meet the challenge of getting staff to behave appropriately. It is not enough to employ people who (we hope) will have the “right personality” or “positive attitudes.” We must be precise about what staff will have to do and about what effects on clients we expect to see. Admonitions alone, however, are unlikely to achieve results. Staff will have to be trained, and their behavior managed, and it is to these topics that the article now turns.

B.

Training Procedures

Having recognized that staff responsibilities include both process and outcome, one is then faced with the problem of how to get staff to behave in that way. If they already have the necessary responses in their repertoires but are not emitting them, this becomes a question of management and contingency arrangement, which is discussed below. If, however, staff need to acquire new repertoires, then it becomes necessary to teach them. Anderson ( 1987) has reviewed research on training techniques published prior to 1984 and it is unnecessary to repeat this review here. Instead I confine myself to a brief comment on some of the more commonly used procedures mentioned by Anderson with some discussion of recent studies and issues that have become important since his review. According to Anderson (1987), staff training on its own almost invariably does not lead to permanent changes in either staff or client behavior. More specifically, Quilitch ( 1975) concluded: [AJvailableresearch does not bear out the optimistic assumptions made about the value of inservice training. In objectively rating the services provided by six institutions for the retarded it has been found that the least effective of the six had the strongest in-service programme. . . . Overall. there was simply no relationship between institutional effectiveness and in-service training. (p. 64)

234

Chris Cullen

1. DIDACTIVE TEACHING

For many years staff training procedures have involved either didactic teaching, a method drawn straight from traditional education systems, or, more recently, a combination of procedures that have attempted to draw (broadly) on behavioral principles. These generally include instructional methods such as lectures and workshops, verbal and written instructions to staff, discussions, textbooks, and films. These are probably still the most common procedures used in staff training. It has been shown consistently that they are liked by staff (presumably when they are done well) and that they result in increases in knowledge; however, these procedures have little effect on what staff do in the work setting (cf. Anderson, 1987). 2. MODELING A second category of procedures may be described as modeling; that is, staff are shown what to do by a third party, the instructor. As a behavioral procedure it has a respectable pedigree but it has been evaluated infrequently in staff training. Wallace, Davis, Liberman, and Baker (1973) found that modeling by professional staff and authority figures was a more effective procedure than instructions or the removal of competing activities in increasing the attendance of the nursing staff of a neuropsychiatric unit at social activities. Gladstone and Spencer (1977) found that modeling increased the rate of contingent praise in a mental retardation setting. What is not clear is how long staff behavior would be maintained in the absence of the model, and it seems highly likely from other research that this would be a major difficulty (cf. Reid et al., 1989).

3. ROLE-PLAYING

Role-playing is a third commonly used training procedure. We might expect this to be particularly effective because it is aimed at getting staff to practice how they ought to behave in the work setting. Unfortunately, though, it has been evaluated relatively infrequently as a separate training procedure. Gardner ( 1972) demonstrated that lectures were more effective in improving knowledge of behavioral principles, but role-playing was more effective in improving actual staff performance. Jones and Eimers (1975) used role-playing with elementary school teachers and showed improvements in student performance. 4. PYRAMID MODELS Most studies have used a combination of training procedures, with the assumption that one (or a few) alone is unlikely to have much effect on staff behavior, especially over the longer term. Even when procedures have been combined the effects of training often do not endure when the external consultants and re-

STAFF TRAINING A N D MANAGEMENT

235

searchers leave the setting. To deal with this problem some studies have tried to build existing agency staff into the system in the hope that their continued presence would help to maintain any changes in staff behavior. Demchak and Browder (1990) reported a “pyramid model” of staff training in group homes for adults with profound disabilities. Page, Iwata, and Reid (1982) had previously used such an approach in an institutional setting, but had shown variable changes in client behavior, even though they had produced increases in the use of correct teaching behaviors by staff. Demchak and Browder (1 990) trained three supervisors to use specific praise and a hierarchy of prompts. They used a combination of modeling, role-play, instructions, and verbal feedback in the training procedure. After they had demonstrated mastery of these skills, the supervisors then trained aides in the use of the skills. The training was most effective for the supervisors; the aides acquired the skills to a lesser degree. Data on client performance showed variable improvements across clients, with some showing greater gains than others. This study provides very tentative support for the pyramid model of staff training, but also allows us to see the “whisper effect.” That is, messages undergo some degradation and change when they are passed down the line. The general who passes by mouth a message from the front that starts off as “send reinforcements; I am going to advance” should not be surprised if what is received at base is “send three- and fourpence; I am going to a dance”! If aides are to carry out a particular task, it would probably be more sensible to teach them directly. Perhaps the role of supervisors ought to be to arrange contingencies that will maintain direct-contact-staff behavior once they have been trained. This has been the subject of a good deal of research during the past two decades, and it is to this I now turn.

C.

Maintenance of Staff Behavior

The maintenance and generalization of trained behavior have been important problems for both practitioners and researchers (cf. Stokes & Baer, 1977). Training procedures (such as those discussed) are available for changing behavior, but that behavior does not necessarily generalize to new settings and is not maintained after a period of time. Staff can be taught to behave appropriately in our training sessions but they do not necessarily behave appropriately when they return to their work settings. In fact it would not be too strong to say that they rarely do so. The field of organizational behavior management is concerned, in part, with methods for obtaining maintenance and generalization of staff training. It is convenient to categorize organizational management procedures under four headings: (1) prompts and other antecedent procedures, (2) feedback and other conse-

236

Chris Cullen

quence procedures, (3) self-management, and (4) organizational procedures that combine elements of each of these categories, including change of the ecosystem (Harchik et al., 1989).

I . ANTECEDENTS Antecedents such as prompts have included instructing staff on what to do. For example, Panyan, Boozer, and Morris (1970) instructed staff to conduct daily self-help training sessions and to keep performance records. This procedure was insufficient to maintain staff behavior, a finding replicated by others (e.g., Patterson, Griffin, & Panyan, 1976; Wallace et al., 1973). Other types of antecedents, such as demonstrations and modeling (Gladstone & Spencer, 1977), have been found to have short-term effects, but the force of evidence indicates that, although antecedents may set the occasion for appropriate staff behavior, they are ineffective on their own (Reid & Whitman, 1983). 2. FEEDBACK AND CONSEQUENCES Feedback in various forms has been an important aspect of research on organizational management procedures, and has been consistently shown to be effective in some respects. Fleming and Sulzer-Azaroff (1989), for example, trained four direct-care staff to work with four residents on various housekeeping tasks, such as bed making, and self-care tasks, such as toothbrushing. Oral and written feedback on performance immediately followed each session. This consisted of handing the staff member a copy of a completed observational checklist; reviewing each item on the checklist, giving praise for correct responses and suggestions for improvement of incorrect responses; responding to any questions or comments; and scheduling the next session. Each staff member also was observed after a 2-month follow-up period. Results indicated significant improvements in the use of a comprehensive set of teaching skills, albeit with some drop off at follow-up. Wilson, Reid, and Korabek-Pinkowski (199 I ) found that verbal feedback presented publicly to individuals during staff meetings resulted in increases in self-help training, although only for those work behaviors for which feedback was received. Although these and other studies that have used performance feedback have demonstrated positive effects, a review by Balcazar, Hopkins, and Suarez (1986) concluded that some forms of feedback have not been shown to be effective. Part of the problem is that the termfeedback is used loosely (cf. Peterson, 1982). In this context, it ought to refer to the procedure of providing information to staff on their level or quality of performance; however, it is usually used in conjunction with other consequences, such as praise (Montegar, Reid, Madsen, & Ewell, 1977), monetary or other concrete rewards (Katz, Johnson, & Gelfand, 1972; Pomerleau et al., 1973), and work schedule incentives (Iwata, Bailey, Brown, Foshee, & Alpern, 1976). Performance feedback has also been given in different

STAFF TRAINING A N D MANAGEMENT

237

ways, such as by written and/or verbal means (Panyan et al., 1970; by public posting of results (Green et al., 1978; Quilitch, 1975); and by video (Panyan & Patterson, 1974). The delay between staff performance of the behavior and receipt of feedback has varied from a few minutes (Fleming & Sulzer-Azaroff, 1989) to several hours (Quilitch, 1975) to several days (Pomerleau et al., 1973). Although it seems clear that consequence procedures, such as feedback in combination with other consequences, do have effects on staff behavior, at least in the short term, it is still not clear what kinds of consequences are most effective in different circumstances. Peterson (1982) has argued that using the ternfeedback in studies such as these is probably a mistake, and draws attention away from consideration of the crucial parameters. Specific operations subsumed under the term are numerous and include public versus private, individual versus group, personal versus mechanical, immediate versus delayed, different schedules of feedback, and cumulative feedback versus feedback for each measure of performance independently. Michael (1980) warned that failure to specify the precise effects of procedures, and their relevant parameters, was essentially “superficial behaviour analysis” (p. 14), and would lead to “a decreasing ability to interpret human behaviour in terms of the complex interacting variables that are so often relevant” (p. 14). A final comment on consequence procedures. Duus (1988) conducted an interesting study showing that publicly posted performance data were effective in improving and increasing interactions between staff and residents, but also that unreinforced aspects of staff behavior also changed. Moreover, these aspects of staff behavior led to an increase in rates of interaction over and above what would be expected from performance feedback alone. This is a somewhat maverick study, standing outside the usual type of literature, but pointing to the need for rather more sophisticated analyses of behavior-environment relations. 3. SELF-MANAGEMENT

A third group of procedures that have been researched are described as selfmanagement approaches. These include getting staff to record their interactions with clients (Davis et al., 1987); to record, graph, and praise themselves on completion of goals (Burgio, Whitman, & Reid, 1983); and to graph and evaluate data (Kissel, Whitman, & Reid, 1983). In each of these there were positive changes in staff behavior and client behavior; however, the procedures were multifaceted and included a number of components. Davis et al. (1987) included role-playing, weekly reading assignments, discussion, group feedback, and project work. In the Burgio et al. (1983) study, staff set daily goals, used selfmonitoring apparatus, graphed their data, and administered self-praise. The external researcher also helped staff to choose a “reasonable number of staffresident interactions as a goal” (p. 43) and provided guidelines based on the frequency of staff-resident interactions observed during baseline. Kissel et al.

238

Chris Cullen

(1983) included teaching staff to use instructions, physical guidance, and contingent reward using “instructional, modeling, rehearsal, and feedback procedures” ( p. 400); watching videos; role-playing with video feedback; and recording and graphing behavior. Azrin and Pye (1989) improved specific employee behaviors by a system of contracting between an employee and her or his supervisor, a unique combination of self-management and externally imposed contingencies. As with feedback procedures it is somewhat difficult to know precisely what are the important behavioral relations when multifaceted self-management procedures are used. It may be, as Baer (1984) suggests, that the success of selfmanagement strategies is largely a function of the procedures’ evoking desired behavior that otherwise would not occur and that can then be reinforced. Certainly this would fit with the fact that effective procedures are almost always multifaceted. Establishing behavior by self-management procedures might act to “bridge the gap” between staff behavior and, say, supervisor feedback, or it might help in those situations where staff behavior is unlikely to be reinforced naturally by changes in client behavior (cf. Reid et al., 1991). 4. CHANGE OF ECOSYSTEM A fourth group of approaches to the problem of maintaining staff behavior are those that change the organization or ecosystem of which staff behavior is a function. A first step toward designing such ecosystems would be an analysis of what staff (and client) behaviors are maintained by current environments. Cullen (1987) found, for example, that the rules and routines in an institution were such that residents could not engage in various self-care and independence behaviors even if they had the skill. For example, serving oneself at mealtimes was not possible because of the longstanding tradition that care staff serve food onto the plates-the most highly qualified staff serving the meat, the less qualified staff, the vegetables! McGuire, VanBuren, Alger, and Thomas (1987) assessed the opportunities for teaching in 17 residential units, finding considerable variability across the units. In such instances it would seem logical first to make changes in the rules and routines of an organization so that more opportunities for staff and clients become available, and to do this before instituting expensive staff training and management interventions. Unfortunately, though, attempts to make permanent beneficial changes by setting new rules have met with mixed fortunes. Meinhold and Mulick (1990) show some of the complex relations between staff (and client) behavior and aspects of the social and physical environment that are subject to rules, routines, and regulations. They argue that the timing of activities during the day and the physical arrangement of day areas are themselves powerful determinants of staff and client behavior. Setting up new rules and regulations may not have an impact on the clients’ well-being. For example, regulations that specify that state finance

STAFF TRAINING AND MANAGEMENT

239

will be made available contingent only on the production of individual habilitation plans (IHPs) may ensure that they are produced, but may not necessarily ensure that the IHP has a real and meaningful impact on the life of the client. In fact, Jacobson (1990) has argued that such regulations are poor ways of ensuring staff compliance and that they may support mediocrity in program practices. In addition to changes in rules, effective ecological or organizational manipulations may include environmental change. It is theoretically possible that changes in the physical environment may lead to appropriate behavior on the part of staff and clients, in much the same way that architects design features such as walkways to encourage the use of particular routes between locations; however, some common, simple approaches are not necessarily effective. For example, Foxx, McMorrow, Bechtel, Busch, Foxx, and Bittle (1986) examined the relative effects of exposure to an austere institutional dayroom, a toy-enriched dayroom, a room containing automated stimulation devices, and a modular room containing wall-mounted stimulation devices. They found only slight differences in object-directed adaptive behavior between the austere and the enriched settings, and concluded that more than straightforward architectural manipulations is needed if client engagement is the goal. Perhaps the best known example of change in the ecosystem is the Teaching Family Model used on a large scale in group homes for delinquent youths (Wolf, Braukmann, & Ramp, 1987) and less frequently for adults with developmental disabilities (Sherman, Sheldon, Morris, Strouse, & Reese, 1984). Part of the job specification for staff of Teaching Family Model programs is that they will interact with clients in particular ways. The system is set up to support such behavior. The advantages of having a whole system that acts as a setting event (Wahler & Fox, 1981) for appropriate staff behavior is that behavior should endure in the absence of external researchers and consultants. Redinger, Alger, and Thomas (1986) also provided an example of change of the ecosystem. They (1) set up an environment with a routine to engage residents in functional training activities, (2) identified specific opportunities to teach signing, (3) set up specific and structured training sessions to augment training within normal daily living activities, and (4) trained staff to be proficient signers with a combination of procedures. They reported that residents who previously had signing skills improved their skills, and had increased opportunities to use these skills in new situations, and those who initially had few skills acquired “functional verbal responses” ( p. 250). So, another multifaceted procedure, this time involving organizational components, seems to be the one that is most effective. The presence of setting events or ecosystems that will maintain newly trained behaviors is essential, a fact that is acknowledged increasingly in individual therapies (cf. LaVigna & Donnellan, 1986) as well as in interventions for groups of people. The right ecosystem will go a long way to ensuring that behavior

240

Chris Cullen

generalizes to new situations and is maintained over time, but there is another issue that is important. It is all very well for external researchers and consultants to prescribe procedures, even effective ones, but will they be “bought” by consumers? This topic has received some attention in recent years.

Ill.

CONCLUDING REMARKS: SOCIAL VALIDITY

An interesting characteristic of most published research studies is that they generally show that some procedures had some effects. Studies that show no relation between interventions and dependent variables usually do not get published; however, in many of the studies reviewed here, often parts of the reports describe interventions that were not effective or that were made effective only by the addition of some extra dimension, such as performance feedback. Furthermore, many professionals and researchers have tried to implement some of the staff training and management procedures described above with little or no success, certainly without the kind of success that would be demanded by journal editors and reviewers. Why were they unsuccessful? There is a good chance, of course, that the procedures have not been applied with as much care and attention as were used by the researchers who carried out and published the successful studies. Perhaps the procedures are less robust than we might like. It may be that the generality (Johnston, 1979) of the procedures is limited, that for reasons yet to be identified the procedures that work so well in North Carolina have less impact in Scotland. This problem-the apparent failure of procedures that have been shown to be effective-was discussed at length by Wolf (1978) in a now classic paper. Wolf (1978) described how, the first time he and his colleagues attempted to replicate the Achievement Place program, the community in which they were trying to work “fired” them. The community was convinced neither about the importance of the changes that could be achieved nor about the appropriateness of the procedures that would be used. Wolf ( 1978) argued for the introduction of measures of social validity into behavioral research to establish ( 1 ) the social significance of intervention goals (i.e., are they agreed as relevant by all concerned?); (2) the social appropriateness of the intervention procedures (i.e., are these acceptable to all concerned?); and (3) the social importance of the achieved outcomes (i.e., was it worth all the effort?). Since the publication of Wolf‘s timely injunction, considerable attention has been paid to social validity issues, and a special section of the Journal of Applied Behaviour Analysis in 1991 (Vol. 24, pp. 179-249) contained a number of useful commentaries on the issue. Perhaps the failure of some staff training and management procedures is partly a function of acceptability and other social validity problems? Although questions about the social validity of management procedures in

STAFF TRAINING A N D MANAGEMENT

24 1

general have been asked relatively infrequently, many studies have asked staff to rate the approach used in those particular studies. Davis, Rawana, and Capponi ( 1989) measured the acceptability of four different staff management procedures by asking staff to evaluate them using a 34-item acceptability rating scale and a number of staff management vignettes. This scale tapped the general acceptability of the procedures’ effectiveness, the likelihood of any negative side effects, “goodness of fit” for the problem, and consistency with conventional standards. The procedures evaluated were instruction and modeling, reinforcement, punishment, and self-management. Although punishment was rated the “most potent,” instruction and modeling were rated as most acceptable, followed by self-management, reinforcement, and punishment, although reinforcement and punishment were not seen to be significantly different. Davis and Russell (1990) extended these findings in an analog study using role-plays of the different management styles. The same order of acceptability was found, with punishment being the least acceptable. Interestingly, approaches incorporating less direct involvement by the supervisor-instruction/modeling and self-management-were more acceptable than either reinforcement or punishment procedures. These approaches also have a more antecedent nature. This observation is somewhat in conflict with the general finding that performance feedback is one of the most potent factors in staff management procedures. Contrast this also with a recent survey of common management practices in residential facilities showing that 460 managers claim to be using performance feedback frequently (as opposed to monetary incentives and time off as consequences) and that such feedback is “effective to some degree” according to 99% of respondents (Green & Reid, 1991); however, the claim by managers for the effectiveness of existing managerial practices has to be viewed in the light of universal experience and recent empirical data showing that staff and client behavior is generally not acceptable (e.g., Parsons, Cash, & Reid, 1989). The other frequently used procedure was claimed to be “staff involvement in management decision making,” which would presumably have some similarities with self-management procedures. There still seems to be a good way to go in determining why effective procedures are not implemented in the workplace. Stolz (1981) analyzed instances in which procedures have been implemented on a wide scale (one such was the Teaching Family Model discussed above) and identified 10 potentially manipulable variables that might be involved in successful implementation. These included research data showing that the intervention was effective, the existence of a pressing management problem, and model programs in existence. Each of these variables applies to the effective staff training and management strategies I have described in this article. Other variables in Stolz’ list might not, such as the adoption of procedures proposed by policymakers rather than researchers, involvement of those who have to implement the programs in the research, and the

242

Chris Cullen

availability of funds for widespread dissemination. Perhaps as a next step we ought to investigate some of these and other factors that might increase the social validity of staff training and management programs.

REFERENCES Alavosius, M. P., & Sulzer-Azaroff, B. (1986).The effects of performance feedback on the safety of client lifting and transfer. Journal of Applied Behaviour Analysis, 19. 261-267. Anderson, S. R . (1987). The management of staff behaviour in residential treatment facilities: A review of training techniques. In J. Hogg & P. Mittler (Eds.), Stafftraining in mental handicap (pp. 66- 122). London: Croom Helm. Arco, L. (1991). Effects of outcome performance feedback on maintenance of client and staff behaviour in a residential setting. Behavioural Residential Treatment, 6 , 231-247. Arco, L., & Bimbrauer, J. S. (1990). Performance feedback and maintenance of staff behaviour in residential settings. Behaviourul Residential Treatment, 5 , 207-2 17. Azrin, N. H.. & Pye, G . E. (1989). Staff management by behavioural contracting. Behavioural Residential Treatment. 4 , 89-98. Baer, D. M. (1984). Does research on self-control need more control? Analysis and fnrervenrion in Balcazar, F., Hopkins, B. L., & Suarez, Y. (1986). A critical objective review of performance feedback. Journal of Organisational Behaviour Management, 7. 65-89. Beardshaw, V. (198I). Conscientious objectors at work. Mental hospital nurses: A case study. London: Social Audit Ltd. Blatt, B., & Kaplan, F. (1966). Christmas in purgatory. Boston: Allyn and Bacon. Briggs, R. M. (1990). Reducing direct-care staff absenteeism: Effects of a combined reinforcement and punishment procedure. Menral Retardation. 28, 163- 168. Bronfenbrenner, U. (1979). The ecology of human development. Cambridge, MA: Harvard Univcrsity Press. Bruininks, R. H. (1990).There is more than a zip code to changes in services. American Journal on Mental Retardation. 95, 13-15. Burgio, L. D., Whitman, T. L., & Reid, D. H. (1983). A participative management approach for improving direct-care staff performance in an institutional setting. Journal of Applied Behaviour Analysis. 16, 37-53. Cam, E. G., & Durand, M. V. (1985). Reducing behaviour problems through functional communication training. Journal of Applied Behaviour Analysis. 18. 1 1 1-126. Cullen, C. (1985). Working with groups of mentally handicapped adults. In F. N. Watts (Ed.), New developmenrs in clinical psychology ( pp. 84-95). Chichester: Wiley. Cullen, C. (1987). Nurse training and institutional constraints. In J. Hogg & P. Mittler (Eds.), Staff training in menral handicap (pp. 335-371). London: Croom Helm. Cullen, C. (1988). A review of staff training: The emperor’s old clothes. The Irish Journal of Psychology, 9, 309-323. Cullen, C., Burton, M., Watts, S., 6r Thomas, M. (1983). A preliminary report on the nature of interactions in a mental handicap institution. Behaviour Research and Therapy, 21, 579-583. Dailey, W., Allen, G., Chinsky, J., & Viet, S. (1974). Attendant behaviour and attitudes towards institutionalised retarded children. American Journal of Mental Deficiency. 78, 586-591. Davis, J. R.,McEichem, M. A,, Christensen, J., & Voort, C. V. (1987). Behavioural skills workshop for staff and supervisor in a community residence for developmentally handicapped adults. Behavioural Residential Treatment, 2 , 25-36.

STAFF TRAINING A N D MANAGEMENT

243

Davis, J. R., Rawana, E. P., & Capponi, D. R. (1989). Acceptability of behavioural staff management techniques. Behavioural Residential Treatment, 4 , 23-44. Davis, J. R., & Russell, R. H. (1990). Behavioural staff management: An analogue study of its acceptability and its behavioural correlates. Behavioural Residential Treatment. 5. 259-270. Demchak, M. A,. & Browder, D. M. (1990). An evaluation of the pyramid model of staff training in group homes for adults with severe handicaps. Education and Training in Mental Retardation, 25, 150-163. Doerner, M., Miltenberger, R. G . . & Bakken, J. (1989). The effects of staff self-management on positive social interactions in a group home setting. Behavioural Residential Treatment, 4 , 313-330. Durand, V. M. (1983). Behavioural ecology of a staff incentive programme: Effects on absenteeism and resident disruptive behaviour. Behaviour Modification, 7, 165-18 I . Duus, R. E. (1988). Response class in the organisational setting: The effects of location-specific feedback. The Psychological Record, 38, 49-65. Edwards, P., & Miltenberger, R. (1991). Bumout among staff members at community residential facilities for persons with mental retardation. Mental Retardation, 29, 125-128. Faw, G. D., Reid, D. H.. Schepis, M. M., Fitzgerald, J. R., & Welty, P. A. (1981). Involving institutional staff in the development and maintenance of sign language skills with profoundly retarded persons. Journal of Applied Behaviour Analysis, 14, 41 1-423. Fleming, R. K., & Sulzer-Azaroff, B. (1989). Enhancing quality of teaching by direct care staff through performance feedback on the job. Behavioural Residential Treatment. 4 , 377-395. Flynn, M. C. (1989). Independent living for adults with mental handicaps: "A place of my own." London: Cassell Educational. Ford, J. E. (198 I). A simple punishment procedure for controlling employee absenteeism. Journal of Organisutionul Behaviour Management. 3 , 7 1-79, Foxx, R. M., McMorrow, M. J., Bechtel, R., Busch, L., Foxx, C. L., & Bittle. R. G. (1986). The lack of effects of enriched and automated environments on the adaptive and maladaptive behaviour of mentally retarded persons. Behavioural Residential Treatment, I , 106- 124. Gardner, J. M. (1972). Teaching behaviour modification to nonprofessionals. Journal of Applied Behuviour Anal.ysis, 5 , 517-521. Gladstone, B. W., & Spencer, J. C. (1977). The effects of modelling on the contingent praise of mental retardation counsellors. Journal of Applied Behaviour Analysis. 10, 75-84. Green, C. W.. & Reid, D. H.(1991). Reinforcing staff performance in residential fac of common managerial practices. Mental Retardation, 29, 195-200. Greene, B. F., Willis, D. S., Levy, R., & Bailey, J. S . (1978). Measuring client gains from staffimplemented programmes. Journal of Applied Behaviour Analysis, 1 1 , 395-412. Harchik, A. E., Sherman, J. A , , Hopkins, B. L., Strouse, M. C., & Sheldon, 1. B. (1989). The use of behavioural techniques by paraprofessional staff: A review and proposal. Behavioural Residential Treatment, 4 , 33 1-357. Hile, M. G . , & Walbran, B. B. (1991). Observing staff-resident interactions: What staff do, what residents receive. Mentul Returdation, 29, 35-41. Iwata, B. A . (1987). Negative reinforcement in applied behaviour analysis: An emerging technology. Joirrnul of Applied Behaviour Analysis. 20, 361-378. Iwata, B. A., Bailey, J. S., Brown, K . M., Foshee. T. J., & Alpem, M. (1976). A performancebased lottery to improve residential care and training by institutional staff. Journal of Applied Behuviour Anulysis, 9, 417-431. Idcobson, 1. W. (1990). Regulations: Can they control staff compliance in human services systems? Menial Retordution. 28, 77-82. Johnston, J. M. ( 1979). On the relation between generalisation and generality. The Behaviour Anal.ysi. 2(2), 1-6.

244

Chris Cullen

Jones, F. H . , & Eimers, R. C. (1975). Role playing to train elementary teachers to use a classroom management “skill package.” Journal of Applied Behaviour Analysis. 8 , 42 1-433. Katz, R. C., Johnson, C. A,, & Gelfand. S . (1972). Modifying the dispensing of reinforcers: Some implications for behaviour modification with hospitalised patients. Behaviour Therupy, 3. 579-588. Kcyes, I. B., & Dean, S . F. (1988). Stress inoculation training for direct contact staff working with mentally retarded persons. Bc,havioural Residen~iulTreatment. 3 . 3 15-323. Kissel. R. C., Whitnian. T. L., & Reid, D. H. (1983). An institutional staff training and selfmanagement programme for developing multiple self-care skills in severelyiprofoundly retarded individuals. Journal ($Applied Behuviour Analysis. 16, 395-415. LaVigna, G . W., & Donnellan, A. M. (1986). Alternatives to punishment: Solving hehuviortr prohlems with rrotr-uver.sive strutegies. New York: Irvington. McGuire. P. S.. VdnBuren. P., Alger, K . , & Thomas. D. R. (1987). Assessing teaching opportunities in a residential trcatmcnt facility. Behuvioural Re.sidential Treatment. 2. 199-2 10. Meinhold, P. M.. & Mulick, J . A. ( 1990). Counter-habilitative contingencies in institutions for people with mental retardation: Ecological and regulatory influences. Mcntal Ratardution, 28. 67-75. Michael, J. (1980). Flight from behaviour analysis. The Behuviour Anul)~.st.3. 1-22. Montegar. C. A.. Reid. D. H., Madscn, C. H., & Ewell. M. D. (1977). Increasing institutional staff to rcsident interactions through In-service training and supervisor approval. Beho,+our Thertrpy, 8,533-540. Page, T. J., Iwata. B. A , , & Reid, D. H. (1982). Pyramidal training: A large-scale application with institutional staff. Jr~itrtrtrlof Applied Behai7iourAtral,v.sis. 15. 335-35 I , Panyan, M., Boozer. H.. & Morris. N . (1970). Feedback to attendants as a reinforcer for applying operant techniques. Jortrnal ($Applied Behuvioitr Analysis. 3 , 1-4. Panyan, M . , & Patterson, E. T. ( 1974). Teaching attendants the applied aspects of behaviour niodification: An empirical approach. Miviral Returddon. I 2 . 30-32. Parsons, M. B., Cash, V. B., & Reid, D. H. (1989). Improving residential treatnient services: Implementation and norm-referenced evaluation of a coniprchensive management system. Journal (J’ Applied Behavior Analysis, 2 2 , 143- 156. Patterson, E. T., Griffin, J. C.. & Panyan, M. C. (1976). Incentive maintenance of self-help skill ~ T h e r u p und Extraining programmes for non-professional personnel. Journal C J Behoviorrr perimetrtul Psy.hiatr.y, 7 , 249-253. Peterson, N . (1982). Feedback is not ii new principle o f behaviour. The Bt,haviour Anulvst. 5. 101102. Pomerleau, 0. F., Bobrove, P. H . , & Smith, R . H. (1973). Rewarding psychiatric aides for the behavioural improvement of assigned patients. Journal of Applied Brhuviour Anulvsis. 6 , 383390. Porterfield, J . . & Blunden, R . (1978). Establishing an activity period and individual skill training within a day setting for profoundly mentally handicapped adults. Journal o f Prtrctrcal Approaches t o Diwlopmental Handicap, 2 , 10- 15. Quilitch. H . R. (1975). A comparison of three staff management procedures. Journal iJ’Applic,d Bc,havioirr Analysis, 8. 59-66. Redinger, R.. Alger, K.. & Thomas. D. R. (1986). Arranging a residential environment for the developmentally dibablcd to generate and niaintain verbal behaviour. Behuvioitrul Rc,.sidenriul Treatment. I . 231-253. Reid. D. H . , Parsons, M. B., & Green, C. W. (1989). Staff management in human scv-vices. Behavioural research und application. Springfield. 1L: Charles C Thomas. Reid. D. H., Phillips, J. F., & Green, C. W. (1991). Teaching persons with profound multiple handicaps: A review of the effects of behavioural research. Joitrnul of Applied Behuviour Analysis. 24, 3 19-336.

STAFF TRAINING AND MANAGEMENT

245

Reid, D. H., Schuh-Wear, C. L., & Brannon, M. E. (1978). Use of a group contingency to decrease staff absenteeism in a state institution. Behaviour Modification, 2, 251-266. Reid, D. H., & Whitman, T. L. (1983). Behavioural staff management in institutions: A critical review of effectiveness and acceptability. Analysis and Intervention in Developmental Disabilities, 3, 131-149. Rice, D. M., & Rosen, M. (1991). Direct-care staff A neglected priority. Mental Retardation, 29, iii-iv. Rusch, F. R . , Hughes, C . , Johnson, J. R . , & Minch, K. E. (1991). Descriptive analysis of interactions between co-workers and supported employees. Mental Retardation, 29, 207-212. Sherman. I. A,. Sheldon, J. B., Morris, K., Strouse, M.. & Reese, R. M. (1984). A community based residential programme for mentally retarded adults: An adaptation of the Teaching Family Model. In S. Paine, G. T. Bellamy, & B. Wilcox (Eds.), Human services ?hat work: From innovation ro standard practice (pp. 167-179). Baltimore, MD: Paul H. Brookes. Shoemaker, J . , & Reid, D. H. (1980). Decreasing chronic absenteeism among institutional staff: Etrects of a low-cost attendance programme. Journal of Organisational Behaviour Management. 2, 317-328. Stokes, T. F., & Baer, D. M. (1977). An implicit technology of generalisation. Journal ofApplied Behaviour Anal.ysis, 10, 349-367. Stolz, S. B. (1981). Adoption of innovations from applied behavioural research: “Does anybody care?” Journal of Applied Behaviour Anal.vsis, 14, 491-505. Sulzer-Azaroff, B. (1990). Strategies for maintaining change over time. Behaviour Change. 7, 3-15. Tennant, L., Hattersley, J., & Cullen, C. (1978). Some comments on the punishment relationship and its relevance to normalization for developmentally retarded people. Mental Retardation, 16, 42-44. wdhler, R . G.. & Fox, J. J. (1981). Setting events in applied behaviour analysis: Toward a conceptual and methodological expansion. Journal of Applied Behaviour Analysis, 14. 327-338. Wallace. C. J., Davis, J. R.. Liberman, R. P., & Baker, V. (1973). Modelling and staff behaviour. Journal of Consulting and Clinical Psychology. 4 1 , 422-425. Williams, P. 11991). Reaction to Zigler, Hodapp, and Edison (1990). American Journal on Mental Retardation. 96, 224-225. Wilson, P. G., Reid, D. H., & Korabek-Pinkowski, C. A. (1991). Analysis of public verbal feedback as a staff management procedure. Behavioural Residential Treatment, 6 , 263-277. Wolf, M. M. (1978). Social validity: The case for subjective measurement or how applied behaviour analysis is finding its heart. Journal of Applied Behaviour Analysis, 11, 203-214. Wolf, M. M., Braukmann, C. J . , & Ramp, K. A. (1987). Serious delinquent behaviour as part of a significantly handicapping condition: Cures and supportive environments. Journal of Applied Behaviour Analysis, 20. 347-359. Woods, P. A., & Cullen, C. (1983). Determinants of staff behaviour in long-term care. Behavioural Psychotherapy. I I . 4-17. Zaharia, E. S., & Baumeister, A. A. (1978). Technician turnover and absenteeism in public residential facilities. American Journal of Mental Deficiency, 82, 580-593. Zigler. E., Hodapp, R. M., & Edison, M. R. (1991). From theory to practice in the care and education of mentally retarded individuals. American Journal on Mental Retardation, 95. 1-12.

This Page Intentionally Left Blank

Quality of Life of People with Developmental Disabilities TREVOR R. PARMENTER UNIT FOR REHABILITATION STUDIES SCHOOL OF EDUCATION MACQUARIE UNIVERSITY NEW SOUTH WALES 2109, AUSTRALIA

I.

INTRODUCTION

The study of quality of life of people with a developmental disability has its roots in the philosophical, economic, social, and political forces that have shaped the delivery of services to that group of citizens. This article, in providing a review of the various approaches that have been made in exploring the quality of life of people with developmental disabilities, analyzes the role these forces have played in the development of the concept. In particular, this article seeks to clarify whether, in Edgerton’s (1990) terms, quality of life is “the challenge-or the shibboleth-of the 1990s,” or are we replacing, in Goode’s (1991) words, “the tyranny of normal” with the “tyranny of quality”? The concept of “quality of life” is deeply personal although, along with other terms that have become catchwords in the field of disability, it is becoming bureaucratized to such an extent that it may be losing much of the richness of its meaning. To some it embraces the notion of liberating people with disabilities from oppressive restraints, both physical and psychological, that have limited their opportunities for active participation in a community. To others, it is an index to assist in the scrutiny of health and welfare budgets where value judgments are made regarding the relative quality or worth of one individual’s life compared with another’s (Richardson, 1991). We are faced with the problem, possibly not restricted to the disability field, of having terms and concepts that have emerged as a result of philosophical debate losing their intrinsic meaning as we search to operationalize them and have them articulated into public policies. Sadly, quality of life may be following in the INTERNATIONAL REVIEW OF RESEARCH IN MENTAL RETARDATION. Vol. 18

247

Copyright 0 1992 by Academic Rens. Inc. All rights of repmduction in any Form reserved.

Trevor R . Parmenter

248

tradition of “normalization,” “self-determination,” “independent living,” “community adjustment,” and “least restrictive alternative,” the essential meanings of which are being replaced by a superficiality of thinking that serves, in Goode’s ( 199I ) terms, “rhetorical, political and professional purposes, but (does) not help people with disabilities achieve a better quality of life” (p. 5). It has long been recognized that philosophical, economic, and social forces have possibly had a much greater impact on policies that have affected the delivery of services to people with disabilities than has patient and detailed research (Baumeister, 1981; Emerson, 1985; Landesman & Butterfield, 1987; Parmenter, 1991). It is to the research community, however, that policy planners look for tools that might measure the effectiveness of their policies. But without first making a detailed examination of the premises on which the philosophies are grounded, researchers may find themselves trapped into a process that aids and abets the aims of the various pressure groups that abound in the welfare field. This article examines the approaches that have been made to researching quality of life in areas not specifically related to developmental disabilities, to establish the generic bases from which much of the current work in developmental disabilities is proceeding. Next, an analysis is made of various theoretical models of quality of life that have been provided in the field of developmental disabilities, together with research that has emerged partly as a result of this model building. Related issues concerning the problems associated with measurement and the more philosophical debate of whether quality of life can be validly assessed are explored. Finally, predictions and recommendations are made concerning the future conceptual and practical problems that must be addressed for us to capture the essential meaning of quality of life as it relates to the lives of those with a developmental disability.

It.

APPROACHES TO THE CONCEPTUALIZATION OF QUALITY OF LIFE

Although the history of the conceptualization of quality of life for people with developmental disabilities is relatively short (Schalock, 1990b), attempts in other domains have a much longer background. For instance, Thorndike (1939) was one of the first scientists to study the quality of life of American cities. In this tradition social indicators, usually environmentally based, were used to provide an empirically based perspective of the collective quality of community life. The area of gerontology also has a long tradition of exploring the well-being and life satisfaction of people who are elderly (Adams, 1941; Edwards & Klemmack, 1973; Riley & Foner, 1968; Spreitzer & Snyder, 1974). Other health-related areas that have been explored in terms of quality-of-life issues include closed head injury (Klonoff, Snow, & Costa, 1986), mental illness

QUALITY OF LIFE AND DEVELOPMENTAL DISABILITIES

249

(Dickey, Gudeman, Hillman, Donatelle, & Grimspoon, 1981; Lehman, 1983), intensive care (Ridley & Wallace, 1990), mobility (Hirst, 1989; Ramund & Stenssman, 1988), postcardiovascular therapies (Wegner, Mattson, & Furberg, 1984), chronic illnesses (Gill, 1984), health indices (Cadman & Goldsmith, 1986; Williams, 1979), and end-stage renal failure (Moms & Jones, 1989). A relatively recent development by health economists is the use of cost utility analysis as a means of quantifying the relative benefits of medical procedures. With a cost-effectiveness analysis approach, medical outputs were equated with the number of lives or life-years saved. Thus, a redistribution of funds to projects with a low cost per life was seen as a means of increasing the total number of lifeyears that may be gained (Drummond, 1981). A major weakness of the costeffective analysis approach was that it treated all life-years as having equal value no matter what the quality of life (Richardson, 1991). This has raised obvious ethical issues, especially in the treatment of neonates with severe abnormalities (Zaner, 1986). Perceived weaknesses in the cost-effective analysis approach have led to the development of the concept of quality-of-life adjusted years (QALYs), (Williams, 1979, 1985). The essence of a QALY is that it assumes a year of healthy life expectancy to be worth 1, but regards a year of unhealthy life expectancy as less than I . As Lee and Miller (1990) have pointed out, in essence, QALYs measure the cost effectiveness of specific medical interventions for decision making at a macro or micro level. Not unexpectedly there is growing criticism of the QALY approach to health care (see Brahams, 1991; Carr-Hill, 1989; Harris, 1987; Loomes & McKenzie, 1989; Rowles, 1989). Loomes and McKenzies’ view that support for QALY is based on its simplicity as a tool for resolving complex choices possibly sums up the opposition very well. Harris, too, took issue with the concept, arguing that QALYs fallaciously value time lived, instead of individual lives; they take an excessively narrow view of what quality of life might be and, finally, they are unjust. Now, although much of the thrust of the work of health economists in this field may not directly impinge on our discussion of the quality of life of people with developmental disabilities, it possibly has a pervasive effect on the way decisions are made regarding resource allocation to devalued groups in society. Various approaches have been adopted in reviewing the concept and measurement of quality of life, one of the most comprehensive being that by Goode (1988a), who suggested a scheme containing seven major categories: social (community and individual), life domains, life events, psychological/psychosocial,overall quality of life, and outcome behaviors. Others (e.g., Brown, Bayer, & MacFarlane, 1989; Dossa, 1989) have classified studies into objective, subjective, and combined measures. A third approach, adapted by Schalock (1990b), has classified studies into either social science or disability categories. The former has been

250

Trevor R . Parmenter

subdivided into what are basically objective measures (social indicators) and subjective measures (psychological indicators). A third subdivision, described as goodness of fit/social policy, reflects the combined approach mentioned above. From the disability perspective Schalock has been guided by U.S. legislative goals for people with disabilities, namely, independence, productivity, and community integration, goals not dissimilar to those advocated by Grimes (1985) in the Australian context. Although this issue will be raised later, it is interesting to note the effects legislative forces are having on the way quality of life of people with disabilities is being perceived in some countries. Bowles (1988a) has suggested that quality of life is metatheoretical by nature as it cannot be measured directly but must be assessed or measured indirectly through a series of “filters” or indicators.

A.

Social Indicators

Objective measures of social indicators have been supported on the basis that they provide a scientific and systematic approach to the study of quality of life. Such measures, Lippman (1976) has suggested, provide a set of indicators that can be applied to community settings, both locally and cross-culturally; however, indicators such as income, marital status, race, and sex have demonstrated only a low to moderate relationship with self-assessments of well-being, life satisfaction, and life quality. On the other hand, Rodgers and Converse (1975) have argued that objective measures, being valid and reliable, are preferable to selfassessments as conditions can be better or worse irrespective of personal opinion. Schneider ( 1975), however, suggested that these should not be used independently of subjective measures. This view was supported by Lewis and Ryan (1986) who observed that there was little relationship between objective social indicators of the environment and subjective indicators of quality of life. Ingelhart and Robier (1986), discussing their study of the relationship between objective and subjective indicators for the United States and eight West European countries, termed the weak relationship between objective living conditions and subjective well-being “the paradox of minimum intergroup variation.” In a careful analysis of Ingelhart and Robier’s findings, Bowles (1988a) cited the work of Zapf (1987), who pointed out that individuals are often under social pressure to suppress feelings of dissatisfaction and that expectations usually adjust to circumstances. In addition, those living under favorable conditions are more likely to be open to new value standards and are, therefore, more likely to express criticism and dissatisfaction. Schneider ( 1975) further criticized the prevalent tendency for objective measures to be overgeneralized to the extent that objective patterns are equated with differences in life experiences. For example, Campbell (1981) in comparing the well-being of groups found that only 20% of the difference could be attributed to objective measures.

QUALITY OF LIFE AND DEVELOPMENTAL DISABILITIES

B.

25 I

Psychological Indicators

Within this classification one may include factors that reflect an individual’s affective dimension, including how a person feels and experiences life. Here Rodgers and Converse (1975) have made a distinction between happiness (an affective component) and satisfaction (a cognitive component). Likewise, Zautra and Goodhart (1979) have pointed out that happiness is a short-term emotional state in contrast to satisfaction, which is more cognitively determined. Satisfaction with a particular state is evaluated in contrast to some external standard of comparison, related in relation to positive or negative effect. More recently, Zautra and Reich (1983) have proposed a two-factor approach to life events and perceptions of life quality, from an analysis of the effects of positive and negative events on psychiatric distress. They concluded that ( I ) positive events produce positive affective states but not negative ones, and (2) negative events influence negative states but not positive ones. A further psychological perspective is a person’s perception of well-being. Here the pioneering work of Flanagan (1978, 1982) is relevant. Flanagan derived a list of 15 factors that define a person’s quality of life from more than 6500 critical incidences collected from nearly 3000 people of various ages, races, backgrounds representing all of the United States. A wide variety of questions were used, for example: Think of the last time you did something very important to you or had an experience that was especially satisfying to you. What did you do or what happened that was so satisfying to you? Why did this experience seem so important or satisfying? (Flanagan, 1978, p. 138).

Flanagan subsequently grouped these 15 quality-of-life components into five general dimensions: physical and material well-being; relations with other people; social, community, and civic activities; personal development and fulfillment; and recreation. An analysis of the responses for the 50- and 70-year-olds in the sample of their overall rating of the quality of life and his or her reports as to how well needs and wants were being met in each of the 15 areas revealed that the six areas showing the largest correlation coefficients with overall quality of life were material comforts, health, work, active recreation, learning, and creative experience. It is important to note that large-scale studies such as these may obscure an individual’s quality-of-life status. For instance, Flanagan suggested that “the effects on each individual’s quality of life should be evaluated in terms of his or her personal values and needs rather than those that some central national authority believes all people have or should have” (p. 146). This observation is particularly relevant for countries such as Australia where legislative and bureaucratic forces are setting the national agenda in disability services (Parmenter, 1991). As indicated above in the work of Zautra and Reich (1983) and Zautra (1983),

252

Trevor R . Parmenter

there are relationships between a person’s perception of well-being and specific life events, such as moving house, death of a loved one, a new job, and divorce. The pioneering work of Holmes and Rahe (1967) tested the hypothesis that adverse events precede and contribute to the onset of illness. Later research extended this hypothesis to the field of mental health, principally neurosis and depression (Dohrenwend, 1973; Dohrenwend & Dohrenwend, 1978). Canadian (Atkinson, 1977) and Australian (Headey, Holmstrom, & Wearing, 1984) studies have also shown some statistically significant relationships between life events scores and changes in well-being. A critical observation by Headey et al. (1984) was that the way people react to life events or the way they perceive life events are possibly as much influenced by their personalities, personal resources, and adaptive styles as the nature of the events themselves. There is, therefore, a need to examine the interactions between events and personal resources. A criticism made of some approaches to the assessment of quality of life is that they do not examine it over time (Headey et al., 1984). Longitudinal studies of components of the past, the present, and the future need to be made to measure temporal changes in attitude (Hall, 1976). For instance, Dalkey, Lewis, and Snyder (1972) suggested that time can be broken down into a sequence of life events that can be categorized into properties or qualities (e.g., sociality, freedom, novelty) that can be scaled. Ratings in terms of behavioral adjustment as well as psychological growth can then be made of major events in a person’s life.

C.

Criticisms of Psychological Indicators

Although psychological indicators have a plausible prime facie case for giving valid assessments of a person’s quality of life, Andrews (1974) has nevertheless drawn our attention to at least four criticisms of this approach. First, validity may be questioned as answers to surveys either vary over time or are biased. People may not have given much thought to their quality of life and, therefore, cannot answer questions about it. A second limitation may lie in the area of the interpretation of the subjective results. Comparisons across groups may not be possible as criteria may change over time. Third, a concern has been expressed about the completeness of the information, and finally there is the problem of utility. For instance, at a policy level the data, which are costly to obtain, may be ignored as the relationship between individual satisfaction and societal welfare is not yet clearly determined. Zautra and Goodhart (1979) have also raised a number of validity problems that may weaken the use of these indicators. For example, scores may be inflated because a person’s responses are skewed toward socially desirable factors. A second concern is that measures of feeling states may be simply indicating idiosyncratic ratings of satisfaction and happiness at a particular point in time. A

QUALITY OF LIFE A N D DEVELOPMENTAL DISABILITIES

253

final caveat is that psychological indicators may not be a good reflection of the realities of external conditions; however, as will be discussed later, these reservations are a clear indication that quality of life is a multidimensional concept that may need a variety of indices to capture its intrinsic meaning.

0. Person-Environment Fit The person-environment fit or goodness-of-fit proposed by Murrell and Norris (1983) defined quality of life as the criterion for establishing the goodness-offit between the person and his or her environment and, in one sense, represents a marriage between objective and subjective indicators of quality of life. Murrell and Norris proposed that the human unit (individual, family, community) may be exposed to stressors (negative life events such as divorce, retirement, illness) within the community that lead to trauma. The provision of resources, such as interpersonal resources and direct services, not only acts as a buffer and protection for the human unit, but may also provide opportunities for growth and advancement. As Schalock (1990a) has pointed out, a central assumption of this model is that the quality of life of a person is a function of the discrepancy between resources and stressors. Likewise, Brown, Bayer, and MacFarlane (1988) have defined quality of life “as the discrepancy between a person’s achieved and their unmet needs and desires. The larger the gap between what people have and what they need and want, the poorer their quality of life” (pp. 11 1-1 12). Schalock (1 990b) has pointed out that the Murrell-Norris model “conceptualizes quality of life as both an outcome from human service programs (application of additional resources should improve a person’s QOL) and the criterion for establishing the goodness-of-fit between a population and its environment. Thus, the better the fit, the higher a person’s QOL” (p. 144). Schalock has therefore identified quality of life as an instrument of social policy, but will we ultimately see quality of life perceived in terms similar to the way Brown et al. (1989) have described normalization in Canada, as “a philosophy without an appropriate and functional technology” ( p. 67)? The person-environment fit approach is a striking reflection of the ecological approach that is increasingly shaping disability policies. In the field of rehabilitation, formerly dominated by medical and psychological models, there is a shift from an emphasis on the individual to a concern with the wider social system (Cottone, 1986). The interdependence of a person with his or her environment is now recognized in the disability field as reflected in the World Health Organization’s classification of impairment, disability, and handicap (World Health Organization, 1980). Schalock (1991) has also noted a “paradigmatic shift” in the disabilities field, reflected in the way people wirh disabilities are taking a more active role in

254

Trevor R. Parmenter

expressing their views about what they want out of life. He has suggested that the new-found paradigm is characterized by supports rather than programs, persons rather than places, person-environment matches, services in natural environments, consumer empowerment, real homes, and real jobs. Within the rubrics of both ecological and symbolic-interactionist theories a number of studies have been conducted that assess a person’s quality of life on the basis of his or her behavior in response to the environment or environments in which the behavior occurs. From an ecological perspective there has been a focus on the interactions that occur between individuals in the same settings in which they operate, highlighting the importance of relationships. There is obviously an interaction between the affective, cognitive, and behavioral components of quality of life. For instance, Andrews and McKennell(l980) argued that actions (behaviors), feelings (affect), and values (cognitions) all interact to determine a person’s level of perceived well-being or quality of life. To remedy what they perceived as a dearth of studies directed toward the behavior dimension, Evans, Bums, Robinson, and Garrett (1985) developed a Quality of Life Questionnaire that was designed to measure a person’s behavior in response to a number of ecological domains that affected him or her, a position somewhat similar to the concept of person-environment fit. The domains selected were general well-being, interpersonal relations, organizational activity, occupational activity, and leisure and recreational activity. Fifteen subdomains were identified and 12 items were developed for each to constitute the Quality of Life Questionnaire. Participants were 298 residents of London, Ontario. A principal-components factor analysis with varimax rotation of the intercorrelations among scales and correlation between each scale and the quality-of-life(QOL) score revealed five factors: (1) occupational/material wellbeing, (2) social well-being, (3) family well-being, (4) personal well-being, and (5) physical well-being. With the exception of physical well-being, all factors had moderate loadings with the overall QOL score. A concurrent validation study indicated that there was a moderate correlation between life satisfaction ratings and QOL scale scores. Particular life satisfaction measures also had moderate correlations with the overall QOL score. On this evidence Evans et al. argued that there was support for the view expressed by several authors (Gutek, Allen, Taylor, Lau, & Majchrzak, 1983; Zautra & Reich, 1983) that functional or rewarding and enriching life experiences are necessary for an individual to report a high level of perceived life satisfaction or subjective well-being. Another research strategy for the study of ecological aspects of quality of life was that of Milbrath (1982) who argued that studies of social ecology and studies of quality of life constituted a “natural marriage” between objective and subjective indicators. His research model allows quality of life to be analyzed for either an individual or a community. His basic premise is that there are interactive

QUALITY OF LIFE AND DEVELOPMENTAL DISABILITIES

255

effects between individual and community experiences of quality of life. His conceptualization of quality of life is a result of two major classes of variables: (1) values, goals, and aspirations; (2) life-styles. Milbrath defined quality of life as the “fulfillment of one’s values, goals, aspirations, and needs” (p. 138). In the context of the quality of life of people with disabilities, the work of both Evans et al. (1985) and Milbrath (1982) has much to offer. A model of quality of life should reflect the values, aspirations, self-perceptions, and other factors of the individual, but it also should accommodate functional behaviors in a range of life domains. There should also be opportunity for societal variables to be incorporated.

111. THEORETICAL MODELS OF QUALITY OF LIFE IN THE AREA OF DEVELOPMENTAL DISABILITY The quest for an empirical definition of quality of life in the field of developmental disabilities undoubtedly received its greatest boost from the policies that have led to the deinstitutionalization of people with mental retardation and mental illness, along with their community integration at the school and work levels. Landesman and Butterfield (1 987) pointed to the significant growth in interest in the scientific study of deinstitutionalization and community placement, reflected in articles published in journals such as the American Journal of Mental Deficiency, where there was a twofold increase in the proportion of articles concerning community placement in the decade 1975-1985. They also referred to the increasing number of volumes of original research on deinstitutionalization and community services such as those by Begab and Richardson (1979, Bruininks, Meyers, Sigford, and Lakin (1981), Edgerton (1984), and Landesman and Vietze (1987). As Knoll (1990) has indicated, there has been an evolution in the development of outcome or program standard measures for services for people with developmental disabilities. The first evolutionary phase, suggested Knoll, was the era of institutional reform (roughly 1965- 1975), when the minimal standards of care were established to protect individuals from harm. The next phase was the era of deinstitutionalization (1976- 1986), when emphasis was placed on defining the characteristics of quality programs. In the current phase, the era of community membership, there is a concentration on community integration, quality of life, and development of individualized support systems. From a scientific perspective efforts have been directed toward the measurement of objective program standards and quantifiable indices of quality. This approach has not been without its difficulties. As Emerson (1985) has cogently observed, evaluation of programs for people with developmental disabilities has produced equivocal results, either because of a lack of attention to the essential

256

Trevor R . Parmenter

elements of what is being evaluated, or because of a failure to recognize the social nature of research, or because of conceptual and methodological difficulties. Too often it has been the superficial elements of community living that have been the focus of attention. Important as changes in the physical environment, residents’ adaptive behavior, or provision of activities for daily living might be, there are more critical aspects that relate to outcomes such as client satisfaction, happiness, social and interpersonal relationships, activity patterns, degree of self-determination, socioeconomic factors, and access to community services. Complex as they might be to identify and measure, outcomes that reflect a person’s interaction with his or her environment are more valid indices of the success or otherwise of community living programs (Bronfenbrenner, 1977; Landesman, 1986; Landesman-Dwyer, 1985; Landesman-Dwyer & Berkson, 1984). Researchers should therefore be guided by the multidimensional and codetermined nature of environments in their design of studies to seek a genuine picture of the degree of community integration of people with disabilities. But, as Knoll (1990, p. 235) has pointed out, a significant factor must not be overlooked in our search for scientifically based outcome studies: ”the definition of program standards and quality is a process that transcends empiricism. This process ultimately appeals to the fundamental values of a society.” This view has also been supported by Emerson (1985) who suggested that the process of deinstitutionalization is basically “an ideologically committed social movement” that must be clearly understood when attempts are made to evaluate the success or failure of the process. Addressing what quality of life actually means for people with developmental disabilities from an epistemological perspective may be of more value than pushing ahead too quickly with the development of indices that may manifest theoretical definitional, operational, and methodological problems (Goode, 1991). Contemporary social policies in many countries have been enshrined in legislation that underpins the way services to people with disabilities are funded and evaluated. There is therefore a need for research that questions the way those policies are being articulated in practice. This conforms with the conceptual role that research plays, in that it provides a framework of concepts, orientations, and generalizations that can inform policy, identify problems and potential solutions, and set the agenda for future policy formulations (Beyer & Trice, 1982). Unfortunately, as Emerson (1985) has noted, the major use of research is symbolic, that is, the nonfunctional use of data to justify predetermined positions. Hence, Goode’s injunction concerning the “tyranny of quality” is timely. The following section analyzes five models of quality of life, two of which have some psychometric support (Halpern, Close, & Nelson, 1986; Schalock, Keith, & Hoffman, 1990) and three that remain to be validated empirically (Brown et al., 1989; Goode, 1991; Parmenter, 1988).

QUALITY OF LIFE A N D DEVELOPMENTAL DISABILITIES

A.

251

Community Adjustment

Community adjustment is a construct that has been developed to encapsulate the outcomes of programs designed to enhance the participation of people with disabilities in the milieu of the general community. Halpern (1985) initially suggested the concept of community adjustment as being a preferred way of looking at the outcomes of transition programs for young people with disabilities in reaction to the “bridges” model of transition proposed by Will (1984). Will suggested that transition is defined as “an outcome-oriented process encompassing a broad array of services and experiences that lead to employment” (p. 1). An alternative approach suggested by Halpern (1985) focused attention on broader dimensions of postschool life. Support for a multidimensional approach to postschool outcomes is found in the work of a number of authors (see Brown & Hughson, 1980; Close & Foss, 1988; Edgar, 1988; Mitchell, 1986; Parmenter, 1980, 186; Ward, Parmenter, Riches, & Hauritz, 1986; Whelan & Speake, 1981). Working from essentially an empirical rather than a theoretical base, Halpern, Close, and Nelson (1986) developed an integrated model of community adjustment that included occupation, residential environment, social support, and personal satisfaction as four key dimensions of community living. To test the model Halpern, Nave, Close, and Nelson (1986) developed a battery of tests that tapped three variables in each of the four dimensions. Under employment were included status, degree of community integration, and financial status; under residential were residential comfort, neighborhood quality, and access to resources; under social support were safety from minor and major abuse and social support; and under satisfaction were general satisfaction, self-satisfaction, and program satisfaction. Measures for each of the 12 variables were field-trialled to test for internal consistency and reliability, as a part of a larger study that studied the community adjustment of adults with mental retardation who were residents of semi-independent living programs (SILPs) in California, Oregon, Washington, and Colorado. To test the model, data were gathered on the 12 variables from 257 residents (141 females, I 16 males; age range from 18-59 years, X 28 years), the majority of whom were classified as mildly retarded. Using both exploratory and confirmatory factor analysis, Halpern et al. found strong support for the four-dimensional model. Inspection of Fig. 1 shows that the level of association between the four factors derived from the analysis using the LlSREL program was, as predicted, quite low. There were, however, significant intercorrelations with the satisfaction factor that correlated .41 with the environment and .61 with social support/safety. Interestingly, as predicted, occupation was quite independent of the other factors. Of some significance was the fact that the proposed model was able to reproduce the original correlation matrix, with the average difference

258

Trevor R . Parmenter

FIG. 1. A model of community ad.justment. Reproduced, with permission, from Halpern, Nave, Close, and Nelson (1986).

between the true correlations and the reproduced values at a low value of .04. The goodness-of-fit index derived from the maximum likelihood technique was .97, providing additional support for the robustness of the model. The Halpern, Nave, Close, and Nelson (1986) model of community adjustment makes a useful contribution to the analysis of quality of life of people with disabilities. It combines both objective and subjective variables and also incorporates aspects of a person’s interaction with the environment. A limitation of the study, recognized by the authors, is that the number of variables within each dimension of the model is relatively small. Also, the location of community integration within the occupation dimension may not be appropriate across other disability areas; however, their premise that opportunities for community integra-

259

QUALITY OF LIFE A N D DEVELOPMENTAL DISABILITIES

tion for people with intellectual disabilities are generally found in the vocational setting was supported by their data for the group studied. This is not altogether surprising considering that work often provides the context for other, more social activities for the population in general.

B.

Conceptualization and Measurement of Quality of Life

In his presentation to the 115th Annual Convention of the American Association on Mental Retardation, Schalock (1991) proposed a multidimensional quality-of-life model (Fig. 2). A person's perceived quality of life, suggested Schalock, results from three aspects of life experiences: personal characteristics, objective life conditions, and the perceptions of significant persons about individuals with disabilities. These aspects may be reflected in the measured indicators of independence-productivity, community integration, and satisfactionthat Schalock has essentially drawn from U.S. legislation pertaining to people with disabilities. VALUES

IAT'IONS

FIG. 2. Quality-of-life model. Reproduced, with permission, from Schalock (1991).

260

Trevor R. Parmenter

These aspects of life experiences are embedded with a number of cultural factors: values, legal foundations, and a paradigm shift in the way society is thinking about issues and solving problems related to people with disabilities. The concept of quality of life, suggested in this model, encompasses both aspects of the macrosystem that represents cultural trends and factors in a society (i.e., “the quality of American life”) and aspects of the microsystem that relate to the individual, family, schools, or habilitation programs (i.e., “the quality of my life”). The inclusion of these cultural factors in Schalock’s model is somewhat similar to the paradigm underpinning research that is being conducted into quality of life in Sweden. In Sweden there is a special law for children, young people, and adults with mental retardation that guarantees a life equal to the lives of others in the society. The basic objectives of the special law are integration and normalization, which ensure that people with mental retardation shall be given the possibility to live like every other citizen and with a mutual sense of community. Thus people with mental retardation in Sweden are by law entitled to a right of self-determination of their lives (Drugge, 1990). Since 1985 Schalock and colleagues have developed, field-tested, revised, and standardized a 40-item Quality of Life Questionnaire (Schalock, Keith, & Hoffman, 1990) based on the model described above. The items in the scale cluster into four factors: 1 . independence, reflected in the opportunity to exert control over one’s en-

vironment and to make choices; 2. productivity, reflected in positive work outcomes, such as income or work that contributed to either a household or a community; 3. community integration, reflected in participation in the same patterns of life as nondisabled people, including social and interpersonal contacts; 4. satisfaction, reflected in the fulfillment of needs or wants and the happiness that goes with that fulfillment. The scale items, encompassing both objective and subjective measures of well-being, can be administered directly in a structured interview format to people who are verbal. In the case of people who are nonverbal, two staff or significant others independently evaluate the person on each item and use the average score for each item (on a 3-point Likert-type scale). The standardization sample consisted of 552 people with mental retardation drawn from mental retardation/developmental disability programs across two regions of Nebraska and from the state of Colorado. The average age of the people in the sample was 37.8 years (47% female, 53% male). An indication of the level of severity of disability can be gauged from the fact that 45% of the sample had not had any formal schooling.

QUALITY OF LIFE AND DEVELOPMENTAL DISABILITIES

26 1

The cross-cultural utility of the scale has been supported by a study that revealed that its factor structure was consistent across data obtained from five countries (Australia, Federal Republic of Germany, Israel, Republic of China, and the United States). Also in each country’s sample a consistent trend emerged, revealing that quality-of-life scores increased as the environments in which people lived became more normalized. Caution must be exercised, however, in drawing firm conclusions from this study as sample sizes were extremely small in three of the countries.

C. Quality of Life for Disabled Persons As a vehicle for analyzing the data obtained from a study of 240 people with mental retardation in five community-based agencies situated across four provinces of Western Canada, Brown et al. (1989)developed a conceptual model of quality of life. The study employed a Client Quality of Life Questionnaire and a Sponsor Questionnaire. For the purposes of the study, Brown et al. defined quality of life as (1) the discrepancy between a person’s achieved and unmet needs and desires and (2) the degree to which individuals increasingly control their environment. Their conceptual model of the measurement of quality of life (Fig. 3) consists of a combination of objective and subjective measures; the former include income, environment, health, and growth and mastery of skills,

FIG. 3. Conceptual model of the measurement of quality of life for disabled persons. Reproduced, with permission, from Brown, Bayer, and MacFarlane (1989).

262

Trevor R . Parmenter TABLE I QUALITY-OF-LIFE VARIABLES FOR DISABLED PEOPLE Objective evaluation Skill attainment Physical environment Level of physical and social integration Training plans Actual support systems Income Provisions Health Philosophy of training agencies Attitudes of training staff Attitudes of community

Subjective evaluation Perceived growth and mastery Safety and security Social involvement and feelings of belonging Independence control Responsibility Self-esteem Expectation Perceived goal attainment Perceived supports Satisfaction level Perceived health (mental and physical) Normalcy of life Pace of life Family stability

and the latter, life satisfaction, psychological well-being, and the person’s perception of skills and needs. Specific variables included under each of the objective and subjective components are provided in Table I. Brown et al. have addressed the quality of environment at both the macro and micro level. At a macro level the political and economic climate and the societal attitudes toward people with disabilities are assessed. The existence of support systems, either formal or informal, is also seen as a crucial element in determining the quality of the environment. On a micro level elements such as safety features of one’s neighborhood, workplace, and leisure activities, as well as the nurturing aspects of these environments, are considered. Other aspects of the environment that can be considered objectively are safety, health, and access. Obviously there may be a discrepancy between the way a person perceives these features and their objective presence or absence in the environment. In their reporting of the data from the study Brown et al. have used the model to guide their discussion, but no attempt was made to draw together the various variables considered into a form of overall quality-of-life index. Rather, the perceptions of the consumers, their sponsors, and various service providers were used as a basis for a comprehensive set of recommendations concerning ways service provisions could be improved for Canadians with a developmental disability. In this respect they have responded well to Emerson’s (1985) plea that evaluation research should address those issues that surround an individual’s quality of life in specific environments, rather than focus solely on location and adaptive behavior repertoires.

QUALITY OF LIFE AND DEVELOPMENTAL DISABILITIES

D.

263

A Process Model

Goode (1988a, 1991) has developed a model somewhat similar to that of Brown et al. (1989). He, too, has drawn on social research and policy in the field of disability, together with current philosophies concerning supports to these people including those that emphasize choices. The model (see Fig. 4) is described as a nonlinear, process model that is client centered. The circles and squares respectively suggest interactive relationships between the subjective and objective variables. Central to this model is the premise that quality of life is specific to environments. For instance, Powers and Goode (1986) have argued that quality of life is basically the product of the relationships between people in each life setting. They have stressed that one’s quality of life is markedly influenced by the quality of life experienced by the people with whom one interacts. For example, Goode has suggested that it is imperative, when assessing the quality of life of people with disabilities in community residential settings, that the quality of life of the staff also be addressed. The subjective elements of the model emphasize the relationships among perceived needs, capabilities, demands, and resources. Goode has stressed that OM.ENVIA. c

Life selling Actors Structure Resources

\

FIG. 4. A process model of quality of life (QOL).R, resources; N, needs; D, demands; A, abilities. Reproduced, with permission, from Goode (1988a, 1991).

Trevor R. Parmenter

264

the individual’s overall assessment of quality of life is influenced if there is a misfit between environmental demands and personal capabilities and the person’s perception of his or her personal needs and environmental and social resources. Goode’s framework has been heavily influenced by an ethnomethodological perspective of disability research, which is in the tradition of the “special constructionist” branch of sociology formulated at the University of California, Berkeley, in the early 1960s. Although not discrediting attempts to quantify the quality of life of people with disabilities, Goode, nevertheless, has perceived some inherent dangers in this activity. For instance, he has expressed concern that an index may produce a standard of quality that can be generalized across populations as a type of metric by which the effectiveness of programs may be judged, hence his term tyranny ofquality. As quality of life is a deeply personal and individual construct, he has urged researchers to place more emphasis on observational methods, which may better portray the richness of the relationships a person forms with others in his or her environment, than on carefully validated and reliable scientific instruments, the content of which may not be sufficient reflection of an individual’s subjective experience of life and their “actual” or “real” social identity. Goode (1984) has made a plea that professionals (including researchers!) must get much closer to the people whom they are studying. For instance, he has claimed that we reify service activities by assuming that the objective data collected about the lives of developmentally disabled people truly reflect a person’s “actual” self. Goode’s ethnomethodological approach to the study of the quality of life of people with developmental disabilities has been applied in a number of countries; an example from Finland is presented later.

E.

A Symbolic-Interactionist Approach to Quality of Life

Much of the work that has been conducted in the development of quality-oflife models for people with a disability has failed to analyze the impact of disability per se. Although the general factors influencing a person’s quality of life may be the same whether the person has a disability or not, the presence of a disability surely adds a significant dimension. The following analysis examines, from a symbolic-interactionist viewpoint, what it means to have a disability; this analysis forms a theoretical base for a proposed model of quality of life (Parmenter, 1988). 1. WHAT IS A DISABILITY? Increasingly disability is being studied from sociological perspectives including structural-functional, conflict, phenomenological, interaction, and ecological. In the light of the almost total absence of sound theoretical bases for much of

QUALITY OF LIFE A N D DEVELOPMENTAL DISABILITIES

265

the research in the disability field, examination of the concept of disability from a symbolic-interactionist viewpoint is proposed. Fundamental to this approach is the principle that human experiences are mediated by interpretation (Bogdan & Kugelmass, 1984). Another basic tenet is that the “self” arises and is maintained in a symbolic and interactive context. For instance, in addressing the question of exactly how a “self” comes into being, Stryker (1959) has suggested that we come to know what we are through others’ response to us. In the context of the present discussion Bogdan and Kugelmass (1984) have pointed out that the word disability is not a symbol for a condition that is already there in advance. Paradoxically, disability is part of the mechanism whereby the condition is created. Physical and/or psychological impairments set the parameters in which the definitions develop, but the way in which people determine their definitions depends on a variety of factors including personal and community attitudes toward people who appear different. Definitions and labels are influenced strongly by the degrees to which people have had the opportunity to interact at a personal level with people with disabling conditions. 2. DISABILITY AS DEVIANCE AND THE LABELING PERSPECTIVE Goffman ( 1963) identified blindness, deafness, epilepsy, and physical disfigurement as examples of stigma. This list was expanded by Clinard (1974) to include mental retardation, cerebral palsy, and stuttering. Some may view the stigmatizing condition as being inherent in the individual, but Clinard (1974) maintained that the emphasis should be on the effects of the imputed impairment, through a process of labeling, on the individual. Schur (1971), for instance, pointed out that in examining deviant human behavior we are seeing the results of a combination of a personally discreditable departure from expectations and the development of certain stigmatizing reactions toward the individual. These reactions may serve to “isolate,” “treat,” “correct,” or “punish” individuals who are engaged in such behaviors. Central to Schur’s position is the notion that deviance is not a static entity; rather it is the dynamic outcome of the complex interactive processes ongoing in society. For persons with a disability there is often a lack of congruence between their desired personal identity and their assigned social identity. Hurst (1984) has commented that stereotypes focus on generalities. This is supported by Scott’s (1972) observation that society has ascribed to people with visual disabilities the attributes of “helplessness, dependency, melancholy, docility, gravity of inner thought, [and] aestheticism,” all traits, he has suggested, “that commonsense views tell us to expect of the blind” (p. 4). The effects of this role-making process are such that the person with a disability will often not develop as an “authentic” person. Unfortunately people who have been assigned an identity or

266

Trevor R. Parmenter

role in society may ultimately fulfill the expectations others have of them, thus reinforcing in the eyes of others the validity of their assessments. The closer the perceptions of self by the person with a disability come to those ascribed by society, the greater is the chance that secondary deviance or career disability will emerge as the major form of role adaptation. Burbach (1981), in distinguishing between primary and secondary deviance, has suggested that people with “primary” deviance do not see their “differentness” as defining them as a person, whereas people with “secondary” deviance see their “differentness” as the crucial defining element in their concept of self. In this respect career disability has been a major focus of attention of a number of writers in the field of rehabilitation (De Jong, 1981; Finkelstein, 1980; Stoddard, 1978). For instance, in his analysis of attitudes and disabled people Finkelstein (1980) has indicted professional groups as having contributed significantly to the social oppression of people with disabilities. In summary, an examination of the processes of social definition or labeling, posited by Schur (1971), has revealed that the deviance attribute of disability is imposed by the social audience or, in the words of Becker (1983), “Social groups create deviance by making the rules whose infraction constitutes deviance” (p. 9). From a symbolic universe perspective, rules arise in society as a result of its need to maintain an orderly view of the world. The anomalous position of people with disabilities in this universe has led to the property of deviance being assigned to these individuals. There often follows a power struggle to determine whose rules should apply. As fragmented and uncoordinated groups of individuals, people with disabilities are relatively powerless in this conflict. In the area of interpersonal relations the lack of congruence between a personal and a socially assigned identity often prevents the person with a disability from developing as an “authentic” human being. This, in turn, has deleterious effects on the quality of interpersonal relations. A brief examination of organizational structures has demonstrated how. in Schur’s (197 1) terms, organizations produce deviance through their attempts to regulate and control those groups that are perceived as a threat to the natural order of things.

3. IMPLICATIONS OF LABELING FOR THE QUALITY-OF-LIFE QUESTION In his illuminating analysis of the labeling issue Burbach (1981) suggested that it is superficial to ask whether we should or should not label anomalous individuals. He pointed out that labeling and categorizing people are the normal processes of apprehending and organizing our world. Of more importance is “how we label (people) and with what consequence” (Burbach, 1981, p. 376). Burbach’s contention is that people with a disability are in a double-bind situation. In addressing the issue of what it means to be disabled they are confronted

QUALITY OF LIFE AND DEVELOPMENTAL DISABILITIES

267

by two messages. One comes from outside and proceeds from the social order. The other, however, comes from within and relates to what they know they can or cannot do, Thus, they have to deal with the negative aspects of their personal condition and at the same time cope with the negative effects of stigmatization and stereotyping. From a philosophical point of view there is a conflict between the existential nature of the person and the social nature of human experience. In trying to establish a coherent meaning for life as well as creating and maintaining self-esteem, the conflict between the messages the person with a disability receives often presents insuperable problems. On the one hand, the person can live a cocoonlike existence built on socially unvalidated meanings or, on the other hand, he or she can conform to the patterns of behavior expected of him or her by society generally. Neither of these approaches leads to a satisfactory resolution to the issue of how these persons define their own meanings. Here, it is proposed, lies the nub of the quality-of-life issue. Quality of life represents the degree to which individuals have met their needs to create their own meanings so they can establish and sustain a viable self in the social world. The resolution proffered by Burbach draws on the basic principles of symbolic interactionism. That is, there is a need for consensuality whereby humans help each other unfold and establish contact and unity in their social existence. Although not without its critics (e.g., Sharp & Green, 1975), symbolic interaction theory has usefully contributed to our understanding of aspects of the social situation of people with disabilities. Barton and Tomlinson (1984) have suggested that this approach has at least two strengths. First, it emphasizes the viewpoint of the participants in social interaction and, second, the perspective explores aspects of social life that have historically either been taken for granted or have been ignored. Other sociological approaches might emphasize economic considerations and the distribution of resources in the society (e.g., structuralist neo-Marxism) or use the traditional Marxist class conflict model to explore the imbalance of power between disabled and nondisabled groups in society. These theories may not be mutually exclusive, for in the context of quality of life it is suggested that the consciousness of self and social identity, status, and social role are obviously preconditions for political activism and social change. 4. A QUALITY-OF-LIFE MODEL

From a symbolic-interactionist/ecological theoretical perspective it would seem essential to include at least three components within a quality-of-life model for people with disabilities. The first would pertain to an individual’s perception of self; the second to the individual’s behavior in response to ecological domains that might affect him or her; and the third to responses the settings might make to the individual. Obviously the model would need to be able to accommodate the interactions that would occur among each of these components. In this section

268

Trevor R . Parmenter

Societal influences Attitudes Values Economic Political Support services Incentives Disincentives Provision of access

J

Self

Functional Behaviors

Social Interactions Social opportunities Relationships Friendship networks Leisurehecreation

Cognitive Beliefs Goals Values Aspirations Knowledge of self Empowerment Knowledge of how the world works

OccupationaMMaterialWell-Being Income Employment Occupational relationships

Affective General life satisfaction Happiness Self-esteem Locus of control Acceptance of disability

Accommodation Comfort Security Utilization of neighborhood Access Knowledge of services Education Possession and use of skills Mobility .

. .

-

..

- .. - -

Personal Life-style Life evenk Perception of personal life-style .

. . .. . ... .

-

.

.

..

. .

-

P I t i 5 A model ot quality ot lite tor people with disabilities Keproduced, with permission, trom Parmenter ( 1988)

the subcomponents that might constitute the basis for instrument development for subsequent testing of the model are described briefly. Figure 5 illustrates the components of the model. 1 . S e v . The subcomponents of self have been arbitrarily classified under cognitive and affective, given that they are highly interactive. A further moderating dimension, personal life-style, has also been included under self. Under the cognitive dimension are included beliefs, goals, values, aspirations, knowledge of self, knowledge of how the world works, and empowerment. The latter refers

QUALITY OF LIFE A N D DEVELOPMENTAL DISABILITIES

269

to strategies such as risk taking and decision making, which may lead to greater levels of autonomy and self-determination. The facets of the affective dimension include general life satisfaction, happiness, self-esteem, locus of control, and acceptance of disability. It is anticipated that life events and perception of one’s personal life-style would affect both the cognitive and affective dimensions. 2. Functional behaviors. These elements, which may be directly observed, have been grouped into four categories. The first, social interaction, includes social opportunities, relationships and friendship networks, leisure/recreational activities, and communications. The second, described as occupationaUmateria1 well-being, incorporates income, employment, and occupational relationships. Under accommodation, the third subcomponent, are included comfort, security, and utilization of neighborhood resources. The fourth subcomponent, access, includes knowledge of and use of services, education, possession and use of skills, and mobility. 3 . Social influences. As the individual with a disability comes to a view of his or her quality of life in the context of the environment in which he or she lives, it is essential that a model incorporate those societal factors that may influence the person’s subjective well-being. Hence the following elements are suggested: community attitudes toward disability, community values, state of the economy, political support provisions, incentives, disincentives, and provision of access. Within these three components the list of subcomponents is presently tentative but, from an examination of the research literature, together with the issues explored under the symbolic-interactionist rubic, and from the results of pilot studies, it would appear they represent important elements of a model. It is proposed that each component interacts with the others. The development of self is largely influenced by both functional and societal factors. Functional behaviors are, in turn, affected by societal influences and the level of selfdevelopment. Societal influences possibly exert a strong moderating force on both of the other components; however, societal influences are also amenable to change, particularly if people with disabilities are assertive and demonstrate skillful control over their lives. The strength of the model lies in its emphasis on how well the person with a disability perceives him- or herself within a community. What is paramount is how that person grows and develops as an autonomous individual, albeit in an interdependent societal framework, with an opportunity to achieve his or her hopes and aspirations. The models described above are in part a reflection of some of the central themes in the study of quality of life outside the field of developmental disability. Described in the models, however, is a much stronger recognition of the need to involve the persons themselves in determining the major determinants of their quality of life. Furthermore, the environmental goodness-of-fit concept, which

270

Trevor R . Parmenter

sits comfortably with the view that disability is largely an environmentally determined phenomenon, is adding a much needed dimension to our understanding of what constitutes a life of quality for people with a developmental disability.

IV.

RECENT EMPIRICAL STUDIES

This section reviews an internationally representative sample of studies that have sought to assess the effectiveness of program outcomes in the context of the quality of life of the individuals concerned. In a number of cases the theoretical models described above have formed a framework for the conduct of the study.

A.

Canada

Brown et al. (1989) investigated quality-of-life outcomes for a sample of 240 intellectually disabled people across five agencies located in the four western provinces of Canada. As a prospective 3-year study, repeated measures related to vocational, social, educational, home living, and leisurehecreational aspects of learning and life-style. These indicators were drawn from the quality-of-life model developed by Brown and colleagues and described above. The results Brown and colleagues found were fairly depressing. On both objective and subjective indicators the clients of the several programs did not show many significant gains over the 3-year period, although the extent of the variance within and between groups makes comparisons difficult to interpret. There was evidence that, although on the average the individuals improved in work skills, there was a lack of progress in social education skills. Despite the fact that many of the clients of the programs had high levels of skills in the vocational and home living areas, there was little evidence of movement into community-based work or living situations, even though there was policy emphasis on community-based work. The study found that in the area of objective indicators there was a high relationship between poor performance and the absence of specific training programs. On subjective measures about half of the clients reported that they enjoyed living where they were and a majority stated they had friends, most of whom were selected from school, agency, or work. Analyses of changes in clients’ perceptions over the study showed that for males significant improvements were reported in three categories: “being more happy now,” “having improved work skills,” and “having better co-worker relations.” Females expressed gains in a larger number of categories: “reading,” ‘‘laundry,’’ “getting along with others,” “leisure time activities,” “making decisions,” “being happy,” “work skills,” and “co-worker relations.” Three specifically subjective questions were asked of the clients: What do you

QUALITY OF LIFE A N D DEVELOPMENTAL DISABILITIES

27 1

worry about? What would you like to change? Do you enjoy living where you are? Many of the clients’ concerns related to their individual health and performance, to starting and maintaining relationships, and to items dealing with the family. The most frequent response to the question What do you want to change? related to the clients’ desire to change where they were living. The next most frequent response was a desire for an increase in leisure activities, followed closely by a request for change in employment and improvement in personal relationships. Other responses included a desire to earn more money, improve health, and have more or different friendships. There was a high level of positive responses to the question Do you enjoy living where you are?, clearly outweighing the concerns and reservations. Brown et al. pointed out the idiosyncratic nature of the reasons given for preferences for a particular residence. For instance, one .person liked where he lived because he had a fireplace in his room. This, suggested Brown et al., demonstrated how important individual variations in preferences were to the concept of quality of life, whether it be for disabled or nondisabled people. Other significant themes that demonstrated the importance of individuals meeting their own personal needs included freedom of choice, the need for family affection and contact, and the opportunity to use leisure time profitably. As a result of their detailed analyses of the various indices measured, Brown et al. (1989) made 104 recommendations for program improvement, ranging from advice concerning the development of agency policy, administration, and staff training to the development of more individualized program planning for the clients. Although the quality-of-life model developed by Brown and colleagues served as a general framework for their investigation, a weakness noted in their report was the absence of a detailed discussion of the results using the framework chosen. For instance, there was little discussion of the concept of quality of life being an interaction between the individual and his or her environment; nor was there a sufficient analysis of the discrepancy model of quality of life put forward earlier, that is, the discrepancy between a person’s achieved and unmet needs and desires. Another aspect of the model that was touched on too briefly was the extent to which an individual increasingly controls aspects of his or her life regardless of the original baseline. One gains the impression that as such reports are directed primarily toward policy planners, there is a natural tendency on the part of researchers to highlight these more objective indices as they tend to fit more comfortably into funding criteria and legislative edicts. Although concepts such as freedom of choice, happiness, empowerment, and satisfaction are readily accepted as important variables, they do not seem to rate as highly as those variables for which ”hard” data can be obtained.

272

B.

Trevor R . Parmenter

New Zealand

A cross-sectional study examined the relationships between employment status, age, quality of life, and well-being for three cohorts of school leavers from classes for students with mild intellectual disabilities in Auckland, New Zealand (Tosswill, k k , & Wilton, 1991). The “early” cohort graduated through 19751979, the “middle” cohort through 1980-1984, and the “late” cohort through 1985-1989. There were 30 individuals in each cohort. Four published measures were used: Affectometer (Kamman & Flett, 1983), Quality of Life Questionnaire (Schalock et al., 1990), Self-Esteem Scale (Rosenburg, 1965) and General Health Questionnaire GHQ. 12 (Goldberg, 1972). The trend in employment status across the three cohorts indicated that, owing to recent high unemployment in the general population, only 23% of recent graduates were in competitive employment, compared with 43% of the early cohort and 47% of the middle cohort. Adults in competitive employment reported a better quality of life, higher self-esteem, higher levels of well-being, and fewer indicators of psychiatric illness than either adults in sheltered work or unemployed adults. The oldest adults experienced a better quality of life and fewer indicators of psychiatric illness than those who left school more recently, adding, suggested the researchers, some support to Edgerton’s ( 1990) conclusion that adaptation tends to improve over time and to the suggestion of a number of authors that securing and holding a job are critical factors in determining a person’s overall quality of life.

C.

Finland

In 1989 the Finnish Association on Mental Retardation launched a research project entitled The Quality of Life and Mental Disability, the goal of which was to study the personal views and experiences of mentally disabled persons, to describe the quality of their life in terms of work, housing, and leisure. A significant factor influencing the overall methodology used in this project was the tradition of Finnish sociological research, which throughout the 1980s focused largely on welfare, the quality of life, and the actual way of life itself. The Quality of Life project had three goals (Myhrman & Ohman, 1989): 1. To describe the quality of a mentally disabled person’s life in Finland by studying the quality of lile in different functional units, such as community residences, institutions, and sheltered work, and by analyzing the interdependence of the quality of life and the quality of services provided. 2. To identify the factors affecting the quality of a mentally disabled person’s life by assessing the needs of a disabled individual, for example, in terms of social relations and activities, ego and self-image, and emotional life by

QUALITY OF LIFE A N D DEVELOPMENTAL DISABILITIES

213

assessing individual capacities and resources, and by assessing the environmental factors affecting an individual’s life. 3. To develop methods for improving the quality of a mentally disabled person’s life by supporting and developing a disabled individual’s capacities and by improving the working methods of staff and other environmental factors. The project is being conducted in three stages. The first stage, the development of a quality-of-life instrument, has followed procedures established by Goode (1988b). The method is based on group discussions including four to six mentally disabled persons, one to two parents, and one to two staff members. Each working group has eight 3-hour sessions. Over the 24 hours of discussion set topics are explored commencing with the question, What is a good life? Other sessions discuss topics such as important needs and ways to fulfill them in different settings. Sessions were analyzed using the matrix method developed by Goode ( 1988b). This study used two matrices. The first was a need/goal matrix consisting of three dimensions. The first dimension included work, housing, and leisure time; the second concerned individual needs such as friendship, self-image, and security (based on a Scandinavian quality-of-life study: Kebbon, 1984); and the third consisted of different stages of integration. The second matrix analyzed resource implications, making it possible to analyze each need separately according to the resources needed to fulfill them in different stages of integration. Work currently in progress is employing the QOL instrument developed out of the content analysis described above, in interviews with loo0 adults with mental retardation across Finland. The essential goal of the project is to test the capacity of the service system to satisfy the expressed needs of its clients.

D.

Sweden

The County Council Vastmanland in Sweden is adopting an approach similar to that described in the Finnish project. Methods to measure the quality of services to people with mental retardation served by the County Council are being based on the people’s own opinions and expressed needs (Drugge, 1990). Central to the philosophy underpinning this development is the concept that disability itself shall be regarded as a lack of ability of society to eliminate the consequences of the disability instead of the disability as perceived within the individual. This approach agrees well with the World Health Organization’s ( 1980) definition of “handicap,” which stresses the interaction between disability and environmental factors. It also parallels the thinking behind the American Association on Mental Retardation’s 1991 draft definition of mental retardation that suggests that “the existence of disabilities in adaptive skills occurs

214

Trevor R . Parmenter

within the context of community environments typical of the individual’s age peers and is indexed to the person’s individualized needs for support” (American Association on Mental Retardation, 1991, p. I). County Vastmanland has a population of 256,000 people, 1326 of whom have mental retardation. The Council is currently developing two projects that are designed to improve the quality of services and the quality of life of those with mental retardation. The first project involves the development of individual plans as a basis for planning the services that the person with mental retardation demands. An individual plan is established for each person with mental retardation by interviewing the person alone or together with an advocate (a parent or someone who knows the person well) in the case of a person with a severe disability. The subsequent plan is computerized so it can be conveniently updated. Objectives obtained are recorded so that the actual quality of performance may be assessed. These data are used to plan future service needs and estimated costs. A second project under development assesses an individual’s quality of life along several dimensions, including respect, commitment, accessibility, nearness, and security. The essential ingredient of quality of life is defined as “the worthiness the person experiences in hidher specific situation” (Drugge, 1990, p. 7). To develop a measurement instrument that would reflect consumer satisfaction or dissatisfaction, reference groups of people with a knowledge and experience in working with people with mental retardation and groups consisting of persons with mental retardation and their parents were formed. The latter groups especially helped to define the level of complexity of the questions and ways in which questions should be asked. For those with severe levels of retardation, rather than asking parents or staff to complete the questionnaire, it was recommended that observation techniques be used to establish valid responses. A second Swedish study (Sonnander, 1990) appraised the quality of life of 217 individuals living in group homes in three representative Swedish counties. The instruments used in the study, which were based on the principle of normalization as enunciated by Nirje (1980), assessed quality of life along four dimensions: activity, relations, self-image, and mood (Naess, 1979). Thus, the personal experiences of the people studied were of paramount importance. The results showed that in general a medium level of normalization had been reached in that all people in the group homes had access to twice as many activities and other services in their community as they used or from which they could benefit. With respect to quality of life, the scores obtained showed that the group as a whole had a rated quality of life around or below the mean on a 5point scale. More than two-thirds of the interviewed people were assigned scores below the mean. Higher scores were attained by people in programs that almost approximated normal community life. An interesting finding was that deficits in quality of life were ascribed to the

QUALITY OF LIFE AND DEVELOPMENTAL DISABILITIES

275

fact that the people were not participating socially in the community at large. Although they lived in the community physically they were on the fringe socially. The implication of this finding is that considerable attention needs to be paid to the development of interpersonal skills in addition to the usual emphasis on practical living skills. There is also a need to help these people develop informal social networks that do not depend on paid professionals.

E.

United States

Thurlow, Bruininks, and Lange (1989) of the Institute on Community Integration at the University of Minnesota compared the postschool status of three groups of people with moderate to severe mental retardation who had been out of school for 7 to 10 years, 3 to 5 years, and 1 to 2 years. Comparisons of former students as a function of time out of school revealed few differences on employment-related variables or on other quality-of-life variables. TWO instruments were used in the study: (1) the Post-School Transition Study Survey Interview (PTS) and ( 2 ) the inventory for Client and Agency Planning (ICAP) (Bruininks, Hill, Weatherman, & Woodcock, 1986). Both instruments were administered by a trained interviewer to a respondent familiar with the subject. The PTS, developed by a task force consisting of the research team, school personnel, and local site coordinators, elicited objective data in the following areas: employment, education, social participation, support payments, social adjustment and living skills, health/physical status, family/household characteristics, living arrangements, service and program information, citizenship status, and miscellaneous information. The ICAP, a tool for managing information in areas for planning and evaluating services for people who are disabled, elicited information on diagnostic and health status, as well as normative scores for adaptive behavior and problem behaviors; information on service level, service history, current placements, and projected service needs; and data on support services and social and leisure activities. Thus, ICAP provides information that can be used to compare the adjustment of different groups to the adult world. It is clear that both instruments basically provide inventories of objective indicators of quality of life; however, for the domains of work, day programs, and living arrangements questions were asked about the subject’s level of satisfaction. Despite the lack of significant differences between the three cohorts on the major variables (unlike in the New Zealand study reported above), there were significant correlations between measures of personal competence and composite outcome measures, suggesting that postschool outcomes are related to variables such as independence and adaptive behavior and, to some extent, to maladaptive behaviors.

276

Trevor R . Parmenter

A principal-components analysis of 21 variables from the PTS and ICAP produced an eight-factor solution, four identified areas of personal competence, and four community adjustment dimensions (Bruininks, McGrew, Thurlow, & Lewis, 1988). In the area of personal competence, the factors identified were personal independence, maladaptive behavior, physical mobility, and physical complications; for community adjustment, the factors were social/recreation/leisure, social and service support, financial independence, and community independence/integration. Thurlow et al. (1989) concluded that in assessing the community adjustment of former school students, major aspects to be considered are functional behaviors, personal competence, and social, economic, and community integration. More recent work by the researchers from the Institute on Community Integration (McGrew & Bruininks, 1991) has supported this conclusion.

F.

United Kingdom

McGuire, Choon, and Akuffo (1 99 1) assessed the quality of life of 19 elderly people with a developmental disability who had been living in two privately owned, government-approved group houses in London for between 1 and 2 years. Prior to living in community-based homes these people had resided in large residential institutions. The majority of residents were considered to have a moderate to severe level of intellectual disability. Quality of life was assessed by a modified version of a life-style questionnaire devised by Raynes, Johnson, Sumpton, and Thorp (1987). The areas assessed included daily activities, involvement in domestic tasks, extent of communitybased activities, and extent to which residents are responsible for, and can make decisions about, their own lives. The results indicated that the involvement of residents in community-based activities was limited, although the authors suggested that the level was possibly as high as it could be for nondisabled, elderly people. In the area of decision making, it was observed that although choices were available (for example, choosing when to go to bed), many of the residents were unable to break away from the institutionalized routines by which they had lived for 50 or more years. One gains the impression from the results of this study that there had been little effort by staff to establish community networks or to implement programs to teach decision-making skills to the residents, weaknesses also noted in the following Australian example.

G.

Australia

Parmenter, Briggs, and Sullivan (1991) investigated the quality of life of 30 residents with severe developmental disabilities who had moved from a hospital

QUALITY OF LIFE A N D DEVELOPMENTAL DISABILITIES

277

setting to seven community houses. With the three-component model of quality of life described above as a guide, a number of specific variables were assessed through a variety of data collection techniques. Variables explored included functional skills acquisition, amount of say the residents have in their lives, individual community-based activities, attitude change by staff and parents toward community living, and parents’ perceptions of their changed roles. Data were collected using interviews with residents (including those who had little verbal communication skills) and participant observation techniques. Staff and parents also completed the same interview schedules as those administered to the residents, to verify the results of the resident interviews. When asked if they felt very happy, just OK, or unhappy, 19 residents indicated that they felt very happy, 10 felt happy, 1 felt just OK, and none felt unhappy. In terms of friendship, 21 residents were happy with the friends they had; 7 were not. The majority of their friends naturally came from their group home; 3 indicated the family home, 4 a staff member, and the remainder school friends. When asked if they were happy with the move from the hospital, 27 indicated “yes” and 3 “no.” Of the latter, two subsequently returned to hospital and the other expressed a desire to live at home with his parents. In terms of functional skills the study revealed that little attention was being given by staff to the development of basic self-care skills and those skills necessary for the routines of daily living. This was attributable to staff shortages, high staff turnover, and lack of staff training. Although residents reported being able to make choices in activities such as selection of menu, clothing, weekend activities, TV programs, time of going to bed and participation in group activities, from an analysis of the interviews and objective observation, it appeared that much more could be done to teach choice and decision-making skills and to provide opportunities where choices could be exercised. In the area of societal influences the data on community access revealed fairly limited participation by individual residents in community activities. This was more a reflection of the inability or inaction of staff in setting up meaningful community networks, than reluctance by community organizations to accept these people. The data assembled in this study support aspects of the quality-of-life model presented earlier. In practical terms, the data demonstrate that it is possible to address issues such as satisfaction, happiness, and decisionmaking in a population that has a severe level of disability. A trend toward increasing levels of empowerment of the residents was noted. A more detailed ethnographic study using longer-term participant observation techniques would possibly throw additional light on the possible changes in the development of a more positive identity by the residents.

278

Trevor R . Parmenter

V.

METHODOLOGICAL ISSUES IN MEASURING THE QUALITY OF LIFE OF PEOPLE WITH DEVELOPMENTAL DISABILITIES

Readers are referred to a comprehensive analysis of measurement issues by Heal and Sigelman (1990), a brief summary of which is presented here. Heal and Sigelman have suggested a methodological taxonomy for assessing quality of life. The first dimension is whether the measures are objective or subjective. The second relates to whether the measure is absolute or relative; that is, it can directly index a person’s quality of life or it can compare it with a standard. Third, the quality of life can be directly reported by the subjects of the study or it can be assessed by someone else. Fourth, the index can be authored or generated by the researcher or by the subjects themselves. Heal and Sigelman (1990) have explored a number of difficulties that may arise depending on the decisions implied by the taxonomy. For instance, as indicated earlier, various subjective measures of well-being do not correlate highly with objective indices of quality of life. In other words, they suggested, “method variance tends to dominate substantive variance in quality of life assessments” (p. 164). There are also difficulties, especially for people with mental retardation, in responding to relative measures that require a comparison of current with past or future quality of life. This problem was evident in the study by Parmenter et al. (1991) reported above. Many individuals had difficulty in responding to the question as to whether they felt happier in their new accommodation compared with their former hospital residence. The use of respondents other than the disabled people themselves also raises questions of reliability, a danger noted in the first Swedish study. The use of independent participant observation techniques does help to mitigate the biases that may occur. An issue central to Goode’s (1991) concern with psychometric approaches to the assessment of quality of life is the question of who generates the list of specific life circumstances to be evaluated. An area for future study would be an assessment of the potential differences between findings based on investigator-authored measures and those based on client-authored measures. There is substantial literature on factors that affect responses in survey research, especially systematic response effects that are a function of the way questions are recorded and the way interviews are conducted. For people with developmental disabilities especially, the risk of acquiescent responding is very high. There is also the problem of socially desirable responding; that is, respondents will tend to present themselves in a favorable light. For instance, studies of quality of life in the general population have shown that people generally rate the quality of their lives above a neutral point, regardless of how “neutral” is anchored or described (Andrews & Withey, 1976; Diener, 1984). In a recent Australian study, Romeo and Cummins (1991) found that a significant number of

QUALITY OF LIFE AND DEVELOPMENTAL DISABILITIES

279

respondents to a quality-of-life questionnaire designed for people with intellectual disabilities were unable to use a 5-point Likert scale. There is also quite a detailed literature concerning difficulties that may be experienced when interviewing people with mental retardation (see Sigelman et al., 1983; Heal & Chadsey-Rusch, 1985). Heal and Sigelman (1990) have suggested that multiple methodologies be employed to reduce these difficulties. For instance, they have urged that greater opportunities be given for people with developmental disabilities to speak for themselves, so they may tell us how they perceive their lives and ways in which they would like their lives to change. In terms of instrument development they have recommended either-or multiplechoice questions accompanied by pictures wherever feasible. There is also a need to build into interview schedules checks for response bias that can be used later to adjust scores to remove the effects of their bias. Finally, information-gathering techniques other than client interviews may be used in place of interviews or to corroborate responses along the lines described by Parmenter et al. (1991).

VI.

CONCLUSIONS

This article has reviewed attempts that have been made to conceptualize and assess quality of life of people, from the perspective of both the generic literature and those in the disability field. The term quafiry of fife has on the surface a seductive simplicity, so much so that it almost seems improper to question its utility as an index of the effectiveness of policies and programs for people with developmental disabilities; however, a number of writers have urged caution in our quest to come up quickly with instruments that can measure the concept. Indeed, Luckasson (1990) has gone so far as to suggest that the disability community should eschew the use of quality of life as an evaluation term, because of the inherent danger of it becoming a justification for the denial of rights to people with disabilities. There are examples of this happening, such as neonates who are born with severe disabilities and who are candidates for euthanasia because their potential for a “quality of life” is very doubtful. The use of an arbitrary metric of quality of life in the field of developmental disabilities could just as easily be applied when resources are being rationed as they are being applied in the general health area in some countries. Luckasson’s concerns are reinforced by lhrnbull and Brunk (1990), who urge us not to overlook the fact that the question of quality of life for people with developmental disabilities is embedded in the larger perspective of what constitutes quality of life for the general population. And this raises the issue of what is “community”? Sadly, many informal mechanisms for facilitating the integration of people with developmental disabilities into the life of a community

280

Trevor R . Parmenter

become formalized into a government-sponsored “program” because some modest levels of funding are required. Just how we solve the problem of providing resources to informal networks that are the lifeblood of genuine communities, and at the same time maintain fiscal accountability, is a serious challenge for policy planners and governments. Also, as Bowles (1988b) has cogently pointed out, disability is a phenomenon with roots in the social, political, and economic processes of our society, rather than in the clinical features of the individual. At this point in time, in particular, it is necessary, as Turnbull and Brunk have suggested, to emphasize the needs of these people for “informal support, friendships, intentional communities, and rights to association between people with and without disabilities” ( 1990, p. 205). There is an inherent danger that we have overbureaucratized the disability world, replete with its laws and armies of well-intentioned professionals, a point forcefully made by Blatt (1979) in his fine essay “Bureaucratizing Values.” In this context Fulcher (1989) has drawn a distinction between the forces of democratism and its opponent, professionalism. Democratism supports the notion that those affected by decisions should take a genuine part in debating the issues and making the decisions. Professionalism, on the other hand, is the view that the experts know best, and has been a major tactic in the historical struggle by professionals to gain control of an area of occupational life. This view parallels that of Turnbull and Brunk who fear that the overbureaucratization of services to people with developmental disabilities will lead to increasing isolation of the individual person from the nondisabled world, despite the multiplicity of individualized plans and other mandated mechanisms that are supposed to ensure quality programs. We have yet some distance to go in ensuring that people with developmental disabilities become OF communities rather than simply IN them. A further worrying aspect is the ascendancy in industrialized countries of economic rationalism, which is leading to a marked reduction in distribution of resources to programs for people with disabilities generally. There appears to be a degree of sophistry in the assumption by governments that disability disappears once the individual has been “integrated” into the community. Soder (1984) has wryly observed that “making the needs of the severely disabled invisible fills an economic function [that results in] strong political pressure for cuts in public sector expenditure” ( p. 33). At least from a Swedish perspective Soder’s main hypothesis is that “the optimistic and active ideology of the 1960s is changing as a result of economic difficulties into a pessimistic and passive one which is contributing to making the need the mentally retarded have of special resources disappear from sight” (p. 16). The strong suspicion that value systems are being misused by policy planners should make us vigilant concerning the possible ossification of the concept of quality of life. On a more optimistic note, Mercer’s (1991) prediction that by the year 2000 there will be a greater acceptance of the multiparadigmatic basis of the concept of

QUALITY OF LIFE AND DEVELOPMENTAL DISABILITIES

28 1

mental retardation will, it is hoped, lead to a greater appreciation of the need to approach the needs of this group from a more individual perspective. The recent developments in defining mental retardation in the context of support needs are a significant step in ensuring that the quality of life of people with developmental disabilities will be viewed more from their perspective, rather than from some bureaucrat’s predetermined notion as to what constitutes “quality.” REFERENCES Adams. D. (1941). Correlates of satisfaction among the elderly. Gerontologist, I / . 64-71, American Association on Mental Retardation ( 1991, May). Classification in mental retardation (Draft). Washington, DC: Author. Andrews, F. M. (1974). Social indicators of perceived life quality. Social Indicators Research, I , 279-299. Andrews, F. M., & McKennell, A. C. (1980). Measures of self-reported well-being: Their atfective, cognitive and other components. Social Indicutors Research, 8, 127-155. Andrews, F. M., & Withey, S. B. (1976). Social indicators of we//-being:American’s perceptions of life qua1it.y. New York: Plenum Press. Atkinson, T. (1977). I s sutisfurtion u good meusure qf the perceived quality of life? Institute of Behavioral Research, York University, Toronto. Mimeo. Barton, L., & Tomlinson, S . (1984). The politics of integration in England. In L. Barton & S. Tomlinson (Eds.), Special education and social interesfs. London: Croom Helm. Baumeister. A . A. (1981). Mental retardation policy and research. The unfulfilled promise. Americun Joiirnal of Mentul Deficiency, 85, 449-456. Becker, H. S . (1983). Outsiders: Studies in the sociology of deviance. New York: Free Press. Begab. M. J., & Richardson, S . A. (Eds.) (1975). The mental1.y retarded and society: A social scienw perspective. New York: Pergamon Press. Beyer, J. M., & Trice, H. H. (1982). The utilization process: A conceptual framework and synthesis of empirical findings. Administrative Science Quarterly. 27, 591-622. Blatt, B. (1979). In und out of’mentul returdation. Essays 011 educability, disability. and human policy. Baltimore, MD: University Park Press. Bogdan. R., & Kugelmass. J. (1984). Case studies of mainstreaming. A symbolic interactionist approach to special schooling. In L. Barton & S. Tomlinson (Eds.), Special education and social interests. London: Croom Helm. Bowles, W. (1988a). Quality of life: Models. marhs and mystev. Unpublished manuscript, Macquarie University. Bowles, W. (1988b). A sociological framework for the study of disability: An introduction. Unpublished manuscript, Macquarie University. Brahams, D. (1991). Rationing health care: Ethical and legal considerations and QALYs. MedicoLegul Journal, 59, 4. Bronfenbrenner. V. (1977). Toward an experimental ecology of human development. American Psvcho/i~gist,32, 5 13-53 1. Brown, R. I . , Bayer. M. B., Kt MacFarlane, C. (1988). Quality of life amongst handicapped adults. In R. I . Brown (Ed.), Quality of life for handicapped people. London: Croom Helm. Brown, R . I . , Bayer. M., & MacFarlane, C. (1989). Rehabilitation programmes: Performance and yualitv ojlife of adults with developmental handicaps. Toronto: Lugus Productions. Brown, R. I . , & Hughson, E. A. (1980). Training the developmentally handicapped adult. Springfield, IL: Charles C. Thomas.

282

Trevor R . Parmenter

Bruininks, R. H., Hill, B. K., Weatherman, R. F., & Woodcock, R. W. (1986). Inventory for client and agency planning. Allen, TX: PDLM Teaching Resources. Bruininks, R. H., McGrew, K. S., Thurlow, M. L., &Lewis, D. R. (1988). Dimensions of community involvement among young adults with intellectual disabilities. Paper presented at the 8th IASSMD Congress, Dublin. Bruininks, R. H., Meyers, C. E., Sigford, B. B., & Lakin, K. D. ( U s . ) (1981). Deinstitutionalization and community adjustment of mentally retarded people (AAMD Monograph No. 4). Washington, DC: American Association on Mental Deficiency. Burbach, H. J. (1981). The labeling process. A sociological analysis. In 1. M. Kauffman & D. P. Hallahan (Eds.). Handbook of special education. Englewood Cliffs, NJ: Prentice-Hall. Cadman, D., & Goldsmith, C. (1986). Construction of social value- or utility-based health indices. The usefulness of factorial experimental design plans. Journal of Chronical Disability. 39, 643-65 1 . Campbell, A. (1981). The sense of well-being in America. New York: McGraw-Hill. Carr-Hill, R. (1989). Assumption of the QALY procedure. Social Science and Medicine, 29. 469-477. C h a r d , M. B. (1974). Sociology of deviant behaviour. New York: Holt, Rinehart & Winston. Close, D. W., & Foss, G . (1988). Approaches to training: The social skills needed for quality of life. In R. 1. Brown (Ed.), Quality of life for handicapped people. London: Croom Helm. Cottone, R. R. (1986). Toward a systematic theoretical framework for vocational rehabilitation. Journal of Applied Rehabilitation Counseling, 17, 4-7. Dalkey, N. C., Lewis, R., & Snyder, D. (1972). Measurement and analysis of the quality of life. In N. C. Dalkey, D. L. Rourke, R. Lewis, & D. Snyder (Eds.), Studies in the quality of life. Lexington, MA: Lexington Books. De Jong, G. (198 I). Environmental accessibility and independent living outcomes. Directions for disability policy and research. University Center for International Rehabilitation, Michigan State University. Dickey, B . , Gudeman, J. E., Hellman, S., Donatelle, A,, & Grimspoon, L. (1981). A follow-up of deinstitutionalized chronic patients four years after discharge. Hospital and Community Psychiatry, 32, 326-330. Diener, E. (1984). Subjective well-being. Psychological Bulletin, 95, 542-575. Dohrenwend, B. S. (1973). Life events on stressors: A methodological inquiry. Journal of Health und Social Behavior. 14, 167-175. Dohrenwend, B. S . , & Dohrenwend, B. P. (1978). Some issues in research on stressful life events. Journal of Nervous and Mental Diseases, 66, 7-15. Dossa, P. A. (1989). Quality of life: Individualism or holism? A critical review of the literature. International Journal of Rehabiliration Research, 12, 12 1 - 136. Drugge, C. (1990). Using the opinion of people with mental retardation to measure quality of services. Vasteras, Sweden: County Council Vastmanland, Social Welfare for People with Mental Retardation. Drummond, M. (198 I). Studies in economicappraisal in heulth care. Oxford: Oxford University Press. Edgar, E. (1988). Employment as an outcome for mildly handicapped students: Current status and future directions. Focus on Exceptional Children, 21. 1-8. Edgerton, R. B. (Ed.) (1984). Lives in process: Mildly retarded adults in u large city (AAMD Monograph No. 6). Washington, DC: American Association on Mental Deficiency. Edgerton, R. B. (1990). Quality of life from a longitudinal research perspective. In R. L. Schalock (Ed.), Quality of life. Perspectives and issues. Washington, DC: American Association on Mental Retardation. Edwards, J. N., & Klemmack, D. L. (1973). Correlates of life satisfaction: A re-examination. Journal of Gerontology. 28, 407-502.

QUALITY OF LIFE AND DEVELOPMENTAL DISABILITIES

283

Emerson, E. B. (1985). Evaluating the impact of deinstitutionalization on the lives of mentally retarded people. American Journal of Mental Deficiency, 90,277-288. Evans, D. R., Bums, I. E., Robinson, W. E., & Garrett, 0. J. (1985). The quality of life questionnaire: A multidimensional measure. American Journal of Community Psychology. 13. 305-322.

Finkelstein, V. (1980). Attitudes and disabled people: Issues for discussion. New York: World Rehabilitation Fund, Inc. Flanagan. J. C. (1978). A research approach to improving our quality of life. American Psychologisr, 33, 138-147. Flanagan, J. C. (1982). Measurement of quality of life: Current state of the art. Archives of Physical Medicine and Rehabilitation. 63, 56-59. Fulcher, G . (1989). Disabling policies? A comparative approach to education policy and disability. London: The Falmer Press. Gill, W. M. (1984). Subjective well-being: Properties of an instrument for measuring this (in the chonically ill). Social Services Medicine, 18, 683-691. Goffman, E. (1963). Stigma: Notes on the management of spoiled identity. Englewood Cliffs, NJ: Prentice-Hall. Goldberg, D. P. (1972). The detection of psychiatric illness by questionnaire. London: Oxford University Press. Goode, D. A. (1984). Socially produced identities, intimacy and the problem of competence among the retarded. In L. Barton & S . Tomlinson (Eds.), Special education and social interests. London: Croom Helm. Goode, D. A. (1988a). Quality of life for persons with disab es. A review and synthesis of the literature. New York: Mental Retardation Institute, Westchester County Medical Center. Goode, D. A. (1988b). Discussing quality of life: The progress andfindings of the work group on qualify of life for persons with disabilities. Valhalla, NY: New York Medical Goode, D. A. (1991). Quality of life research: A change agentfor personsfor dis presented to the 1991 American Association on Mental Retardation National Meeting Round Table Program, Washington, DC. Grimes, D. (1985). New directions. Report of the handicapped programs review. Canberra, Australian Govt. Publishing Service. Gutek, B. A,, Allen, H., Taylor, T. R., Lau, R. R., & Majchrzak, A. (1983). The importance of internal referents as determinantsof satisfaction.Journal of Community Psychology, I I , 1 I 1- 120. Hall, J. (1976). Subjective measure of quality of life in Britain: 1971 to 1975. Some developments and trends. Social Trends, 7, 25-33. Halpern, A. S . (1985). Transition: A look at the foundations. Exceptional Children, 51, 479-486. Halpern, A. S., Close, D. W., & Nelson, D. J. (1986). On my own: The impact of semi-independent living programs for adults with mental retardation. Baltimore, MD: Paul H.Brookes. Halpern, A. S., Nave, G., Close, D. W., Nelson, D. J. (1986). An empirical analysis of the dimensions of community adjustment for adults with mental retardation. Australia and New Zealand Journal of Developmental Disabilities, 12, 147- 157. Hams, I. (1987). QALYfying the value of life. Journal of Medical Ethics, 13, 117-123. Headey, B., Holmstrom, E., & Wearing, A. (1984). The impact of life events and changes in domain satisfactions on well-being. Social Indicators Research, 1.5, 203-227. Heal, L. W., & Chadsey-Rusch, J. (1985). The lifestyle satisfaction scale (LSS): Assessing individual's satisfaction with residence, community setting, and associated services. Applied Reseurch in Mental Returdation, 6 , 475-490. Heal, L. W., & Sigelman, C. (1990). Methodological issues in measuring the quality of life of individuals with mental retardation. In R. L. Schalock (Ed.), Quality of life. Perspectives and issues. Washington, DC: American Association on Mental Retardation.

284

Trevor R . Parmenter

Hirst, M. (1989). Patterns of impairment and disability related to social handicap in young people with cerebral palsy and spina bifida. Journal of Biosocial Science, 21, 1-12. Holmes, T. H., & Rahe, R. H. (1967). The social readjustment scale. Journal of Psychosomatic Researrh, 1 1 , 213-218. Hurst, A. (1984). Adolescence and physical impairment: An interactionist view. In L. Barton & S . Tomlinson (Eds. ), Special education and social interest. London: Croom Helm. Ingelhart, R., & Robier, J. R. (1986). Aspirations adapt to situations-But why are the Belgians so much happier than the French? In F. M. Andrews (Ed.), Research on the quality of life, Ann Arbor: Institute of Social Research, University of Michigan. Kamman, R., & Flett, R. (1983). Sourrebook for measuring well-being with aflectmeter 2 . Dunedin: Why Not? Foundation. Kebbon, L. (1984). Noramliserng och linskvaliter. Uppsala: Projekt Mental Retardation. Klonoff, P.,Snow, W., & Costa, L. (1986). Quality of life in patients 2 to 4 years after closed head injury. Neurosurgery. 19. 735-743. Knoll, J. A. (1990). Defining quality in residential serices. In V. J. Bradley & H. A. Bersani (Eds.), Quality assurance for individuals with developmental disab es. It's everybody's business. Baltimore, MD: Paul H. Brookes. Landesman, S. (1986). Quality of life and personal life satisfaction: Definition and measurement issues. Mental Retardation, 24, 141- 143. Landesman, S., & Butterfield, E. C. (1987). Normalization and deinstitutionalization of mentally retarded individuals. Controversy and facts. American Psychologist, 4 2 , 809-8 16. Landesman, S., & Vietze, P. (Eds.) (1987). Living environments and mental retardation. Washington, DC: American Association on Mental Deficiency. Landesman-Dwyer. S . (1985). Describing and evaluating residential environments. In R. H. Bruininks and K. C. Lakin (Eds.), Living and learning in the least restrictive environment. Baltimore, MD: Paul H. Brookes. Landesman-Dwyer. S., & Berkson, C. (1984). Friendships and social behavior. In 1. Wortis (Ed.), Mental retardation and developmental disabilities: A n annual review. New York: Plenum Press. Lee, R. G . , & Miller, F. A. (1990). The doctor's changing role in allocating US and British Medical Services. Law, Medicine and Health Care, 18, 69. Lehman, A. F. (1983). The well-being of chronic mental patients. Archives of General Psvchiatry. 40, 369-373. Lewis, S . , & Ryan, L. (1986). The quality of community and the quality of life. Soriologiral Spectrum. 6. 397-410. Lippman. L. (1976). Indicators of societal concern for mentally retarded persons. Social Indicators Research, 3, 181-215. Loomes, G., & McKenzie, L. (1989). The use of QALYs in health care decision making. Social Science and Medicine, 28, 299-308. Luckasson, R. (1990). A lawyer's perspective on quality of life. In R. L. Schalock (Ed.), Quality of life. Perspectives and issues. Washington, DC: American Association on Mental Retardation. McGrew, K. S., & Bruininks, R. H . (1991, March). Dimensions of personal competence and adjustment in the communify. Paper presented to International Conference on Mental Retardation, Hong Kong, March. McGuire, B. E., Choon, G . , & Akuffo, E. (1991). Community living for elderly people with an 1 disability: A pilot study. Ausrralia and New Zealand Journal of Developmental in D S, 17, 25-33. Mercer, 1. R. (1991). The impact of changing paradigms of disability on mental retardation in the year 2000. In Rowitz (Ed.), Menlal retardation in the year 2OOO. New York: Springer. Milbrath, L. W. (1982). A conceptualization and research strategy for the study of ecological aspects of the quality of life. Social Indicators Research. 10, 133-157.

QUALITY OF LIFE AND DEVELOPMENTAL DISABILITIES

285

Mitchell, D. R. (1986). A developmental systems approach to planning and evaluating services for persons with handicaps. In R. 1. Brown (Ed.), Management and administration of rehabilitation programmes. London: Croom Helm. Moms, P. L. P., &Jones, B. (1989). Life satisfaction across treatment methods for patients with endstage renal failure. The Medical Journal of Australia, 150, 428-432. Murrell, S. A,, & Nonis, F. H. (1983). Quality of life as the criteria for need assessment and community psychology. Journal of Community Psychology. I I , 88-97. Myhrman, A , , & Ohman, A. (1989). Quality of life for mentally handicapped adults in Finland. International Journal of Rehabilitation Research, 12, 465-466. Naess, S. (1979). Livskvalitet: Om a ha det godt i byen ogpa landet. INAS Rapport, Oslo, 79:2. Nirje, B. (1980). The normalization principle. In R. J. Flynn & K. E. Witsch (Eds.),Normalization. social integration and community services. Baltimore, MD: University Park Press. Parmenter, T. R. (1980). Vocational trainingfor independent living. New York: World Rehabilitation Fund, Inc. Parmenter, T. R. (1986). Bridges from school to working life for handicapped youth: The viewpoint from Australia. New York: World Rehabilitation Fund, Inc. Parmenter, T. R. (1988). An analysis of the dimensions of quality of life for people with physical disabilities. In R. Brown (Ed.), Quality of life for handicapped people. London: Croom Helm. Parmenter, T. R. (1991). Has social policy left research behind? Australia and New Zealand Journal of Developmental Disab Parmenter, T. R., Briggs, L., & Sullivan, R. (1991). Quality of life: Intellectual disabilities and community living. Evaluation Journal of Australasia. 3, 12- 15. Powers, J., & Goode, D. (1986). Partnerships for people: Enhancing the quality of life of persons with disabilities through incentives management. Unpublished manuscript, The Rose F. Kennedy Center, Albert Einstein College of Medicine, Bronx, NY. Ramund, B., & Stenssman, R. (1988). Quality of life and evaluation of functions among people with severely impaired mobility and non-disabled controls. Scandinavia Journal of Psychology, 29. 17-144.

Raynes, N. U., Johnson, M., Sumpton, R. C., & Thorp, D. (1987). Comparison of the daily lives of four young adults who are mentally retarded. Journal of Mental Deficiency Research, 36. 303-3 10. Richardson, J . (1991). Economic assessment of health care: Theory and practice. Australian Economic Review, 1st quarter, 4-21. Ridley, S . A,, & Wallace, P. G. M. (1990). Quality of life after intensive care. Anaesthesia, 45, 808-8 13.

Riley, M. W., & Foner, A. (1968). Aging and society: An inventory of researchfindings. New York: Russell Sage Foundation. Rodgers. W. L., & Converse, P. E. (1975). Measures of the perceived quality of life. Social Indicators Research. 2, 127- 152. Romeo, Y., & Cummins, R. A. (1991). Com Qol-ID: Results of a pilot study of quality of life for people with intellectual disabilities living in the community. Paper presented to 1991 National Congress on Intellectual Disability, Deakin Universty, Geelong. Rosenburg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Rowles, J. (1989). Castigating QALYs. Journal of Medical Ethics. 15, 143-147. Schalock, R. L. (Ed.) (1990a). Quality of life: Perspectives and issues. Washington, DC: American Association on Mental Retardation. Schalock, R. L. (1990b). Attempts to conceptualize and measure quality of life. In R. L. Schalock (Ed.), Quality of life: Perspectives and issues. Washington, DC: American Association on Mental Retardation.

Trevor R . Parmenter

286

Schalock, R. L. (1991). The concept of quality of life in the lives of persons with mental retardation. Paper presented to 115th Annual Convention of the American Association on Mental Retardation, Washington, DC. Schalock, R. L., Keith, K. D., & Hoffman, K. (1990). 1990 Quality of life questionnaire: Standardization manual. Hastings, NE: Mid-Nebraska Mental Retardation Services. Schneider, M. (1975). The quality of life in large American cities: Objective and subjective social indicators. Social Indicators Research, I , 495-509. Schur, E. M. (1971). Labeling deviant behavior: Its sociological implications. New York: Harper & Row. Scott, R. A. (1972). A proposed framework for analyzing deviance as a property of social order. In R. A. Scott & J. D. Douglas (Eds.), Theoretical perspectives on deviance. New York: Basic Books. Sharp, R., & Green, A. (1975). Education and social control. London: Routledge & Kegan Paul. Sigelman, C. K., Schrenrock, C. J., Budd, E. C., Winer, J. L., Spanhel, C. L., Martin, P. W., Hromas. S., & Bersberg, G. J. (1983). Communicating with mentally retarded persons: Asking questions and getting answers. Lubbock: Research and Training Center in Mental Retardation, Texas Tech University. Siider, M. (1984). The mentally retarded: Ideologies of care and surplus population. In L. Barton & S. Tomlinson (Eds.), Special education and social interests. London: Croom Helm. Sonnander, K. (1990). Normalization and quality of life. Paper presented at Young Adult Institute Congress, New York. Spreitzer, E., & Snyder, E. E. (1974). Correlates of life satisfaction among the aged. Journal of Gerontology, 29, 454-458. Stoddard, S . (1978). Independent living: Concepts and programs. American Rehabilitation, 3, 2-5. Stryker, S. (1959). Symbolic interaction as an approach to family research. Marriage and Family Living, 21, 111-119. Thomdike, E. L. (1939). Your city. New York Harcourt, Brace & Co. Thurlow, M. L., Bruininks, R. H., & Lange, C. M. (1989). Assessing post-school outcomes for srudenrs with moderate to severe mental retardation, Minneapolis: Institute on Community Integration, University of Minnesota. Tosswill, J., 'hck, B., Wilton, K. (1991). Employment status, age and the quality of life and well being of adults with mild intellectual disability. In Crossing Boundaries: Presenr realiriesfuture possibilities. Proceedings of 15th National Conference of the Australian Association of Special Education Inc., Brisbane (pp. 477-482). 'hrnbull, R., & B ~ n k C. , L. (1990). Quality of life and public policy. In R. L. Schalock (Ed.), Quality of life. Perspectives and issues. Washington, DC: American Association on Mental Retardat ion. Ward, J., Parmenter, T. R., Riches, V., & Hauritz, M. (1986). A summative report of a work preparation program for mildly intellectually handicapped disabled school leaven. Australian Disability Review, 3, 7-15. Wegner, N. K., Mattson, M. E., & Furberg, C. D. (Eds.) (1984). Assessment of quality of life trials in cardiovascular therapies. New York: Le Jacq Publishing. Whelan, E., & Speake, B. (1981). Getting to work. London: Souvenir Press. Will, M. (1984). OSERS programming for the transition of youth with disabilities: Bridges from school to working life. Washington, DC:Office of Special Education and Rehabilitation Services. Williams, A. (1979). A note on '*trying to value life." Journal of Public Economies, 1 2 , 257-258. Williams, A. (1985). Economies of coronary artery bypass grafting. British Medical Journal, 291, 326- 329.

Williams, R. G. A. (1979). Theories and measurement in disability. Epidemiology and Community Health. 33, 32-47.

QUALITY OF LIFE A N D DEVELOPMENTAL DISABILITIES

287

World Health Organization ( 1 980). Iniernuiional classification of diseases (9th review). New York: Author. Zoner, R. M.(1986). Soundings from uncertain places: Difficult pregnancies and imperiled infants. In P. R. Dokecki & R. M. Zaner (Eds.),Ethics of dealing with persons with severe handicaps. Baltimore, MD: Paul H. Brookes. Zapf, W. (1987). German social report-Living conditions and subjective well-being, 1978- 1984. Social Indicators Research, 19, 25-38. Zautra, A. J. (1983). Social resources and the quality of life. American Journal of Community Psychology, 11, 275-290. Zautra, A. J., & Goodhart, D. (1979). Quality of life indicators: A review of the literature. Community Menial Health Review, 4-10. Zautra, A. J., & Reich, J. W. (1983). Life events and perceptions of life quality: Developments in a two-factor approach. Journal of Community Psychology, I I , 121- 132.

This Page Intentionally Left Blank

Index

A Abecedarian Project, new morbidity and, 165 Absenteeism, staff training and, 229 Acceptability rating scale, staff training and, 24 1 Accuracy scores, stimulus control analysis and, 69-70, 72, 95, 98 Achenbach Youth Self-Report, new morbidity and, 176 Achievement Place program, staff training and, 240 Acquisition stimulus control analysis and, 56 stimulus organization and, 33 Adaptation. modifiability of life path and, 154 Adaptive behavior new morbidity and, 169, 178, 190 quality of life and, 252, 256, 262, 273, 275 stimulus control analysis and, 62 Adaptive Behavior Scale Down syndrome and, 2 I3 staff training and, 226 Adjustment, modifiability of life path and, 142, 146-149 Adolescent parenting, new morbidity and, see New morbidity, adolescent parenting and Adversity, modifiability of life path and, 138139, 141-142, 148, 150 Affect, quality of life and, 254, 268 Aftereffects, perceptual deficits and, 5 Age. see also Chronological age; Mean chronological age; Mental age Down syndrome and, 198, 218-220 behavior, 214-217 cognitive aspects, 203-210 self-help. 210-214 modifiability of life path and, 139-140, 145, 147, 149. 151

new morbidity and, 167, 173 perceptual deficits and, 17-18, 20 quality of life and, 25 I , 274 staff training and, 226 stimulus organization and, 33, 35, 42 Aggressive behavior, Down syndrome and, 214 Alertness, modifiability of life path and, 145 Ambivalent attachments, new morbidity and, 183-185, 187 American Association on Mental Retardation perceptual deficits and, I5 quality of life and, 259, 273 Ames window, perceptual deficits and, I 1 Antecedents. staff training and, 236 Anxiety modifiability of life path and, 146-147 new morbidity and, 161, 176, 185, 188 APGAR scores, new morbidity and. 169 Applied behavior analysis, staff training and, 229 Arbitrary assignment, stimulus control analysis and, 81, 83 Arbitrary matching-to-sample, stimulus control and, 91-95 Arousal, sustained attention and. 117, 127-128 Asymmetry, stimulus organization and, 46 Attachment, new morbidity and, 178, 189 cognitive development, 182- I88 Attention new morbidity and, 159, 186-187 perceptual deficits and, 10- 1 1 , 17 stimulus control analysis and, 70 stimulus organization and, 34, 40, 45-46 Attention-deficithyperactivity disorder, sustained attention and, 113 Attention-engineering manipulation, stimulus organization and, 34 Attentional deficits, stimulus organization and, 29, 41 289

290

INDEX

Auditory analysis, stimulus control analysis and, 84 Auditory intervention, stimulus organization and, 41-42 Auditory stimuli, organization, 32 Auditory tests, sustained attention and, I17 Australia, quality of life studies in, 250-251, 276-277 Autism, perceptual deficits and, 24 Automatic processes, perceptual deficits and, I1

Avoidance responding, staff training and. 230 Avoidant attachments, new morbidity and, 183-185, 187

B Backward masking, perceptual deficits and, 5 Bakan test, sustained attention and, 113 Bayley Scales of Infant Development, new morbidity and, 169, 181, 188 Bayley test, Down syndrome and, 200, 202 Behavior Down syndrome and, 214-218, 221) modifiability of life path and, 139-141, 143, 146, 152 new morbidity and, 159, 181, 185 environmental factors, 169, 175, 177-179 quality of life and, 254, 276 staff training and, 231-240 stimulus control analysis and, 56, 58-59 definition, 63 functions. 75-76, 80. 82 theoretical issues, 82-84 stimulus organization and, 29-31 Behavior replacement strategy, staff training and, 230 Bimodal relational information, stimulus organization and, 40-42 Binocular fusion, perceptual deficits and, 2 Binocular vision, perceptual deficits and, 2-5, 8, 16 Biological factors, new morbidity and, 169, 171 Blissymbols, stimulus control analysis and, 57 Blocking, stimulus control analysis and, 99 Brain, perceptual deficits and, 1 1 Brain damage perceptual deficits and, 13 sustained attention and, I15

Bureaucracy, quality of life and, 280-281 Burnout, staff training and, 229-230

,.

L

Canada, quality of life studies in, 270-271 Carey Scale, new morbidity and, 181 Category learning, stimulus organization and, 48 Cattell Infant Scale, modifiability of life path and, 144 Causality, new morbidity and, 189-190 Chain effects, modifiability of life path and, 154 Child temperament, new morbidity and, 168, 177- 178 Chronological age, see also Mean chronological age stimulus organization and, 33, 42, 44 sustained attention and, 115 Cognitive development, new morbidity and, I89 attachment, 182-188 environmental factors, 168 high risks, 162-163, 165 Cognitive processes Down syndrome and, 203-210, 219 new morbidity and, 160, 190 environmental factors, 167, 169-175 high risks, 161, 164, 166-167 perceptual deficits and, 1 1 , 13, 24 quality of life and, 25 1, 254, 268 stimulus organization and, 29-32, 39, 46 sustained attention and, I13 future research, 128, 130 vigilance, 116-1 17, 119, 121-122 Cognitive readiness, new morbidity and, 171175, 181-182, 184 Colorado Adoption Project, new morbidity and, 170 Community integration, quality of life and, 250, 276, 280 models, 255-260, 263 Community services, quality of life and, 274275, 277, 279-280 Comparison stimuli, stimulus control analysis and, 61, 80 Computer-assisted instruction, stimulus control analysis and, 100-101 Computer-based stimulation, perceptual deficits and, 10

INDEX

29 1

Conceptual skills, stimulus organization and, 34, 37 Conditional discrimination, stimulus control analysis and, 74-75, 79-81 Conditional functions, stimulus control analysis and, 74 Conditional stimuli, stimulus control analysis and, 61 Conduct disorders, modifiability of life path and, 146-147 Congenital defects, modifiability of life path and, 144 Connectionism, stimulus organization and, 30, 48-49 Consistency, Down syndrome and, 217 Constructed-response matching-to-sample (CRMTS), stimulus control analysis and, 95, 97-99 Context-specific rules, stimulus organization and, 35 Contextual stimuli, stimulus control analysis and, 62 Contiguity, stimulus organization and, 33 Contingency new morbidity and, 186, 188 staff training and, 232-233, 235, 238 stimulus control analysis and analytical framework, 58-59, 61-62 research, 64, 73-75, 94 Continuous Performance Test (CPT), sustained attention and, 113, 115-116, 127-128, 130 Cooperativeness, Down syndrome and, 216 Coping, Down syndrome and, 213 Correlation, perceptual deficits and, 19-21 Cortical mechanism, perceptual deficits and, 5 Cortical neurons, perceptual deficits and, 10, 12 Cues new morbidity and, 162 perceptual deficits and, 2, 4, 8, 1 I stimulus control analysis and, 89, 98 stimulus organization and, 38, 46-47 sustained attention and, 119

D Deafness, stimulus control analysis and, 83, 97 Deinstitutionalization. stimulus control analysis and, 57

Delay interval, sustained attention and, 123124, 129 Delayed construction method, stimulus control analysis and, 99 Delayed-recognition tasks, sustained attention and, 124 Democratism, quality of life and, 280 Demographics modifiability of life path and, 145 new morbidity and, 163, 165, 168, 172175 Density, stimulus organization and, 45 Density reduction, perceptual deficits and, 7, 15, 20, 22 Density values, perceptual deficits and, 17, 19, 21 Depression, new morbidity and, 175-177, 179, 190 Deprivation modifiability of life path and, 143, 149I50

new morbidity and, 170 Depth constancy, perceptual deficits and, 16 Depth perception, deficits, 2-3, 8, 16 Detection, sustained attention and future research, 127-130 methodology, 112-113, 115 vigilance, 116, 118, 120-121, 123, 127 Detection latency stimulus organization and, 46 sustained attention and, 129 Detection processes, stimulus organization and, 33, 47 Detection threshold, perceptual deficits and, 17, 20, 23 Developmental delays, new morbidity and, see New morbidity, adolescent parenting and Developmental disabilities, quality of life and, see Quality of life Developmental lag, sustained attention and, 118, 127 Deviance, quality of life and, 265-266 Diagnosis, perceptual deficits and, 15 Didactive teaching, staff training and, 234 Digit recall, stimulus organization and, 33 Disability, meaning of, quality of life and, 264-266 Disadvantage, escape from, modifiability of life path and, 149-152 Discrepancy, stimulus organization and, 41

292

INDEX

Discrimination perceptual deficits and, 1, 11-12, I 4 research progress. 15- 17, 20-2 I stimulus control analysis and, 56-58 analytical framework, 58-59, 61--62 functions, 73-75, 79-81 identity matching, 95, 97-98 research, 63-64 techniques for developing control, 87-91, 94 theoretical issues, 85-86 topographies, 64-67, 69-72 stimulus organization and, 30, 33, 43, 4748 sustained attention and, 118, 120-122, 127 Discrimination reversals, stimulus control analysis and, 89-91 Discrimination threshold, perceptual deficits and, 17, 23 Disorganized attachments, new morbidity and, 183-184, 187 Disoriented attachments, new morbidity and, 183- I84 D i spdrit y detection of, deficits and, 4-5, 12 stimulus organization and, 42-43 Disposition, modifiability of life path and, 153 Distractibility, sustained attention and, 128 Distractor stimuli, organization, 46 Dopamine, perceptual deficits and, 24 Down syndrome, 197-198, 218-220 intervention studies, 199-201 new morbidity and, 182 nonintervention research, 198- 199 Surrey study, 201-203 behavior, 214-218 cognitive aspects, 203-210 self-help, 210-214 Drug use, new morbidity and, 160 Dyslexia, perceptual deficits and, 13

E Early intervention, new morbidity and, 165 Ecology, quality of life and, 253-254, 267 Ecosystem, change of, staff training and, 238240 Education Down syndrome and, 205-206, 209, 219

modifiability of life path and, 139, 141142, 151-153 new morbidity and, 159 environmental factors, 168, 172-175 high risks, 161, 164-165 quality of life and, 260. 270, 275 Element density, perceptual deficits and, 10, 16 Emotion new morbidity and, 160, 189-190 cognitive development, 181, 188 environmental factors, 167, 170, 172, 176, 179 high risks, 162-164, 166-167 quality of life and, 251 Emotional deprivation, modifiability of life path and, 148, 154 Emotional disorders, modifiability of life path and, 146-147 Employment status, quality of life and, 272 Encoding stimulus organization and, 43-45 sustained attention and, 114-1 16, 120-121, 127 Environment modifiability of life path and, 140-141, 150, 152-153 new morbidity and, 165, 170-171, 182, 188, 190 cognitive development, 185- 186 quality of life and conceptualization, 248, 253-255 empirical studies, 273-274 models, 258, 261-262, 264, 269-270 staff training and, 231, 237-238 stimulus control analysis and, 59, 65 Equivalence classes, stimulus control analysis and, 83-86 Event rate, sustained attention and, 119, 121122, 129 Excitation, perceptual deficits and, 13, 23 Exclusion, stimulus control analysis and, 80. 94, 99 Extraction, stimulus control analysis and, 72

F Fading, perceptual deficits and, 16 False alarms, sustained attention and, 118, 122-124, 127-128

293

INDEX

Feature positive effect, stimulus organization and, 32 Feedback modifiability of life path and, 141 perceptual deficits and, 12, 14, 19 staff training and, 232-233, 235-238, 24024 I sustained attention and. 1 14, I16 Feedfonvard, perceptual deficits and, 12 Filling-in process, perceptual deficits and, 7, 11, 16 Finland, quality of life studies in, 272-273 Forced-choice methods, perceptual deficits and, 14, 16-17, 19 Frame number, perceptual deficits and, 8, 10. 17, 19, 23 Free play, new morbidity and, 185-186 Front-end processing, stimulus organization and, 30 Function transfers, stimulus control analysis and, 86-87 Functional behavior, quality of life and, 268269, 277 Functional play, new morbidity and, 187 Functional verbal responses, staff training and, 239 Functioning level, stimulus control analysis and, 72, 81, 89

G General Health Questionnaire, quality of life and, 272 Generalization, stimulus organization and, 34, 37-38, 41 Generalization of behavior, staff training and, 235 Genetic factors modifiability of life path and, 141 new morbidity and, 165, 170- I7 I . I89 Gestalt psychology, stimulus organization and, 32

H Habilitation quality of life and, 260 staff training and, 239

Habituation new morbidity and, 186, 189 sustained attention and, 127 Handicap, defined, quality of life and, 273 Health, Behavior and Skills, Down syndrome and, 203 Heart disease, Down syndrome and, 213

I Identity quality of life and, 277 social, quality of life and, 264-267 Identity changes, new morbidity and, 176 Identity matching, stimulus control analysis and functions, 76-77, 79, 89-92 printed words, 95-98 Identity relations, stimulus control analysis and, 76 Independence Down syndrome and, 210-214 quality of life and, 250, 259-260, 275-276 staff training and, 238 Index of Self-Sufficiency, Down syndrome and, 213 Individual habilitation plans (IHR), staff training and, 239 Infant characteristics, new morbidity and, 168-170 Information processing new morbidity and, 183, 186-188 stimulus organization and, 30, 45-46 sustained attention and, 122, 128, 130-131 Information theory, stimulus organization and, 32 Inhibitors, perceptual deficits and, 13 Instability, new morbidity and, 160- 161, 165, 176 Institutionalization modifiability of life path and, 137, 149 quality of life and, 256, 276 staff training and, 225-226, 235 sustained attention and, 115, 130 Instruction staff training and, 241 sustained attention and, 114 Instructional functions, stimulus control analysis and, 74, 80

294

INDEX

Instructional stimuli, stimulus control analysis and, 91 Integration processes, perceptual deficits and, 1 I Intellectual disability definition, 56 services, staff training for, see Staff training and management stimulus control analysis and, see Stimulus control analysis Intelligence modifiability of life path and, 150 new morbidity and, 181, 189 environmental factors, 168, 172 high risks, 163-165 maternal intelligence, 170- 171 perceptual deficits and, 18-20 stimulus organization and, 30, 32, 34, 43, 45, 48 sustained attention and, I 11, 114 future research, 128, 130 vigilance, 115, 117, 118, 121-122, 124 Intensity fading, stimulus control analysis and, 89 Interpolate< quotients, Down syndrome and, I99 Intervention Down syndrome and, 197-201, 207-208 staff training and, 23 I, 240 stimulus organization and, 35-36 Inventory for Client and Agency Planning (ICAP), quality of life and, 275-276 IQ, see also Mean 1Q Down syndrome and, 198, 215, 218-220 cognitive aspects, 203-209 self-help, 210-214 modifiability of life path and, 138, 143, 145 new morbidity and, 181, 188 environmental factors, 167-168, 170171, 173-175 high risks, 163, 165 perceptual deficits and, 15 staff training and, 226 stimulus control analysis and, 83 sustained attention and, 116-120 Isolation, modifiability of life path and, 142143

K Kinematograms random element, perceptual deficits and, 5-8

ecological significance of stimuli, 8-9 methodology, 14 research, 10-12, 16 stimulus organization and, 45

L Labeling, quality of life and, 265-267 Language Down syndrome and, 208-209 modifiability of life path and, 143 new morbidity and, 163,469, 171, 188 stimulus control analysis and, 83 Language deficits, stimulus control analysis and, 55 Lateral geniculate nucleus, stimulus organization and, 47 Learning, relational, stimulus organization and, see Stimulus organization Learning by exclusion, stimulus control analysis and, 94 Learning disabilities new morbidity and, 162 sustained attention and, 113 L i t e r International Performance Scale, Down syndrome and, 202, 208, 218 Letter recognition, sustained attention and, 119-120 Lexigrams, stimulus control analysis and, 57 Life path, modifiability of, 137-142. 152-154 adjustment difficulties, 146- 149 escape from disadvantage, 149- 152 recovery from perinatal problems, 144-146 removal from isolation, 142- 144 Life-style, quality of life and, 255, 268-269, 276 Luminance, perceptual deficits and, 5, 14, 21

M Magnocellular layers, stimulus organization and, 47 Maintenance stimulus control analysis and, 56 stimulus organization and, 37-38 Maintenance of behavior, staff training and, 235-240 Malaise scale, Down syndrome and, 206

295

INDEX

Management Down syndrome and, 214-218 staff, see Staff training and management Matching stimuli, organization, 39 Matching-to-sample stimulus control analysis and functions, 76-77, 80-81 identity matching, 95, 97-98 techniques for developing control, 88-94 theoretical issues, 82-87 topographies, 68-69, 73 word:picture relations, 99 stimulus organization and, 32-33, 36, 38, 42, 44 Maternal characteristics, new morbidity and, I68 MBA scale, new morbidity and, 173-174 Mean chronological age, sustained attention and, 115-118, 121-122, 124 Mean IQ, sustained attention and, 115-1 18, 121-122, 124 Mediational processes, stimulus organization and, 30-31, 34, 38.40 Memory perceptual deficits and, 14 stimulus control analysis and, 70 stimulus organization and, 32-33, 35, 42 sustained attention and, 114, 121-123. 126129 Memory load, sustained attention and, 112, 114-115 vigilance, 119-121, 124, 129 Memory trace, sustained attention and, 129 Mental age Down syndrome and, 199, 213, 220 stimulus control analysis and, 82, 95, 97 stimulus organization and, 33-35, 39 Menill-Palmer Scale of Mental Tests, Down syndrome and, 202, 218 Methodology, perceptual deficits and, 13- 15 Middle temporal area, perceptual deficits and, 12 MIS scale, new morbidity and, 173-175, 181 Modeling, staff training and, 234-235, 241 Modifiability of life path, see Life path, modifiability of Mortality, Down syndrome and, 209 Motivation Down syndrome and, 208 modifiability of life path and, 151 new morbidity and, 159, 186

perceptual deficits and, 17 sustained attention and, 126, 130-131 Motor skills, new morbidity and, 169

N Naming, stimulus control analysis and, 61 Naming tests, stimulus control analysis and, 84, 86 National Child Development Study, modifiability of life path and, 150-151 Neocortex, perceptual deficits and, 12 Neurological disease, stimulus control analysis and, 68 Neurons, perceptual deficits and, 12-13, 24 Neurophysiology, stimulus organization and, 30, 47-48 New morbidity, adolescent parenting and, 159-160 cognitive development, attachment and, 182- 183 attachment defined, 183-184 attention, 187- 188 implications, 188 information processing, 186 parenting, 184-185 social consequences, 185- I86 cognitive readiness, 181-182 environmental factors, 167- 168 child temperament, 177- 178 cognitive readiness, 171-175 maternal intellignece. 170- 171 Notre Dame Parenting Project, 168-170 personal adjustment, 175- 177 social support, 178-181 high risks, 160-161 childrearing, 161-164 children, 164-165 model, 166-167 unraveling, 188- 189 complex causality, 189- 190 defining mental retardation, I90 New morbidity, modifiability of life path and, 153 New Zealand, quality of life studies in, 272 Newcastle Study, modifiability of life path and, 149 Nonintervention research, Down syndrome and, 198-199

296

INDEX

Nonverbal instructional methods, stimulus control analysis and, see Stimulus control analysis Normalization, quality of life and, 274 Notre Dame Parenting Project, new morbidity and, 163 cognitive development, 187 environmental factors, 167- 17 I , 176 Novel behavior, stimulus control analysis and, 59 Novelty stimulus control analysis and, 80, 94 stimulus organization and, 34

0

Oddity tasks stimulus control analysis and, 69 stimulus organization and, 32-41. 44 One-trial discrimination learning, stimulus control analysis and, 88-89 Ontogeny, perceptual deficits and, 11 Ordinal relations, stimulus control analysis and, 81-82 Organizational behavior management, staff training and, 229, 235 Organizational procedures, staff training and, 236

P Paired associate learning, stimulus organization and, 33 Parenting skills, new morbidity and, see New morbidity, adolescent parenting and Parenting Stress Index, new morbidity and, 174, 181 Parvocellular layers, stimulus organization and, 47 Pattern recognition, perceptual deficits and, 9 Peabody Picture Vocabulary Test Down syndrome and, 218 new morbidity and, 169, 171, 188 perceptual deficits and, 18-19 stimulus control analysis and, 91 Peer group modifiability of life path and, 152-153 new morbidity and, 180, 190 quality of life and, 274

Perception stimulus organization and, 30-31 early stages of processing, 45-46 global stimulus, 34, 40-41 stimulus specificity, 43 sustained attention and, 120-121, 123, 128I29 Perceptual deficits, I , 23-24 methodology, 13-15 random element kinematograms, 5-6 correlation, 7 density reduction, 7 frame number, 8 random element stereograms, 2-3 depth, 3-4 disparity, 4-5 locus of binocular interaction, 5 random element stimuli ecological significance, 8-9 models, 13 physiological data, I I - 12 psychophysical data, 10- I 1 research, 9- 10 research progress, 15-22 Perceptual salience, stimulus organization and, 35, 37 Perinatal Longitudinal Project, new morbidity and, 164 Perinatal problems, modifiability of life path and, 142, 144-146 Perry Preschool Program, modifiability of life path and, 153 Personal adjustment, new morbidity and, 168, 175-177, 181, 184 Personal care, staff training and, 231 Personal factors, new morbidity and, 165, 182, 188 environmental factors, 167- 168. 174- 179 Personality Down syndrome and, 214, 220 new morbidity and, 178 Physiology, perceptual deficits and, 10- 13 Pop-out effect, stimulus organization and, 46 Position alternation, stimulus organization and, 40 Position preservation, stimulus organization and, 40 Post-School Transition Study Survey Interview, quality of life and, 275-276 Postnatal assessment, new morbidity and, 173, 189

INDEX

297

Poverty, new morbidity and, 159-161, 165 Preattentive processes, stimulus organization and, 45-46 Predisposed salience, stimulus organization and, 38-40, 46 Prematurity modifiability of life path and, 145-146 new morbidity and, 164 Prenatal assessment, new morbidity and, 173, 175, 189 Prenatal problems, modifiability of life path and, 142. 144 Proactive inhibition, sustained attention and, 123, 126 Probability-matching strategies, sustained attention and, 129 Productivity, quality of life and, 250, 260 Professionalism, quality of life and, 280 Prompts, staff training and, 235 Psychoanatomy, perceptual deficits and, 5 Psychological factors, new morbidity and, 167, I76 Psychological indicators, quality of life and, 25 1-253 hychophysics, perceptual deficits and, 10- 1 I , 14, 23 Punishment Down syndrome and, 216-217, 220 new morbidity and. 162-163, 180-181 staff training and, 229, 241 Pyramid models, staff training and, 234-235

Q Quality of life, 247-248, 279-281 conceptualization, 248-250 person-environment fit, 253-255 psychological indicators, 25 1-253 social indicators, 250 empirical studies, 270 Australia, 276-277 Canada, 270-271 Finland, 272-273 New Zealand, 272 Sweden, 273-275 United Kingdom, 276 United States, 275-276 methodology, 278-279 models, 255-256, 261-262 community adjustment, 257-259

measurement, 259-261 process model, 263-264 symbolic interactionism, 264-270 staff training and, 227 Quality-of-life adjusted years, 249 Quality of Life Questionnaire, 254, 261, 272, 279

R Race new morbidity and, 163. 173-175, 180 quality of life and, 251 Random E acuity test, perceptual deficits and, 15

Random element kinematograms, see Kinematograms, random element Random element stereograms, perceptual deficits and, see Stereograms Random element stimuli, perceptual deficits and, 8-13 Raven’s Coloured Progressive Matrices, Down syndrome and, 208, 218 Reaction time, sustained attention and, 130 Redundancy, stimulus organization and, 33 Reflexivity, stimulus control analysis and, 77, 79 Reinforcement modifiability of life path and, 141 new morbidity and, 177 staff training and, 238, 241 stimulus control analysis and, 56-57, 59 techniques for developing control, 87, 89, 94, 98 topographies, 64, 70 stimulus organization and, 39-40 Relational frame theory, stimulus control analysis and, 83 Relational learning, stimulus organization and, see Stimulus organization, relational learning and Repetition, stimulus control analysis and, 92 Residential environment, quality of life and, 257, 261, 276 Residential services, staff training and, 225226, 230, 232, 238, 241 Resilience modifiability of life path and, 146 new morbidity and, 175

298

INDEX

Response bias perceptual deficits and, 13-14 sustained attention and, 126 Response strategy, sustained attention and, 114, 129 Retention, stimulus control analysis and, 56, 92 Retina perceptual deficits and, 5 , 12 stimulus organization and, 48 Retinal disparity, perceptual deficits and, 2-3 Retrieval, stimulus organization and, 33 Reversal of letters, stimulus control analysis and, 95 Reynell Developmental Language Scale, Down syndrome and, 202 Risk factors modifiability of life path and, 141, 146, 148, 153 new morbidity and, 160-166, 189 cognitive development, 182, 185 environmental factors, 167, 169, 172, I79 Role-playing, staff training and, 234-235, 238, 241 Rudimentary reading, stimulus control analysis and, 57-58, 62, 87, 99-101 Rule-based skills, stimulus organization and, 31. 38

S Salience perceptual deficits and, 16 stimulus organization and, 44 Salience, perceptual, stimulus organization and, 35, 37 Salience, predisposed, stimulus organization and, 38-40, 46 Sample stimulus, stimulus control analysis and, 80 Sample stimulus control shaping, stimulus control analysis and, 91-94, 98 Satisfaction, quality of life and, 256, 277 conceptualization, 25 1, 254 models, 259-260, 262, 269 Scanning new morbidity and, 189 sustained attention and, 129

Search tasks, stimulus organization and, 46 Secure attachments, new morbidity and, 183I87 Selection-based communication, stimulus control analysis and, 57-58 Selective attention, stimulus organization and, 38 Self, quality of life and, 265, 268-269 Self-care skills quality of life and, 277 staff training and, 232, 238 Self-esteem new morbidity and, 176 quality of life and, 269, 272 Self-help Down syndrome and, 210-214, 219-220 staff training and, 236 Self-management, staff training and, 236-238, 24 1 Self-regulatory skills, new morbidity and, 159, 183, 186, 188-189 Self-righting, modifiability of life path and, 138, 144, 146 Semantics, stimulus control analysis and, 99 Semi-independent living programs, quality of life and, 257 Separation, new morbidity and, 187 Sequence classes, stimulus control analysis and, 81-82, 100 Serial order, stimulus control analysis and, 8 I Sex Down syndrome and, 209, 213, 215, 218219 staff training and, 226 Shaping, stimulus control analysis and, 88-95, 97-99, 101 Short-term memory, sustained attention and, 114, 120, 122 Signal detection perceptual deficits and, 14 theory, sustained attention and, 113-1 14, I18 Signal probability, sustained attention and, I 14 Situated cognition, stimulus organization and, 32 Size disparity, stimulus control analysis and, 81

Social class Down syndrome and, 205-207, 209, 212 modifiability of life path and, 142

INDEX

Social identity, quality of life and, 264-267 Social indicators, quality of life and, 250 Social networks, new morbidity and, 180-181 Social support new morbidity and, 181, 184 environmental factors, 167-168, 172, 178- I81 high risks, 164, 166-167 quality of life and, 257, 274 Social validity, staff training and, 240-242 Socialization new morbidity and, 169 staff training and, 230-231 Societal influences, quality of life and, 268270, 277 Socioeconomic status modifiability of life path and, 145-146 new morbidity and cognitive development, 186 environmental factors, 168- 169, 173175, 180 high risks, 163-165, 167 perceptual deficits and, 18, 20 quality of life and, 256 Spatial contiguity, stimulus control analysis and, 81 Spatial distance, stimulus organization and, 48 Spatial factors, perceptual deficits and, 23 Spatial matching, perceptual deficits and, 6 Speech perception, stimulus organization and, 48 Staff training and management, 225-227, 240-242 model, 227-229 absenteeism, 229 burnout, 229-230 client training, 23 1-232 maintenance of staff behavior, 235-240 outcome, 232-233 procedures, 233-235 staff-client interactions, 230-23 I Stanford-Binet test, new morbidity and, 169, 188 S tereograms perceptual deficits and, 2-5 ecological significance of stimuli, 8-9 methodology, 14 random element kinematograms, 5-8 research, 10-12 research progress, 15-16 stimulus organization and, 45

299 Stereopsis, perceptual deficits and, 2-5, 11, 17 Stimulation devices, staff training and, 239 Stimuli new morbidity and, 186 sustained attention and methodology, 113-1 14 vigilance, 117, 119-122, 128 Stimulus alternation, organization, 40 Stimulus conspicuity, sustained attention and, I29 Stimulus contiguity, organization, 36-38 Stimulus control analysis, 55, 101 analytical framework, 58-59 four-term relations, 61 higher-term relations, 62 three-tern relations, 59-61 two-tern relations, 59 definition, 56 research, 56, 62, 82 identity matching, 89-91, 95-98 matching-to-sample, 91-94 one-trial discrimination learning, 88-89 rudimentary reading, 99-100 stimulus control topographies, 64-73 stimulus definition, 63-64 stimulus functions, 73-82 technical support, 100-101 techniques for developing control, 87-88 theoretical issues, 82-87 word:picture relations, 98-99 teaching technology, 56-58 topographies, 64-73 Stimulus definition, stimulus control analysis and, 63-64, 73 Stimulus discriminability, sustained attention and, 126-127 Stimulus encoding, sustained attention and, 114, 121 Stimulus equivalence, stimulus control analysis and functions, 76-80, 82 theoretical issues, 82-87 Stimulus generalization, stimulus control analysis and, 66, 76 Stimulus hierarchy, stimulus control analysis and, 14 Stimulus organization, relational learning and, 29-33.49 connectionism. 48-49

300

INDEX

early stages of processing, 44-47 global stimulus, 33-34 bimodal relational information. 40-42 contiguity, 36-38 oddity paradigm, 34-36 salience, 38-40 neurophysiology, 47-48 perception, 30-3 I stimulus specificity, 42-44 Stimulus preservation, organization, 39-40 Stimulus repetition, organization, 40 Stranger anxiety, new morbidity and, 183 Stranger Situation test, new morbidity and, 183-184, 187 Stress Down syndrome and, 216 modifiability of life path and, 141. 146 new morbidity and cognitive development, 183, 187 cognitive readiness, 181 environmental factors, 174- 175, 177179 high risks, 161-164, 167 quality of life and, 253 Supervision, staff training and, 230, 241 Support systems, quality of life and, 262 Surrey study, Down syndrome and, 201218 Sustained attention, 1 I 1 - I 12 methodology laboratory tests, 1 12-1 13 measurement issues, 113- I15 vigilance, 115-127 future research, 127-131 Sweden, quality of life studies in, 273.275 Symbolic interactionism, quality of life and, 254, 264-270 Symbols stimulus control analysis and, 57 sustained attention and, I13 Symmetry stimulus control analysis and, 77, 79, 92 stimulus organization and, 33-34, 36, 4244

Target detection, sustained attention and methodology, 113-1 14 vigilance, 116-118, 121, 123, 128-129 Target probability, sustained attention and, I 19 Task complexity, sustained attention and, I 12, I19 Task type, sustained attention and, I12 Taxonomy quality of life and, 278 sustained attention and, 114, 120-121 Teaching Family Model, staff training and. 239 Teaching technology, stimulus control analysis and. 56-58 Temperament Down syndrome and, 214, 217 new morbidity and, 168, 177-178 Temporal correlation perceptual deficits and, 16, 21, 23 stimulus organization and, 45 Texture segregation, stimulus organization and, 46 Thalamus, perceptual deficits and, 12 Total independence, Down syndrome and, 210, 212 Transfer, stimulus organization and, 34-38 Transfer of stimulus functions, stimulus control analysis and, 85-87 Transition, quality of life and, 257, 275-276 Transitivity, stimulus control analysis and, 77, 79-80 Transmitted Deprivation Program, modifiability of life path and, 149, 151 Transposition, stimulus control analysis and, 81-82 Transposition judgments, stimulus organization and, 34 Trial initiation, stimulus control analysis and, 70-7 I Troublesome children, modifiability of life path and, 153 Twins, modifiability of life path and, 142143, 145 S r a n n y of quality, quality of life and, 247, 256, 264

U T Target density, sustained attention and, 124, 126, 129

United Kingdom, quality of life studies in, 276 United States, quality of life studies in, 250, 275-276

30 1

INDEX

V Values, quality of life and, 254-255, 268 Vehicle dimension, stimulus organization and, 38-40, 46 Verbal skills, stimulus control analysis and, 55 Vigilance, sustained attention and, 1 11-127 future research, 127-131 Vineland Adaptive Behavior Scales new morbidity and, 169 perceptual deficits and, 15 Vineland Social Maturity Scale, Down syndrome and, 213-214 Vision, perceptual deficits and ecological significance of stimuli, 9 methodology, 14-15 random element stereograms, 2-5 research, 12-13, 15-16 VISTECH, perceptual deficits and, 15 Visual detection, sustained attention and, 117, I19

Visual processes, stimulus organization and, 47-48 Visual stimuli, organization, 29, 31-33, 4042, 46 Vulnerability, modifiability of life path and, I50

W Wedge model, modifiability of life path and, 140 Well-being, quality of life and, 252, 254, 269, 272 WISC-R, modifiability of life path and, 143 Wisconsin General Test Apparatus, stimulus control analysis and, 70 Working memory, sustained attention and, 128 World Health Organization, quality of life and, 253, 273

This Page Intentionally Left Blank

Contents of Previous Volumes

Volume 1

Learning and Memory of Mental Retardates HERMAN H. SPIT2

A Functional Analysis of Retarded Development SIDNEY W. BIJOU

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

Classical Conditioning and Discrimination Learning Research with the Mentally Retarded LEONARD E. ROSS

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

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

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

Research on Personality Structure in the Retardate EDWARD ZIGLER

Programming Perception and Learning for Retarded Children MURRAY SIDMAN AND LAWRENCE T. STODDARD

Experience and the Development of Adaptive Behavior H. CARL HAYWOOD AND JACK T. TAPP

Programming Instruction Techniques for the Mentally Retarded FRANCES M. GREENE

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

Some Aspects of the Research on Mental Retardation in Norway IVAR ARNWOT BJORGEN

The Behavior of Moderately and Severely Retarded Persons JOSEPH E. SPRADLIN AND FREDERIC L. GIRARDEAU

Research on Mental Deficiency During the Last Decade in France R. LAFON AND J. CHABANIER Psychotherapeutic Procedures with the Retarded MANNY STERNLIGHT

Author Index-Subject Index

Volume 2

Author Index-Subject Index

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

Volume 3 Incentive Motivation in the Mental Retardate PAUL S. SIEGEL

The Role of Input Organization in the 303

304

CONTENTS OF PREVIOUS VOLUMES

Development of Lateral and Choice-Sequence Preferences IRMA R. GERJUOY AND JOHN J. WINTERS, JR.

Mental Retardation in Animals GILBERT W. MEIER

Studies in the Experimental Development of Left-Right Concepts in Retarded Children Using Fading Techniques SIDNEY W. BIJOU

Author Index-Subject Index

Verbal Learning and Memory Research with Retardates: An Attempt to Assess Developmental Trends L. R. GOULET

Volume 5

Research and Theory in Short-Term Memory KEITH G. SCO'IT AND MARCIA STRONG SCO'IT Reaction Time and Mental Retardation ALFRED A. BAUMEISTER AND GEORGE KELLAS 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. 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

Audiologic Aspects of Mental Retardation LYLE L. LLOYD

Medical-Behavioral Research in Retardation JOHN M. BELMONT Recognition Memory: A Research Strategy and a Summary of Initial Findings KEITH G. S C O I T 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 FREDERlC L. GIRARDEAU German Theory and Research on Mental Retardation: Emphasis on Structure LOTHAR R. SCHMIDT AND PAUL B. BALTES Author Index-Subject Index

Volume 6 Cultural Deprivation and Cognitive Competence J. P. DAS Stereotyped Acts ALFRED A. BAUMEISTER AND REX FOREHAND

305

CONTENTS OF PREVIOUS VOLUMES Research on the Vocational Habilitation of the Retarded: The Present, the Future MARC W. GOLD

Retardation with Special Emphasis on Evolution HERMAN H. SPIT2

Consolidating Facts into the Schematized Learning and Memory System of Educable Retardates HERMAN H. SPlTZ

The Role of the Social Agent in Language Acquisition: Implications for Language Intervention GERALD J. MAHONEY AND PAMELA B. SEELY

An Attention-Retention Theory of Retardate Discrimination Learning MARY ANN FISHER AND DAVID ZEAMAN 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 KERI M. WILTON AND FREDERIC J. BOERSMA

Cognitive Theory and Mental Development EARL C. BUTTERFIELD AND DONALD 1. DICKERSON A Decade of Experimental Research in Mental Retardation in India ARUN K. SEN The Conditioning of Skeletal and Autonomic Responses: Normal-Retardate Stimulus Trace Differences SUSAN M. ROSS AND LEONARD E. ROSS Malnutrition and Cognitive Functioning 1. 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

Placement of the Retarded in the Community: Prognosis and Outcome RONALD B. McCARVER AND ELLIS M. CRAIG

The Processing of Information from ShortTerm Visual Store: Developmental and Intellectual Differences LEONARD E. ROSS AND THOMAS B. WARD

Physical and Motor Development of Retarded Persons ROBERT H. BRUININKS

Information Processing in Mentally Retarded Individuals KEITH E. STANOVICH

Subject Index

Mediational Processes in the Retarded: Implications for Teaching Reading CLESSEN J. MARTIN

Volume 8

Rychophysiology in Mental Retardation J. CLAUSEN

Self-injurious Behavior ALFRED A. BAUMEISTER AND JOHN PAUL ROLLINGS Toward a Relative Psychology of Mental

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

CONTENTS OF PREVIOUS VOLUMES 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 Mainstreaming Mentally Retarded Children: A Review of Research LOUISE CORMAN AND JAY GOTTLIEB Savants: Mentally Retarded Individuals with Special Skills A. LEWIS HILL

Individual Differences in Short-Term Memory RONALD L. COHEN Inhibition and Individual Differences in Inhibitory Processes in Retarded Children PETER L. C. EVANS Stereotyped Mannerisms in Mentally Retarded Persons: Animal Models and Theoretical Analyses MARK H. LEWIS AND ALFRED A. BAUMEISTER An Investigation of Automated Methods for Teaching Severely Retarded Individuals LAWRENCE T. STODDARD Social Reinforcement of the Work Behavior of Retarded and Nonretarded Persons LEONIA K. WATERS

Subject Index

Social Competence and Interpersonal Relations between Retarded and Nonretarded Children ANGELA R. TAYLOR

Volume 10

The Functional Analysis of Imitation WILLIAM R. McCULLER AND CHARLES L. SALZBERG

The Visual Scanning and Fixation Behavior of the Retarded LEONARD E. ROSS AND SUSAN hl. 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 Rumination NIRBHAY N. SINGH Subject Index

Volume 11 Cognitive Development of the LeamingDisabled Child JOHN W. HAGEN, CRAIG R. BARCLAY, AND BE’ITINA SCHWETHELM

Index

Volume 12 An Overview of the Social Policy of Deinstitutionalization BARRY WILLER AND JAMES INTAGLIATA Community Attitudes toward Community Placement of Mentally Retarded Persons CYNTHIA OKOLO AND SAMUEL 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 AND RONALD B. McCARVER Issues in Adjustment of Mentally Retarded Individuals to Residential Relocation TAMAR HELLER

307

CONTENTS OF PREVIOUS VOLUMES Salient Dimensions of Home Environment Relevant to Child Development KAZUO NIHIRA, IRIS TAN MINK, AND C. EDWARD MEYERS Current Trends and Changes in Institutions for the Mentally Retarded R. K. EYMAN, S . A. BORTHWICK, AND G. TARJAN Methodological Considerations in Research on Residential Alternatives for Developmentally Disabled Persons LAIRD W. HEAL AND GLENN T. FUJIURA 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 Sustained Attention in the Mentally Retarded: The Vigilance Paradigm JOEL B. WARM AND DANIEL B. BERCH Communication and Cues in the Functional Cognition of the Mentally Retarded JAMES E. TURNURE Metamemory: An Aspect of Metacognition in the Mentally Retarded ELAINE M. JUSTICE Inspection Time and Mild Mental Retardation T. NE'TTELBECK Mild Mental Retardation and Memory Scanning C. J. PHILLIPS AND T.NE'lTELBECK

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 AND LOUISE C. SPEAR Index

Volume 14 Intrinsic Motivation and Behavior Effectiveness in Retarded Persons H. CARL HAYWOOD AND HARVEY N. SWITZKY The Rehearsal Deficit Hypothesis NORMAN W. BRAY AND LISA A. TURNER Molar Variability and the Mentally Retarded STUART A. SMITH AND PAUL S. SIEGEL Computer-Assisted Instruction for the Mentally Retarded FRANCES A. CONNERS, DAVID R. CARUSO, AND DOUGLAS K. DETTERMAN Procedures and Parameters of Errorless Discrimination Training with Developmentally Impaired Individuals GlULIO E. LANClONl AND PAUL M. SMEETS Reading Acquisition and Remediation in the Mentally Retarded NIRBHAY N. SINGH AND JUDY SINGH Families with a Mentally Retarded Child BERNARD FARBER AND UlUlS ROWITZ

Cognitive Determinants of Reading in Mentally Retarded Individuals KEITH E. STANOVICH

Social Competence and Employment of Retarded Persons CHARLES L. SALZBERG, MARILYN LIKINS, E. KATHRYN McCONAUGHY, AND BENJAMIN LIGNUGARIS/KRAR

Comprehension and Mental Retardation LINDA HICKSON BILSKY

Toward a Taxonomy of Home Environments SHARON LANDESMAN

308 Behavioral Treatment of the Sexually Deviant Behavior of Mentally Retarded Individuals R. M. FOXX. R. G. BITTLE, D. R. BECHTEL, AND 1. R. LIVESAY Behavioral Approaches to Toilet Training for Retarded Persons S . BETTISON Index

Volume 15 Mental Retardation as a Thinking Disorder: The Rationalist Alternative to Empiricism HERMAN H. SPIT2 Developmental Impact of Nutrition on Pregnancy, Infancy, and Childhood: Public Health Issues in the United States ERNEST0 POLLlTT The Cognitive Approach to Motivation in Retarded Individuals SHULAMITH KREITLER AND HANS KREITLER

CONTENTS OF PREVlOUS VOLUMES Delay: Lead and Cadmium as Prototypes STEPHEN R. SCHROEDER The Role of Methylmercury Toxicity in Mental Retardation GARY J. MYERS AND DAVID 0. MARSH Attentional Resource Allocation and Mental Retardation EDWARD C. MERRILL Individual Differences in Cognitive and Social Problem-Solving Skills as a Function of Intelligence ELIZABETH J. SHORT AND STEVEN W. EVANS Social Intelligence, Social Competence, and Interpersonal Competence JANE L. MATHIAS Conceptual Relationships between Family Research and Mental Retardation ZOLINDA STONEMAN Index

Mental Retardation, Analogical Reasoning, and the Componential Method J. McCONAGHY Application of Self-control Strategies to Facilitate Independence in Vocational and Instructional Settings JAMES E. MARTIN, DONALD L. BURGER, SUSAN ELIAS-BURGER, AND DENNIS E. MITHAUG Family Stress Associated with a Developmentally Handicapped Child PATRICIA M. MINNES Physical Fitness of Mentally Retarded Individuals E. KATHRYN McCONAUGHY AND CHARLES L. SALZBERG Index

Volume 16 Methodological Issues in Specifying Neurotoxic Risk Factors for Developmental

Volume 17 The Structure and Development of Adaptive Behaviors KEITH F. WIDAMAN, SHARON A. BORTHWICK-DUFFY, AND TODD D. LITTLE Perspectives on Early Language from Typical Development and Down Syndrome MICHAEL P. LYNCH AND REBECCA E. EILERS The Development of Verbal Communication in Persons with Moderate to Mild Mental Retardation LEONARD ABBEDUTO Assessment and Evaluation of Exceptional Children in the Soviet Union MICHAEL M. GERBER, VALERY PERELMAN, AND NORMA LOPEZREYNA

309

CONTENTS OF PREVIOUS VOLUMES

Constraints on the Problem Solving of Persons with Mental Retardation RALPH P. FERRETTI AND AL R. CAVALIER

Long-Term Memory and Mental Retardation JAMES E. TURNURE Index

This Page Intentionally Left Blank