Behavior Modification in Child Treatment

An Experimental and Clinical Approach

Robert M. Browning

• Publisher: Routledge
• Year: 2017
• ISBN:
  9781351314541
• Edition: 1
• Topic: Psychology
• Subtopic: History & Theory in Psychology
• Index: Psychology

This book is the first attempt to validate behavior modification techniques in a carefully controlled experimental treatment environment for emotionally disturbed children. Such special settings permit carefully conducted research experiments can be carried out. This is the first book to synthesize scientific and clinical approaches to human behavior, indicating that behavior modification may one day be as much an applied science as engineering or medicine.

This experimental approach introduces scientific rigor to the clinical setting, as evidenced by precise measurement of behavior variables, detailed specification of treatment procedures, and the use of sophisticated experimental designs to provide objective evaluation of the effectiveness of treatment programs. In this approach, series of idiographic (single-subject) case studies are conducted in a precise manner with each patient-subject admitted to the treatment program. The general research methodology is similar to that used in the broad area of operant conditioning, and most work reported in the book was conducted within a learning theory or behavior-modification framework.

Browning and Stover discuss the general problems of developing and controlling a total therapeutic milieu, presenting practical discussions of problems of data collection, decisions about treatment programs to be used, staffing problems, and documental opinion on the relative values of various treatment techniques. Throughout attention is devoted to developing a method for answering common questions of parent, child-care worker, and professional. The authors conducted symposia on the material contained in this book at various national and regional meetings and have lectured extensively on college campuses. It is a ground-breaking study.

CHAPTER ONE

The Experimental-Clinical Method

The scientific method must be the basis for treatment of psychologically disturbed persons. The clinician should identify and measure behavioral problems in the same manner that the physical scientist approaches any natural phenomena. As an applied scientist, the clinician should combine basic psychological knowledge with current observations of his patient to form testable treatment hypotheses. Factual understanding of the patient’s behavior derived from experimental tests of these hypotheses should then be the basis for the design and execution of a total treatment program. Finally, the clinician should be capable of objectively evaluating his treatment procedures and, consequently, his understanding of the patient. This technological ideal is the experimental-clinical method.

Experimental methods are seldom employed in current clinical practice. Most clinicians have difficulty accepting science as even one of the boundaries of practice, and would vehemently reject the proposition that scientific methodology should stand as a central pillar of treatment. Instead, they state that the complexities of human feeling and action cannot be measured and that the treatment of unique individual behavior is the proper province of art.

Traditionally, there has been a marked difference between the scientific and the clinical approaches to the problem of psychological disturbance. Arising from philosophical questions about the nature of man, and motivated by a search for knowledge, the scientific orientation has become formalized in the pure research of academic psychology. In contrast, the clinical approach has its roots in the humanistic tradition and is motivated by the desire to alleviate human suffering by treating disturbances of individual functioning.

As clinical psychologists, the authors have been acutely aware of the continuing antithesis of scientific and clinical approaches to human behavior. Our training included a heavy emphasis upon academic psychology, experimental method, statistical design, and the other accoutrements of pure behavioral research. However, we were also deeply immersed in clinical problems and traditional treatment methods. In our training, as in our current view of the field, we could seldom find useful relationships between the experimental approach of the university and the practice of the clinic. As clinicians, we have been deeply concerned about the complexities of individual human behavior and we have shared with our clinical colleagues the conviction that the uniqueness of an individual personality can never be comprehended within the rigid rules of the currently revered behavioral science methodologies. As scientists, however, we have valued a degree of preciseness, experimental control, and predictiveness that the vagaries and ambiguities of the usual clinical methods are unable to provide. The great masses of scientific data contained in archival journals relate only peripherally to the human problems that arise daily in a clinic. There is a plethora of studies on basic learning processes, unconscious motivation, perception, and the physiological correlates of behavior, but the gap between these experimental studies and the patients in our offices is great. Integrity prohibits their direct application with anything less than intense skepticism. We regret the absence of a meaningful technological literature and are unwilling to accept in its place the lore of traditional clinical practice which, be it analytic, phenomenological, or behavioral, is replete with subjective evaluation, dogmatic belief, and mystical thinking. Thus, we have found ourselves in conflict between the apparently antithetical values of behavioral science and clinical practice.

Despite the apparent contrasts, the motives of clinical and experimental approaches may be complementary. Emphasis upon the disparities has hindered development of meaningful research on clinically critical aspects of behavior, and treatment methods have scarcely improved since the historically prominent psychiatrists first attacked the problems. A technical integration of craft and art, such as exists in engineering or medicine, could enhance the goals of both scientist and practitioner. Accomplishment of an integrated technology would mean the rejection of contrived manipulations, complex experimental designs, esoteric statistics, and overly sophisticated measurements. Equivalent cynicism must be directed at treatment methods that rely upon intuition and hunches while failing to accommodate objective results. From clinicians though, we accept the positive value of working to assist suffering persons find more satisfactory behavioral adjustments; from scientists we accept the value of efforts precisely to predict and control the behavior of clinic patients. The scientific and clinical traditions have much to offer humanity; but their current separateness can lead neither to efficient treatment of disturbed individuals nor to complete understanding of human behavior.

We hope that an eventual integration of scientific method and therapeutic craft will permit knowledgeable treatment. This book reports our initial attempts to work within the experimental-clinical method. It is our first approximation to the ideal of making clinical practice an applied science.

Science Versus Practice

Before proceeding with a detailed description of experimental-clinical methods, we shall more specifically examine the differences between science and practice. The clinician and the experimenter differ in their use of basic theory, the independent variables each espouses, their definitions of dependent variables, their attitudes toward prediction and control of behavior, their conception of the variabilities of behavior, and their goals concerning the generality of knowledge about human behavior.

Clinicians have preferred comprehensive theories of total personality; experimentalists have insisted upon precisely formulated conceptions. The broadly encompassing and thoroughly literary clinical theories generally fail to meet basic scientific criteria. Their concepts are difficult to define in terms of the operations that could be used to measure them. The interrelationships of concepts in clinical theories are usually difficult to specify. Even when operational criteria are applied to measure the strength of a cathexis, the quality of identity formation, or the extent of libido drive, the resulting definition fails to capture the essence of the concept as a clinician would employ it in his daily practice. By far the most unacceptable aspect of many clinical theories is that they in no way specify the conditions under which they would by proven inaccurate. This, of course, is a basic criterion of any scientifically acceptable theory. At the other extreme, many of the most scientific theories are but miniature theories relating to highly restricted fields such as immediate memory or psychophysical judgment. Although relevant as building blocks for a pure science of behavior, these limited theories offer little promise to the practicing clinician. The clinician holds no antipathy for precisely constructed theory, nor the experimentalist for comprehensive theory. However, the present state of the science indicates that the achievement of both goals in a common theoretical framework will not occur soon.

The argument between global and molecular theories is irrelevant to clinical practice. If molecular theories are available and seemingly valid for dealing with even limited aspects of a given patient’s behavior, they should be applied and used to give treatment direction. Where precise notions are not available, the clinician can do no harm by falling back upon more global notions of personality so long as he recognizes their limitations, directs his work toward the development of more scientifically acceptable approaches, and, most important, refuses to adopt status quo attitudes about sloppy theorizing.

In accepting scientifically weak theories, the clinician permits himself to follow a poorly blazed trail often meandering far afield from the proper goal of objective practical treatment procedures. Since clinical theories rely upon concepts that are not easily related to touchable, visible realities, the clinician often accepts diffuse definitions of dependent variables. The patient comes to him with a referral complaint—specifically, a behavior or set of behaviors disturbing to himself or others. Rather than defining observable behaviors, typical clinical theories emphasize feeling states, cognitions, complexes, and other hypothetical constructs.

The clinician seldom contracts with the patient to change the patient’s behavior. He does not promise to eliminate rages or produce a higher rate of smiling. In fact, the advice has often been that it would be dangerous to make a direct attack upon “symptoms” of underlying disorders. The underlying disorders are considered the cause, and the “symptomatic” behavior, the effect; but seldom is the mechanism by which cause and effect are supposed to be related clearly specified.

As clinicians, we maintain that a change in the referral behavior is the appropriate goal of treatment. As scientists, we assert that referral complaints are appropriate dependent variables. To meet scientific criteria, some observable and specifiable behavior must serve as the dependent variable under study, and definable treatment manipulations must be designed to produce change in that variable. Clinicians often begin with this point of view, but shortly find themselves concerned wtih more abstract concepts that are only vaguely related to the original behavior problem. Early therapy notes may discuss the referral behaviors in detail, but subsequent accounts will be filled with discussion of changes in other areas. Seldom is the relationship between the original dependent variable and the apparent new variables of interest made explicit.

The clinician is also guilty of an inability adequately to define his independent variables or treatment manipulations. Once again the problem is related to the clinician’s acceptance of imprecise theory. The favored forms of treatment, such as milieu therapy, family therapy, and psychotherapy often defy operational definition. Thus, no two practitioners may be expected to perform similar treatment procedures. How could one hope to replicate another’s interpretations? How can one describe precisely what is meant by encouraging dependency, providing structure, permitting catharsis, or supporting the ego? Favored treatments are so global in their description that it would be difficult to duplicate them. Further, in general practice treatment methods are so confounded that assessment of their interactive effects would be virtually impossible. A controlled attempt is seldom made to evaluate the effects of a particular kind of interpretation or reward procedure on a patient’s behavior. The impreciseness of description of independent variables in common clinical practice has led to comments like “Psychotherapy is an undefined technique applied to unspecified problems with unpredictable outcomes. For this technique we recommend rigorous training” (Raimy, 1950, p. 93). If treatment is to be a technology, it will be necessary to specify more clearly what the therapist does.

One can be equally critical of the scientist. His productivity to date has given us little that would successfully objectify the clinical task. He busies himself with watching college sophomores pursue rotors with metal sticks which electrically and quite operationally measure performance. His dependent variables are usually irrelevant to the complex human issues faced in the clinic. The clinician seldom cares how well his patient can memorize a list of nonsense syllables, discriminate the weights of tin cans, or salivate at the sight of the word “meat.” The contrived manipulations of most psychological experiments seem to have limited clinical value. Do messages mixed with static coming through earphones to a college sophomore in a communications laboratory have any significant relationship to the communication breakdown that leads to the dissolution of a family? Does being told that your score on an experimental task is subnormal compare with the experience of having a frustrated teacher shout “stupid” at you? Is a hypnotically induced sense of guilt similar to the feeling that arises in a child when his father commits suicide? Although similarities may exist, the distance between such contrived laboratory situations and the impact of the human events that comprise clinical cases is far too great to permit meaningful application. The findings of the laboratory often seem useless to the clinician precisely because they are the result of laboratory manipulations. He demands that research be performed in the context of real situations with real people under real stresses.

The artificiality of some laboratory research is not sufficient reason to explain the clinician’s usual indifference to behavioral science. After all, many competent studies have been conducted under highly naturalistic conditions. Even naturalistic studies fail to provide the clinician with the kind of information he needs in his daily practice. For example, the extensive literature on early deprivation (e.g. Yarrow, 1961) clearly indicates an association between maternal deprivation and later deviances in intellectual and social functioning. However, the literature does not tell the clinician how a given individual will be or has been affected by deprivation, nor how to formulate specific treatment programs to alter the behavior of a particular patient. The clinician must ask, not “what are the general effects of deprivation?” but rather, “what are the specific effects of deprivation for this child?” Because the clinician sees his patient as a unique individual with highly idiosyncratic problems, studies providing general conclusions about the average effect of some independent variable seem to offer little promise of meaningful clinical application.

This brings us to an essential difference between clinical and experimental approaches to human behavior. The clinician and the experimentalist seek and require totally different kinds of predictions. Experimental psychology has long been enamored with predictions based upon groups of subjects. To write a paper acceptable to the prominent journals that maintain the standards of the science, a researcher must provide a study consisting (at its simplest level) of a randomly selected pool of subjects divided into two groups, each of which receives specifiably different treatment manipulations. If the null hypothesis (that the population means for the two groups are the same) is rejected, the researcher generally concludes that the treatment manipulations have important significance with respect to the dependent variable. The procedure, of course, overlooks the fact that “there is really no good reason to expect the null hypothesis to be true in any population” (Bakan, 1966, p. 426). As a review of much of the published literature will quickly reveal, experimenters often find statistically significant differences, but the extent of correlation between their dependent and independent variables is so small that the finding may have virtually no psychological significance. Anyone who has conducted a typical group study is aware that examination of individual data often results in the conclusion that the treatment manipulation had little or no effect upon many of the subjects. The statistics, even if they could be shown to be logically related to the usual conclusions from such studies, would still leave us with the finding that the treatment effect works for most but not all subjects.

The clinician cannot be satisfied with a finding that works for most individuals. He must choose a treatment technique that will have a maximum effect upon the behavior he wished to change in a particular person. Further, his interest and concern are centered upon the oddities of each patient’s behavior. For example, most of the children described in this volume had normal hearing, so their hearing capacity was of no special interest to us. One of them, however, was nearly deaf, and this was of enormous interest in our clinical investigations of his behavior. By the very nature of their work, clinicians are confronted with behaviors so unusual that they defy any grouping. The experimentalist often criticizes clinical studies because they do not meet the criteria of random sampling. Indeed, random sampling of subjects is virtually impossible when one is studying deviant behavior. Clinically relevant behaviors are unique entities which can seldom be lumped together across groups of individuals. Every child seen in the clinic has a unique learning history which accounts for his abnormal behavior. For this reason the clinician needs information which will permit accurate clinical prediction about the relevant behaviors of each individual child.

Statistical and clinical prediction cannot be compared fairly because their end goals differ. Actuarial prediction is intended to make statements such as, “most persons of this type will respond in X manner to Y treatment.” Clinical prediction has as its goal statements such as, “this person will respond in X manner to most circumstances of Y type.” The actuarial prediction has applied value for the program administrator or planner. For example, a welfare director may base his planning upon a prediction that 50% of the juveniles in a particular ghetto area will become delinquent unless assisted by prevention programs. Such predictions are of little use, however, to the clinician who works in a prevention program in that ghetto area. His responsibility is to single individuals and he must be able to predict relationships between specific environmental interventions and delinquent behaviors for single individuals on his case load.

The inability of research to provide clinically applicable data has often been simply attributed to the relative youth of behavioral science. As research continues, this argument runs, sufficient knowledge will be gained to permit individual clinical prediction. Ultimately we shall know enough about the myriad variables which affect human behavior that we shall be able accurately to measure these with respect to a single person, and predict what specific treatment will alter his behavior. The history of the physical sciences would certainly lend support to this view; a century ago few would have believed it possible to understand and control circumstances so as to permit placing a satellite in exact orbit about the earth. However, even if extensive basic research will improve prediction of individual behavior, that improvement is many decades away from today’s clinician. Our viewpoint is that, even given vast amounts of basic knowledge, the clinician will still be grossly limited in understanding the highly idiosyncratic problems of individual patients. The inevitable uniqueness of a person’s developmental history means that major aspects of his behavior will be lawful only onto himself. Gordon Allport, whose writings have enormously influenced our own position, put it succinctly: “The true goal of clinical psychology is not to predict the aggregate, but to foretell ‘what any one man will do.’ In researching this ideal, actuarial prediction may sometimes help, universal and group norms are useful, but they do not go the whole distance” (Allport, 1961, p. 21).

The experimentalist would have his findings apply to a broad class of persons while the clinician wants his findings to apply to a sphere as limited as all of a particular kind of behavior in a single individual. Because the respective universes to which they must generalize their findings differ, clinicians and experimentalists should and do sample differently. The experimentalist samples from a population of persons; the clinician, from a population of behaviors in a single person. The experimental psychologist aims his research at understanding the variability of behavior between groups of persons who differ because of some treatment manipulation or because they were selected according to some organismic factor. His basic measure is the mean score for a group of individuals, and differences between individuals within a given group are classed as error variance. The clinician is interested in understanding the conditions of the experimentalist’s error variance. The clinician’s subject is the one who produces the statistically rare score. Consequently, the clinician’s basic measure (if only he would use measures) would be an individual’s average response over many repeated occasions; and his error variance would be differences in that individual’s behavior under a given level of a specified independent variable. His interest would be to understand the conditions necessary for that individual’s behavior to differ from its average. His findings would be generalized to permit the control of future behaviors of his subject.

Clinicians and behavioral scientists do not, as is sometimes supposed, differ in their wish to improve their knowledge or understanding of human behavior, nor in their wish to control or predict behavior. Rather, it appears that they differ in terms of the kind of knowledge they sel...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• Preface
• 1. The Experimental-Clinical Method
• 2. The Residential Setting and Treatment Groups
• 3. Measurement
• 4. Same-Subject Experimental Designs
• 5. Treatment Techniques
• 6. Unexpected Values of Social Reinforcers
• 7. Case Studies
• 8. Conclusion
• References
• Index