Behavioral Diagnosis PDF

Behavioral Diagnosis FREDERICK H. KANFER GEORGE SASLOW University of Oregon Medical School, Portland, Oregon This chapter is an enlarged and updated version of an earlier paper published...

Behavioral Diagnosis FREDERICK H. KANFER GEORGE SASLOW University of Oregon Medical School, Portland, Oregon This chapter is an enlarged and updated version of an earlier paper published in the Archives of General Psychiatry, 1965, 12, p.529-538. During the past decade criticism of conventional psychiatric diagnosis has been so widespread that many clinicians now use diagnostic labels sparingly and apologetically. In particular, these labels are sometimes useless from the viewpoint of treatment and prognosis. One study investigating the reasons for disagreements among experienced psychiatric diagnosticians (Ward, Beck, Mendelson, Mock, & Erbaugh, 1962) reported that 62.5 percent of the differences were due to inadequacies in the nosological system, the widely used 1952 APA classification. The continued adherence to the nosological terms of this classificatory scheme suggests some utility of the present categorization of behavior disorders, despite its apparently low reliability (Ash, 1949; Rotter, 1954); its limited prognostic value (Freedman, 1958; Windle, 1952); and its multiple feebly related assumptive supports. The theories of personality underlying the present APA diagnostic classification system do permit some limited generalizations about a patient's probable behavior under very general conditions. But with the well-documented overlap of symptoms in different groups (eg. Wittenborn, Holzberg & Simon, 1953) and poor interjudge reliabilities of categories (Kostlan, 1954 Schmidt & Fonda, 1956) the ubiquity of enduring behavioral characteristics within diagnostic groups is doubtful, at best. However, even if better group separation and improved rater reliability were established, there is still no good basis for assuming even minimal utility of the currently used system for assignment to treatment. Bannister, Salmon, & Leiberman (1964) examined the relationship between diagnosis and treatment in 1000 psychiatric patients. To test their hypotheses at three levels of diagnostic specificity, all patients were classified three times: (1) as psychotic, neurotic, or organic; (2) according to such commonly used categories as schizophrenics, affectives, hysterics, reactive depressives; and (3) according to the finer differentiation made by the International Classification of Diseases for the World Health Organization in 1948. For each level of classification a matrix was constructed by diagnostic categories and 14 treatment categories. The analyses reveal that the best prediction of treatment for a given diagnosis at level I would result in 18.2 percent correct guesses, for level II in 32.9 percent, and for level III in 30.6 percent in 1,000 cases. The authors conclude that "the findings are not consistent with the notion that each particular diagnosis leads logically (or habitually) to particular treatment. It suggests that variables other than diagnosis may be as important as, or more important than, diagnosis in predicating choice of treatment" (1964, p. 731). We propose here to examine some sources of dissatisfaction with the present approach to diagnosis, to describe a framework for a behavioral analysis of individual patients which implies both suggestions for treatment and outcome criteria for the single case, and to indicate the conditions for collecting the data for such an analysis. The most desirable classification system would be one which, from knowledge of only a few salient characteristics of a person, permits highly accurate predictions for many crucial behaviors, including responses to specific treatments, the probability of occurrence of various unacceptable behaviors, and the degree of social effectiveness. While such goals may be 1 unrealistic at the present time, exploration of an empirical and nontraditional approach to diagnosis may contribute new behavioral dimensions for an eventual conceptualization of the behavior disorders nearer to what is desired and clinically useful. The feasibility of handling diagnostic information by data-processing machines has already been demonstrated for some medical diseases (eg, Brodman, van Woerkom, Erd mann, & Goldstein, 1959; Ledley & Lusted, 1959). The large remaining problem is the provision of meaningful observations and treatment categories to the computer for empirical evaluation of their relationships. Problems in Current Diagnostic Systems Numerous criticisms have dealt with the internal consistency, the explicitness, the precision, and the reliability of psychiatric classifications. It seems to us that the most important fault lies in our lack of sufficient knowledge to categorize behavior along those pertinent dimensions which permit prediction of responses to social expectations, social stresses, life crises, or psychiatric treatment. This limitation obviates anything but a crude and tentative approximation to a taxonomy of effective and ineffective behavior. A reasonable expectation of a practical diagnostic schema is that the taxonomic system be closely related to the conceptual and empirical framework of treatment. Zigler & Phillips, in discussing the requirement for an adequate system of classification, suggest that an etiologically oriented closed system of diagnosis is premature. Instead, they believe that an empirical attack is needed using "symptoms broadly defined as meaningful and discernible behaviors, as basis of the classificatory system" (1961, p. 616). But symptoms of a class of responses are defined, after all, only by their nuisance value to the patient's social environment or to himself as a social being. They are also notoriously reliable in predicting the patient's particular etiological history or his response to treatment. An alternate approach lies in an attempt to identify classes of dependent variables in human behavior which would allow inference about the particular contemporary controlling factors, the social stimuli, the physiological stimuli, and the reinforcing stimuli, of which they are a function. In the present early stage of the art of psychological prognostication, it appears most reasonable to develop a program of analysis which is closely related to subsequent treatment. A classification scheme which implies a program for behavioral change is the only one which has not only utility but the potential for experimental validation. The task of assessment and prognosis can be reduced to efforts which answer the following three questions: (1) Which specific behavior patterns require change in their frequency of occurrence, their intensity, their duration or in the conditions under which they occur? (2) What are the conditions under which this behavior was acquired, and what factors are currently maintaining it? (3) What are the best practical means which can produce the desired changes in this individual (manipulation of the environment, the behavior or the self-attitudes of the patient)? The investigation of the history of the problematic behavior is mainly of academic interest, except as it contributes information about the probable efficacy of a specific treatment method. 2 EXPECTATIONS OF CURRENT DIAGNOSTIC SYSTEMS In traditional medicine, a diagnostic statement about a patient has often been viewed as an essential prerequisite to treatment because a diagnosis suggests that the physician has some knowledge of the origin of the difficulty and of the probable future course of the illness. Further, in medicine, diagnosis frequently brings together the accumulated knowledge about the pathological process which leads to the manifestation of the symptoms, and the experiences which others have had in the past in treating patients with such a disease process. Modern medicine recognizes that any particular disease need not have a single cause, or even a small number of antecedent conditions. Nevertheless the diagnostic label attempts to define at least the necessary conditions which are most relevant in considering a treatment program. Some diagnostic classification system is also invaluable as a basis for many social decisions involving entire populations. For example, planning for treatment facilities, research efforts, and educational programs takes into account the distribution frequencies of specified syndromes in the general population. Ledley & Lusted (1959) give an excellent conception of the traditional model in medicine by their analysis of the reasoning underlying it. The authors differentiate between a disease complex and a symptom complex. While disease complex describes known pathological processes and their correlated signs, symptom complex represents particular signs present in a particular patient. The bridge between disease and symptom complexes is provided by available medical knowledge, and the final diagnosis is tantamount to labeling disease complex. However, the current gaps in medical knowledge necessitate the use of probability statements when relating diseases to symptoms, admitting that there is some possibility of error in the diagnosis. Once the diagnosis is established, decisions about treatment still depend on many other factors including social, moral, and economic conditions. Ledley & Lusted (1959) thus separate the clinical diagnosis into a two-step process. A statistical procedure is suggested to facilitate the primary or diagnostic labeling process. However, the choice of treatment depends not only on the diagnosis proper. Treatment decisions are also influenced by the moral, ethical, social, and economic conditions of the individual patient, his family, and the society in which he lives. The proper assignment of the weight to be given to each of the values must in the last analysis be left to the physician's judgment (Ledley & Lusted, 1959). The Ledley & Lusted model presumes available methods for the observation of relevant behavior (the symptom complex) and some scientific knowledge relating it to known antecedents or correlates (the disease process), But contemporary psychological theory has as yet little scientific knowledge of behavior pathology and therefore no guidelines for the observer to suggest what is to be observed. A patient who presents himself for diagnosis maybe examined by five observers with diverse orientations to behavior pathology. One observer might attempt to define the problem in terms of the patient’s biological constitution, another in terms of his early life experiences, a third in terms of his interpersonal experiences, a fourth in terms of his unconscious system, and a fifth in terms of faulty communication systems. Each framework influences the observer to elicit data selectively and to accept or reject information as it can be accommodated in his conceptual system. As a result, lack of agreement exists even on what is to be observed as a basis for diagnosis. In the absence of a unique psychiatric model for psychiatric illness, there tends to be a widespread use of the medical model with demonstrated utility for other types of disorders, even when its relevance 3 may be questionable. In contrast to this solution, Szasz (1960) has expressed the view that the medical model may be totally inadequate because psychiatry should be concerned with problems of living and not with diseases of the brain or other biological organs. Szasz argues that mental illness is a myth, whose function it is to disguise and thus render more palatable the bitter pill of moral conflict in human relations" (1960, p. 118). The limitations of the somatic model have been discussed even in some areas of medicine for which the model seems most appropriate. For example, in the nomenclature for diagnosis of disease of the heart and blood vessels, the criteria committee of the New York Heart Association (1953) suggests the use of multiple criteria for cardiovascular diseases, including a statement of the patient's functional capacity. The committee suggests that the functional capacity be "estimated by appraising the patient's ability to perform physical activity" (p. 80) and decided largely by inference from his history. Further, [it] should not be influenced by the character of the structural lesion or by an opinion as to treatment or prognosis" (p. 81). This approach makes it clear that a comprehensive assessment of a patient, regardless of the physical disease which he suffers, must also take into account the social expectations of his life settings, his social effectiveness, and the ways in which physiological, anatomical and psychological factors interact to produce a particular behavior pattern in an individual patient. In cardiovascular disease (as, we believe, in psychiatric problems) the patient's functional capacity is evidently not predictable from any single one of a number of highly relevant contributing factors. The importance of the total matrix in which the evaluation is made is indicated by the committee's conclusions that a patient's functional capacity must be observed in its actual manifestations. When medical illness is looked upon in such a comprehensive way, it is possible to find a model which may be common to both medical and psychiatric problems (Guze, Matarazzo, & Saslow, 1953). In terms of such an inclusive model which neglects no dimension relevant to the problem, we can understand that even an acute schizophrenic patient may perform quite adequately in some social interactions or areas of talent, such as musical or literary, while the usual diagnostic label has no value in predicting the total range of his behavioral potentials. MULTIPLE DIAGNOSIS A widely used practical solution and circumvention of the difficulty inherent in the application of the medical model to psychiatric diagnosis is offered by Noyes & Kolb (1963). They suggest that the clinician construct a diagnostic formulation consisting of three parts: (1) a genetic diagnosis incorporating the constitutional, somatic, and historical-traumatic factors representing the primary sources of the determinants of mental illness; (2) a dynamic diagnosis which describes the mechanisms and techniques unconsciously used by the individual to manage anxiety and enhance self-esteem, ie, that traces the psychopathological processes; and (3) a clinical diagnosis which attempts to convey connotations concerning the reaction syndrome, the probable course of the disorder, and the methods of treatment which will most probably prove beneficial. Noyes & Kolb's multiple criteria can be arranged along three simpler dimensions of diagnosis which may have some practical value to the clinician (1) etiological, (2) behavioral, and (3) predictive diagnosis, or prognosis. The kind of information which is conveyed by each type of diagnostic label is somewhat different and specifically adapted to the purpose for which the diagnosis is used. By this triple-label approach Noyes & Kolb attempt to counter the criticism aimed at use of any single classificatory system. Confusion in a single system of classification as used by other 4 authors is due in part to the fact that a diagnostic formulation intended to describe current behavior, for example, may be found useless in an attempt to predict the response to specific treatment, to "postdict" the patient's past personal history and development, or to permit collection of adequate frequency data on hospital populations. CLASSIFICATION BY ETIOLOGY The Kraepelinian system and portions of the 1952 APA classification emphasize etiological factors. They share the assumption that common etiological factors lead to similar symptoms and respond to similar treatment. This dimension of diagnosis is fruitful for immediate treatment when dealing with behavior disorders which are mainly associated with biological condition or state. When a patient is known to suffer from excessive intake of alcohol, his hallucinatory behavior, lack of motor coordination, poor judgment, and other behavioral evidence of disorganization can often be related directly to some antecedent condition such as the toxic effect of alcohol on the central nervous system, liver, etc. For these cases, classification by etiology also has some implications for prognosis and treatment. Acute hallucinations and other disorganized behavior due to alcohol usually clear up when the alcohol level in the bloodstream falls. Similar examples can be drawn from any class of behavior disorders in which a change in behavior is associated primarily or exclusively with a single, particular antecedent factor. Under these conditions this factor can be called a pathogen, and the situation closely approximates the condition described by the traditional medical model. Utilization of this dimension as a basis for psychiatric diagnosis, however, has many problems apart from the rarity with which a specified condition can be shown to have direct "causal" relationship to a pathogen. Among the current areas of ignorance in the fields of psychology and psychiatry, the etiology of most common disturbances probably takes first place. In the present state of knowledge, no specific family environment, no dramatic traumatic experience, no known constitutional abnormality has yet been found which results in the same pattern of disordered behavior. While current research efforts have aimed at investigating family patterns of schizophrenic patients, and several studies suggest a relationship between the mother's behavior and schizophrenic process in the child (Jackson, 1960), it is not at all clear why the presence of these same factors in other families fails to yield a similar incidence of schizophrenia. Further, patients may exhibit behavior diagnosed as schizophrenic when there is no evidence of the postulated mother-child relationship (Frank, 1965; Freedman, 1958). CLASSIFICATION BY SYMPTOMS A clinical diagnosis often is a simple expanded statement about the way in which a person behaves. On the assumption that a variety of behaviors are correlated and consistent in any given individual, it becomes more economical to assign the individual to a class of persons than to list and categorize all his behaviors and to describe his operations in thousands of isolated situations. The utility of such a system rests heavily on the availability of empirical evidence concerning probable correlations among various behaviors (response-response relationships) and the further assumption that the frequency of occurrence of such behaviors is relatively independent of specific stimulus conditions and of specific reinforcement. There are two major limitations to such a system. The first is that diagnosis by symptoms, 5 as we have indicated in an earlier section, is often misleading because it implies common etiological actors. The second limitation is that the current approach to diagnosis by symptoms tends to center on a group of behaviors which is often irrelevant with regard to the patient's total life pattern. These behaviors may be of interest only for historical reasons, or because they are popularly associated with deviancy and disorder. For example, occasional mild delusions interfere little or not at all in the social interactions or occupational effectiveness of many ambulatory patients. Nevertheless, admission of their occurrence is often sufficient for a diagnosis of psychosis. Refinement of diagnosis by symptoms beyond current usage appears possible, as shown for example by Lorr, Klett & McNair (1963), but this does not remove the above limitations. Utilization of a symptom-descriptive approach frequently focuses attention on by-products of larger behavior patterns and results in attempted treatment of behaviors which may be simple consequences of other more important aspects of the patient's life. Much of the emphasis on the patient's subjective experiences, feelings, and moods results in such a syndrome-oriented classification. Subsequent therapeutic efforts are then made to change the feelings, anxieties, and moods (or at least the patient's report about them) rather than to investigate the life conditions, interpersonal reactions, and environmental factors which produce and maintain these habitual response patterns. The questionable value of a treatment approach, founded on the assumption that psychiatric intervention must focus on disordered subjective states, has recently been shown by McPartland & Richart (1966). The authors compiled information from 393 applicants for psychiatric treatment at a public municipally supported center, with regard to (1) clinical features, such as depression, suspicion, and feelings of worthlessness, and (2) problems of living, related to the patient's environment, or plights, such as marital problems, financial problems, and physical or work problems. They found that admissions to some form of treatment (inpatient or nonresidential) were significantly higher for patients who were described as delusional, hallucinating, disorganized and confused, incoherent, suicidal, and hostile. Presence or absence of the remaining eight clinical features (agitated, depressed, fearful-anxious, withdrawn/vague, suspicious, feeling worthless, hypochondriacal) was not significantly selective for admission to treatment. Of the series of presented plights, only one (traumatic event) showed a significant relationship to the decision to treat. The central question of the study addressed itself to the relative persistence of plights and clinical features in a six-month follow-up. A sample of 50 applicants was interviewed in their homes. The sample was matched to the larger base sample, in which approximately three-fourths had undergone treatment. It was found that the eight elements which showed least relative reduction over the six-month interval were plights (ranging from 45 to 27 percent reduction). The greatest reductions were found in clinical features (e.g. suicidal, hostile, hallucinating, etc.), with a range downward from 95 percent. McPartland & Richart (1966) conclude that, in their population of disadvantaged people who use public clinics, clearing of the clinical picture is not usually followed by a less problematic way of life. They suggest that the problems of life are unlikely to be the derivatives of disordered thoughts and feelings. It is more plausible to interpret the data as indicating that these disorders are products of life stresses. They comment on the limited role played in clinical decisions by the complex problems of life, and suggest that the current paradigm, treating "inner" symptoms to produce fundamental improvements in the problem areas of patient's lives, may be a poor one. 6 CLASSIFICATION BY PROGNOSIS To date, the least effort has been devoted to construction of a classification system which assigns patients to the same category on the basis of their similar response to specific treatments. The proper question raised for such a classification system concerns the manner in which a patient will react to treatments, regardless of his current behavior or past history. The numerous studies attempting to establish prognostic signs from current behaviors by behavioral observations, by projective personality tests, or by somatic tests all represent efforts to categorize the patients on this dimension. Windle (1952) has called attention to the low degree of predictability afforded by personality (projective) test scores, and has pointed out the difficulties encountered in evaluating research in this area due to the inadequate description of the population sampled and of the improvement criteria. Summaries in the Annual Review of Psychology of the yearly crops of outcome research in psychotherapy and of prognostle sign studies attest to the repeated failures to find useful predictors of response to treatment. The lack of reliable relationships between, on the one hand, diagnostic categories, test data, demographic variables, or any other measure of the patient's performance, and on the other hand, duration of illness, response to specific treatment, or degree of recovery precludes the construction of a simple empiric framework for a diagnostic-prognostic classification system based only on an array of symptoms. None of the currently used dimensions for diagnosis is directly related to methods of possible modification of a patient's behavior. Since the etiological model clearly stresses causative factors, it is much more compatible with a personality theory which strongly emphasizes genetic-developmental factors. Classification by symptoms facilitates social-administrative decisions about patients by providing some basis for judging the degree of deviation from social and ethical norms. Such classification is compatible with a personality theory founded on the normal curve hypothesis and concerned with characterization by comparison with a fictitious average. The prognostic-predictive approach appears to have the most direct practical applicability. If continued research were to support certain early findings, being able to predict outcome of mental illness from a patient's premorbid social competence score (Zigler able predict outcome and Phillips, 1961), from his score on an ego-strength scale (Barron, 1953), or from many of the other signs and single variables which have been shown to have some predictive powers, it would indeed be comforting. It is unfortunate that these powers are frequently dissipated in cross-validation. As Fulkerson & Barry (1961) have indicated, single predictors have not yet shown much success. A Functional (Behavioral-Analytic) Approach The growing literature on behavior modification derived from learning theory (Bandura, 1961, Ferster, 1963; Kanfer, 1961, Krasner, 1962a: Ullmann Krasner, 1965, Wolpe, 1958, and others) suggests the necessity for a change in effective diagnostic procedures. The learning approach differs in its closer articulation between diagnosis and treatment, with continuing reconsideration of therapeutic efforts on the basis of information obtained during the entire operation. This approach sacrifices the taxonomic features of the usual diagnostic enterprise for greater specificity and heavier contributions of the obtained observations to direct use in the therapeutic intervention. A functional analysis of behavior endeavors to ascertain the explicit environmental and historical variables which control the observed 7 behaviors. In discussing this approach, Ferster (1965) has emphasized that “a functional analysis of behavior has the advantage that it specifies other causes of behavior in the form of explicit environmental events which can be objectively identified and which are potentially manipulable." Such an undertaking makes the assumption that description of the problematic behavior. its controlling factors, and the means by which it can be changed are the most appropriate "explanations” for the patient's actions. It stresses that the unit of analysis is a relationship between the environment and behavior, with attention not only on antecedent variables but also on the impact of behavioral acts on the patient's environment. The patient is considered to be a member of several social systems differing in significance to him (eg, his family, friends, coworkers, clubs), and it is assumed that behavior contributes to the maintenance or disruption of these systems, just as the group norms of these systems affect his behavior. Lindsley (1964) has expanded the traditional SR analysis to include the components deemed necessary for a full behavioral analysis. He describes the sequence as stimulus (S), response (R), contingency (K), consequence (C). This listing suggests separate consideration of each unit in the analysis of operant behaviors. The S component includes antecedent events, the R refers to observed (or reliably reported) behaviors, K describes the schedules or contingency-related conditions, and C refers to events following R, be they environmental or organismic. Our present view expands Lindsley's behavioral view by one additional component, the biological condition of the organism (O), in order to permit inclusion of variables especially relevant to populations with psychological and biological dysfunctions. Several other characteristics of the type of analysis proposed here need to be listed. 1. Since the patient operates in a complex of systems, not only psychological events but all events, including biological, economic, and social, must be admitted as potential variables in the analysis, without prior judgment about their order of importance. 2. It is assumed that many of the delicate social interactional patterns are operant responses. Operant responses are maintained by their effect on the environment. Therefore, an account of the consequences of the patient's behaviors on the environment can provide a rough organization of the patient’s instrumental acts into classes defined by the outcome of these acts. The verification of specific antecedent stimuli may often be neither feasible nor necessary. 3. The potential range of the behavioral repertoire is limited by the individual's biological, social, and intellectual competence, by his past history of reinforcement for his behaviors, and by the current norms of his membership groups. Consequently, knowledge of the patient's history, of the limits of his capacities, and of the norms of his membership and reference groups are essential for effective therapeutic planning 4. By emphasis on the importance of individual learning histories, it is recognized that numerous common features appear in the behavior of individuals reared in the same cultural environment and exposed to similar learning experiences. However, the approach stresses the need for individual assessment and for construction of therapeutic plans which are suited to the unique environmental and behavioral characteristics encountered with each patient. Therefore, no specific limited catalog of treatment techniques is envisioned. Rather, the functional analysis provides a method of analysis and some procedural guidelines. These may be combined into unique constellations of operations and priorities in the individual case. The totality of the patient's problems as initially presented does not have to be dealt 8 with immediately, or by the time treatment is terminated. An effective assessment procedure requires only that some hierarchy of problems is established so that priorities be assigned for treatment of various maladaptive patterns or for specific situational interventions. As some of the patient's problematic behaviors are treated in accordance with the initial decisions about priorities, both priorities and treatment operations are changed by a method of successive approximation. The results of the initial operations determine the choice of appropriate strategies for succeeding problems. It is thus clear that a behavioral analysis is not limited only to formulating an initial strategy but that reanalyses takes place as the treatment plan proceeds. 5. A functional analysis does not inevitably lead to therapeutic interventions in a patient's psychological functioning. In fact therapeutic interventions may consist solely of modifications of the patient's physical or social environment, or of the behavior of persons other than the patient, or of other variables believed to play a role in the maintenance of the presenting problem. The present approach shares with many psychological theories the assumption that psychotherapy is not an effort aimed at removal of intrapsychic conflicts, nor at a change in the personality structure by therapeutic interactions of intense nonverbal nature (eg, transference, self-actualization). Instead, we adopt the assumption that the job of "psychological” treatment involves the utilization of a variety of methods to devise a program which controls the patient's environment, or enables him to control his behavior and the consequences of his behavior in such a way that the presenting problem is resolved. We hypothesize that the essential ingredients of a psychotherapeutic endeavor with the patient usually involve two separate efforts: (1) to change the patient's approach to perceiving, classifying, and organizing sensory events, including perceptions of himself, and (2) to change the response patterns which he has established in relation to social objects and to himself over the years (Kanfer, 1961). It is necessary to indicate what the theoretical view presented here does not espouse in order to understand the differences from other procedures. It does not rest upon the assumption that (1) insight is a sine qua non of psychotherapy, (2) changes in thoughts or ideas inevitably lead to ultimate changes in actions, (3) verbal therapeutic sessions serve as replications of and equivalents for actual life situations, and (4) a symptom can be removed only by uprooting its cause or origin. In the absence of these assumptions it becomes unnecessary to conceptualize behavior disorder in etiological terms, in psychodynamic terms, or in terms of a specifiable disease process. While psychotherapy by verbal means may be sufficient in some instances, the combination of behavior modification in life situations and in verbal interactions serves to extend the armamentarium of the therapist. Therefore verbal psychotherapy is seen as an adjunct in the implementation of therapeutic behavior changes in the patient's total life pattern, not as an end in itself, nor as the sole vehicle for increasing psychological effectiveness. In embracing this view of behavioral modification, there is a further commitment to a constant interplay between assessment and therapeutic strategies. An initial diagnostic formulation seeks to ascertain the major variables which can be directly controlled or modified during treatment. During successive treatment stages additional information is collected about the patient's behavior repertoire, his reinforcement history in pertinent controlling stimuli in his social and physical environment, and the sociological limitations within which both patient and therapist have to operate. While a behavioral approach stresses the importance of environmental variables, it is not oblivious to the importance of a person's capacity for modifying his environment and his own behaviors 9 in relation to it. Ultimately most patients leave psychotherapeutic supervision. The changes which can be initiated by a therapist must be maintained by the patient himself. Consequently, proper assessment must include an evaluation of the patient's potentials for self-regulation and for self-initiation, i.e., active participation in the treatment process and in subsequent readjustments to life circumstances. To maximize the probability of successful treatment the clinician's task may necessitate at first direct intervention in the patient's environmental circumstances, and later the modification of the behavior of other people significant in his life, as well as the control of reinforcing stimuli which are available either through self-administration or by contingency upon the behavior of others. These procedures complement the verbal interactions of traditional psychotherapy. They require that the clinician, at the invitation of the patient or his family, participate more fully in planning the total life pattern of the patient outside the clinician's office, until there is reason to believe that either the environmental conditions or the patient's ability to control them have so changed that the therapist's interventions in the patient's environment are no changed that the therapist's Jonger necessary. When a behavioral analysis reveals that a patient's problems lie mainly in his dissatisfactions with or uncertainties about his self-attitudes, or a loss of meaning or purpose, there is further a serious question, ably discussed by Schofield (1964), whether such patients should be treated by the traditional professional groups, the psychiatry, psychology, and social work members of the "mental healing" professions. These cases constitute mostly self-referred persons, sensitive to their own plight, psychologically sophisticated, and often functioning effectively in their daily responsibilities. One solution suggested by Schofield lies in recognizing the contributions of "invisible" psychotherapists, ie., clergymen, teachers, counsellors, work supervisors, even neighbors and friends, who often provide effective help. We agree heartily with Schofield when he says (p. 140): Because of the established impossibility of meeting the demand for therapeutic conversation within the present and probably future supply of currently sanctioned psychotherapists, it is essential that careful consideration be given to all possible means of increasing the supply and quality of help from these invisible therapists. The exploration of such resources early during the behavioral analysis permits the clinician to determine the degree to which clinical intervention can be supplemented by these resources. Such knowledge should also curtail excessive use of the clinician's time and of other psychiatric resources by gradual shifting of therapeutic actions to qualified but nonprofessional (ie non clinical) persons in the patient's natural environment. Again, the underlying stress on the problem not as "internal" but as interactional, not as "pathological" but as related to the patient's life pattern and his physical and social environment, dictates greater concern in diagnosis with matters other than the patient's mental status or emotional-developmental history. This analysis is consistent with earlier formulations of the principles of comprehensive medicine (Guze, Matarazzo, & Saslow, 1953: Saslow 1952) which emphasized the joint operation of biological, social, and psychological factors in psychiatric disorders. The language and orientation of the proposed approach are rooted in contemporary learning theory. The conceptual framework is consonant with the view that the course of psychiatric disorders can be modified by systematic application of scientific principles from the fields relevant to the patient's habitual mode of living, eg, biology, psychology, sociology, clinical medicine. The analysis proposed here is not intended to lead to assignment of the patient to diagnostic categories. It should serve as a basis for making decisions about specific 10 therapeutic interventions, regardless of the presenting problem. The compilation of data under as many of the headings as are relevant should yield a good basis for decisions about the areas in which intervention is needed, the particular targets of the intervention, the treatment methods to be used, and the series of goals at which treatment should aim. I. Initial analysis of the problem situation A preliminary formulation attempts to sort out the behaviors which are brought to the clinician's attention with regard to their eventual place in the treatment procedures. The patient’s repertoire may be conspicuously different from what is required for adequate adjustment to his circumstances because of the unusual frequency with which various acts occur. Further, account is also taken of the extent of the behavioral repertoire which is nonproblematic and the presence of behaviors representing special strengths, qualitatively or quantitatively, which would be available as resources in treatment. Although the classification of behaviors into excesses and deficits is conceptually useful, it is clear that in all but extreme cases humans have a rich and continually changing repertoire and the interrelationships between the items in the repertoire cannot be ignored. The initial classification may not retain its appropriateness, once changes in the patient's life conditions and his overall activities begin to occur. Since no objective frequency tables are available for reference, behavior items can be viewed either as excesses or as deficits, depending on the vantage point from which the imbalance is observed. For instance, excessive withdrawal and deficient social initiative, or excessive response to emotional stimulation and deficient self-controlling behavior, may be complementary. However, a decision about the starting point for treatment is necessary, committing the clinician to a set of priorities for treatment which can later be reviewed and changed. Preference for viewing behavior as excessive or deficient is often determined by cultural valuation of the behavior, based on its consequences to other people. For example, in a child a physical blow in response to a critical remark can be regarded as excessive aggressive behavior, leading to the decision to reduce its occurrence by constraints or punishment. A clearly different consequence results from regarding the same behavior as a deficit in self control, to be remedied by training the child to substitute acceptable alternative responses. Another choice resulting from the analysis may lie in a dual approach. The therapist may decide to train the child to discriminate among cues which indicate potential acceptance or rejection of the behavior by social members, e.g., the differential reinforcement potentials for physical blows in such sports as boxing, as contrasted to its consequences in conversation. Concurrently, major effort could be spent on remedying the relevant deficits in alternate social behaviors. A. Behavioral excess. A class of related behaviors occurs and is described as problematic by the patient or an informant because of excess in (1) frequency, (2) intensity, (3) duration, or (4) occurrence under conditions when its socially sanctioned frequency approaches zero. Compulsive hand washing, combativeness, prolonged excitement, and sexual exhibitionism are examples of behavioral excesses along one or another of these four dimensions. Less obvious, because they often do not constitute the major presenting complaint and appear only in the course of the behavioral analysis, are examples of socially unacceptable solitary, affectionate, or other private 11  behaviors. For instance, a housewife showing excessive solitary preoccupation can do so by excessive homemaking activities, (1) several hours a day, (2) seven days weekly for most of the waking day, (3) to the extent that phone calls or doorbells are unanswered and family needs are unattended. From this example it is clear that both duration and intensity values of the behavior may jointly determine the characterization of the behavior as excessive. B. Behavioral deficit. A class of responses is described as problematic by someone because it fails to occur (1) with sufficient frequency, (2) with adequate intensity, (3) in appropriate form, or (4) under socially expected conditions. Examples are: reduced social responsiveness (withdrawal), amnesias, fatigue syndromes, and restrictions in sexual or somatic function (e.g., impotence, writer's cramp). Other examples of behavioral deficits can be found in depressed patients who have no appropriate behavior in a new social environment, e.g., after changes from a rural to an urban area, from marital to single status, or from one socioeconomic level to another. "Inadequate" persons often are also found to have large gaps in their social or intellectual repertoires which prevent appropriate actions. C. Behavioral asset. Behavioral assets are nonproblematic behaviors. What does the patient do well? What are his adequate social behaviors? What are his special talents or assets? The content of life experiences which can be used to execute a therapeutic program is unlimited. Any segment of the patient’s activities can be used as an arena for building up new behaviors. In fact, his natural work and play activities provide a better starting point for behavior change than can ever be provided in a synthetic activity or relationship. For example, a person with musical talent, skill in a craft, physical skill, or social appeal can be helped to use his strengths as vehicles for changing behavior, relationships and for acquiring new behaviors in areas in which some successful outcomes are highly probable. While a therapeutic goal ultimately be the acquisition of specific social or self-evaluative behaviors, the learning can be programmed with many different tasks and in areas in which the patient has already acquired competence. II. Clarification of problem situation A. Assign the classes of problematic responses to group a or b above as study of the patient proceeds. B. Which persons or groups object to these behaviors? Which persons or groups support them? Who persuaded or coerced the patient to come to the clinician? C. What consequences does the problem have for the patient and significant others? What consequences would removal of the problem have for the patient or others? D. Under what conditions do the problematic behaviors occur (biological, symbolic, social, vocational, etc.)? E. What satisfactions would continue for the patient if his problematic behavior were sustained? What satisfactions would the patient gain if as a result of psychiatric intervention, his problematic behavior were changed? What positive or aversive effects would occur for significant others if the patient's problematic behavior were changed? How would the patient continue to live if therapy were unsuccessful, i.e., if nothing in his behavior changed? F. What new problems in living would successful therapy pose for the patient? 12  G. To what extent is the patient as a sole informant capable of helping in development of a therapy program? The questions raised here are derived from the assumption that maladjusted behavior requires continued support. It cannot be banished from patient's life for all future circumstances. Change in it is related closely to the environment in which the person needs to live. Elimination of the problematic behavior is also impossible as long as powerful, and often undefined reinforcing events operate. The answers to the above questions can help to bring about an early decision about the optimal goals within practical realm of the clinician and within the inevitably fixed boundaries of the patient's life pattern. III. Motivational Analysis A. How does the patient rank various incentives in their importance to him? Basing judgment on the patient's probable expenditure of time, energy, physical discomfort, which of the following reinforcing events are relatively most effective in initiating or maintaining his behavior: achievement of recognition, sympathy, friendships, money, good health, sexual satisfaction, intellectual competence, social approval, work satisfaction, control over others, securing dependency, etc.? B. How frequent and regular have been his successes with these reinforcers? What are his present expectations of success for each? Under what circumstances was reinforcement achieved for each of these incentives? C. Under what specific conditions do each of these reinforcers arouse goal-directed behavior (biological, symbolic, social, vocational)? D. Do his actions in relation to these goals correspond with his verbal statements? How does any definable discrepancy affect goals and procedures in therapy? E. Which persons or groups have the most effective and widespread control over his current behavior? F. Can the patient relate reinforcement contingencies to his own behavior, or does he assign reinforcement to random uncontrollable factors, “superstitious" behavior, belief in luck, fate, miracles, etc.)? G. What are the major aversive stimuli for this patient, (1) in immediate day-to-day life, (2) in the future? Are there bodily sensations, conviction of illness, or fears of illness which serve as important aversive stimuli for changes? What are his fears, the consequences which he avoids and dreads, the risks which he does not take? H. Would a treatment program require that the patient give up current satisfactions associated with his problem, e.g., invalid status in the family or on the job; gratifications possibly due to 13  unemployment; life restrictions and special privileges justified by his "nervous" status; Illness as justification for failure to fulfill expectations of himself or others? L. Which events of known reinforcing value can be utilized for learning new interpersonal skills or self-attitudes during treatment? In what areas and by what means can positive consequences be arranged to follow desirable behaviors, replacing earlier aversive consequences? IV. Developmental analysis A. Biological changes 1. What are the limitations in the patient's biological equipment thst may affect his current behavior, like defective vision and presence of illnesses, such as stroke, poliomyelitis, mononucleosis, glandular imbalances)? How do these limitations initiate or maintain undesirable behaviors (eg, behavioral constrictions due to fatigue, fear of overexertion, avoidance of social exposure of these deficits)? Can the patient's self-limiting expectations of the interfering consequences be changed? 2. When and how did biological deviations or limitations develop? What consequences did they have on his life-pattern and on his self-attitudes? What was done about them, by whom? Has he developed specific consistent response patterns toward some body structure or function? 3. How do these biological conditions limit response to treatment or resolution of his problems? B. Sociological changes. 1. What are the most characteristic features of the patient's present sieciocultural milieu (with regard to urban versus rural environment, religious affiliation, socioeconomic status, ethnic affiliation, educational-intellectual affiliation, etc.)? Are his attitudes congruent with this milieu? For instance how is a college orientation of an adolescent accepted by his peer group in a poor neighborhood? How does the home and neighborhood environment respond to a patient's religious, social, and sexual activities and fantasies? 2. Have there been changes in this milieu which are pertinent to his current behavior? If so, how long ago, how permanently, and under what conditions did such changes occur? What immediate consequences did they have for the behavior of the patient? For example, what impact on a wife did a husband's rapid promotion have? Or a marriage into a different socioeconomic or religious group? Or a move from a rural southern community to an urban northern part of the United States? 3. Does the patient view these changes as brought about by himself, by significant persons, or by fortuitous circumstances? What attitudes does he have about these changes? 4. Are the patient's roles in various social settings congruent with one another? For example, is there role conflict between value systems of the patient's early and adult social environments? Are there behavioral deficits due to the changes (eg, an inability to cope with new social demands, sexual standards, or affectional requirements, associated with rapid acquisition or loss of wealth, or geographic relocation)? If the roles are incongruent, is incongruence among these roles pertinent to his problem? Does the problematic behavior occur in all or only some of these different settings? 5. How can identified sociological factors in the problematic behavior be brought into relation with a treatment program? 14  C. Behavioral changes. 1. Prior to the time of referral did the patient's behavior show deviations in behavioral patterns compared with developmental and social norms? If so, what was the nature of changes in social behaviors, in routine self-care behaviors, in verbal statements toward self and others? Under what condition were these changes first noted? 2. Do identified biological, social, or sociological events in the patient's life seem relevant to these behavior changes? 3. Were these changes characterized by (a) emergence of new behaviors (b) change in intensity or frequency of established behaviors, or (c) nonoccurrence of previous behaviors? 4. Under what conditions and in which social settings were these behavioral changes first noted? Have they extended to other social settings since the problematic behavior was first noted? 5. Were the behavioral changes associated with the patient's exposure to significant individuals or groups from whom he learned new patterns of reinforcement and the behavior necessary to achieve them? Can the problematic behaviors be traced to a model in the patient's social environment from whom he has learned these responses? V. Analysis of self-control A. In what situations can the patient control those behaviors which are problematic? How does he achieve such control, by manipulation of self or others? B. Have any of the problematic behaviors been followed by aversive consequences by others, e.g., social retribution, jail, ostracism, probation, etc.? Have these consequences reduced the frequency of the problematic behavior or only the conditions under which it occurs? Have these events modified the patient's self-controlling behavior? C. Has the patient acquired some measure of self-control in avoiding situations which are conducive to the execution of his problematic behavior? Does he do this by avoidance or by substitution of alternate instrumental behaviors leading to similar satisfactions? D. Is there correspondence between the patient's verbalized degree of self-control and observations by others? Can the patient match his behavior and his intentions? E. What conditions, persons, or reinforcers tend to change his self controlling behavior (e.g., a child behaves acceptably at school but not at home, or vice versa)? F. To what extent can the patient's self-controlling behavior be used in the treatment program? Is constant supervision or drug administration necessary? VI. Analysis of social relationships A. Who are the most significant people in the patient's current environment? To which persons or groups is he most responsive? Who facilitates constructive behaviors? Who provokes antagonistic 15  or problematic behaviors? Can these relationships be categorized according to dimensions which clarify the patient's behavioral patterns (eg, does a patient respond in a submissive or hostile way to all older men)? B. In these relationships, by use of what reinforcers do the participants influence each other? For example, analysis may reveal a father who always calls out a delinquent son whose public punishment would be embarrassing to the father. Is the cessation of positive reinforcement or onset of punishment clearly signaled? C. What does the patient expect of these people in words and in action? On what does he base his verbal expectations? D. What do these people expect of the patient? Is there consistency between the patient's and others' expectations for him? E. How can the people who can influence the patient participate in treatment? VII. Analysis of the social-cultural-physical environment. A. What are the norms in the patient's social milieu for the behaviors about which there is a complaint? B. Are these norms similar in various environments in which the patient interacts, eg, home and school, friends and parents, work and social mille etc.? If not, what are the major differences in behaviors supported in one but not in other environments? C. What are the limitations in the patient's environment which reduce his opportunities for continued reinforcement; are social, intellectual, sexual, vocational, economic, religious, moral, or physical restrictions imposed by his environment? D. In which portion of the environment is the patient's problematic behavior most apparent, most troublesome, or most accepted? Can the congruence of several environments be increased or can the patient be helped by removal from dissonant environments? Does his milieu permit or discourage self-evaluation? E. Does his milieu regard psychological procedures as appropriate for helping him solve his problems? Is there support in his milieu for the changes in attitudes and values which successful psychotherapy may require? The preceding outline has as its purpose the definition of a patient's problem in a manner which suggests specific treatment operations and also suggests specific behaviors as targets for modification. It may also lead to the major conclusion that no such operations are possible. Therefore, the formulation is action oriented. It can be used as a guide for the initial collection of information, as a device for organizing available data, or as a design for treatment. 16  The formulation of a treatment plan follows from this type of analysis because knowledge of the reinforcing conditions suggests the motivational controls at the disposal of the clinician for the modification of the patient's behavior. The analysis of specific problem behaviors also provides a series of goals for psychotherapy or other treatment, and for the evaluation of treatment progress. Knowledge of the patient's biological, social, and cultural conditions should help to determine what resources can be used, and what limitations must be considered in a treatment plan. The various categories attempt to call attention to important variables affecting the patient's current behavior. Therefore, they aim to elicit descriptions of low-level abstraction. Answers to these specific questions are best phrased by describing classes of events reported by the patient or observed by others, or from critical incidents described by an informant. The analysis does not exclude description of the patient's habitual verbal-symbolic behaviors. However, in using verbal behaviors as the basis for this analysis, one should be cautious not to "explain" verbal processes in terms of postulated internal mechanisms without adequate supportive evidence, nor should inference be made about nonobserved processes or events without corroborative evidence. The analysis includes many items which are not known or not applicable for given patient. Lack of information on some items does not necessarily indicate incompleteness of the analysis. These lacks must be noted nevertheless because they often contribute to the better understanding of what the patient needs to learn to become an autonomous person. Just as important is an inventory of his existing socially effective behavioral repertoire which can be put in the service of any treatment procedure. This approach is not a substitute for assignment of the patient to traditional diagnostic categories. Such labeling may be desirable for statistical, administrative or research purposes. But the current analysis is intended to replace other diagnostic formulations purporting to serve as a basis for making decisions about specific therapeutic interventions. Methods of Data Collection for a Functional Analysis Traditional diagnostic approaches have utilized as the main sources of information the patient's verbal report, his nonverbal behavior during an interview, and his performance on psychological tests. These observations are sufficient if one regards behavior problems only as a property of the patient's particular pattern of associations or his personality structure. A mental disorder would be expected to reveal itself by stylistic characteristics in the patient's behavior repertoire. However, if one views behavior disorders as sets of response patterns which are learned under particular conditions and maintained by definable environmental and internal stimuli, an assessment of the patient's behavior output by the usual methods is insufficient unless it also describes the conditions under which it occurs. This view requires an expansion of the clinician's sources of observations to include the stimulation fields in which the patient lives, and the variations of patient behavior as a function of exposure to these various stimulational variables. Therefore, the resourceful clinician need not limit himself to test findings, interview observations in the clinician's office, or referral histories alone in the formulation of the specific case. Nor need he regard himself as hopelessly handicapped when the patient has little observational or communicative skill in verbally reconstructing his life experiences for the clinician. Regardless of the patient's communicative skills the data must consist of description of the 17 patient's behavior in relationship to varying environmental conditions. These comments are not intended to propose that the scope of clinical diagnosis be infinitely expanded. Rather, we suggest that the clinican make a selection from the numerous investigative avenues explicitly opened by the present conceptual model. For instance, a single home visit or telephone call can be far more economical than spending several sessions with one family member in an office. An interview permitting direct observation of the patient in interaction with a significant person in his life may be more fruitful than extended verbal description of this relationship by the patient. A survey of his current sociocultural milieu could yield more valuable clues for a therapy program than prolonged attention to his childhood experiences. A behavioral analysis excludes no data relating to a patient's past or present experiences as irrelevant. However, the relative merit of any information (e.g., growing up in a broken home or having had homosexual experiences) lies in its relation to the independent variables which can be identified as controlling the current behavior which requires modification. The observation that a patient has hallucinated on occasions may be important only if it has bearing on his present problem. If looked upon in isolation it may be misleading and result in emphasis on classification rather than treatment. In the psychiatric interview the model here described opposes acceptance of the content of the verbal self-report as equivalent to actual events or experiences. However, verbal reports provide information concerning the patient’s verbal construction of his environment and of his person, his recall of past experiences, and his fantasies about them. While these self-descriptions do not represent data about events which actually occur internally, they do represent current behaviors of the patient and indicate the verbal chains and repertoire which the patient has built up. Therefore, the verbal behavior may be useful for description of the way in which a patient construes his world. To make the most of the approach here described, variations on traditional interview procedures may be obtained by such techniques as role playing, discussion, and interpretation of current life events, or controlled free association. In addition to the use of the clinician's own person as a controlled stimulus object in interview situations, observations of interaction with significant others can be used for the analysis of variations in frequency of various behaviours as a function of the person with whom the patient interacts. For example, prescribed standard roles for nurses and attendants may be used, and members of the patient's family or his friends may be used to obtain data relevant to the patient's habitual interpersonal response pattern. Such observations are especially useful if in a later interview the patient is asked to describe, discuss, and report these observed sessions. Confrontations with sound or video tape recordings for comparisons between the patient's report and the actual session as witnessed by the observer may provide information about the patient's perception of himself and others as well as his habitual behavior toward peers, authority figures, and other significant people in his life. Except in working with children or family units, insufficient use has been made of material obtained from other informants in interviews about the patient. These reports can aid the observer to recognize behavioral domains which the patient's report deviates from or agrees with the descriptions provided by others. Such information is also useful for contrasting the patient's reports about his presumptive effects on another person to the stated effects by that person. If a patient's interpersonal 18 problems extend to areas in which social contact is not clearly defined, contributions by informants other than the patient are essential. It must be noted that verbal reports by other informants may be no more congruent with actual events than the patient's own reports and need also to be related to each informant's own credibility. If such crucial figures as parents, spouses, and employers can be interviewed, the clinician is also provided with some information about those people with whom the patient must interact repeatedly and with whom interpersonal problems may have developed. Some observation of the patient's daily work behavior represents an excellent source of information, if it can be made available. Observation of the patient by the clinician or his staff may be preferable to descriptions by peers or supervisors. Work observations are especially important for patients whose complaints include difficulties in their daily work activity or who describe work situations as contributing factors to their problem. While freer use of technique may be hampered by cultural attitudes toward psychiatric treatment in the marginally adjusted, such observations may be freely accessible in hospital situations or in sheltered work situations. With use of behavior-rating to guide treatment. This analysis is performed on the group and its environment rather than on individuals. Many recent innovations, combining treatments for individual problems with "social actions," have used this form of diagnosis. The current concern with mental health problems in the poor (eg Pearl & Riessman, 1964; Riessman, Cohen, & Pearl, 1964) with delinquency (Slack & Schwitzgebel, 1960), and similar social problems has introduced behavioral analysis methods to the diagnosis of general conditions and to the culturally prevalent contingencies for particular problems. From such analyses the clinician can draw inferences about the conditions maintaining individual problem behaviors, integrating both the societal and individual variables which would require change for therapeutic success. In this section we have mentioned only some of the numerous life situations, test instruments, and behavioral laboratory analogs which can be evaluated in order to provide information about the patient. Criteria for their use lie in economy, accessibility to the clinician, and relevance to the patient problem. While gathering data from a patient in the office may be more convenient, acquiring firsthand information about the actual conditions under which the patient lives and works may be far more valuable and economical. Such familiarity may be obtained either by utilization of informants or by the clinician's entry into the home, the job situation, or the social environment in which the patient lives. The view of psychotherapy as readjustment of the patient's social and physical environment and as learning and unlearning of self-attitudes and behavior patterns suggests full utilization of observational techniques similar to those employed in laboratory and field studies in the behavioral sciences. Wider use of laboratory methods derived from animal and learning experiments is also indicated for assessment of the performance of defined responses. The introduction of new and rigorous methods of behavior assessment and modification may have consequences for the scope and nature of the clinician's activities which are by no means clear at this time. Certainly, the emphasis on behavioral engineering approaches, aimed at changing specific response patterns rather than hypothetical disease processes or personality structures represents radical departure from the traditional attacks on behavior disorders and will require continuous reevaluation of all facets of clinical work and training. The present type of approach to behavioral analysis also reveal a similarity of the total clinical strategy to the domain of policy-process models, as described by Bauer (1966) for situations in which no single best solution is available. Clinical strategy seems to be closer to the social process of policy formation, in which constant consideration of 19 the interests of different parties is required and a solution can only be achieved by "negotiation” or compromise between the rationally most desirable and operationally most feasible alternatives, than to a rational process aimed toward restoring a patient to a predefined state of psychological health. Chapter Summary Currently used psychiatric classification systems have failed to provide a reliable method for categorization of psychiatric patients, or a scheme which permits prediction of response to psychiatric treatment as a result of assignment to diagnostic categories. In this chapter we have examined some of the problems encountered in the use of current diagnostic systems, and the shortcomings associated with them. Even with further technical refinement, important limitations on effective use of current diagnostic categories exist. These are associated with the divergence of complaints presented by psychiatric patients, the current lack of understanding of the genesis of psychiatric disorders, the haphazard selectivity of behavioral observations in arriving at a diagnostic label, and the lack of correlation between diagnostic categories and specific psychiatric treatments. Recent interest in the direct modification of problematic behaviors by application of learning principles has suggested a new approach to diagnosis, the use of behavioral analysis of individual case data. This procedure sacrifices the taxonomic features of the usual diagnostic enterprise but promises to make greater contributions to the formulation of specific treatment plans in individual cases. We have described a behavioral analytic approach which presumes that therapeutic intervention can be based on a comprehensive knowledge of two sets of variables which maintain problematic behaviors: those inferred from the patient's history and those in his current situation. We have indicated a set of specific guidelines which help the clinician obtain the type of information from assessment procedures on which subsequent recommendations for therapeutic interventions can be based. The guidelines suggest examination of the following areas in each case: 1. A detailed description of the particular behavioral excesses or deficits which represent the patient's complaints, and of the behavioral assets which may be available for utilization in a treatment program. 2. A clarification of the problem situation in which the variables are sought that maintain the patient's current problem behaviors, Attention is also given to the consequences of psychiatric intervention on the current adjustment balance of the patient in his social environment. 3. A motivational analysis which attempts to survey the various incentives and aversive conditions representing the dominant motivational factors in the patient. 4. A developmental analysis suggests consideration of biological, sociological, and behavioral changes in the patient's history which may have relevance for his present complaint and for a treatment program. 5. An analysis of self-control, which provides assessment of the degree of self-control exercised by the patient in his daily life. 20  Rating scales or other simple measurement devices, brief samples of patient behavior in work situations can be obtained by minimally trained observers. The patient himself may be asked to provide samples of his own behavior by making tape recordings of segments of interactions in his family, at work, or in other situations during his everyday life. A television monitoring system for the patient's behavior is an excellent technique from a theoretical viewpoint, but it is extremely cumbersome and expensive. Use of recordings for diagnostic and therapeutic purposes has been reported by some investigators (Bach, in Alexander, 1963: Cameron, Levy, Ban, & Rubenstein, 1964) Playback of the recordings and a recording of the patient's reactions to the playback can further be used in interviews to clarify the patient's behavior toward others and his reaction to himself as a social organism. Such feedback can also be used with video tape recordings. The availability of sophisticated electronic telemetering devices makes possible innovations in observational methods along dimensions barely considered by behavioral scientists until recently (Schwitzgebel, Schwitzgebel. Pahnke, & Hurd, 1964). Psychological tests represent problems to be solved under specified interactional conditions. Between the highly standardized intelligence tests and the unstructured and ambiguous projective tests lies a dimension of structure along which more and more responsibility for providing appropriate responses falls on the patient. By comparison with interview procedures most psychological tests provide a relatively greater standardization of stimulus conditions. In addition to the specific answers given on intelligence tests or on projective tests, these tests also provide a behavioral sample of the patient's reaction to a problem situation in a relatively stressful interpersonal setting. Therefore, not only can psychological tests provide quantitative test scores but they can also be treated as a miniature life experience, yielding information about the patient's interpersonal behavior and variations in his behavior as a function of the nature of the stimulus conditions. The advent of conditioning therapies has introduced new techniques for diagnostic analysis. In these methods, the specificity of therapeutic operations requires prior determination of the stimulus conditions and responses which are to be attacked in therapy. Initial explorations in desensitization therapy (Frank 1964, Rachman, 1983; Wolpe, 1958) complement the usual anamnesis with questions about specific fear-arousing situations, or with questionnaires designed to elicit this information (Wolpe & Lang, 1964), or with observations of the critical behavior in its natural setting (Paul, 1966). The construction of desensitization hierarchies for use as therapeutic implements illustrates well the use of diagnostic procedures with direct relevance for therapeutic intervention and demonstrates the continuing overlap between the diagnostic and treatment portions of the clinical enterprise. Among these novel techniques are standardized behavioral tests with quantifiable response measures ascertaining the strength of the "problematic" response on presentation of critical stimuli. In treatment of sexual deviations. for example, Freund (1963, 1965), and Solyom & Miller (1965) use pictures of nude males and females as the critical stimuli. The extent of response deviation is assessed by the magnitude of such responses as the galvanic skin response, the volume change of the male genital, or plethysmograph response to pictures of adults or children of either sex. A study by Hess, Seltzer. & Shlien (1965) lends to the validity of this approach. Clear-cut differences in pupil size changes were found for adult males of known homosexuality and heterosexuality. 21  Operant conditioning techniques have originated in the animal laboratory. Therefore, it is not surprising that their application to clinical problems has been accompanied by increased utilization of diagnostic methods based directly on long-established laboratory learning procedures. Conceptualization of the therapeutic intervention as a learning process results naturally in procedures which parallel the usual procedures of establishing base lines for at given response prior to learning. Most of the reports of operant conditioning treatments include some description of the pretherapeutic problem behavior, measured under specified conditions. Typically, reports present numerical or graphic records of frequency or intensity of a defined verbal or motoric response class, often obtained in the natural setting in which later behavior modification is attempted. Reports of behavior modification with autistic children in institutions (Ferster & DeMyer, 1962; Lovaas, Berberich, Perloff, & Schaeffer, 1966) or at home (Wolf, Risley, & Mees, 1964), with a hyperactive child in the classroom (Patterson, Jones, Whittier, & Wright, 1965), with a mother-child dyad in the home (Hawkins, Peterson, Schweid. & Bijou, 1966), with adult schizophrenics on a ward (Ayllon & Michael, 1950), with a patient with multiple tics (Barrett, 1962), with juvenile delinquents in a natural setting (Schwitzgebel, 1967), or with numerous other patients, differing in age, complaint, and setting, all are characterized by the use of direct pretherapy and posttherapy measures of the criterion response. Despite problems of statistical control associated with the use of this method in the single case, it represents an approach which may eventually lead to reliable diagnostic devices with the absolutely indispensable pertinence to both assessment and treatment. The outstanding advantage of a behavioral analysis lies in its inherent potential for verification of the effects of treatment by posttherapy reevaluation under the same conditions In treatment of groups with similar problematic behaviors, preparation of a treatment program often requires the prior analysis of the setting in which the behavior occurs. An initial step in diagnosis comprises observation of the response reinforcement which the environment provides, definition of the range of undesirable behaviors, and definition of the potential controls available. 22