...... wwm mes-30mm: ' A WY” 6% macaw AND wcomusmc f; A- rms mam‘mmm, ' , "1’. ";’ T37; “Inn for “I. Dogma 0E pk D 7 MICHICAN HATE UNWMITY ‘ ”Wm J Karl LIBRARY Michigan Stat: University “a. ‘5‘“ > 5'. (Tan _ THESIS This is to certify that the thesis entitled commmn PSYCHOTHERAPIES : A STUDY OF PROCESS AND OUTCOME USING A TIME SAMPLING TECHNIQUE presented by Norman J. Karl has been accepted towards fulfillment of the requirements for Pkg . degree in JaIcholngy I if ' , / 1; {9/1 (, : ll "ill/(x (-fll "s. If N" ’ .W'N". Major professor Date MM 0-169 Wm n-‘IH _ Mail’s ABSTRACT COMPLETED PSYCHOTHERAPIES: A STUDY OF PROCESS AND OUTCOME USING A TIME SAMPLING TECHNIQUE by Norman J. Karl The present study investigates the form of the pro— cess observed for therapy clients as it relates to aspects of ”Aggression Conflict." And, it examines the process of therapy as it changes during the course of a single thera— peutic interview. Three groups of §s were identified on the basis of their responses to a Semantic Differential questionnaire. ”Aggression Conflict" represents the condition in which an § manifested positive Active—Potent hostility scores, but nega— tively evaluated such feelings no such conflict (NAC), gs who or behavior. gs exhibiting resolved such conflict over the course of therapy (ACR), and §s who achieved no such resolution with therapy (ACU), This study attempts to for those §s who resolve their pared to those who do not. An comprised the sample. predict the process involved ”Aggression Conflict,” as com— attempt is also made to characterize the differences observed in content material for Norman J. Karl _§s initially identified as belonging to the NAC or AC categories. In addition, a major modification in sampling tech- niques is introduced, and an effort is made to confirm the findings of other authors with the present methodolOgy. Ten— minute time segments are randomly selected for every inter— view of each subject included in the study. It was hypothesized that those_§s resolving their "Aggression Conflict" should find greater approach, on the part of the therapist to hostility content, than do the ACU _§s. Further hypotheses predict the form of the roles as— sumed by the therapist over the course of therapy, and also attempt to differentiate the three client groups on the basis of variations in the role patterns which they manifest during therapy. Attention is also given to changes in the form of the types of verbal behavior engaged in by therapists as the therapeutic process progresses. A content analysis is made of ten—minute segments se— lected randomly from 330 tape—recorded psychotherapy inter- views at a university counseling center. These interviews represent the total sample available for 25 self-referred fe- male clients, who were categorized into NAC, ACU, or ACR groups. The results were generally negative. NOne of the tests which employed the "Aggression Conflict" scale as a Norman J. Karl means of identifying groups were statistically significant. It seems that this outcome measure does not represent an ade— quate index of the manifest content under investigation. Interestingly enough, however, the time sampling pro— cedures employed in the present study suggest a marked tendency for certain segments of therapy to be characterized more by some content categories than others. This finding has major implications for therapy research, in that sampling might best be performed in future research on that segment of the interview which is known to include more of the particu— lar content under investigation. A significant decrease is found in hostility content over the first, middle, and last sessions. There is some evidence to indicate that it is not al- ways a warranted procedure to assume that first, middle, and last points in the therapies of short-term clients have the same respective points for long-term clients. While this is the case for some variables, it is not for others. Pilot data must establish the legitimacy of this assumption for the particular variables under investigation. The possibility of predicting the differences in content between the two groups is raised as a major goal for future research. Contradictory findings observed between this and other investigations argue against the legitimacy of Norman J. Karl generalizing to larger populations on the basis of results obtained on a single sample. COMPLETED PSYCHOTHERAPIES: A STUDY OF PROCESS AND OUTCOME USING A TIME SAMPLING TECHNIQUE BY NOrman J. Karl A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1967 TO ELLY ii ACKNOWLEDGMENTS The author wishes to express his deep appreciation to Dr. Norman Abeles, thesis committee chairman, for his many suggestions, criticisms, and continual interest throughout this entire project. Dr. Abeles' ideas were frequently in— corporated into the present design. His generous availa— bility for consultation was an invaluable asset. A special expression of indebtedness is due Dr. William J. Mueller for making the tape library available, and for development of the Aggression Conflict scale employed in the present study. Dr. Mueller"s suggestions also proved helpful in performing the statistical analyses. A word of gratitude is also extended to Dr. Dozier Thornton and Dr. Gary Stollak for their constructive efforts. The writer is extremely grateful to Mr. Arthur Dale Shears for his collaboration in the present investigation. His efforts as secondary judge were arduous, and helped es- tablish both manuals as reliable instruments. To the author's wife, an expression of extreme warmth and appreciation for her support in those thousands of in— stances during the graduate program when it seemed that the next hurdle was unattainable. iii TABLE OF CONTENTS Page DEDICATION . . . . . . . . . . . . . . . . . . . . . . ii ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . iii LIST OF TABLES . . . . . . . . . . . . . . . . . . . . vi LIST OF FIGURES . . . . . . . . . . . . . . . . . . . vii LIST OF APPENDICES . . . . . . . . . . . . . . . . . . viii Chapter I. INTRODUCTION . . . . . . . . . . . . . . . . . l Therapist Variables . . . . . . . . . . . 2 Learning Theory . . . . . . . . . . 5 Verbal Conditioning Studies . . . . . . . 9 Learning of Roles in Therapy . . . . . . . 13 Content— Analysis . . . . . . . . . 15 Purpose of the Present Study . . . . . . . 18 Description of Terms . . . . . . . . . . 22 II. STATEMENT OF HYPOTHESES . . . . . . . . . . . 25 Therapist Approach-Avoidance . . . . . . . 25 Aggression Conflict . . . . . . . . . . . 26 Role Statements . . . . . . . . . . . . . 27 Therapist Propositions . . . . . . . . . . 30 III. METHOD . . . . . . . . . . . . . . . . . . . . 32 Subjects . . . . . . . . . . . . . . . . . 32 Therapists . . . . . . . . . . . . . . . 33 Sampling Procedure . . . . . . . . . . . . 35 Coding Procedure . . . . . . . . . . . . . 35 Scoring Reliability . . . . . . . . . . . 37 Semantic Differential . . . . . . . . . . 40 iv Chapter IV. RESULTS Total Frequencies Hypotheses Tape Segment Sampled Further Considerations Phase of Treatment Aggression Conflict Personality Data Duration of Therapy V. DISCUSSION Inter-Judge Reliability . Therapist Approach and Avoidance Aggression Conflict Role Systems Therapist Propositions Tape Segment Sampled Phase of Treatment Duration of Therapy Personality Data Implications for Further Research VI. SUMMARY REFERENCES APPENDICES Page 42 42 42 53 55 55 59 6O 61 66 66 67 67 69 72 73 75 76 77 78 8O 83 87 Table 10. LIST OF TABLES Summary of client and therapist characteristics . Duration of therapy and therapist experience Inter—judge reliability coefficients and the standard scores and p values associated with the differences between a pair of correlations . . . Total number of units and relative pro— portions of each measure found in the total sample . . . . Kruskal—Wallis H values obtained in com— paring ACR, ACU, and NAC_§s Total frequencies for measures grouped by tape segment sampled . . . . F and p values relating to the changes in each measure over the course of therapy - first, middle, and last interviews for all §s . . . . . . . . . . . . . . . . . . . F and p values relating to the changes in each measure over the course of therapy — first six interviews for all §s . . X2 and p values comparing first six inter— views for Long-term and Short—term §s X2 and p values comparing first, middle and last interviews for Long—term and Short— term §s . . . . . . . . . . . . . Page 34 35 39 43 45 54 56 58 62 63 LIST OF FIGURES Figure 1. Frequency of client and therapist primary role statements over the course of therapy 2. Frequency of client and therapist secondary role statements over the course of therapy 3. Frequency of therapist propositions over the course of therapy . vii Page 48 50 52 LIST OF APPENDICES Appendix Page A Scoring Manuals . . . . . . . . . . . . . . . 87 B Mann—Whitney U values comparing two groups of-Ss AC—NAC_§s . . . . . . . . . . . . . . . . 103 ACR—NAC_§s . . . . . . . . . . . . . . . . 104 ACU—NAC_§s . . . . . . . . . . . . . . . . 105 ACU—ACR_§s . . . . . . . . . . . . . . . . 106 C Mann—Whitney U values comparing AC and NAC _§s on first sessions for certain measures . 107 D Kruskal—Wallis H values comparing ACR, NAC, and ACU §s on last sessions for certain measures . . . . . . . . . . . . . . . . . . 108 E Analyses of variance for response cate— gories 2 first six interviews . . . . . . . 109 F Analyses of variance for response cate- gories — first, middle, and last interviews . . . . . . . . . . . . . . . . . 110 viii CHAPTER I INTRODUCTION The pursuit of psychotherapy as a scientific disci— pline is little more than 75 years old. Controlled research in this area is of much more recent vintage yet. The methods of psychological investigation were not vigorously applied to the process of psychotherapy until around 1940, when interest emerged as a function of the work of Carl Rogers and his students (Strupp, 1962a). The burgeoning literature which is presently con— cerned with psychotherapy research has its antecedents in the development of new theories, values, experimental attitudes, equipment, and methods. The introduction of the sound re— cording of the interview, the development of methods of content-analysis, and the appearance of psychological theories such as learning theory, enabled the development of many new possibilities for psychotherapy research (Auld and Murray, 1955). In addition, inquiry arose as a direct effect of Eysenck's contention that psychotherapy might not be at all effective, and out of a need to decide which processes of psychotherapy Ci.e., psychoanalytic, Rogerian, Sullivanian, etc.), warranted or lent themselves to investigation (Frank, 1961). Even today there exists considerable doubt as to whether the efficacy of the psychotherapeutic process has been scientifically demonstrated. Cross (1964), suggests that two areas of investigation hold the greatest promise in the attempt to comment on the above criticism. He advocates the continued development and refinement of process research, and of further outcome research which emphasizes careful measurement and the elimination of spontaneous remission ef— fects in the design. Some authors (cf. Winder, 1957), feel that the only way in which precision in psychotherapy will be achieved is through the study of ongoing process. Winder (1957), views outcome as representing an arbitrarily chosen point in process. The present study is geared toward shedding more light on the process variables in therapy. Therapist Variables Freud (1959), originally developed the concept of counter—transference to describe the phenomenon in which difficulties in the therapeutic process emanated from person— al qualities inherent in the therapist. Glover's survey of British psychoanalysts in the 1930's similarly supported the notion of the importance of the therapist as a source of variance in the therapeutic out— come. He found that the therapeutic practice of individuals who had all been trained in orthodox psychoanalysis varied considerably. It seems inevitable that, whether one is -3.- interested in the problem of outcome or process in psycho- therapy, one must focus upon what it is that the therapist actually does during the therapeutic hour (Strupp, 1962b). At present, the literature abounds with theoretical considerations pertaining to the role of the therapist. Personality characteristics (Strupp, 1958; Mills and Abeles, 1965); level of experience, theoretical orientation and atti— tudes towards the patient (Strupp, 1960); and physical and be— havioral qualities (Binder et a1., 1957; Goldstein, 1962), have all been cited as important therapist variables which enter into the nature of the relationship between client and therapist in such a manner that therapeutic endeavors are substantially enhanced or hindered. The research evidence lends clear support to the sig— nificant impact of the person of the therapist. Strupp's (1958), investigations of the therapeutic process seem to indicate that maximal effectiveness is achieved in therapy where the therapist can relate to the client in a warm, em- phatic manner, and to demonstrate to the client that the self-defeating qualities he exhibited in previous patterns of interpersonal relations were no longer appropriate. Barnes (1963), found that therapists have conflict in a particular area (e.g., hostility, dependency, sex), re- sponded to their clients conflictual statements in the same area with relatively greater avoidance reactions than non— conflicted therapists. The conflicted therapists also tended to approach conflictual material less frequently as therapy progressed. Cutler (1958), similarly found a systematic re- lationship between the therapist's conflicts and his tendency to over— or under—report the occurrence of the same behavior in his clients, and also found that the therapist handled less adequately client material which was conflict-relevant for himself. This latter finding appears to be of great im- port, for, as Kell and Mueller (1966), emphasize, client change may be dependent on the therapist's ability to explore the client's conflict-associated affect. The activity versus passivity characteristics of the therapist has also come under scrutiny. Murray's (1956) re— sults indicate that the rate of increase of client statements about the therapist is clearly related to the extent to which the therapist responds to material in an active rather than a passive manner. It appears, then, that there is a plethora of evi— dence which identifies the therapist as a major variable in the therapeutic process. Sarason (1954), has been extremely emphatic in his position that the complete investigation of a clinical or experimental situation must give consideration to the ways in which the social and personal characteristics of the interviewer or experimenter influence the behavior of the client or subject. Luborsky and Strupp (1962), in summarizing current research problems in psychotherapy, pro- pose the basic question to be: How does the therapist influence the patient? They suggest that it is crucial to know how patients interpret therapeutic maneuvers, and whether the activities of the therapist can be reduced to roles and operations which are shared with other members of the culture. This trend toward viewing the therapist's role as more flexible, magnifies the importance of utilizing pro— cess analysis as a means of focusing increasingly on the therapist—client interaction (Winder, 1957). Learning Theory The learning theory model has come into continued favor in research on the psychotherapy process. It becomes increasingly apparent in such research that the therapist is often serving the role of elicitor, rewarder, and punisher of the client's behavior and effects. Bandura and Walters (1963) express doubt that the "schools‘' which incorporate as basic tenets the uncondition- ally accepting attitudes of the therapist toward the client actually achieve their goal. They propose that, instead, the therapist expresses approval and disapproval in many subtle ways during the treatment process. These authors fail to mention, however, that psychotherapists employing a learning theory model may be as guilty as client—centered therapists in the inadvertent expression of subtle negative and positive sanctions. It seems that mild forms of approval and disap— proval, avoidance, initiations, and topic transitions are characteristically used by most therapists to channel client's responses. They have proved to be useful categories for the analysis of therapist—patient interaction in controlled re- search studies (Bandura and Walters, 1963). Bandura (1961), further maintains that most of the conventional forms of therapy employ extinction, and primary and secondary rewards. Extinction effects are most likely operative when the client expresses thoughts or feelings that provoke anxiety or guilt and the therapist does not withdraw, disapprove, or criticize. Similarly, once the client has come to want the interest and approval of the therapist, pri- mary and secondary rewards may play an indispensable role in the treatment process. As a matter of fact, the possibility that the therapist reinforces the client's behavior is such a strong one, that London (1964) warns against the possibility of the therapist inadvertently teaching his own moral, ethi— cal or social value system. It appears that some of the modification of behavior which occurs as a result of the client—therapist interaction is due to the unwitting application of reinforcement pro— cedures by the therapist. The therapist may employ cues to modify, control, guide or manipulate the client's behavior without being aware of it (Krasner, 1958). Bandura (1961), Bandura and Walters (1963), and Krasner (1962), all point to the same conclusion: the therapist should be aware of his own behavior so that he can set out with deliberation to modify behavior. In terms of the learning theory paradigm, this would suggest that a more deliberate application of our understanding of the process of learning to psychotherapy would increase the effectiveness and efficiency of the results. If our conception of psychotherapy is one of a pro— cess which, in part, involves reinforcement procedures, then one area which becomes of interest for study is the patient response class and the immediately consequent reinforcing cue behavior on the part of the therapist. Many investi- gators have shown that such interactions are lawful and pre— dictable. Studies of verbal conditioning conducted in the laboratory have shown that a great variety of response classes which appear to be relevant to psychotherapy are in- fluenceable under the proper reinforcement conditions (Krasner, 1962). Kopplin (1965) has shown that the therapist's re— sponse is more influential than the client's own preceding statement as a stimulus for provoking inner feelings. Simi— larly, the same investigator reports that clients tend more often than not to continue discussion of therapist—introduced topics in their subsequent statement. Frank and Sweetland (1962), employing a variety of response categories, also at— test to the selective influence of the therapist's statements on client's responses. They conclude that client statements appear to be quite amenable to therapeutic modification. Many investigations of psychotherapy process have evolved as a result of the content—analysis methods developed by Bandura, Lipsher, and Miller (1960). These authors at- tempted to determine the specific mechanisms involved in the effects of therapist's anxieties on their psychotherapeutic work. As a result of their research efforts, the notion of the therapists' approach or avoidance to client response cate— gories has appeared as a prominent technique in several subse- quent research endeavors. Bandura et a1. (1960), confirmed their prediction that approach responses were more effective in getting clients to express their hostility than were avoidance responses. Winder et a1. (1962), confirmed these findings for the ex- pression of client hostility, and further found that the same general function applied to the expression of client de- pendency statements. Using the Bandura system of content— analysis, with greater or lesser degrees of modification, several investigators (Kopplin, 1963; Kopplin, 1965; Caracena, 1963; Lerman, 1963; Varble, 1964), have confirmed these ini- tial findings. It seems that where a therapist positively reinforces a client's verbal response class by approaching these expressions, there is a high probability of continu— ation of responses in this class. Conversely, where negative reinforcement in the form of avoidance obtains, there is a substantial decrease of responses in that class. Furthermore, this statement seems to be applicable to response classes other than dependency and hostility as well. The same gener- al effects are found for the expression of affiliation, self- assertion, sex, and vague affect (Kopplin, 1965), and for in— dependence and intellectual defenses (Murray, 1956). The above findings have major implications not only for changing behavior in therapy, but can also provide a means through which the client can be retained in treatment in the early phases of therapy. One method may be the ini— tial reinforcement of dependency by approaching statements in this response class. Skinner (1953), Dollard and Miller (1950), and Ferster (1958), all tend to advocate that the therapist induce the client to express his existing dependency needs more fully in the initial psychotherapeutic stages, theorizing that this procedure would assure continuation in treatment. Verbal Conditioning Studies It was primarily due to the efforts of Skinner that verbal behavior came under experimental scrutiny. Since that time, many laboratory studies on verbal conditioning have provided the clinician with aids in dealing with the verbal behavior of clients. Basic to Skinner's approach, and to that of subsequent studies, is the acceptance of the verbal behavior as representing a response in and of itself, and the arbitrary division of this behavior into response classes for purposes of investigation (Greenspoon, 1962). 10 As Kanfer (1964) points out, studies pertaining to therapeutic interviews have typically been confined to naturalistic observations or to partial standardization procedures by controlling therapist's temporal or content patterns of speech. The verbal conditioning experiment, on the other hand, may be viewed as a miniature representation of the interpersonal therapeutic situation, but with a greater degree of control of major variables. Ullmann et a1. (1961), regard the verbal conditioning situation as one in which means are provided for the reduction and isolation of interpersonal variables, resulting in greater power for their systematic study. The research on verbal conditioning lends support to the results of the process analyses of therapy cited above—- it seems clear that verbal behavior is modifiable. The major problem is whether, in fact, the verbal conditioning which occurs in psychotherapy generalizes to appropriate verbal be— havior in other settings (Greenspoon, 1962). It seems that the latter question will vex psychotherapists at all times, no matter what their orientation, unless the type of research is done which actually has objective answers to this question—— answers which are derived from situations external to the therapeutic one. The following presentation of research evidence represents but a small sample of verbal conditioning studies which tend to indicate the effectiveness of the reinforcement ll procedure. Waskow (1962), employed a procedure in which she differentially reinforced three experimental groups for the expression of feelings, a combination of feeling and content, and content alone. She found a significant difference in the predicted direction in the amount of feeling and/or content expression. Ullmann et a1. (1961), similarly provide evi- dence for the contention that changes in criterion measures ascribed to psychotherapy can be demonstrated to be associ— ated with specific behaviors on the part of the therapist (or experimenter). They report significant gains in interperson— al adequacy in group therapy for those individuals who re- ceived positive-personal reinforcement. Kanfer (1964), demonstrated the influence on listening versus speaking be— havior of a varying ratio of reinforcement, while Buss and Durkee (1958), describe the ease with which intensely hostile verbalizations could be conditioned in experimental subjects. Binder et a1 (1957), found that the mere use of the word "good,” served as a reinforcing agent for a response class of behavior which consisted of the use of hostile words in sentences. Breger and McGaugh (1965), criticize the proposition that studies of verbal conditioning represent adequate ana— logues to psychotherapy. They suggest that there is not suf- ficient understanding of the verbal conditioning process to permit it to be legitimately utilized as a basic explanation of the much more complex phenomenon of psychotherapy. 12 It would similarly be naive to claim that a concept of reinforcement can sufficiently account for much of the complex process occurring during psychotherapy. In the first place, the phenomenon of reinforcement is itself little understood. Experimental investigation provides strong evi- dence that the law of effect is a rather weak law of per— formance. Some question also arises as to whether reward is a necessary condition for learning at all. And, it has been a major, if not impossible, task to define a reinforcing stimulus independently of the particular learning situation in which it occurs (Breger and McGaugh, 1965). Breger and McGaugh (1965) further contend that psy— chologists of most persuasions would accept the concept of reinforcement in its delimited sense, i.e., as it applies to the particular situation. These authors maintain that view— ing reinforcement in this manner relegates the concept to a trivial one, for it then merely suggests that certain events, when presented, tend to increase the probability of the oc- currence of the response which they have succeeded. What needs to be done in order to make the concept of reinforce— ment a meaningful one, is to specify the commonalities of various reinforcing events. More specific to psychotherapy, the issues which need to be clarified pertain to which thera— pist behaviors are rewarding, in what manner they achieve this function, for what kinds of clients, and with what ef— fects. The present study is designed to shed some light on these latter issues. 13 Learning of Roles in Therapy There is some evidence to suggest that, in part, what the client learns in therapy is a ”patient role.” Since the present study attempts to investigate some of the role features of the therapeutic process, a brief introduction to this viewpoint is presented. Lennard and Bernstein (1960, 1967) have been among the foremost proponents of this type of conceptualization of the therapist-client relationship. They indicate that the patient must be ”inducted" into the ”patient role" before a therapeutic system can begin its operation. In other words, the client must come to acquire, relatively early in therapy, a set of role expectations which are reciprocal to the thera- pist's. They further argue that the extent to which the patient can transfer the role patterns learned in therapy to other relationships is the extent to which therapy has achieved its goal of helping the client learn the "learning of roles." The therapeutic encounter thus permits the patient to learn not only a set of expectations appropriate to the therapist, but also the principles underlying the learning of role expectancies in general. Once learned in therapy, these principles can be subsequently transferred to extra- therapy conditions. ”For example, having learned in therapy that he must accept responsibility for activity and partici— pation, he is now sensitized to expectations with respect to 14 activity and participation in other relationships.” (Lennard and Bernstein, 1960, pp. 29—30.) Although it might seem that the therapist is imposing his system of values on the client, research indicates that the client is typically equally concerned with clarifying and defining his role in relation to the therapist. Lennard and Bernstein (1960), report that both therapist and client de- vote about an equal amount of time in attempting to define the roles in their relationship. ”Teaching” the client his role is, therefore, a cru- cial part of the therapeutic process, and may even determine whether treatment will continue. Kelly has observed that a patient may conceptualize his approaching therapeutic contact in many ways, and that this initial set of assumptions will effect his behavior during therapy, and especially in early interviews. If incompatible role expectations obtain, there will be a deleterious effect on the therapeutic relationship. Some therapists advocate the detailed and deliberate dis- cussion of therapist and client role expectations early in therapy (Goldstein, 1962). There is considerable discussion of the role system early in treatment, and it continues to disturb the sequence of development throughout therapy if the therapist has not concentrated enough of his effort on role induction (Lennard and Bernstein, 1967). 15 Content-Analysis Since Lasswell devised a system, in 1938, for cate— gorizing what the client said during the psychoanalytic interview (Frank, 1961), a myriad of content—analysis systems have appeared in the literature. It appears that whether one supports or castigates this procedure depends more on whether he considers himself a researcher or a critic. Although there are, to be sure, many shortcomings associated with the content—analysis method, it seems to this author that the merits of the procedure far outweigh the disadvantages. As Holt (1965) indicates, the method is systematic, orderly and comprehensive, and can be demon— strated to be objective in that different persons can learn to reliably perform this operation. In addition, the content- analysis procedure permits the investigator to refrain from interfering with the natural flow of behavior, which would be the case if he used a questionnaire or carefully controlled experimental variables (Dollard and Auld, 1959). Critics of the method level serious charges against content-analysis techniques, many of which seem justified. Harway and Iker (1964), point out that almost all content— analysis systems depend on a priori classifications. They suggest that an alternative means of investigation of pro— cess would be the factor analysis of verbatim transcripts, and rotation of the corresponding correlation matrix by 16 means of computer programs. This method is seen by some as the procedure of choice for the measurement of significant regularities and structure in the verbal content of inter— views (Starkweather and Decker, 1964). Computer programming is also suggested to overcome the unreliability of judges as measuring instruments, the relatively small amount of data scorers can handle, and the need to re—establish reliability with each classification system or shift in sampling (Harway and Iker, 1964). A crucial considerabion in content—analysis studies arises when an isomorphism is assumed to exist between quantitative indices derived from verbal symbols, and under— lying psychic processes (Strupp, 1962a). It must be kept in mind at all times that the method represents a way of ob— taining measures on the manifest content of the therapeutic interaction, and is, therefore, limited to the conscious part of therapy (Murray, 1956). Sargent (1961) proposes that if changes in psychotherapy are taken as manifesting them- selves on different levels (i.e., behavioral, subjective, or intrapsychic), then intralevel validation must be based upon measures which are appropriate to the particular level under investigation. Another general criticism of content—analysis re- search revolves about the fact that the procedure is, of necessity, a limited one. It does not capture the naunces of non—verbal cues such as facial expressions, posture, voice l7 tones, etc. It is highly possible that these non—verbal cues influence the therapy process even though verbal be— havior is the point of focus (Kanfer, 1964; Varble, 1964). Limited though the approach may be, it seems self- evident that great emphasis is in fact placed upon verbal be— havior in the therapeutic relationship. The primary in— fluence which the therapist exerts on the client is transmitted through communications. And, although the rela— tive effects of verbal as opposed to non—verbal cues have not yet been determined, one cannot deny the importance of the verbal aspects (Barnes, 1963). Lastly, the question of the validity of content- analysis is still a cause of major concern. There is a great distinction between talking about one's hostility or de- pendency, and ”being hostile” or "being dependent” in the outside world. Research which establishes such validity re— mains yet to be done. In the final analysis, Auld and Murray (1955) present one of the best rebuttals to the critics of content—analysis procedures. They regard it as unfair to expect any single content—analysis system to adequately describe the multiplex behaviors which occur from moment to moment in psychotherapy. Instead, if various systems dealt with certain aspects of this complex situation adequately, and were reliable and valid for their limited purposes, then, they maintain, we would begin to get the large number of measures which a 18 truly adequate descriptive and causal analysis of the psycho- therapy process would require. Purpose of the Present Study The present study is designed to answer some of the general challenges which appear in the literature on psycho— therapy research, to see if the findings of previous experi— menters can be confirmed with the application of a modified methodology, and to determine if process can be predicted on the basis of outcome measures obtained prior to and during therapy. The specific primary goals of this investigation are further elaborated in this section, in the Statement of Hypotheses section, and in the Methodology section. Rogers (1963) contends that one major implication which can be derived from the great diversity in present—day psychotherapeutic techniques and research, is the necessity for movement in the direction of looking at the facts. To look at the facts demands research—~it is to this goal that the present study is dedicated in a general sense. Frank (1961) similarly points to the inevitable pro- gression toward research in psychotherapy. On the one hand, part of the demand characteristics for such research are societal in nature, for as more people become aware of and seek therapeutic intervention as an avenue of help, the greater is the requirement for adequate, efficient, and ef- fective methods. On the other hand, if psychotherapy is to 19 be recognized as a scientific endeavor, then therapists must as scientists, develop techniques which will enable psycho— therapy to become an increasingly reliable and valid process. Reliability, validity, and general scientific knowledge are rarely derived from one piece of "break—through” research. MOre typically, experiments provide valuable leads to future research projects (Holt, 1965). As Marsden (1965) has pointed out, however, most research in content—analysis is not replicated. Instead, initial trusts are made, sys- tems are developed, and then lie buried in the literature. The system developed by Bandura et a1. (1960), and refined by Winder et a1. (1962), however, has generated a great deal of research. The present study employs the Bandura—Winder content- analysis system, but introduces a major modification in sampling methods normally described in the literature. And, attempts are made to replicate, with this modified sampling procedure, some of the results obtained with the Bandura— Winder system, and some of the findings reported when using the Lennard and Bernstein (1960) system. The present study focuses upon Lennard and Bernstein's Primary and Secondary Role System categories, and also utilizes their Evaluative, Descriptive, and Prescriptive modes of therapist interaction response categories. Both Bandura's and Lennard and Bern— stein's content systems are described more fully in later sections. 20 The author's basic tenet in embarking upon this re- search endeavor is the belief in the legitimacy of working with a small number of clearly operationalized variables. It s felt that a careful analysis of certain client response categories (i.e., hostility and dependency), of the thera— pist's interactional influence on these categories, and of the nature of the processes involved in change (as derived from an outcome measure), will lend more impetus to future research in psychotherapy, than will a study dealing with many variables whose definitions are ill defined, and whose associations with independent variables are, at best, vague. Varble (1964), in referring to the studies which have emanated from the Bandura—Winder system (Caracena, 1963; Barnes, 1963; Lerman, 1963; Kopplin, 1963, 1965), points to one major limitation which they share, i.e., the use of only one or two interviews of each client. In the original study conducted by Bandura et_al. (1960), only two—thirds of the total amount of therapy tapes were selected for investi— gation. Varble (1964) attempted to correct this shortcoming by sampling the lst, 25th, 50th, 75th, and 100th percentiles of the total number of interviews for each client. The present study attempts to replicate the findings of other ex- perimenters while sampling randomly from each interview of every client. To the author's knowledge, this procedure has not been attempted in psychotherapy research, and is seen to represent a major methodological consideration. One primary 21 question here, then, is to determine whether the form of the function(s) found in early interviews, or in interviews se- lected over the course of therapy, holds true for random samples taken from each session of the psychotherapeutic process. Some might argue that sampling from each therapy session also has limitations in that the entire process of treatment is not actually being examined. To this criticism, Sargent's (1961) statement seems applicable. She proposes that the notion that full hour by hour data are necessary to (\J support inferences regarding pervasive invariance is in it- self an assumption. Besides the goals of replication and the confirmation .7 of the invariance of the therapist's influence on the client in spite of the sampling technique used, the present study attempts to identify the process of change as it relates to an outcome measure. Strupp (1962b) contends that since the broad aim of research in psychotherapy is to identify the con— ditions which facilitate favorable client outcome, the out— come measures provide the essential, logical criterion for the evaluation of the conditions which are or are not ef- fective in the therapist's efforts. To date, the literature on psychotherapy research reports a paucity of studies which attempt to relate process variables to outcome measures, and to indicate in what manner the process differs for different outcomes. The present investigation attempts to rectify, in part, this flaw. 22 The outcome measure, "Aggression Conflict,” chosen for this study was developed by Mueller and Grater (1965, 1966), and its characteristics are more fully described in subsequent sections. Varble (1964), supports the position that, since psychotherapy attempts to help clients resolve emotional conflicts, it would seem that the manner in which the process does so would represent a meaningful approach. Since the ”Aggression Conflict" method was applied before and during psychotherapy with the §S included in the present sample, one can apply a content-analysis system to identify the means through which conflicts over aggression are resolved. Employing an outcome measure also serves to ease an- other source of criticism frequently leveled at content- analysis studies. Frank (1961) sees as a major drawback to research in psychotherapy process the element of inadequate controls—-there is great difficulty in selecting subjects that are at all comparable. With the ”Aggression Conflict" measures taken on each subject before, during, and at the termination of treatment, there is some measure of an own- control element introduced (Cross, 1964). Description of Terms Prior to presenting a statement of predictions to be made on the basis of this analysis, it seems advisable to de- scribe a bit more fully some of the terms which will be em— ployed in the following sections. 23 Aggression Conflict. Refers to the concept developed by Mueller and Grater (1965). _§s were considered to be ex— periencing aggression conflict if their scores on the Seman- tic Differential expressed negative evaluation of aggression but simultaneously regarded it as strong and active. The factor analytic derivation of this measure, and follow—up data on reliability, is provided in the Methodology section. Role Systems. Refer to Lennard and Bernstein's (1960) categories which are concerned with client or thera— pist content relating to their roles and expectations with regard to each other. The role systems of interest in the present investigation are: 1) Primary Systems——in which client or therapist statements refer to their respective roles during the process of therapy, and to the purposes, goals, and achievements of treatment; 2) Secondary Systems—— included here are statements relating to the client and therapist in a manner other than their primary roles as client and therapist. What's involved in this category centers about references to the nature of the interpersonal relation— ship between the client and therapist. Transference phenomena are classified as secondary systems. Therapist Propositions. While more complete de— scriptions of Bandura's and Lennard and Bernstein's scoring systems are provided in the Methodology section and in the Appendices, a brief description of Lennard and Bernstein's 24 "Therapist Proposition" classifications is felt to be help— ful as an orientation to the subsequent discussion of hypotheses. Lennard and Bernstein (1960), collapsed six of Bales' Interaction Process categories into three scales, in an at— tempt to describe the pragmatic significance of an act in re- lation to either the preceding or subsequent acts. These three categories are employed in the present study in order to identify more closely the types of verbal behavior en- gaged in by therapists. Descriptive propositions seek or give information. They similarly ask for or offer clarification, reiteration, or orientation. Evaluative propositions seek or convey appraisal or statements containing value judgments. In addition, they inquire about or offer opinion, expression of affect, or analysis. Prescriptive propositions seek or convey directives, including giving or asking for suggestions. CHAPTER II STATEMENT OF HYPOTHESES Therapist Approach-Avoidance Hypothesis I predicts that therapist approach to hostile or dependent content should have the function of in— creasing such discussion in the immediately following state— ment during the early phases of therapy, while having a lesser effect as the therapy process continues. It is felt that as the clients get a chance to talk about their hostile and dependent feelings, and have them accepted for perhaps the first time, they will have decreasing need to do so over the course of therapy. Conversely, it is predicted that therapist avoidance tendencies to these prOpositions will have the effect of inhibiting continuation in a given con— tent area in the immediately subsequent statement, but will concomitantly result in a lack of decrease (and possible in— tensification) of the client's need to discuss hostile or dependent material over the process of therapy. The essential question here is whether the results of Bandura et a1. (1960), and others who have replicated their findings, will obtain with use of a different sampling technique. 25 26 Aggression Conflict Hypothesis II predicts that the frequency of material relating to aggressive content should be greater during the initial stages of therapy for the ”Aggression Conflict” (here- after referred to as AC) group, than the ”Non—Aggression Con— flict" (hereafter referred to as NAC) group. It is felt that the conflict manifested in the Semantic Differential will, of necessity, intrude most frequently into the verbal content of therapy for the AC gs. Hypothesis III predicts that the frequency of de— pendent material will also be greater for the AC §s during early therapy sessions. .Ss experiencing "Aggression Con— flict” have been described by Mueller and Grater (1965) as being more anxious and manifesting lower ego strength than I§s who do not experience this conflict. Thus, it would seem that the person conflicted about aggression would also have a meager self-representation, and would, therefore, be de— pendent on others for guidance and feedback. Hypothesis IV predicts that the frequency of both statements pertaining to aggressive and dependent content should be lower, towards the final phases of therapy for both the NAC group and the group of §S who have resolved their ”Aggression Conflict” (hereafter referred to as ACR, where resolved), as compared to those §s who have never resolved their ”Aggression Conflict” (hereafter referred to as ACU, 27 where unresolved) during therapy. This prediction emanates from the contention that the ACR group, having successfully come to grips with their conflict will manifest decreasing need to discuss material related to the formerly conflictual area. ACU_§s, on the other hand, have not resolved the con- flict, and should exhibit the need to refer to the trouble- some area throughout therapy. NAC §S, never having experi- enced the conflict to begin with, should show little need to fluctuate, over the therapy process, in their expression of aggressive content. Hyppthesis V predicts that a greater proportion of therapist approach responses to aggressive and dependent ma— terial is more characteristic of the ACR cases in comparison to the ACU cases. It is felt that, in part, the mechanism which permits the client to resolve conflicts is his oppor— tunity to discuss these conflictual feelings during the therapy situation. The occurrence of discussion of such ma— terial seems to be a minimum condition for conflict resolution (Barnes, 1963). Winder (1957) maintains that reorganization of verbal behavior is especially dependent on the discussion of such material during therapy. _391e Statements The author's interest in role characteristics of therapy pertains more to theoretical issues, rather than to those questions which deal specifically with therapist effect 28 on response categories or the outcome measure. It seems cer- tain that "rules" governing behavior are important to the flow and continuation of the therapy process. Further, it has been shown that therapy is frequently disrupted when dis— equilibrious expectancies are maintained. On the basis of these findings, it is considered important to determine whether the form of the client—therapist role functions is maintained where different therapist and client samples are studied. There are major differences between therapists and clients in Lennard and Bernstein's (1960) sample and the present one. For, while the former researchers employed therapists with strong analytic orientations and patients from the general population, the latter investigation in- cluded therapists with more interpersonal orientations and college student clients. The following ”role statement” hy- potheses, therefore, are mainly concerned with the replica- bility of Lennard and Bernstein's results. Hypothesis VI predicts that both therapists and clients will engate in a relatively equal amount of primary role statements. This hypothesis relates to an attempt to replicate Lennard and Bernstein's (1960) results with the present sample and tape sampling procedure. While such role statements were found to generally decrease over the course of therapy by these authors, there was a definite correlation between client and therapist discussion of primary roles. 29 Hypothesis VII predicts, in an attempt to confirm Lennard and Bernstein's findings, that there is a steady downward trend, over the course of therapy, in the frequency of primary role response category statements for both client and therapist. This function should vary, however, for the AC and NAC_§s, in a manner suggested by the following two corollary hypotheses: l) AC_§s are characterized by Mueller and Grater (1965) as manifesting greater anxiety than the NAC_§s, and since it is assumed that AC §s exhibit greater dependency, it is predicted that the AC §s will manifest a greater degree of primary role statements during the early phases of therapy as compared to the NAC_§s. It is felt that it is of much greater importance for an anxious— dependent client to clarify the nature of his "client role" than it is for a client not as much characterized by these behaviors; 2) it is predicted that primary role statements remain concomitantly higher for the ACU group throughout therapy than for either the ACR or NAC groups. Since it might be assumed that the ACU_§s have never satisfactorily resolved their aggression conflict and the predicted associ— 7 ated dependency needs, they will manifest the greatest con— tinual need for client—therapist role clarification. Hypothesis VIII predicts, in accordance with Lennard and Bernstein's (1960) results, that there is a relatively constant level of secondary role communication during psycho— therapy (i.e., communications referring to the client or 3O therapist in transference roles). Since the focus on the interpersonal relationship seems to be at a higher level of maturity than discussion centering about primary role con- siderations, and since the NAC and ACR gs appear to be oper- ating at a higher level of ego integration, it seems that secondary role statements will be highest throughout therapy for NAC §S, increasing for ACR Ss, and minimal for ACU_§s. Therapist Propositions Hypothesis IX predicts that therapist evaluative propositions lead to the greatest continuation of hostile or dependent content, with descriptive propositions yielding a lesser degree and prescriptive propositions producing the least amount of continuation. Hypothesis X predicts that there is an initially high degree of therapist's descriptive propositions, with a lesser degree of prescriptive and least degree of evaluative propositions early in therapy. These latter two should in- crease and the former decrease over the course of therapy. This function should obtain, since it seems that, in terms of therapist interaction involvement, descriptive statements are least interactive, prescriptive statements more inter— active, and evaluative statements most interactive. Pre- sumably, the therapist's interaction involvement with the client increases as therapy progresses. Some support for 31 this hypothesis is provided by Lennard and Bernstein (1960), who report a slight decrease in descriptive propositions, with a concomitant small increment in evaluative propositions over therapy. CHAPTER III METHOD Subjects The clients selected for inclusion in the present study were all female undergraduate students who completed a minimum of six therapy interviews at the Michigan State Uni— versity Counseling Center. Female §s were chosen to repre- sent the sample due to Mueller and Grater's (1965) finding that female_§s clustered well on those Semantic Differential concepts which went into defining the ”Aggression Conflict" variable, while, for the male_§s in their sample, the factor analysis of these concepts yielded a rather unsatisfactory clustering. They attribute this phenomenon to the greater degree of discrimination with which the female §s responded to the Semantic Differential. A total of 25 SS comprised the present sample. The clients are late adolescents who are self- referrals seeking help with personal—social problems. An in- take interview determined that a program of therapy was warranted. The_§s included in the present study represent 25 of the total of 54 randomly selected clients who were asked to 32 33 participate in the research program being conducted at the Counseling Center. A tape library was developed on the psychotherapies of these 54 clients, and it is from this library that the current sample is drawn. In addition to the recording of all interviews, these clients participated in the following program of research testing: Personal Data Sheet, MMPI, and Semantic Differential were obtained prior to treatment and subsequent to termination, and the Semantic Differential was obtained at monthly intervals during the therapy process. Therapists Clients were assigned to therapists on the basis of matching availability of hours. Meetings were generally con- ducted on a weekly basis, except in the case of the practicum therapists, who saw their clients twice a week. Final as- signment of a case depended upon the therapist's willingness to work with the client in question. Such judgments were made on the basis of intake interview data and the personal impressions of the intake interviewer. Some bias is intro— duced into the library sample since not all therapists at the Counseling Center participated in this research. The psychotherapists involved in the present sample fell into three groups: staff counselors are represented by 6 Ph.D. level counseling or clinical psychologists with 3 to 20 years of psychotherapy experience; the intern group was 34 comprised of a total of 14 advanced doctoral candidates in clinical or counseling psychology who have received a mini— mum of one year of intensive psychotherapy supervision; the 5 practicum students represent a group of less advanced graduate students who are treating their first cases as part of an introductory practicum course offered at the Counseling Center. Since the present study restricted its sample to fe- male clients only, matching for therapist sex differences was impossible. Seventeen of the twenty—five therapists are male. Therapist sex and experience level group, and client categories within these two groups are summarized in Table 1. Table 2 indicates the duration of client continuation in therapy according to the experience level group of the therapist. Table 1. Summary of client and therapist characteristics. Male Therapists Female Therapists Client Character— istics Staff Intern Practicum Staff Intern Practicum TOTAL ACR 0 2 0 l 1 1 5 ACU 2 5 2 0 0 1 10 NAC 2 4 0 l 2 l 10 TOTAL 4 ll 2 2 3 3 25 35 Table 2. Duration of therapy and therapist experience. Number of Therapy Sessions Therapist Ex— perience Level 6—10 11—15 16—20 21+ Total Staff 3 0 l 2 6 Intern 5 3 5 l 14 Practicum 1 l 2 1 5 Total 9 4 8 4 25 Sampling Procedure All therapy tapes for each of the 25 SS were included in the present investigation. Ten-minute samples were randomly selected. A total of 330 such ten-minute samples represents the population presently under investigation. Tape-recordings rather than transcripts represented the units of analysis. There does not appear to be any clear— cut evidence to indicate superiority of either method. While Kbpplin (1965) maintains that inter—judge reliability with tape—recordings necessitates a great deal of training, Harway et al. found no difference between transcript and tape— recording methods in their investigation (Winder, 1957). CodingiProcedure The coding procedure employed in this investigation derives, in part, from two sources: 1) the schema developed 36 by Bandura et a1. (1960), and modified by Winder et a1. (1962), and Kopplin (1965), is applied to measure client "dependency," ”hostility," and ”other" response categories, and therapist ”approach" and ”avoidance“ behavior. The many investigations previously cited which have employed this system, and have replicated the basic parameters, serve to establish the Bandura-Winder system as a reliable procedure for the investi— gation of the therapy process; 2) Lennard and Bernstein's (1960) method is employed to identify client and therapist primary and secondary role systems, and the types of thera— pist propositions manifested. The manual describing the operational definitions of these response categories, and providing examples of these categories is contained in Ap- pendix A. The basic model subsumed under both coding procedures derives from the Bandura—Winder system, in that focus is turned to the interaction sequence. Scoring is as follows: a client statement, a therapist response, and the subsequent response on the part of the client. By definition, a scoring unit is represented by the entire verbal content of a speaker with the preceding and succeeding verbalizations of the other person acting as the boundary for that unit. The scoring sequence employed may be summarized as follows: the judge listens to a client verbalization and classifies it first according to the hostility, dependency, or "other" content response category. Then the judge scores 37 the client statement according to the primary or secondary role system involved, if any. Therapist responses are then scored as approach, or avoidance. The therapist's verbali- zation is similarly scored for presence of primary or second— ary role statements, and then according to the type of propo— sition employed (i.e., descriptive, evaluative, or prescriptive). Scoring Reliability Scoring reliability centers about two distinct con— siderations: the determination of units, and the scoring of those units. Kopplin (1965) and Dollard and Auld (1959) have emphasized the need to show that the scoring unit employed is a reliable and steady one. This source of variance was eliminated in the present study, since Judge A indicated to Judge B where a particular client or therapist unit began by transcribing the first few words of each sequence. Judge B stayed within these confines when scoring his responses. Secondly, the reliability of the content, interaction pro— cess, and therapists“ propositions had to be determined. Common practice indicates that product—moment correlation co— efficients varying from .80 to .85 are considered adequate (Dollard and Auld, 1959), and were strived for in this investigation. The units and response categories were scored by two raters—~the author, Judge A, and a reliability scorer, Judge 38 B, a first year graduate student in clinical psychology who received remuneration for his part in the study. The two judges initially scored several tapes together in order to train themselves in the use of the system. This procedure obtained until the raters achieved an .80 — .85 product— moment correlation coefficient. Training tapes were not in— cluded in the data analyzed as part of the study. The subsequent sample on which reliability was de— termined included approximately one-half of the total of 330 ten-minute segments sampled. The first 75 and last 74 tapes were rated by both scorers in an attempt to determine any consistent changes in rater A's scoring over the course of coding. Each judge rated these randomly selected interview segments independently. Judge A rated the remainder of the 330 ten—minute segments. Reliability rating was effected in a rather intensive ten—week scoring schedule. Inter—judge coefficients and the significance of the differences between pairs of correlations are presented in Table 3. It is evident from consulting this table that the scoring categories represent highly reliable measures. In most cases where reliability was lower, this obtained due to the small frequencies observed in the particular category under consideration. Nevertheless, inter-judge reliability estimates for both the first 75 and the last 74 tape segments sampled were all highly significant, with p < .002 for all 39 Table 3. Inter-judge reliability coefficients and the standard scores and p values associated with the differences between a pair of correlations. * * Measure Ia IIb Zd p Dependency .83 .80 .53 .56 Hostility .92 .89 1.00 .31 Client Other .98 .98 -— -- Client Primary Statements .91 .96 2.51 .01 Client Secondary Statements .84 .84 —— __ Client Other Statements .99 .98 2.09 .04 Approach .99 .99 —— __ Avoidance .79 .83 .64 .52 Ther. Primary Statements .93 .97 2.60 .01 Ther. Secondary Statements .92 .95 1.40 .16 Ther. Other .98 .97 1.23 .22 Descriptive .91 .93 .78 .43 Evaluative .98 .97 1.23 " .22 Prescriptive .81 .74 1.06 .29 aCorrelation coefficients for first 75 tape segments. bCorrelation coefficients for last 74 segments. *All coefficients, p < .002. 40 measures. Similarly, the Z scores obtained for the differ- ence between two coefficients for a given measure indicates the high degree of consistency over time in the placement of a response in a given category. While some of the Z scores d did achieve statistical significance, which would suggest that a change in scoring procedures between the two judges did in fact occur, it can be seen from Table 3 that these differences occurred where correlations were so high, that a difference of .01 or .02 was critical enough to produce a significant Zd' Semantic Differential The ”Aggression Conflict" concept developed by Mueller and Grater (1965) provided the basis for the outcome measure employed. As previously defined, AC is said to ex— ist where aggression is seen as potent and active, but is negatively evaluated. The resolution of AC (i.e., ACR) re— volves about the change to more positive evaluations of ag- gressive feelings. The Semantic Differential employed to define this construct represents the instrument prepared originally by Osgood. It contains sixteen scales and 21 concepts prepared in the form of a differential. Mueller and Grater (1965) factor analyzed these concepts, and found five which clustered together for fema1e_§s, to yield a more reliable composite aggression score--My Feelings of Anger Directed at Others, 41 My Feelings of Anger Directed at Myself, Hate, Guilt, and_My Mixed—up Feelings About My Behavior. These authors also indicate that the Semantic Differential method which they em— ployed has previously been shown to be quite sensitive to attitude change. In further investigations employing the AC cluster (Mueller and Grater, 1966), it was reported that test—retest reliability of .75 obtained for the active—potent factor scores, while a .62 correlation obtained for the evaluative factor scores. On the basis of these data, the authors posit the legitimacy of employing the AC scale in investigations which study changes in the same subject's aggression scores over time. CHAPTER IV RESULTS Total Freguencies Table 4 depicts the total number of responses in- cluded in the 330 ten—minute tape segments sampled, and the proportion of the total sample represented by a particular measure. It should be noted that the differences in the total number of units for clients and therapists is due to inaudible therapist statements, and to the fact that the client's statement sometimes completed the unit (i.e., when the units at the end of the tape were sampled). Hypotheses Hypothesis I predicts that therapist approach to hostile or dependent client content should have the effect of eliciting a continuation of such material in the immediately proceeding statement, while therapist avoidance responses to such content should produce the opposite result. It can be seen from Table 4 that the total number of therapist avoid— ance responses was exceedingly low. The number of such re— sponses which followed a hostility or dependency statement represented a total of 18 such events, therefore, preventing 42 43 Table 4. Total number of units and relative proportions of each measure found in the total sample. Measure Frequency Proportion Total Units Client Dependency 564 .0944 Hostility 476 .0799 Other 4919 .8255 Primary Statements 783 .0893 Secondary Statements 106 .0178 Other Statements 5321 .8929 5959 Therapist Approach 5853 .9845 Avoidance 92 .0155 Primary Statements 783 .1317 Secondary Statements 254 .0427 Other Statements 4908 .8256 Descriptive 965 .1623 Evaluative 4726 .7950 Prescriptive 254 .0427 5945 44 an analysis which would serve either to confirm or deny this hypothesis. Hypotheses II, III, IV, V, and parts of the subse— quent hypotheses pertain to predictions regarding §s defined along the AC-NAC dimension. Before dealing with each hy— pothesis individually, a general summary statement of the over-all differences found among these three groups (i.e., NAC, ACR, ACU), appears in Table 5. The results indicate that the three groups do not differ significantly from each other on any of the response category measures. In addition to these over-all comparisons, individual comparisons be— tween ACU—NAC, ACU—ACR, ACR—NAC, and AC—NAC_§s were performed on all measures. These findings appear in Appendix B. Here again it can be seen that none of the groups varied appreci— ably from one another on any of the measures involved. Mann— Whitney U tests failed to achieve the .05 level of signifi— cance for any of the content categories under investigation. Hypothesis II predicts a greater frequency of initial aggressive content for AC as compared to NAC gs, while_Hy- pothesis III predicts the same response pattern for dependent material. Reference to Appendix C reveals that such a function did not, in fact, obtain. The AC and NAC §S were not found to differ along the aggressive or dependent content measures when comparisons for these two groups were made for the first session and the first three sessions. It should be noted, however, that the results are in the predicted 45 direction for the Hostility measure. The Mann—Whitney U values obtained for this response category is 15, which falls between .10 > p > .05. It can be tentatively maintained that aggressive content seems to be somewhat higher during the initial phases of therapy for the AC §s as compared to the NAC.§s. Table 5. Kruskal—Wallis H values obtained in comparing ACR, ACU, and NAC_§s. Measure H Duration of Treatment 2.18 Number of Statements 1.60 Dependency 1.60 Hostility 2.09 Client Primary Statements 1.30 Cleint Secondary Statements .30 Approach 1.70 Avoidance 3.10 Therapist Primary Statements .20 Therapist Secondary Statements 1.00 Descriptive Propositions 2.74 Evaluative Propositions 1.40 Prescriptive Propositions 1.40 *All H values are not significant. 46 Hypothesis IV predicts that both aggressive and de— pendent content are lower, during the final phases of therapy, for the NAC and ACR_§s than it is for the ACU_§s. Kruskal— Wallis H values pertaining to this hypothesis for both the last session and last three sessions of therapy appear in Ap- pendix D. Examination of these results clearly indicates a lack of evidence for confirmation of this hypothesis. The ACR, NAC, and ACU §s are not found to differ in terms of the frequency of responses in either the dependent or hostile categories at the end of therapy. Hypothesis V predicts that the ACU_§s are character— ized, as a group, by a greater frequency of therapist avoid— ance responses to their hostile and dependent statements in comparison to the ACR_§s. The same difficulty obtained in testing this prediction as was manifested in testing Hypothe— sis I, i.e., the frequencies were too small to permit ade— quate statistical analyses. The results of the present in— vestigation, therefore, provide no evidence to comment upon the issue of relative therapist avoidance responses to ACU and ACR SS! dependent and hostile statements. Hypothesis VI predicts that both clients and thera— pists will engage in.a relatively equal amount of primary role statements, i.e., statements pertaining to the structure of the therapeutic process. This hypothesis is not supported. The present experiment yields approximately 13% therapist 47 primary statements as compared to 9% client primary state— ments. Apparently, then, therapists in the present sample engage in a relatively larger amount of statements which en- deavor to structure the therapeutic relationship. (The fre— quencies involved in the current study provide an X2 = 47.8, p < .001.) Hypothesis VII predicts a steady downward trend, over the course of therapy in both client and therapist pri- mary statements. The results clearly contradict this pre— diction. Figure 1 reveals that while there is a decrease in the middle of therapy for client primary statements, there is a sharp increase at the end of therapy (F = 4.78, p < .05). Therapists, on the other hand, show a slight increase in pri— mary role statements from the beginning to the middle of therapy, but manifest the same sharp increase observed be— tween the middle and end of therapy. The two corollary hypotheses predict: 1) that AC_§s manifest a greater degree of primary role statements during the early phases of therapy as compared to the NAC §S. This hypothesis is not supported. Mann—Whitney U tests (see Ap- pendix C) comparing AC and NAC S8 for the first session and for the first three sessions, reach 49.5 and 24.5, re— spectively—~both values yield p > .10. 2) that primary role statements remain higher throughout the course of therapy for the ACU §S as contrasted to the ACR and NAC SS. Reference 48 135 130 125 120 1.15 / 110 / 105 / 100 /,/ 95 9O 85 80 75 70 65 6O 55 50 45 4O 35 3O 25 20 15 10 Frequency of Primary Role Statements lst Middle Last PHASE OF THERAPY Client Therapist ————————— Figure 1. Frequency of client and therapist primary role statements over the course of therapy. 49 to Table 5 indicates that the predicted relationship does not exist. Kruskal—Wallis analysis finds H = 1.30, p > .50. Hypothesis VIII predicts a relatively constant level of secondary role statements throughout psychotherapy. Sup- port of this prediction is untenable. While the frequencies for client secondary statements were too small to analyze statistically there is some support for the notion that the form of the function for clients is initially one of an in- creasing nature, which then levels out over the latter half of therapy. Therapist secondary role statements similarly increase from the first to the middle part of therapy but subsequently decrease (X2 = 14.68, p < .001). The form of the function is presented in Figure 2. As was the case for primary role statements, there was a significant tendency for therapists to engage in a greater degree of secondary role statements than did clients (X2 = 60.8, p < .001). Reference to Table 5 reveals no sup— port for the prediction of the existence of differences in the level of secondary role statements for ACU, ACR, or NAC §s over the course of therapy. Hypothesis IX predicts that therapist evaluative propositions produce the greatest percentage of continuation of hostile or dependent content, with descriptive propositions encouraging less continuation, and prescriptive propositions accounting for the least degree of continuation. This 50 U) u C.‘ (D E B 20 B 19 m 18 o 17 H O 15 m 15 5‘ 14 ,0 13 “g 12 \ \ 8 11 \ o 10 \\\\ U) 9 \ O 7 5* 6 c: 5 3% 4 3' 3 .. 2 ‘“ 1 o lst Middle Last PHASE OF THERAPY Client Therapist __________ Figure 2.“ Frequency of client and therapist secondary role statements over the course of therapy. 51 hypothesis is not supported. The X2 value for continuation of dependency responses in relation to the type of therapist proposition involved, is 3.14, p > .20; with X2 = 1.79, p > .30 for continuation of hostility responses. Hypothesis X predicts an initially high degree of therapist' descriptive propositions, with a subsequent de— crease. Prescriptive and evaluative propositions are pre— dicted to be initially low, and increasing over the thera- peutic process. For the present sample, there is an initial decrease in descriptive propositions from the first to middle sessions, with a concomitant increase from the middle to last sessions (F = 3.92, p < .05). These results are found tabled in a subsequent section (see Table 7). Although there is a trend in the reverse direction for evaluative propositions, viz., an increase from first to middle and a decrease from middle to last, significance is not achieved (F = 2.48, .10 > p > .05). The form of the relationship appears to represent a relatively stable phenomenon throughout the course of therapy for prescriptive propositions. Figure 3 depicts the relationship observed for the therapist propo- sitions under consideration. Thus, in summary of the ten hypotheses under investi— gation, eight were unconfirmed, one was tentatively supported, and one yielded frequencies too small to be analyzed. Two corollary hypotheses were similarly unconfirmed. Frequency of Propositions 52 370 355 ’\ 340 , ,~—’ \ \\ 325__ 1..” ’ I \ \ 310 \.\\ 295 .\ \ 280 \C 265 250 235 220 205 190 175 170 165 150 135 120 105 90 755IIIINEITE““‘-~el11‘\‘\\\§Jflflf’,,,,lse»’*”””’gfl#fifl/g 6O 7 45 . ‘ , . 1 . . 3O . v ‘ ‘ o . , . . 7 ‘ 15 lst Middle Last PHASE OF THERAPY Descriptive Evaluative ——————————— Prescriptive ........ Figure 3. Frequency of therapist propositions over the course of therapy. 53 Tape Segment Sampled In addition to the hypotheses under investigation, a significant feature of the present study centered about the method of randomly sampling different time segments of therapy sessions. The question here asks what, if any, changes oc- cur in the form of the process during the therapeutic hour itself. The results of this examination are found in Table 6. It is evident that there are major differences oc- curring in the relative frequencies for a given response cate— gory on the basis of the segment of the therapeutic hour under investigation. It seems that dependency statements tend to occur most frequently in the last segment of the therapy hour, and least often during the middle portion. Conversely, the middle sessions appear to have the greatest degree of hostility statements, with the least amount ap— pearing in the initial and final stages. Client primary role statements are found to occur more frequently during the first and last phases of the hour, with the low point be- ing manifested during the next—to—last segment. Differences were found to be non—significant for both the client secondary role statements and the therapist avoidance categories. Therapist primary role statements do, however, show a steady decline through the fourth ten—minute segment with a sharp rise during the last ten minutes. Therapist secondary role communications, on the other hand, are least frequently 54 .ucoEmow Hmspfl>apcfl onp CH onsmmoE pmcu Mom ooconngooo mufl mo coflunomoum onp >9 .wusoemom Ham now ousmmofi amasoflpnmm m How monouusooo mo humongoum some map mcflwamflpasfi an pocflmuflo moflocosquMIlpcoEmom Mom mucoEoumuw mo Hogan: Hmuou CH moocoHoMMHo How pcsooom on popmsflpm moflocosqoume Hoo. V o>.mm Hm.mHH H©.H¢ mo.om Hm.m¢ oa.mm o>flumfluomonm Hoo. V mo.mm m©.mmm mm.maoa m¢.¢voa ¢©.mooa no.0mm o>Hpm5Hm>m Hoo. V mm.¢> H¢.>Hm Ho.mma OH.N¢H m©.S©H hm.omm o>HpQHHoon H00. V mm.ma hm.mm mn.mo mo.©m om.H© mo.¢m humpcooom pmflmmnona 400. V mH.em em.mmm om.mm do.ema em.een oe.men sumsnnm umedmnmae No. V mm.mH mo.mm mm.ma mm.ma ma.m mm.HN oocmpflo>¢ .m.c om. Hm.®mHH mm.nmaa H©.NONH mv.>oma Hm.mmHH Lomosmm< .m.: mm.m mm.mm NN.©H no.mm vw.hm mw.ma humocooom peoHHO Hoo. V mm.mm HN.©mH om.nm Hm.om No.0m hm.ava humeflnm pcoflao mo. V om.m m¢.mm m©.moa om.oaa HN.©OH mo.hh muflaflumom Ho. V Nv.ma mm.mva an.mma mm.moa nm.mm m¢.woa mocopcomom Q mx m w m N H ousmmoz pcofimom mama .poHQEmm DCoEmom ommp >9 oomsonm monsmmoE How emoflocosqonw HmDOB .o oHQmB 55 expressed during the first ten—minute portion of the hour, and then increase and level off to a relatively stable inci— dence of occurrence for the remainder of the session. Thera— pist descriptive propositions occur with the greatest fre- quency during the first and last portions of the hour, with a lower, stable function during the middle part. Evaluative propositions appear in a direct inverse relationship to de- scriptive propositions. Prescriptive propositions seem to occur with equal frequency during the first four parts of the session, and increase sharply to reach their peak during the final phase. The strength of the fluctuations during the hour is a rather marked and consistent one, a phenomenon which is maintained across several measures. It is important to note as well that the total number of therapist and client state— ments are significantly higher during the first ten minutes than during any other phase. Further Considerations Phase of Treatment Several examinations were performed which yielded interesting data, but which were not included in the original design. Tables 7 and 8 represent two of these analyses. Table 7 includes analysis of variance data pertaining to the changes which occur for each response measure over the course 56 Table 7. F and p values relating to the changes in each measure over the course of therapy—-first, middle, and_last interviews for a11.§s. . a Se581on Measure First Middle Last F p Dependency 54 45 42 .61 n.s. Hostility 51 31 31 2.48 < .01 Client Primary statements 48 34 89 5.59 < .01 Client Secondary Statementsb Approach .428 431 397 .53 n.s. AvoidanceC 3 9 7 n.s. Therapist Primary Statements 63 67 114 4.78 < .05 Therapist Secondary Statementsd o 15 8 < .001 Descriptive 93 68 103 3.92 < .05 Evaluative 316 347 274 2.48 n.s. Prescriptivee 22 24 26 n.s. 5A total of 7 (of the sample of 75) tape segments were missing--missing cells were filled by use of row and column means. bFrequencies too small to be computed. sz = 2.95. dx2 = 14.68 e 57 of the first, middle, and last interviews for all gs. Com- pared to this, Table 8 provides analysis of variance infor— mation on the changes in these response measures for the first six sessions for all.§s. Of primary importance in considering Tables 7 and 8 is the observation that none of the results which appear in sampling the first six interviews are lost by sampling the first, middle, and last interviews. Further, it appears that changes in process can be more readily observed when sampling according to the latter method. Many more changes achieved significance when employing this technique. Several of the response measures change in a prescribed fashion over the course of therapy. Hostility tends to decrease markedly from the first to the middle session and then achieve a plateau. As mentioned previously, client primary statements tend to decrease and then experience a sharp increase during the last session, while there is a consistent increase in therapist primary statements from the first to last session. Therapist secondary statements appear, on the other hand, to increase sharply from the first to the middle portion of therapy, and experience a subsequent decline. Both Tables 7 and 8 indicate a significant downward trend for therapist' descriptive propositions over the first half of therapy, with a marked upward movement occurring in the final session. 58 Table 8. F and p values relating to the changes in each measure over the course of thera interviews for all §S. —— —‘__ py-—first six . a 8e581on Measure 1 2 3 4 5 6 F p Dependency 53 40 47 63 49 37 1.65 n.s. Hostility 50 42 15 43 39 33 1.50 n.s. Client Primary 43 24 39 22 30 33 .85 n.s. Client Secondaryb Approach 432 470 438 467 434 429 .50 n.s. Avoidanceb Therapist Primary 60 52 68 45 48 Therapist SecondaryC o 3 -10 26 6 Descriptive 93 80 67 64 37 Evaluative 317 384 356 390 382 Prescriptived 22 15 18 l4 13 49 .56 n.s. 14 < .001 100 2.70 < .05 308 2.10 n.s. l6 n.s. aA total of 16 (of 150 comprising segments were missing—~missing cells were row and column means. b . FrequenC1es too small to compute. x2 = 35.61. x2 = 3.27 the sample) tape filled by use of 59 Aggression Conflictl Recasting the data such that "Aggression Conflict“ was regarded as a variable for each_§l similarly failed to differentiate the NAC from the AC_§s on the two measures (i.e., dependency and hostility) under consideration. A further analysis compared high Active—Potent hostility_§s with low Active—Potent_§s obtained on the Seman— tic Differential test administered prior to therapy. Hostile and dependent content categories for the first interview were 1Due to apparent inability of the AC measure to differentiate groups on an over-all manifest level, it was felt that the AC measure might be viewed in a different light. (AC and NAC was determined on the basis of initial testing, while ACU and ACR was determined according to Seman- tic Differential data obtained subsequent to the final session. The "Aggression Conflict" measure represented a variable, however, in that while an individual may have been classified in the NAC group on the basis of initial test scores, his Semantic Differential protocols may have mani— fested such conflict (i.e., positive active-potent and nega- tive evaluation of aggression) on tests administered during the middle phase of therapy; the reverse might also have been true for an‘g initially classified as AC. _§s were, there- fore, re-categorized as AC or NAC on the basis of the test scores attained during the middle phases of testing, and the interview immediately preceding test administration was ana— lyzed for hostile and dependent content. So, for instance, all §s who achieved a positive active—potent score and nega— tive evaluation score for aggression on the second testing (which was administered after the fourth interview), were categorized as AC_§s regardless of their original classifi— cation. The fourth interview (i.e., the one immediately pre— vious to testing) was analyzed for hostile and dependent con— tent. The same procedure obtained for the 3rd, 4th, and 5th Semantic Differential testing and the 8th, 12th, and 16th interviews, respectively. Analysis of the first interview was provided for in the original design, and no significant differences were found between the NAC and AC_§s (see Ap- pendix B). 6O analyzed along the high—low Active—Potent dimension. No sig— nificant differences resulted, i.e., the high Active—Potent group did not use the hostile or dependent content category with substantially greater frequency than low Active—Potent _§s on the initial interview. Analysis of increased vs. de— creased Active—Potent scores from one testing to the next similarly did not serve to separate the groups on the rela- tive frequency of hostile or dependent content employed. Personality Data Post hoc analysis of pre-therapy MMPI data revealed that AC_§s tended much more frequently to fall into the ”de— viant" (i.e., three or more MMPI scales over 70), category than did NAC_§s. While 9 NAC and 1 AC §S fell into the ”normal” classification, 6 NAC and 8 AC_§s comprised the "deviant” group (p < .02, Fisher's Exact Test). "Normal” vs. ”deviant” comparisons did not, however, serve to separate the groups on any of the response measures included in the study. Mann—Whitney U tests failed to achieve significance for any of the variables under inspection. There was a tendency, however, for the ”deviant” S5 to produce a higher degree of dependency (U = 37.5, .05 < p < .10), and hostility (U = 37, .05 < p < .10) statements. In addition there appeared to be a greater number of descriptive therapist propositions for ”deviant”_§s (U = 38.5, .05 < p < .10). 61 Duration of Therapy The question of differences between long—term (four— teen or more interviews) and short-term (twelve or fewer) therapy clients served as a basis for further analysis in the present investigation. No differences were found between the two groups on either the "normal“~-"deviant” (p > .28, Fisher's Exact Test), or the AC—NAC dimension (p > .50, Fisher's Exact Test). Another issue connected with long-term and short— term clients centers about the comparability of results ob— tained on the two groups both in terms of session by session differences, and changes observed over the first, middle, and last portions of the therapeutic process. In other words, is it justifiable to assume that the first N sessions (or the first, middle, and last interviews) of long—term and short-term clients manifest the same form. Tables 9 and 10 present X2 analyses of the differences observed between the longand short-term therapy groups. The results appearing in Table 9 indicate that the form of the process for the first six sessions differs between long—term and short-term clients on the client primary state— ments, therapist primary statements, and approach measures. While the first six sessions for long-term'Ss are character- ized by an alternating decrease and increase in the frequency of client primary statements, the short-term'Ss manifest a 62 Table 9. X2 and p values comparing first six interviews for long—term and short—terms gs. Session Measure 1 2 3 4 5 6 X2 p Dependency Long Term 28 19 25 30 25 20 Short Term 25 21 22 33 24 17 .74 n.s. Hostility Long Term 38 29 7 25 23 20 Short Term 12 13 8 18 16 13 6.40 n.s. Client Primary Long Term 27 13 30 19 25 19 ' Short Term 16 ll 9 3 5 14 15.64 < .01 Client Secondary* Approach Long Term 211 258 246 264 260 210 Short Term 163 142 144 167 101 142 22.10 < .001 Avoidance* Therapist Primary Long Term 41 29 52 34 32 24 Short Term 19 23 16 ll 16 25 14.8 < .02 Therapist * Secondary Descriptive Long Term 56 50 47 32 25 53 Short Term 37 30 20 32 12 47 8.62 n.s. Evaluative Long Term 180 241 216 243 245 199 Short Term 137 143 140 147 137 109 5.66 n.s. Prescriptive Long Term 16 9 l3 8 9 8 Short Term 6 6 5 6 3 8 3.64 n s *Frequencies too small to compute. 63 Table 10. X2 and p values comparing first, middle, and last interviews for long-term and short—term_§s. Session Measure First 'Middle Last X2 p Dependency Long Term 28 15 19 Short Term 25 28 16 4.00 n.s. Hostility Long Term 38 23 18 Short Term 12 9 8 .44 n.s. Client Primary Long Term 27 28 69 Short Term 16 5 14 7.95 < .02 Client Secondary* Approach Long Term 211 240 222 Short Term 163 174 107 10.36 < .01 Avoidance* Therapist Primary Long Term 41 49 79 Short Term l9 17 28 .73 n.s. Therapist Secondary* Descriptive Long Term 56 34 63 Short Term 37 30 31 3.12 n.s. Evaluative Long Term 180 198 176 Short Term 137 135 70 14.07 < .001 Prescriptive Long Term 16 15 9 7 1 17 n 3 Short Term 6 9 *Frequencies too small to compute. 64 steady decreasing function with heightened frequencies during the fifth and sixth sessions. Similarly, although therapist approach responses alternatingly increase and decrease for both groups, the form of the pattern characteristically varies with the particular session under consideration, thus pro- ducing significant differences for the two types of §S. The same type of phenomenon is observed in contrasting therapist primary statements, and similarly results in significant differences in the patterns shown over the first six sessions for this measure. Table 10, in comparing the nature of the process for the first, middle, and last interviews, indicates that differences do occur according to the particular variable under investigation. Thus, long and short—term.§s appear to differ on the following measures: 1) Client Primary State- ments——while long—term_§s appear to show a relatively stable function for the first to the middle interview, and a subse— quent sharp increment in the last interview, short-term.§s exhibit a decrease from the first to middle session with a concomitant increase in the last session; 2) Approach——while approach responses increase from the first to the middle session and subsequently decrease in the last session for both groups, the form of the increase from the first to the middle session is much steeper for the long—term‘Ss, whereas the subsequent decrease in the last session attains a more 65 marked decline for the short-term_§s; 3) Evaluative--evalu— ative propositions appear to maintain relative stability for long—term_§s over the course of therapy, with a slight in- crease from the first to the middle session followed by a slight decrease in the last session. While evaluative propo— sitions similarly seem to maintain stability from the first to the middle session for short—termigs, a sharp decrease is noted between the middle and last session. CHAPTER V DISCUSSION Certain of the content categories employed in this research were selected in order to investigate the nature of the therapeutic process as related to aspects of "Aggression Conflict" and others were used in an attempt to confirm the results described by other authors with the present sample and sampling technique. Essentially at issue are the questions of generalizability of the findings to other popu— lations, and the attempt to predict the form the therapy process takes for those individuals identified along an ”Ag- gression Conflict" dimension. Inter-Judge Reliability The high reliability attained for all content cate— gories is primarily attributed to the explicitness of the scoring manuals included in this design. While latent mean— ing is inevitably sacrificed, the ability to score a unit with such high independent inter—judge agreement seems to es— tablish the utility of this approach for further psycho- therapy research. With adequate training, then, use of either manual can surmount the problem of poor inter-judge 66 67 reliability typically encountered with clinical judgments, diagnoses, and other scoring manuals. Therapist Approach and Avoidance The small number of avoidance responses precluded statistical analysis and further comment upon the question of the relative frequency of continuation of hostile or de- pendent content when therapist approach or avoidance to a client statement obtained. Clearly, the present investi- gation differs from others employing the Bandura-Winder scoring system (c.f. Kopplin, 1965; Bandura et al., 1960), in that such analysis was performed. One possible cause of this difference in frequency of avoidance responses emanates from the exacting requirements the present investigation maintained before scoring a response in this category. Interestingly, the frequencies of avoidance responses re- corded in the other studies alluded to are not reported in the literature, thus preventing more accurate comparisons. Aggression Conflict The results tend to shed serious doubt on the pre— dictive ability of the hypotheses relating to the Aggression Conflict outcome measure. This measure failed to differenti- ate ACU, ACR, and NAC_§s from one another on any of the measures under consideration. Although it does seem that initial aggressive content is higher for AC as compared to 68 NAC_§s, the tentative nature of this finding is strongly em- phasized since the p value fell between .10 > p > .05. The ”Aggression Conflict” measure thus grants some support for the notion that the conflict manifested initially in the Semantic Differential scale will intrude more frequently in to the manifest verbal content of the AC_§s. Such con- tentions are not confirmed, however, for the dependency measure, or for changes in the relative frequencies of hostile content, over the course of the therapy process, for the three groups. Thus, while theory would support the soundness of the hypotheses, the assumptions underlying these predictions must be investigated. Probably the major assumption centered about the notion that the "Aggression Conflict” measure repre- sented a reliable indicator of manifest behavior. On the basis of this assumption, others naturally followed-~i.e., that there would be an associated high degree of dependency for conflicted_§s, and that they would tend to relate to the therapist in certain ways. Since the initial assumption does not appear to be warranted, those based upon it simi— larly do not hold up under experimental scrutiny. Recasting of the data in terms of considering "Ag- gression Conflict" as a variable for each subject, and ex— amining the content categories for those interviews which im- mediately preceded Semantic Differential testing, similarly proved unfruitful, as did examination of the data along the 69 Active—Potent dimension. These findings further augur for the apparent inability of the outcome measure to act as an index of manifest behavior. For the present study, there— fore, little can be said about the therapist variables which produced resolution of conflict of aggression (e.g., differ- ential avoidance), nor for the differences in the forms of the processes manifested for the three groups over the thera- peutic endeavor. The difficulty of developing outcome measures as adequate predictors of change, as discussed in the literature, is confirmed in the present investigation. The question must be raised as to whether, in the long-run, outcome must be defined as a variable for the particular client undergoing treatment. And, further consideration must be given to the issue as to whether change must not be considered only in terms of the ongoing context, i.e., the individual client—therapist interaction. Role Systems The present study indicates that a significantly greater proportion of therapist time is given to primary role statements as compared to clients. These results are in direct contradiction to those reported by Lennard and Bernstein (1960). While these authors report an average .incidence of Primary statements in the amount of 10% for therapists and 9% for clients, the current experiment yields approximately 13% Therapist Primary Statements as compared to 7O 9% Client Primary statements. It seems that the therapists included in this experimental population spend more time than their clients (or than did the therapists in Lennard and Bernstein's sample) in discussing the structure of the therapeutic relationship. This finding may derive from the fact that clients in psychoanalysis are generally more sophisticated populations who are better attuned to the therapeutic process, therefore, needing less ”instruction" on the part of the therapist. College students, on the other hand, may represent a more naive group who have to undergo greater "role induction." Another contradictory finding appears between the results of the present study and those reported by Lennard and Bernstein. While the latter authors report a steady de- crease of both client and therapist primary statements as therapy progresses, the present investigation finds a sharp increase between the middle and last sessions of therapy. A possible factor involved in the difference might be the rela— tive lengths of therapy for the clients in the respective samples. While the clients included in this study usually had a maximum of 20—25 interviews, Lennard and Bernstein's group spent many times this amount in a more traditional type of psychoanalytic therapy. Lennard and Bernstein's de— creasing trend may have been a function of the increasing length of therapy. More probable is that Freudian techniques traditionally structure the nature of the client—therapist 71 roles very early in therapy, and there is little room for deviation. The interpersonal orientation of counseling center therapists might provide for greater variability along these lines. Another possibility centers about the notion that the strength of the effect found in the present study might be due to sampling the last session, rather than, for instance, the next to last session. Both therapists and clients are attuned to the last session and spend more of their efforts on clarification rather than more interpretive aspects. It should be kept in mind, however, that Lennard and Bernstein's population also included last session data and still did not manifest the function found in the present investigation. Whatever the underlying mechanisms involved in the observed differences between the two samples, it is clear that generalizing from one set of results to larger populations of therapeutic interactions is often an un- warranted maneuver. This latter argument is further supported by the failure of the present findings to confirm Lennard and Bern— stein's indications of a relatively constant level of Secondary Role statements. Rather than manifesting a con— stant function, there appears to be a significant proportion of Secondary Role statements appearing during the middle phases of therapy for both clients and subjects. It appears, for the present sample, that the transference relationship is most frequently alluded to during the middle stages of 72 therapy. Assuming this to be the point of most intense client—therapist transference interaction, it seems feasible that such intensity would require time to build up, and would also necessitate client—therapist resolution prior to termination. As was the case with Primary Role statements, Second- ary Role statements appear with much greater frequency for therapists than they do for clients. This finding is not at all unexpected for the present sample of therapists. Counsel— ing center orientation focuses, to a great degree, on the nature of the interpersonal relationship between the client and therapist. The therapist often points to his own feel- ings toward the client, and to how the client is interacting with him as a means of effecting change. Therapist Propositions These content categories provide more interest as process variables within a given interview and across inter- views, than they do when regarded in the terms of their rela— tive abilities to produce continuation of dependent or hostile client content. No particular therapist proposition seems better-suited than the others in eliciting or causing the continuation of such content. While prescriptive propositions occur in a rather constant pattern over therapy, descriptive propositions ap— pear with the greatest frequency during the first and last sessions, and with least frequency during the middle session. 73 The inverse relationship appears to be the case for thera- pist evaluative propositions. These findings are undoubtedly related to the total structuring of the interview discussed previously in connection with Primary Role statements. For while the therapist is structuring the nature of the thera- peutic relationship in the first interview, and refers to progress made during therapy in the last interview, he must inevitably be supplying and asking for information, which are classified in the descriptive category. During the middle phases of therapy, we are much more likely to see the more familiar reflective, explorative, interpretive operations of the therapist at work——phenomena which would be categorized as evaluative propositions. Tape Segment Sampled Possibly the most significant findings of the present investigation derive from employing the time-sampling method— ology. The results reflect a marked tendency for certain segments of therapy to be characterized more by some content categories than others. These findings suggest some im— portant implications for future sampling techniques. If, for example, a particular investigation is primarily in— terested in studying dependency phenomena, the characteristic method of doing so would be to sample one or two tapes of N subjects. This inevitably results in partial exploration, at best, since the entire process of therapy is not being 74 sampled for a given subject. The results of the present in- vestigation indicate that greater efficiency is obtained by sampling the last segment (where dependency is under con— sideration). This would permit the investigation of a vari— able precisely at the point where it appears in its most potent form and would add power to the investigation. This time—saving procedure would further permit all therapy tapes for a given_§ to be included in the sample. What remains yet to be explored is whether the content (relating to a given variable), appearing during the second ten minutes is comparable to the type of expressions manifested during the last ten minutes. Whatever the answer to this latter question time-sampling still appears to represent a significant new step in assessing psychotherapy interviews. One other result warrants further comment. Thera- pists and clients engage in many more statements during the first ten minutes of the interview than during any other phase. This finding would seem to suggest that the client— therapist interaction is most intense during the initial part of the session, in terms of client-therapist verbal ex— change. Here again, where frequencies of statements is under experimental scrutiny, these results indicate that sampling the first ten—minutes of the interview would prove most profitable. 75 Phase of Treatment This section pertains to the changes occurring in the various content categories over interviews. While the changes which occur for therapist propositions and the thera- pist and client Primary statements have been discussed, there remains one more important observed change——i.e., the nature of the process for the Hostility category. The ini— tial session is characterized by a much greater degree of hostile content than either the middle or last interviews. A possible explanation of this finding centers about the nature of the surroundings in which the present clients find themselves. The large university setting with its attendant impersonality presents very few outlets for these clients to ventilate everyday frustrations. Upon finding an atmosphere in which they can express such feelings, perhaps for the first time, and have these expressions accepted in a non— evaluative fashion may ultimately result in the diminution of hostile statements. The findings obtained by sampling over the first, middle, and final prgtions of therapy point to the serious flaw in those studies employing first interview data alone.“ It is evident that changes occur over therapy. And, any generalizations made to all of therapy on the basis of the first interview alone is clearly unwarranted. First, middle, and last interview data is minimally sufficient for such generalizations. The present investigation 76 sheds some light on the feasability of these generalizations by sampling the first six sessions for all 25_§s. None of the significant changes which obtained as a result of this analysis failed to appear when first, middle, and last data was used. Further, many significant changes were observed with first, middle, and last interview data, suggesting its greater power as a technique. The conclusive answer to the question as to whether or not first, middle and last inter— view sampling is in fact representative of the entire process of therapy is still dependent upon more comprehensive exami— nation. Duration of Therapy The tentative nature of conclusions drawn from first, middle, and last interviews is especially emphasized when long— and short—term therapy clients are included in the sample. Some psychotherapy researchers assume that the form of therapy is equivalent at the respective first, middle, and last points of the process for both short— and long—term clients. The results of the present investigation indicate that there are often enough deviations from this assumption to question its application as a general procedure. While the form is similar for both long- and short—term clients on certain variables, significant differences occur frequently enough (i.e., on three of eight measures) to argue against its haphazard application. While such a method is 77 permissible, it would seem that adequate pilot analysis be performed in order to establish the validity of the as— sumption that the process for a given variable is parallel at the first, middle, and last points for a group of long— and short—term therapy clients. An interesting consideration centers about the possi— bility of future prediction of the form of certain aspects of therapy for long— and short—term clients. If the present results could be replicated with other samples, it would lend further support to the notion that clients who are in therapy for a short time represent a very different population than those who remain in psychotherapy on a long—term basis. And, if the exact form of the differences between these two groups can be demonstrated with other samples, then the strength of the phenomena is greatly enhanced. Personality Data MMPI data obtained prior to testing appeard to differ- entiate groups only somewhat more clearly on a manifest level than did the AC-NAC categorization. While none of the differ— ences achieved statistical significance, there was a tendency for the "deviant" group to express more hostile and dependent verbalizations during therapy. Further investigation in terms of personality data obtained prior to therapy might be a fruitful approach in future research. If predictions con— cerning the form of the process of therapy for a client could 78 be made on the basis of personality variables, the impli- cations for therapy in terms of effecting changes in client behavior are considerable. Implications for Further Research Outcome measures need to be developed which serve to identify the effectiveness of the therapist's endeavors. Further, process variables must be isolated which are shown to effect such outcomes in a predicted manner. Replication of the present study with a different sample, but with the same sampling procedure is essential. If confirmation of the results presented in the present in- vestigation is obtained with other samples, the generaliza- bility of these findings are considerably enhanced. Further, given that this phenomenon represents a stable function, sampling in future research can be more powerfully effected. Additional information needs to be obtained concern- ing the nature of the differences observed in the therapy process for long- and short-term_§s. Do these_§s differ on pre— and post—therapy personality profiles? Does the form look different when using scoring systems other than the ones employed in the present investigation? Further data also needs to be gathered comparing Lennard and Bernstein's (1960) sample with other subjects. The present comparison yielded quite different results from those reported by these authors. Is this a function of the 79 sample, the sampling methods employed, or some complex inter— action of both? How does the process vary as a result of the orientation of the therapist? It is through the answers to these questions that the inches of progress in psychotherapy research may be transformed into yards. CHAPTER VI S UMMARY The present study investigates the form of the pro- cess observed for therapy clients as it relates to aspects of ”Aggression Conflict." And, it examines the process of therapy as it changes during the course of a single thera— peutic interview. Three groups of_§s were identified on the basis of their responses to a Semantic Differential questionnaire. ”Aggression Conflict” represents the condition in which an_§ manifested positive Active—Potent hostility scores, but negatively evaluated such feelings or behavior. .§S exhibit— ing no such conflict (NAC),_§s who resolved such conflict over the course of therapy (ACR), and SS who achieved no such resolution with therapy (ACU), comprised the sample. This study attempts to predict the process involved for those_§s who resolve their ”Aggression Conflict," as com— pared to those who do not. An Attempt is also made to characterize the differences observed in content material for _§s initially identified as belonging to the NAC or AC categories. 80 81 In addition, a major modification in sampling tech- niques is introduced, and an effort is made to confirm the findings of other authors with the present methodolOgy. Ten- minute time segments are randomly selected for every inter— view of each subject included in the study. It was hypothesized that those §s resolving their ”Ag— ression Conflict" should find greater approach, on the part of the therapist to hostility content, than do the ACU_§s. Further hypotheses predict the form of the roles assumed by the therapist over the course of therapy, and also attempt to differentiate the three client groups on the basis of variations in the role patterns which they manifest during therapy. Attention is also given to changes in the form of the types of verbal behavior engaged in by therapists as the therapeutic process progresses. A content analysis is made of ten—minute segments se— lected randomly from 330 tape—recorded psychotherapy inter— views at a university counseling center. These interviews represent the total sample available for 25 self-referred fe— male clients, who were categorized into NAC, ACU, or ACR groups. The results were generally negative. NOne of the tests which employed the ”Aggression Conflict” scale as a means of identifying groups were statistically significant. It seems that this outcome measure does not represent an ade— quate index of the manifest content under investigation. 82 Interestingly enough, however, the time sampling pro- cedures employed in the present study suggest a marked tendency for certain segments of therapy to be characterized more by some content categories than others. This finding has major implications for therapy research, in that sampling might best be performed in future research on that segment of the interview which is known to include more of the particu- lar content under investigation. A significant decrease is found in hostility content over the first, middle, and last sessions. There is some evidence to indicate that it is not al- ways a warranted procedure to assume that first, middle, and last points in the therapies of short-term clients have the same respective points for long—term clients. While this is the case for some variables, it is not for others. Pilot data must establish the legitimacy of this assumption for the particular variables under investigation. The possi— bility of predicting the differences in content between the two groups is raised as a major goal for future research. Contradictory findings observed between this and other investigations argue against the legitimacy of general— izing to larger populations on the basis of results obtained on a single sample. REFERENCES Auld, F. Jr., and Murray, E.J. Content—analysis studies of psychotherapy. Psychol. Bull., l955,_§2, 377—395. Bandura A. Psychotherapy as a learning process. Psychol. Bull., 1961, 28, 143-159. Bandura, A., Lipsher, D.H., and Miller, Paula E. Psycho- therapists' approach-avoidance reaction to patients' expressions of hostility, J. consult. Psychol., 1960, £44) 1-80 Bandura, A., and Walters, R.H. Social Learning and Person- ality Development. New York: Holt, Rinehart and Winston, 1963. Barnes, E.J. Psychotherapists conflicts, defense preferences, and verbal reactions to certain classes of client ex- pressions. Unpublished doctoral dissertation, Michi- gan State Univ., 1963. Binder, A., McConnell, D., and Sjoholm, Nancy A. Verbal con— ditioning as a function of experimenter character— istics. J. abn. soc. Psychol., 1957, 55, 309-314. Breger, L. and McGaugh, J.L. Critique and reformulation of "learning theory" approaches to psychotherapy and neurosis. Psychol. Bull., l965,_§3, 338—358. Buss, A.H., and Durkee, Ann. Conditioning of hostile verbali- zations in a situation resembling a clinical inter— view. J. consult. Psychol., 1958,_22, 415-418. Caracena, P.F., Verbal reinforcement of client dependency in the initial stage of psychotherapy. Unpublished doctoral dissertation, Michigan State University, 1963. Cross, H.J. The outcome of psychotherapy: A selected analy— sis of research findings. J. consult. Psychol., 1964, 28, 413—417. Cutler, R.L. Countertransference effects in psychotherapy. J. consult. Psychol., 1958,_22, 349—356. 83 84 Dollard, J. and Auld, F. Jr., Scoring Human Motives: A Manual. New Haven: Yale University Press, 1959. Dollard, J. and Miller, N.E. Personality and Psychotherapy. New York: McGraw—Hill, 1950. Ferster, C.B. Reinforcement on punishment in the control of human behavior by social agencies. Psychiat. Res. Rep., 1958, 19, 101-118. Frank, G.H. On the history of the objective investigation of the process of psychotherapy. J. Psychol., 1961, _51, 89—95. Frank, G.H. and Sweetland, A. A study of the process of psychotherapy: The verbal interaction. J. consult. Psychol., 1962, 26, 135—138. Freud, S. The dynamics of transference. In Collected Papers, Vol. II, New York: Basic Books, 1959. Goldstein, A.P. Therapist—Patient Expectancies in Psycho— therapy. New York: McMillan Co., 1962. Greenspoon, J. Verbal conditioning and clinical psychology. In A.J. Bachrach (ed.), Experimental Foundations of Clinical Psychology. New York: Basic Books, 1962. Harway, N.I., and Iker, H.P. Computer analysis of content in psychotherapy. Psychol. Rep., 1964, 14, 720—722. Holt, R.R. Experimental methods in clinical psychology. In B.B. WOlman (ed.), Handbook of Clinical Psychology. New York: McGraw—Hill, 1965. Luborsky, L. and Strupp, H.H. Research problems in psycho- therapy: A three—year follow—up. In H.H. Strupp and L. Luborsky (eds.), Research in Psychotherapy, Vol. 2, Baltimore: Trench—Bray, 1962. Kanfer, F.H. Control of communication in dyads by reinforce— ment. Psychol. Rep., 1964, 15, 131-138. Kell, B.L., and Mueller, W.J. Impact and Change: A Study of Counseling Relationships. New York: Appleton- Century-Crofts, 1966. Kopplin, D.A. Hostility of patients and psychotherapists' approach-avoidance responses in the initial stage of psychotherapy. Unpublished master's thesis, Michigan State Univ., 1963. 85 Kopplin, D.A. Eliciting responses in client-therapist inter- action: A content analysis of initial psychothera- peutic interviews. Unpublished doctoral dissertation, Michigan State Univ., 1965. Krasner, L. Studies of conditioning of verbal behavior. Psychol. Bull., 1958, 55, 148-170. Krasner, L. The therapist as a social reinforcement machine. In H.H. Strupp, and L. Luborsky (eds.), Research in Psychotherapy. Washington, D.C.: American Psycho- logical Association, 1962. Lennard, H.L., and Bernstein, A. The Anatomy of Psycho— therapy. New York: Columbia University Press, 1960. Lennard, H.L., and Bernstein, A. Role learning in psycho— therapy. Psychother. theor° res. prac., 1967, 4, 1—6. Lerman, Hannah. A study of some effects of the therapist's personality and behavior and of the clients' reactions in psychotherapy. Unpublished doctoral dissertation, Michigan State Univ., 1963. London, P. The Modes and Morals of Psychotherapy. New York: Holt, Rinehart and Winston, 1964. Marsden, G. Content-analysis studies of therapeutic inter— views: 1954-1964. Psychol. Bull., 1965, 63, 298—321. Mills, D.H. and Abeles, N. Counselor needs for affiliation and nurturance as related to liking for clients and counseling process. J. couns. Psychol., l965,_12, 353—358. Mueller, W.J. and Grater, H.A. Aggression conflict, anxiety, and ego strength. J. consult. Psyghol., 1965, 29, 130-134. Mueller, W.J., and Grater, H.A. A stability study of the ag— gression conflict scale. J. consult. Psychol., 1966, _39, 357—358. Murray, E.J. A content—analysis method for studying psycho— therapy. Psychol. MonOgr., 1956, ZQ_(13, Whole No. 420). Rogers, C.R. Psychotherapy today or where do we go from here. Amer. J. Psychother., 1963, 11, 5-16. 86 Sarason, S.B. The Clinical Interaction. New York: Harper, 1954. Sargent, Helen D. Intrapsychic change: Methodological problems in psychotherapy research. Psychiat., 1961, ‘2—4—3 93-1089 Skinner, B.F. Science and Human Behavior. New York: MacMillan, 1953. Starkweather, J.A. and Decker, J.B. Computer analysis of interview content. Psychol. Rep., 1964, 15, 875-882. Strupp, H.H. The psychotherapist's contribution to the treat— ment process. Behav. Sci., 1958, 3, 34-67. Strupp, H.H. Nature of psychotherapist's contribution to treatment process. Arch. gen. Psychiat., 1960, 3, 219—231. Strupp: H.H. Patient-doctor relationships: Psychotherapist in the therapeutic process. In A.J. Bachrach (ed.), Experimental Foundations of Clinical Psychology. New Yerk: Basic Books, 1962a. Strupp: H.H. The therapist's contribution to the treatment process: Beginnings and vagaries of a research pro— gram. In H.H. Strupp, and L. Luborsky (eds.),_§§- search in Psychotherapy. Washington, D.C.: American Psychological Association, 1962b. Ullmann, L. P., Krasner, L., and Collins, Beverly J. Modifi- cation of behavior through verbal conditioning: Ef- fects in group therapy. J. abn. soc. Psychol., 1961, pg, 128—132. Varble, D.L. An exploratory analysis of hostility in psycho— therapy. Unpublished doctoral dissertation, Michigan State Univ., 1964. Waskow, Irene E. Reinforcement in a therapy-like situation through selective responding to feelings or content. J. consult. Psychol., l962,_26, ll-19. Winder, C.L. Psychotherapy. In P.R. Farnsworth and Q. McNemar (eds.), Annual Review of Psychology. Palo Alto: Annual Reviews, 1957, Vol. 8. Winder, C.L., Ahmad, F.Z., Bandura, A., and Rau, Lucy C. De— pendency of patients, psychotherapists' responses, and aspects of psychotherapy. J. consult. Psychol., 1962, 26, 129-134. APPENDIX A Scoring Manuals (Client dependency and hostility and therapist approach and avoidance response categories represent manuals used in the following studies: Winder, C.L., Ahmad, F.Z., Bandura, A., and Rau, L.C., Dependency of patients, psychotherapist's re- sponses, and aspects of psychotherapy: J. consult. Psychol., 1962, pp, 129-134; Bandura, A., Lipsher, D.H., and Miller, P.E., Psychotherapists' approach—avoidance reactions to patients' expressions of hostility, J. consult. Psychol., l960,_24, 1-8; and modified by KOpplin, D.A. Eliciting re- sponses in client—therapist interactions: (A content analysis of initial psychotherapeutic interviews. Unpublished doctor— al dissertation, Michigan State Univ., 1965. Client and therapist primary and secondary role categories and the thera— pist proposition category derived from scoring systems pre— sented in: Lennard, H.L., and Bernstein, A. The Anatomy of Psychotherapy. New York: Columbia University Press, 1962). I. Scoring Unit and Interaction Sequence. A. Definition. A unit is the total verbalization of one speaker bounded by the preceding and succeed— ing speeches of the other speaker with the ex— ception of interruptions. There are three types of scoring units: the "client statement" (C St), the "therapist re- sponse" (T R), and the "client response" (C R). A sequence of these three units composes an "inter— action sequence.” The client response not only completes the first interaction sequence but also initiates the next sequence and thereby become a new client statement. Example: C. I can't understand how you can stand me. (C St) T. You seem to be very aware of my feelings. (T R) 87 88 C. I am always sensitive to your feelings. (C R) Pauses. If a speaker pauses between statements, his statements are not scored as separate units. The verbalization before and after the pause is considered one unit. Therapist silences are scored as prescribed under III, A, 2, e, of this manual. There are no client silences in this system. Interruptions. Statements of either therapist or client which interrupt the other speaker will be scored only if the content and/or temporal conti— nuity of the other speaker is altered by the interruption. Then the interrupting verbalization becomes another unit and is scored. A non—scored interruption is never taken into account in the continuation of the other speaker. Interruption scored as one unit: C. I asked him to help me and—— T. Why was that? C. -—he refused to even try. Non-interruption scored as 3 units, one inter- action sequence: C. I asked him to help me and—— T. Why was that? C. I don't know. Verbalizations such as "Um hmm," "yes,“ "I see” are ignored in scoring unless they are so strongly stated as to convey more than a listening or re- ceptive attitude. Client requests for the therapist to repeat his response are considered interruptions and are not scored. Therapists" requests of this sort are scored as units (as approach or avoidance of the client statement) except for simple requests to the client to repeat a few words. 89 II. Categories of Client Statements. A. Interpgrsonal Categories: There are three major categories: Dependency, Hostility, and Other. They are scored as exhaustive categories. All discriminations are made on the basis of what is explicitly Verbalized by the speaker in the unit under consideration. l. Hostility: statements include description or expression of unfavorable, critical, sarcastic, depreciatory remarks; oppositional attitudes; antagonism, argument, expression of dislike, disagreement, resentment, resistance, irri- tation, annoyance, anger; expression of aggres- sion and punitive behavior, and aggressive domination. Hostility which the client di— rects at himself is not scored as hostility. The following examples group under a series of types of hostility. These types aid in identi- fication of hostility statements, but are not differentiated in the scoring. a. Anger: expresses or describes feelings or actions which indicates anger. C. I'm just plain mad: C. My uncle was furious at my aunt. b. Dislike: expresses dislike or describes actions which would usually indicate dislike. C. I just don't get interested in them and would rather be somewhere else. C. I've never ever felt I liked them and I don't suspect I ever will. C. He hates editorials. c. Resentment: expresses or describes a per- sistent negative attitude which does or might change to anger on a specific occasion. C. They are so smug; I go cold whenever I think about having to listen to their 'our dog' and 'our son'. Boy: 90 C. I was always jealous of my brother; he was their favorite. d. Antagonism: expresses or describes an— tipathy or enmity. C. It's really nothing definite, but we always seem at odds somehow. C. There is always this feeling of be— ing enemies. e. Opposition: expresses or describes oppo- sitional feelings or behavior. C. If he wants to do one thing, I want to do another. C. It always seems she is against things. She is even against things she wants. C. No, I don't feel that way (in re- sponse to T's assertion). f. Critical attitudes: ,expresses negative evaluations or describes actions which usually imply negative evaluations. C. If I don't think the actors are do— ing very well, I just get up and walk out. C. There is something to be critical about in almost everything anyone says or does. g. Aggressive actions: acts so as to hurt another person or persons either physical- ly or psychologically. C. He deserves to suffer and I'm making it that way every way I can. C. I can remember Mother saying: 'We slap those little hands to make it hurt'. Dependency: dependency statements include ex- pressions of needs to depend on someone; let someone else take the initiative; to be told what to do; to be helped; to be cured by an 91 outside agent; description of dependent be— havior; approval seeking and concern about disapproval; dependent agreement with others; accepting nurturant actions from others; making personal security contingent on an- other; expresses concern about parental plans and expectations regarding the client; ex— presses need to confide in, write to or com— municate with parents; includes discussion of relationship between the client and therapist. The types listed below aid in identification, but are not differentiated in scoring. a . Problem Description: States problem in coming to therapy; gives reason for seek— ing help; expresses a dependent status or a general concern about dependency. C. I wanted to be more sure of myself. That's why I came. C. I wanted to talk over with you my reasons for dropping out of school next quarter. C. Part of the reason I'm here is that everything's all fouled up at home. C. I depend on her——I am tied to her. C. I want to be babied and comforted. Help—seeking: Asks for help; reports ask— ing for help; describes help-seeking be— havior; describes going to see a teacher, therapist or other professional. C. I asked him to help me out in this situation. C. I try to do it when he can see it's too hard for me. Approval—seeking: Requests approval or acceptance; asks if something has the ap— proval of another; reports having done so with others; tries to please another; con— fides in parents for support or approval; expresses fear of "hurting" parents and superiors; asks for or needs support or 92 security; seeks acceptance or approval by achievement; expresses or describes some activity geared to meet his need. C. I hope you will tell me if that is what you want. C. If there was any homework, I did it so Dad would know I was studying like a good girl. C. Is it all right if I talk about my girl's problem? C. That's the way I see it, is that wrong? C. I asked him if I were doing the right thing. Succorance: A wish to be taken care of: to seek another's help when troubled, de— pressed or hurt; to seek encouragement from others; to have others be sympathetic and understanding about personal problems; to go home to see one's parents, feel close to one's parents in the sense of be- ing loved; to receive nurturant behavior from others (especially parents, authori- ties and therapist). C. It looks as if it'll be another lonely weekend, and who's going to cheer me up? C. Instead of studying, I go talk with the guys about my problems. C. I went home just for the day; I told my mother to make a home—cooked meal because that is what I was coming home for. C. My father isn't rich, but he's putting me through college, and he gives me all the money I want; he's always given me everything I've ever wanted. 93 Information—seeking: Asks for cognitive, factual or evaluative information; ex— presses a desire for information from others; arranges to be the recipient of information. C. I asked him why he thought a girl might do something like that. C. I came over here to see about tests you have to offer. I want to know what they say. C. I'm planning to change my major, I'd like to know how to do it. Dependent agreement with another: Re- sponds with dependent agreement with others, rapidly accepts the therapist's reflection; often illustrates therapist's remarks with examples, draws a parallel example to indi— cate agreement; may accept preceding state— ment on authority. C. Oh, yesl You're absolutely right about that. C. Immediately I felt he was right and I had never thought about it that way. Concern about disapproval: Expresses fear, concern, or unusual sensitivity about dis- approval of others, describes unusual dis- tress about an instance of disapproval, insecurity, or lack of support. Little or no action is taken to do something about the concern. C. She didn't ever say a thing but I kept on wondering what she doesn't like about me. C. My parents will be so upset about my grades, I don't even want to go home. C. It seems like I always expect I won't be liked. C. I can't understand how you can stand me when I smoke. 94 h. Initiative-seeking: Asks the therapist or others to initiate action, to take the responsibility for starting something (to start discussion, determine the topic); arranges to be a recipient of therapist's initiative; may solicit suggestions. C. Why don't you say what we should talk about now? C. If you think I should keep on a more definite track, you should tell me. C. I got my advisor to pick my courses for next term. C. Tell me what to do in these circum- stances. 2. Other: ,Inc1udes all content of client's verbalizations not classified as one of the previous two categories. When other is integral to another category, do not score. To score other, the other state— ments must be distinct from the discussion of another major category. C. I grew up in Chicago. C. Everyone ought to have a liberal education. III. Categories of Therapist Responses. A. Approach-Avoidance Response: Therapist responses to each scored client interpersonal category are divided first into two mutually exclusive classes, approach and avoidance responses. When both ap- proach and avoidance are present, score only the portion which is designed to elicit a response from the client. Examples of approach and avoid- ance responses are grouped in sub—types to aid scoring. But the judge must decide first that a response is an approach or avoidance before he considers the various sub-types. 1. Approach responses: An approach response is any verbalization by the therapist which seems designed to elicit from the client further ex- pression or elaboration of the dependent, hos— tile, or other expression which was scored in 95 the client's immediately preceding statement. Approach is to the major category, not neces- sarily to the specific subcategories or the particular content of the client's preceding speech. The following subcategories are exhaustive. a. Exploration (probing): Includes remarks or questions that encourage the client to describe or express his feelings, atti- tudes, or actions further; asks for further clarification, elaboration, descriptive information, continuance; calls for details or examples; probing opinions which direct the client to reconsider by more careful thinking a previous statement. Should de— mand more than a yes or no answer; if not, may be a "label.” C. How do I feel? I feel idiotic. T. What do you mean, you feel idiotic? C. I can't understand his behavior. T. What is it about his behavior you can't understand? T. Tell me more how you felt. T. I don't completely follow that se— quence. b. Reflection: Repeats or restates a portion of the client's verbalization of feeling, attitude or action. May use phrases of synonymous meaning. Therapist may some- times agree with his own previous response; if the client had agreed or accepted the first therapist statement, the second therapist statement is scored as a re- flection of the client statement. Thera— pist finishes client statement in an ob— vious manner. C. I wanted to spend the entire day with him- T. You wanted to be together. C. His doing that stupid doodling upsets me. T. It really gets under your skin. 96 Labeling: The therapist gives a name to the feeling, attitude or action contained in the client's verbalization. May be a tentative and broad statement not clearly aimed at exploration. Includes "bare” interpretation, i.e., those not explained to the client. May be a question easily answered by yes or no. It may be more than a simple clause, but it is a state- ment of fact, opinion or situation without elaboration. C. I just don't want to talk about that anymore. T. What I said annoyed you. C. She told me never to come back and I really did have a reaction. T. You had some strong feelings about that—-maybe disappointment or anger. Interpretation: Points out and explains patterns or relationships in the client's feelings, attitudes and behavior: ex- plains the antecedents of them, shows the similarities and discriminations in the client's feelings and reactions in diverse situations or at separate times. C. I had to know if Barb thought what I said was right. T. This is what you felt earlier about your mother. Support: Expresses sympathy, reassurance, approval, agreement or understanding of client's feelings, attitudes or behavior. Includes strongly emphasized, "Mm Hmm," "Yes"; offers explicit permissiveness. C. It's hard for me to just start talking. T. I think I know what you mean. C. I hate to ask favors from people. T. I can understand that would be difficult for you. 97 C. But, this, I don't know whether I am cheating myself or not. Well, I want to feel, you see. T. You're feeling. C. May I just be quiet for a moment? T. Certainly. C. I have my girlfriend's problems on my mind. Could we talk about them? T. Why don't we talk about that? f. Information: Gives factual information or therapist opinion to general, direct, or implied questions; includes general remarks about the counseling procedure; tells the client what to do; points out that the client's feelings are natural or common (generalization). C. Shall I take tests? T. I feel in this instance tests are not needed. C. What's counseling all about? T. It's a chance for a person to say just what's on his mind. C. I don't like to talk about it. T. Mary, we have to deal with this somehow. Avoidance Responses: The following sub- categories are exhaustive. An avoidance re— sponse is any verbalization by the therapist which seems designed to inhibit, discourage or divert further expression of the dependent, hostile, or other categories. The therapist attempts to inhibit the feelings, attitudes or behavior described or expressed in the im— mediately preceding client statement which de— termined its placement under the major cate— gory. Avoidance is avoidance of the major category, not specific subcategories or psychic states. a. Disapproval: Therapist is critical, sar- castic or antagonistic toward the client or his statements, feelings or attitudes, expressing rejection in some way. May 98 point out contradictions or challenge statements. (NOte: remember you must de— cide response as an avoidance before you consider sub-types). C. Why don't you make statements? Make a statement. Don't ask another question. T. It seems that you came here for a reason. C. Well, I wonder what I do now? T. What do you think are the possibili— ties? You seem to have raised a number of logical possibilities in our discussion. C. I'm mad at him: that's how I feel. T. You aren't thinking of how she may feel. Topic Transition: Therapist changes or introduces a new topic of discussion not in the immediately preceding client verbal— ization. Usually fails to acknowledge even a minor portion of the statement. C. Those kids were asking too much. It would have taken too much of my time. T. We seem to have gotten away from what we were talking about earlier C. My mother never seemed interested in me. T. And what does your father do for a living? Ignoring: Therapist responds only to a minor part of the client response or re- sponds to content, ignoring affect. He misses the point of the client statement May under— or over—estimate affect. May approach the general topic but blatently ignore the affect verbalized. C. You've been through this with other people so help me out, will you? T. You are a little uneasy. C. T. 99 You can see I don't know what to do and I want you to give me advice. Just say whatever you feel is im— portant about that. My older sister gets me so mad I could scream. Mm-hmm. How old did you say she was? We went out for Chinese food; he's so easy to get along with. Is he from New York? Mislabeling: Therapist names attitudes, feelings or actions which are not present in the actual verbalization preceding the response. C. I just felt crushed when she said that. T. Really burned you up, huh? C. I don't know how I felt-—confused, lost-— T. I wonder if what you felt was resentment. C. I may drop out of Honors College. T. You mean Arts and Letters. C. No--Honors College. Silence: Scored when it is apparent that the client expects a response from the therapist but none is forthcoming within five seconds after the client stops talk- ing. If the therapist approaches after five seconds have elapsed, silence cannot be scored and the therapist's response is merely "delayed.” C. If you think I should keep on a more definite track, tell me because I'm just rambling. (five second silence) It is very confusing to know what to do. 100 Role System: We are concerned here in a limited sense with what the statement refers to, with what it is about. The emphasis is in classifying communication as referring to the various systems or role relationships in which the patient and therapist participates. Of particular concern here is the identification of content in which the therapist or the patient discuss their roles and expectations with regard to therapy. (1) Primary System (the treatment). Included in this cate— gory are patient or therapist propositions that refer to their roles during treatment and the process of therapy, and to the purposes, goals, and accomplishments of therapy. Examples: Therapist: There's no homework, no reference work, it's all done here. Therapist: Talk about it anyway even though you think it's unimportant. Therapist: Could you just take a moment to tell me what the specific areas are in which there has been help? Patient: Isn't a therapist able to tell me a thing like this? Patient: I want to be cured. Patient: Was there some misunderstanding about the ap- pointment time? (2) Secondary System, Included in this category are pospo— sitions in which the manifest content refers to therapist and patient in other than their primary roles as patient and therapist. Transference phenomena readily fall into this category. Therapist: Are you projecting doing to me what you did to your father? Therapist. I'm not angry with you. Patient: I want you to treat me as a child. Patient: I was thinking of asking you out to dinner. 101 Therapist Propositions. Bales has developed a system of classification, which permits the scoring of every interaction in terms of one of twelve categories. As he says, "this classification still does not catch 'content' in the usual sense of the term when 'content' is usually taken to mean the 'subject matter'; that is, the reference of the symbols used is the interaction in short 'what' is being talked about. The method employed in the present study tries to classify rather what we might call the 'process significance' of the single interaction; prior acts and acts expected to come." Six of the Bales' categories were condensed into three. The categories used follow: (1) Descriptive Propositions ask for or convey infor- mation. They give or ask for orientation, repetition, or clarification. Examples: "I see that it is six o'clock." ”What does the statement refer to?” "I was angry.” ”Do you have anything in mind?" ”For instance?" (a) There should be no doubt involved. Proposition refers to a "fact." (b) The speaker is in a position to know the fact. (c) Affect is not considered. If an emotion is de— scribed as a fact, it is coded here. (d) There should be a minimum of interpretation, analysis, opinion, or inference involved. (e) General, indeterminate questions are coded here. (2) Evaluative Propositions ask for or convey appraisal or statements of value. They give or ask for opinions, expression of feeling, or analysis. Examples: ”What do you think is wrong with you?” "That's fine." ”I think . . . ,” ”It seems . . . ,” " . . . perhaps . . ." ”You are upset." 102 (a) If doubt is involved, the proposition is coded here. (b) If the speaker is not in a position to know, but is only guessing or inferring, the propo— sition is coded here. (c) The only kind of affect that is necessarily coded here is the expression of preference, wish, value judgement, etc. (d) Propositions involving interpretation, analysis, or inference are coded here. (e) Only questions specifically asking for evalu— ation, interpretation, analysis, opinion, etc., are coded here. (3) Prescriptive Propositions express or ask for di- rectives. They give or ask for suggestions. Examples: "I would prefer that you didn't." "Where should we begin?” "We will have to stop in a few minutes." ”That's something we must consider too." “I would like to hear it all from you." (a) Propositions defining what should or must be done are coded here. (b) Usually therapist "wishes" imply a prescription for the patient and should be coded here. APPENDIX B Mann—Whitney U values for measures comparing two groups of_§s AC—NAC_§s Measure u* Duration of Treatment 11 Number of Statements 12 Dependency l6 Hostility 12.5 Client Primary Statements 10 Client Secondary Statements 20 Approach 12 Avoidance ll Therapist Primary Statements 20.5 Therapist Secondary Statements 21.5 Descriptive 12 Evaluative 12 Prescriptive l6 *Not significant. 103 104 Mann—Whitney U values for measures comparing two groups of SS ACR- NAC _SS Measure u* Duration of Treatment 11 Number of Statements 12 Dependency l6 Hostility 12.5 Client Primary Statements 10 Client Secondary Statements 20 Approach 12 Avoidance ll Therapist Primary Statements 20.5 Therapist Secondary Statements 21.5 Descriptive 12 Evaluative 12 Prescriptive 16 *Not significant. 105 Mann—Whitney U values for measures comparing two groups of_§s ACU— NAC __s_s Measure u* Duration of Treatment 43 Number of Statements 42 Dependency 37 Hostility 37 Client Primary Statements 47. Client Secondary Statements 47 Approach 41. Avoidance 37. Therapist Primary Statements 49 Therapist Secondary Statements 47 Descriptive 39. Evaluative 42. Prescriptive ‘ 47 *Not significant. 106 Mann—Whitney U values for measures comparing two groups of_§s ACU— ACR _ss Measure Duration of Treatment Number of Statements Dependency Hostility Client Primary Statements Client Secondary Statements Approach Avoidance Therapist Primary Statements Therapist Secondary Statements Descriptive Evaluative Prescriptive 20 24 21 25 22 23 24 17 23 l7 16. 24 17 *NOt significant. APPENDIX C Mann—Whitney U values comparing AC and NAC_§s on First sessions for certain measures First Sgssion Measure u Dependency 48 Hostility 49.5 Client Primary Statements 49.5 First Threp Sessions u Dependency 28 Hostility 15 Client Primary Statements 24.5 *Not significant. 107 APPENDIX D Kruskal—Wallis H values comparing ACR, NAC, and ACU §s on last sessions for certain measures Last Spssion Measure H Dependency . .14 Hostility 1.63 Last Three Sessions H* Dependengy . . .34 Hostility .59 *Not significant. 108 APPENDIX E Analyses of variance for response categories - First six interviews Measure .Source 88 DF MS Dependency Sessions 17.47 5 3.49 n.s. _§s 118.02 24 4.92 .§s x Sessions 254.70 120 2.12 Hostility Sessions 29.36 5 5.87 n.s. 1§s 108.11 24 4.50 .§s x Sessions 469.97 120 3.91 Client Primary Sessions 13.55 5 2.71 n.s. _§s 132.29 24 5.51 _§s x Sessions 383.39 120 3.19 Approach Sessions 68.16 5 13.63 n.s. _§s 4205.00 24 175.21 ‘§s x Sessions 3274.84 120 27.29 Therapist Primary Sessions 15.09 5 3.02 n.s. _Ss 277.44 24 11.56 .Ss x Sessions 645.24 120 5.38 Descriptive Sessions 103.58 5 20.72 2.70, _ss 424.98 24 17.70 p < “05 .Ss x Sessions 919.92 120 7.67 Evaluative Sessions 257.63 5 51.53 n.s. _§s 3358.70 24 139.95 _§s X Sessions 2949.54 120 24.58 109 APPENDIX F Analyses of variance for response categories - first, middl e, and last interviews Measure Source ss DF MS F Dependency Sessions 3.12 2 1.56 .s. _Ss 71.92 24 3.00 _§s x Sessions 122.88 48 2.56 Hostility Sessions 10.67 2 5.34 .48, p < .01 .Ss 126.75 24 5.28 .§s x Sessions 103.33 48 2.15 Client Primary Sessions 65.36 2 32.68 .59, p < .01 _§s 575.12 24 23.92 _Ss x Sessions 280.64 48 5.85 Approach, Sessions 28.53 2 14.27 .s. _Ss 1129.52 24 47.06 .§s x Sessions 1286.32 48 26.80 Therapist Primary Sessions 64.35 2 32.18 .78, p < .05 _§s 458.86 24 19.11 _§s X Sessions 322.98 48 6.73 Descriptive Sessions 26.00 2 13.00 .92, p < .05 _§s 357.39 24 14.89 I§s x Sessions 159.33 48 3.32 Evaluative Seséions 107.39 2 53.70 .s. .§s 1051.42 24 43.81 _§s x Sessions 1037.94 48 21.62 llO ““111111131fillifijfluilulflllll“