c. LIBRfi £33.13" Michigan State Univcnity This is to certify that the ‘ thesis entitled THE EFFECTS OF PRE-COUNSELING EXPOSURE TO A MODEL CLIENT ON CLIENT INTERVIEW BEHAVIOR presented by MELVIN L. SCHERPENISSE has been accepted towards fulfillment of the requirements for Ph .D degree in qunsgling Date February 6, 1976 0-7639 we av “1 1 EMS & SUNS 800K BINDERY INC ‘ LIBR mv BINDER 1 33;:4m"" IIGMAI llllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll 3 1293 10040 2571 ABSTRACT THE EFFECTS OF PRE-COUNSELING EXPOSURE TO A MODEL CLIENT ON CLIENT INTERVIEW BEHAVIOR BY Melvin L. Scherpenisse The beginning client for counseling is presented with a unique situation. He is burdened with personal problems and seeks help for these problems from a stranger, his counselor. He often has little notion what is expected from him in his role as a client. Thus the counseling situation can be as problematic for him as the very problems that impelled him to enter counsel- ing. Perhaps if beginning clients were educated in appropriate client behaviors the counseling process would be facilitated. Theoretically, modeling or imitation learning promises to teach appropriate, effective client inter— view behavior and thus reduce the ambiguity of the counseling process and increase the progress of the counseling process. A review of the literature indicates that exposure to a model does affect certain aspects of Melvin L. Scherpenisse client counseling behavior. However, most research regarding pre-counseling presentation of a model has been in college settings. Little research has been done with clients from community mental health centers who primarily had sought counseling and only secondarily volunteered to participate in research. What are the effects of pre-counseling exposure to a model on clients in the naturalistic setting of a rural, small town, com— munity mental health center? Thirty female clients who had never had a pre- vious counseling experience were assigned randomly to either the experimental treatment or the control treat- ment and also assigned to one of three female counselors. The subjects were from a rural, small town community mental health center. The experimental treatment consisted of reading a written, one-page statement describing effective client interview behavior and then listening to an 11-minute audio recording of a female client engaging in good client counseling behavior with a female counselor. The control treatment consisted of reading a one-page statement describing the history of the mental health center and then listening to an 11—minute audio recording of one woman asking guestions about mental health which were answered by another woman. Melvin L. Scherpenisse Audio recordings were made of the first inter- view. To measure longitudinal effects the third coun- seling session was recorded. Four dependent variables were selected to measure the relative effectiveness of the treatments. Two inde- pendent judges were trained to rate the recorded inter- views on the Client Interview Behavior Scale which was developed by the author. The counselors also rated the counseling sessions on the Client Interview Behavior Scale. The independent judges rated the amount of time each subject required to first state a personal problem. The total number of appointments kept was tallied for each subject. Multivariate analysis of variance was performed on the data of the initial interview. Significant dif- ferences (d = 0.05) were not found between treatment means for any of the dependent variables measured: judges' ratings on the Client Interview Behavior Scale, counselors' ratings on the Client Interview Behavior Scale, the number of appointments kept and the amount of time required to first state a personal problem. No interaction among counselors and treatments was found. Because of the high attrition between the first and third interviews and because the counselor for the second and third interview was not necessarily the same as the first interview, the independent variable of Melvin L. Scherpenisse counselor was collapsed for third interview data. Analysis of variance was performed on the dependent variables: judges' ratings on the Client Interview Behavior Scale, counselors' ratings on the Client Interview Behavior Scale, and the time required to first state a personal problem. No significant differences between treatment means were found. Reliability coefficients for the two independent judges on the Client Interview Behavior Scale were approx- imately 0.80. Reliability coefficients for the two judges' ratings of the amount of time required by each subject to first state a personal problem were approx- imately 0.25. The failure to find significant differences between treatments was discussed in the following areas: (1) the construction of the model, (2) the choice of control treatment, (3) the difficulty in obtaining client volunteers, (4) the validity of the dependent variables, and (5) the characteristics of the population from which the sample was drawn. THE EFFECTS OF PRE-COUNSELING EXPOSURE TO A MODEL CLIENT ON CLIENT INTERVIEW BEHAVIOR BY Melvin L? Scherpenisse A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling, Personnel Services and Educational Psychology 1976 ACKNOWLEDGMENTS I would like to thank the members of my guidance committee, Richard Johnson, Ph.D.; Maryellen McSweeney, Ph.D.; James Engelkes, Ph.D.; and Andrew Barclay, Ph.D., who provided much support and encouragement throughout this research. I would also like to thank the many counselors who gave of their time and energy in various aspects of this study. Without their assistance this research could not have been accomplished. ii TABLE OF CONTENTS Chapter Page I. THE PROBLEM AND RATIONALE . . . . . . . 1 Review of the Literature. . . . . . . 5 Bandura's Theory of Observational Learning . . . . . . . . . . 5 Other Theories of Observational Learning . . . . . . . . . . 9 The Problems of the Client . . . . . 13 Treatment Methods Related to Modeling. . 16 The Use of Modeling in Counseling . . . 18 smary C O O O I O O O O O O O 28 II. DESIGN OF THE STUDY . . . . . . . . . 30 Rating Scale for Client Interview Behavior . . . . . . . . . . . 30 Weighting the Rating Scale . . . . . 32 Reliability of the Rating Scale. . . . 33 Intercorrelations Within the Client Interview Behavior Scale . . . . . 35 Sample. . . . . . . . . . . . . 36 Power Test. . . . . . . . . . . 36 Subjects . . . . . . . . . . . 36 Treatments . . . . . . . . . . . 37 Preorganizer . . . . . . . . . . 37 The Model . . . . . . . . . . . 38 The Control Treatment . . . . . . . 39 The Counselors . . . . . . . . . . 40 The Independent Raters . . . . . . . 40 Training the Raters. . . . . . . . 41 iii Chapter Criteria Measures . . . . . . . . Judges' Ratings on Client Interview Behavior Scale. . . . . . . . Problem Statement Time. . . . . Counselor Ratings on the Client Inter- view Behavior Scale . . . . . . Appointments Kept . . . . . . . Procedure . . . . . . . . . . Research Hypotheses . . . . . . . Statistical Model . . . . . . . . Multivariate Analysis of Variance . . Order of Presentation in Terms of Potential Significance for the F Step Down Ratios . . . . . . Summary . . . . . . . . . . . III. ANALYSIS OF RESULTS . . . . . . . . First Interview. . . . . . . . . Reliability of Judges . . . . . Counselor - Judge Reliability . . . Counselor Main Effects. . . . . . Third Interview. . . . . . . . . Reliability of Judges . . . . . . Counselor Judge Reliability . . . . Summary . . . . . . . . . . . IV. SUMMARY OF RESULTS, DISCUSSION OF RESULTS, CONCLUSIONS AND IMPLICATIONS . . . . Summary of Results. . . . . . . . Discussion of Results. . . . . . . The Sample. . . Procedure . . . The Model . . Interaction . . The Control Treatment . . . . . . iv Page 42 42 42 43 43 43 44 45 45 46 47 49 49 57 59 60 64 69 70 7O 72 72 72 72 77 78 80 80 Chapter Page Client Interview Behavior Scale Reliability and Validity . . . . . 81 Problem Statement Time . . . . . . 82 Number of Appointments Kept. . . . . 83 Third Interview. . . . . . . . . 84 Dependent Variable. . . . . . . . 85 Conclusions. . . . . . . . . . . 85 Implications for Further Research . . . 86 APPENDICES APPENDIX A. DATA FOR WEIGHTING CLIENT INTERVIEW BEHAVIOR SCALE. . . . . . . . . . 87 B. PILOT STUDY DATA. . . . . . . . . . 88 C. PARTICIPATION REQUEST FORM . . . . . . 91 D . DEMOGRAPHIC DATA O O I I O 0 I o o o 9 3 E. CHARACTERISTICS OF THE GOOD CLIENT. . . . 94 F. PREORGANIZERS AND THEIR RANKINGS . . . . 95 G. RATINGS OF POTENTIAL MODEL TAPES . . . . 101 H. EXPERIMENTAL TREATMENT. . . . . . . . 102 I. CONTROL TREATMENT . . . . . . . . . 113 J. EXPERIMENTAL DATA . . . . . . . . . 121 REFERENCES I O O O O O C I O I O O O O 1 2 3 LIST OF TABLES Table Page 2.1 Hoyt's Estimated Reliability . . . . . . 34 2.2 Client Interview Behavior Scale Subtest Intercorrelation Matrix . . . . . . . 35 3.1 Variable Matrix and Cell Means . . . . . 50 3.2 Treatment Combined Means . . . . . . . 52 3.3 F Ratio for Multivariate Test of Equality of Mean Vectors for Main Effects Associated with Treatment . . . . . . 54 3.4 F Ratio for Multivariate Test of Equality of Mean Vectors for Main Effects Associated with Treatment . . . . . . . . . . 55 3.5 F Ratio for Multivariate Test of Equality of Mean Vectors for Interaction Effects Among Counselors and Treatment. . . . . 56 3.6 F Ratio for Multivariate Test of Equality of Mean Vectors for Interaction Effects Among Counselors and Treatment. . . . . 58 3.7 Counselor - Judges Reliability . . . . . 60 3.8 Counselor Combined Means . . . . . . . 61 3.9 F Ratio for Multivariate Test of Equality of Mean Vectors for Main Effects Associated with Counselors. . . . . . . . . . 62 3.10 F Ratio for Multivariate Test of Equality of Mean Vectors for Main Effects Associated with Counselors. . . . . . . . . . 63 3.11 Variable Matrix and Cell Means . . . . . 66 3.12 Analysis of Variance - Judges' Rating . . . 67 vi Table Page 3.13 Analysis of Variance - Counselor Ratings . . 68 3.14 Analysis of Variance - Problem Statement Time . . . . . . . . . . . . . 68 4.1 Interview Means . . . . . . . . . . 75 vii CHAPTER I THE PROBLEM AND RATIONALE Much has been debated about counseling, that vague and nebulous process to which many subjectively commit their allegiance but to which many others objec- tively raise suspicions of blind faith. That counseling works would be conceded; but it is also generally agreed that counseling produces greatest results with intelli- gent, verbally skilled, middle class clients whose psy— chological problems are not too severe. The concensus implies that counseling works best with those who need it least. The problem is how can counseling be made more effective for those clients who do not enter coun- seling with the entire array of skills, behaviors, and attitudes that enhance the probability of counseling success. Intuitively, the uniqueness of the counseling situation seems to be a deterrent in the process's efficiency. Is it not possible that the first set of interviews function significantly to teach the client appropriate interview behavior: behavior which will facilitate problem presentation, actualize personal resources, and effect resolution? The general indirect approach used by many counselors would make learning the appropriate client role difficult. The client emits the desired behavior largely by chance and learns grad- ually by nonspecific reinforcement criteria which way of doing things yields the best consequences. As long as counselors teach and clients learn appropriate interview behavior, it would be better if such learning took place by explicit means rather than by implicit and vague means. Lennard and Bernstein (1960) analyzed the first 50 sessions of psychotherapy with eight neurotic cases considered typical of office practice. During the first three sessions 20% of all therapist communications and 15% of all patient communications were classified as primary role system communications. By the fourth month, the percentages were less than 8%. This decrease in role discussion suggests that counselors, at least implicitly, use the initial interviews to educate their clients in appropriate interview behavior. The clients studied by Lennard and Bernstein (1960) were typical of office practice. Undoubtedly these were rather psychologically sophisticated clients who had some idea of the client role. In a later study of very disturbed clients, Lennard and Bernstein (1967) found only 5% of therapist communication and only .003% of client communication during the first four interviews dealt with primary role system communication. As the authors state, the reduction of expectational discrepan— cies is a crucial functional requirement of a counseling interactional system. In addition, the greater the dis- turbance of the client, the greater his need for role information, and the greater his deficits in obtaining this information. Given that knowledge of his role as a client would improve his counseling process, what kind of behaviors comprise a good client? Every counselor has, however vaguely, a mental image that defines a good client. Although there are individual differences among counselors in their defi- nitions of the good client characteristics, five factors emerge as crucial (Parloff, 1956; Goodstein & Grigg, 1959; Rogers, 1967; Rogers & Dymond, 1954; Snyder, 1961). The good client engages in self-exploration. He talks, states the problem, and discusses his attitudes, beliefs, feelings, and motives. The good client assumes responsi- bility. He sees himself as the one who will make the primary changes and does not blame others for his present predicament or expect changes in others to effect solutions for his problem. The good client has trust in his counselor and perceives his counselor as one who will help. The good client is motivated. He is willing to incur anxiety in order to effect change, does not demand easy solutions, and is not overly dependent upon his counselor. The good client is honest. He tries neither to maximize nor minimize his problems, does not try to gain sympathy by exacerbating his symptoms nor try to impress the counselor with psychological assets. The problem remains: few, if any, clients incorporate all these desirable characteristics. The primary purpose of counseling is to change behavior, whether adding new behaviors to the client's repertoire or modifying existent behavior. An effective means of accelerating behavioral change is vicarious learning. Practically all learning which results from direct experiences can occur vicariously through the observation of another person's behavior and its conse- quences. Modeling procedures can eliminate behavioral deficits, reduce excessive fears and inhibitions, trans- mit self-control, and facilitate learning of social interaction (Bandura, 1969). The specific concerns of this investigator are the uses of modeling techniques to increase effective client behavior in the counseling interview itself. Many, if not all, clients come to their first counseling session with past histories of being punished for self- disclosure, fears of being rejected or having weaknesses, or confusion regarding how to behave with a benign helper. For some clients the counseling interview pre- sents as many difficulties as the problem situation which impelled them to seek help. If successful coun- seling involves the transfer of behaviors learned in the interview to the outside world, then increasing adaptive client interview behavior should increase adaptive client reality-behavior. Therefore, it is proposed that modeling techniques teach beginning clients the interview behaviors that will produce the greatest benefits from the counseling relationship. Review of the Literature Bandura's Theoryiof Obser- vational Learning The theory assumed in this paper will be that of Albert Bandura (1969). He purports that observational learning can be understood by positing two representa- tional systems, imaginal and verbal. Modeling stimuli are coded into images or words in the memory system where they function as mediators for response retrieval and reproduction if a given stimulus situation is dis- criminative for the reward of the given response. Through the process of sensory conditioning model- ing stimuli elicit percuptual responses that have been sequentially associated with the modeling stimuli. For example, mention of a person's name will elicit the perceptual image of that person even though he is not physically present. Recall of the imaginal mediators serve as templates for reproduction of matching responses. The great speed and retention of modeled contents in man is a function of man's ability to verbally code observed events. In fact, verbal coding could be con- ceptualized as a model itself. Simply telling a person what to do often is as effective as showing him. Bandura (1969) elaborates on the major subpro- cesses that govern modeling. The attentional processes are important because stimulus contiguity must be accom- panied by discriminative observation in order to produce learning. In social learning, model attributes play an important role. On the basis of past experience, sub- jects discriminate among model attributes that differen- tiate probabilities of reinforcement. Models are more likely to command attention if they have demonstrated high competence, are purported experts, are celebrities, and possess status—conferring symbols. Characteristics such as age, sex, social power, and ethnic status affect the degree of attention given to the model. Characteristics of observers also are associated with different observational patterns. Dependency, self- esteem, level of competence, and socio-economic status are significant variables. Persons who have been fre- quently rewarded for engaging in matching response are likely to give high attention to modeling cues. Attentional behavior can be increased and focused by appropriate incentive conditions. If the person is informed that he will later have the opportunity to be rewarded for matching the model's behavior, such an incentive usually will override the effects of other variables. The second basic component in observational learn- ing is the retention process. In tasks that rely strongly on symbolic functions, covert rehearsal increases the retention of the acquired matching responses. Maccoby (1959) suggests that covert rehearsal plays an important role in some social interactions. Fantasy role-playing serves as a defensive function in which the person attempts to control his behavior toward models who possess controlling power so that he may avoid their potential punishment and gain their potential rewards. He would not practice covertly for engagement with a powerless model. Likewise the beginning client is likely to ruminate potential personal behaviors that will result in his counselor's potential favorable reactions and avoid his potential aversive reactions. Again in the retention process, symbolic coding operations are crucial. In fact the effectiveness of covert rehearsal may be more attributable to increased symbolic organization rather than increased associative strength. ...\. awn ngu . o.. ‘ -.-.,. ‘m by fin. 'V .‘l‘ “r L [U (j The third component of modeling phenomena involves the transfer of the symbolic representations into overt performances. Persons lacking some of the necessary behavioral components will produce partial factors of the model's behavior, but gradually, in structured and logical steps, the complete behavior set can be estab- lished. Accurate imitation of a good client-model's cues may be difficult since the counseling process attends much to internal cues that are not readily observable. However, the model's general verbal cues can differen— tiate general areas of affect and activity in which the client can discriminate his own personal and unique behaviors, e.g., the observer can learn to express anger toward his mother even though the model expressed anger toward his father. The last major component of the modeling phenome- non is the incentive or motivational process. Incentive operates in the attention, retention, and reproduction phases. Observers are not passive robots directed by environmental cues, but active actors who select behaviors to imitate according to their utilitarian value. The modeling process is not automatic but can be adversely affected by failures in sensory registration, poor trans- formation of modeled events to symbolic modes of repre- sentation, retention difficulties, motor deficiencies, or unfavorable conditions of reinforcement. Other Theories of Observational Learning Associative and classical conditioning theories. Holt (1931), Piaget (1952), and Allport (1924) explain modeling as a function of the temporal contiguity between the modeling stimuli and the imitator's matching response. When the model and observer engage in mutual repetitive behavior, strong associative stimuli are set up so that if the model emits a novel response the observer will imitate that behavior. Bandura (1969) comments that although the associ— ative theories may describe the process in which an observer imitates with a response already in his behavioral repertoire, they fail to account for the psychological processes involved in the acquisition of novel responses. Furthermore, modeling sequences do not typically begin at the lowest level of desired behavior. In teach— ing a child a new verbal response, the response itself is presented, not some approximation. The associative theories cannot explain the rapidity of language develop- ment. Reinforcement theories. The reinforcement theories assume that the occurrence of observational learning is contingent upon reinforcement of the imitated behavior. Miller and Dollard (1941) posit that a moti- vated subject must be positively reinforced as he matches 10 a model's response in order for observational learning to occur. The observer is assumed to emit random, trial and error responses which are gradually differentiated through selective reinforcement process. Skinner (1953) agrees closely with Dollard and Miller but focuses his explanation on a stimulus matching concept. When a person is positively reinforced for correct matching and negatively reinforced or punished for aberrant matching, the behavior of a model will become discriminative stimuli for reinforcement. Gewirtz and Stingle (1968) conceptualize modeling as a matching to sample phenomenon. In the matching to sample paradigm a person chooses from an array of stimuli the stimulus that has a common characteristic with the sample stimulus. The problem with using a matching to sample para- digm to explain modeling is that the accurate stimulus discrimination is not equivalent to performing the associated behavior. To know what is not necessarily to know how. In answer to the general criticism that persons continue to respond appropriately in settings absent of reinforcement, the reinforcement theorists counter that if accurate reproduction of modeling stimuli is con- sistently rewarded, then the behavioral similarity per se acquires secondary reward value. 11 Again, the same criticism that was levied against the classical conditioning theorists can be applied to the reinforcement theorists. While reinforcement can explain how modeling responses are controlled, it does not explain their acquisition in the first instance. To rely on trial and error, chance occurrences of behavior before reinforcement can be used to solidify or control behavior would necessitate long time periods to form desired behavioral sets. Additionally, if learning is trial and error followed by reinforcement, what processes determine which responses are to be emitted "randomly"? Affective feedback theory. The role of rein- forcement is also critical in the sensory feedback theory of Mowrer (1960). He adds classical conditioning variables to his reinforcement theory of modeling. He classifies two types of imitation learning: the subject is directly reinforced and the subject is vicariously reinforced. When the subject is directly reinforced by the model, the repeated association of matching the model's responses and rewarding experiences take on positive value for the subject. Because of stimulus generali- zation, the subject later can produce self-rewarding experiences by matching the model's behavior. Thus classical conditioning spans the gap between the actual 12 reinforcement in the presence of the model and the con— ditioned reinforcement in the absence of the model. In the second empathetic type of imitative learn- ing, the model not the observer, receives reinforcement. The observer intuitively identifies with model's sensory and emotional states following the reinforcement. The observer vicariously experiences the sensory effects of the reinforcement that the model receives. The observer then will be predisposed to match the model's behavior so that he can experience the reward directly. The main difference between Bandura's theory and the other theories seems to be other theories' emphasis on some type of reinforcement. Bandura does include the effects of reinforcement and thus his theory seems more comprehensive. Beginning clients exposed to models of appropriate client interview behavior certainly will experience some rewards for matching the model's behavior. If being exposed to a model reduces client confusion, gives him a sense of structure, and provides information regarding how to act, then these effects would reduce his anxiety level and could be considered reinforcing. However, Bandura's concepts of attentional processes, retentional processes, and transference of symbolic coding to overt performance provide a more inclusive theoretical frame- work by which to understand the modeling process. 13 The Problems of the Client The one thing that all clients have in common is that they have a problem which they hope counseling can help to some degree. They are unique in their psycholog— ical gestalts and interpersonal histories. Patterson (1958) believes that clients enter counseling with varied expectations concerning the role they will play in counseling. In turn, the role that they assume affects the role that they expect the coun- selor to play. An attitude of passive dependency with its deference to authority will significantly influence the counseling process, usually negatively. Or a client may perceive counseling as an educational process and expect the counselor to behave didactically as a teacher. In general, Patterson feels that clients interact with counselors with attitudes and behaviors historically rooted in their past experiences with authority figures. Goodstein and Grigg (1959) in a rather adamant article argue that unconscious determinants are only part of the processes that influence client expectations. These authors imply that clients' expectations do affect the counseling relationship but are too individual to be treated generally. They state that it is necessary for clients to learn to accept responsibility and to learn to be independent. Direct handling by the counselor of 14 the client's expectations is considered essential to realizing these goals of independence and responsibility. Bordin (1955) places the emphasis on the type of problem as determining the client's expectations. He compared clients with personal adjustment problems to clients with education-vocational problems. The latter more than the former perceived the counselor as a source of information. Kelley (1955) believes that how the client pre— conceives psycho-therapy will affect both the role he will play and the role he expects the therapist to play. Although these authors disagree as to the spe- cific sets with which clients enter counseling, they do agree that the client's perceptual set does influence the counseling process. Not only does the client's pre- conception effect counseling, but also his behavioral sets affect his counselor. Parloff (1956) underscores this point with his finding that the counselor relates most satisfactorily to those clients who most clearly approximate his conception of an "Ideal Client." Many clients then are in a dilemma. They are unclear about their own and their counselor's expec- tations, and their lack of good counseling behaviors may make it more difficult for their counselors to relate to them. o-r '5‘ - ‘u6 .3... a: 'ogu 'l‘ (I) F: 5.. (I! 'S Lvd -‘l I 15 Orne and Wender (1968) claim that "anticipatory socialization" of clients before psychotherapy will improve psychotherapeutic outcome. Anticipatory social- ization is the process by which the client is given information describing the psychotherapy interaction; what to expect from the psychotherapist and what the psychotherapist expects from the client. These authors believe that the counseling relationship is a social relationship. Any social relationship is likely to be less frustrating and more satisfying if the participant is aware of the expectations of the situation. Heller (1969) states that modeling procedures have the ability to remove inhibiting anxiety responses and to teach new behaviors. As applied to counseling, modeling would ease client fears of self-revelation and provide information on how to behave in the counseling setting. A client who is not motivated to change provides a very negative stimulus to many, if not most counselors. Lack of motivation for change traditionally has been considered a low prognostic indication, but early termi- nation with little change is often a result of inadequate preparation for counseling rather than the result of a nebulous entity named no-motivation. Warren and Rice (1972) gave training in counseling behavior to high drop- out, low prognosis clients and obtained positive results. N! Ah‘lfi 16 Thus, exposure to a model of good client behaviors would benefit the client. Observing a model would give him information about what to expect from the counselor, what the counselor expects of him, and what behaviors will enable him to gain the most help from the counseling process. Such information would lower his anxiety level. A lowered anxiety state would enable the client to draw upon his personal resources. With the inhibitory effects of anxiety lessened, the client would be better able to use his cognitive, emotional, and behavioral skills. Grigg and Goodstein (1957), studying clients three years after they had terminated counseling, found that clients who reported having felt relaxed and comfortable during counseling perceived the most favorable results from their counseling. Treatment Methods Related to Modeling Role therapy. Kelly's (1955) fixed role therapy is essentially based upon modeling procedures. The therapist draws up a personality sketch which the client enacts. The roles typically involve behavior sets that contrast with the client's neurotic sets. For example, a passive person may be asked to assume an assertive role. The brief experimental simulation minimizes the initial threat of making massive changes in the life style. The therapy involves much rehearsal of the new 17 behavior and the role-reversal with the therapist. Thus the client learns skillful ways of relating to others and by role-reversal he experiences how others will experience or reach to the behavior being modeled. Corsini (1966) states that role playing is not only beneficial to the person engaged in the role playing but it also benefits persons watching. Since the observers are likely to have the same problem, knowing that others have the same problem and seeing how others handle the problem will be beneficial. Behavioral rehearsal. Lazurus (1963, 1966) used a procedure called play acting. A client would rehearse the behaviors which he ordinarily would have difficulty performing in real life. Thus a person having difficulty asserting himself toward a parent would practice several different assertive scenes with the parent numerous times before having an actual assertive interaction with that parent. The question could be raised whether or not a person engaging in this play acting is serving as his own model. Counseling training. Carkhuff (1971) emphasizes the importance of training the client in how to act in the counseling situation. In fact, he feels that the training is counseling. The same behaviors that allow 18 the client to function at high levels in counseling also permit the client to function at high levels in his personal life. Counselor as model. The integrity therapy espoused by Mowrer (1964) and Drakeford (1967) emphasizes that the person is morally accountable for his objec- tionable, irresponsible, and wrongful behavior. A major characteristic of an integrity therapist is that he con- sistently strives to model self-disclosure and personal responsibility. Kell and Mueller (1966) underline the importance of the counselor setting examples of mature psychological behavior. They emphasize that the counselor himself is an important model for the client. The client is likely to imitate the thinking, feeling, and risk-taking pro- cesses of the counselor. The Use of Modeling in Counseling Group counseling. Truax and Carkhuff (1965) developed a vicarious therapy pretraining (VTP) audio tape which consisted of actual group therapy interaction in which the group members were exploring, relatively deeply, their problems and feelings. Two studies subjected inpatient groups to VTP (Truax & Carkhuff, 1965; Truax et al., 1966). These researchers concluded that groups given VTP showed 19 greater depth of self-exploration in early therapy sessions when compared to groups not given VTP. The VTP patients also showed ideal self-concept change in the direction of greater social awareness, while the non-VTP patient's ideal self-concept changed in the opposite direction. In a replication study with outpatient group therapy, Truax and Wargo (1969) found that clients exposed to vicarious therapy pretraining (VTP) before beginning group therapy showed greater personality change as measured by a lO-test battery than did non-VTP clients. The authors also concluded that the VTP clients engaged in more self-exploration which resulted in better thera- peutic outcome. Also studying group therapy, Swartz and Hawkins (1965) presented one group with models that used verbali- zations characterized as affective statements. A second group was presented with models emitting statements characterized largely as nonaffective statements. A third group received no models. Although the group therapists reinforced affective statements in all groups, the first group showed the greatest increase over time in affective statements. Group two showed less affective statements over time than did the control group. Whalen (1969) used four treatment conditions: (1) a film showing interpersonal openness, accompanied 20 by detailed written instructions (F-Dl); (2) the film showing interpersonal openness, accompanied by minimal written instructions (F-Ml); (3) detailed instructions only (D1); (4) minimal instructions only (Ml). After receiving one of the four conditions, male college stu- dents participated in leaderless, group discussion. She concluded that the F-Dl groups exhibited more inter- personal openness than the other three groups. Both the film model and the detailed instructions were necessary for modeling to have effective results, thus showing the value of instructing the subjects in what to look for in the model. In an extensive study, Krumboltz and Thoreson (1964) compared individual reinforcement counseling, group reinforcement counseling, individual reinforcement counseling preceded by presentation to an audio taped model reinforced by his counselor (model—reinforcement counseling), group model reinforcement counseling, indi- vidual control film discussion, group control film dis- cussion, and inactive control. The model-reinforcement and reinforcement counseling procedures produced the most information seeking behavior outside the counseling interviews. Model-reinforcement counseling surpassed reinforcement counseling for boys exposed to a male model. Group counseling equaled individual counseling, but many interactions were found. 21 This investigator does not think that the study adequately tested model reinforcement counseling against reinforcement counseling. The model, a high school male, discussed law, college athletics, and ROTC. Such a narrow array of model topics is unlikely to be relevant to a majority of boys and of even less interest to girls. Higgins et al. (1970) compared the effectiveness of three treatment sequences in teaching the direct, mutual communication skills usually used in encounter groups: (1) media therapy: programmed text, video models, live models, and feedback; (2) programmed text and video models; (3) reading material only. The full treatment group showed the most improvement in their communication skills. This study appears to indicate that modeling has an effect in teaching communication skills related to counseling; however, the models are buried with too many other variables to state confidently the exact effect of the models. Individual counseling. Marlat et al. (1970) designed a study to determine the effects of exposing observers to a live, problem discussing model who received either encouraging, discouraging, or neutral feedback from a third person. An unstructured interview was used to assess imitative behavior (problem state- ments). The neutral feedback condition and, to a lesser extent, the encouraging feedback condition yielded more 22 problem statements compared to the negative feedback and neutral conditions. The authors interpreted the results with the concept that an individual will use a model's behavior to reduce response uncertainty when faced with an ambiguous task. In a follow-up study Marlat (1971) studied matching behavior as a function of four conditions: (1) high ambiguity task with presentation of a model, (2) low ambiguity task with presentation of model, (3) high ambituity task--no model, (4) low ambiguity task--no model. Subjects exposed to one of these con- ditions were then interviewed. The number of problem admission statements was analyzed for each condition. The model condition produced more problem admission statements than the no model condition. Furthermore, the high ambituity-model condition produced more problem admission statements than the low ambituity-model con- dition. Marlat suggests that the greater the ambituity of a situation, the more a person will match the responses of a model. Hoehn-Saric et a1. (1964) studied a procedure called The Role Induction Interview which was an informal interchange between patient and interviewer regarding: a general explanation of psycho—therapy: a description of what the client could expect from the therapist and what the therapist would expect from the client; a preparation “no .415. v ‘6 3.51:1 23 for typical events in the process of therapy, such as anxiety, resistance, etc.; an idea of what the client could expect in terms of improvement within a given time period. The rationale was that subjects who received The Role Induction Interview would present themselves as better clients, which would cause the therapists to respond to them more favorably, and thus initiate a mutually positive feedback cycle resulting in a more favorable outcome. The authors concluded that the role induction subjects exhibited significantly more "good client" behaviors than the controls. Also, an analysis of the tapes from the first sessions indicated that the therapist gave control subjects significantly more role induction than the experimental subjects. Using a color-sound film model of a 20-minute counseling session, Long (1968) found that clients exposed to this pre-training model did better on the criteria of quantity of words spoken, number of first person pronouns used, the number of present tense verbs used, and the scores on a depth of self-exploration scale than did the control group. Even though Long did obtain significant results using the dependent variables that he used, this investi- gator feels that guantity of words and language syntax 24 are arbitrary. The statement, "I think counseling is not for me,” would qualify as good client behavior under Long's criteria. Granted the example is extreme, but Carnes and Robinson (1950) also had difficulty with client talk time. After analyzing 78 counseling interviews, they found a low positive correlation between amount of client talk and growth of client insight and between amount of client talk and the client-counselor working relationship. A higher positive correlation was found between amount of client talk and client responsibility for the progress of the interview. However, the authors noted that the causal relationship between the amount of client talk and desirable interview outcome was not clear. There- fore, the use of client talk time as a criterion of counseling outcome or effectiveness is rather tenuous. In an attempt to improve self-disclosure during the initial interview, Goi (1971) subjected clients to either an audio tape containing instructions about and examples of self-disclosure, an audio tape with general counseling information, or no treatment. He obtained no significant differences among the treatments. How- ever, the subjects were college students seeking coun- seling at their university counseling center. It may be postulated that these subjects did have high prior levels 25 of self-disclosure skills which would be increased minimally by exposure to a model. Brody (1968) tested the effectiveness of three modeling procedures in modifying the frequency of self- referent affect (SRA) responses in quasi counseling interviews with nonvolunteer subjects. Sixty female college freshmen and sophomores were asked to express their feelings about their first year of college. They were assigned to one of four treatment conditions: (1) passive listening to a model- ing tape in which the model described her feelings toward her first year of college. The subjects in this group listened to the tape in the presence of the experimenter who was the model on the tape; (2) modeling interaction, in which the experimenter emitted 15 SRA modeling state- ments while verbally interacting with the subjects; (3) modeling reinforcement in which the experimenter reinforced with modeling statements on an intermittent schedule 15 SRA responses by each subject in this group; (4) no treatment. Each interview consisted of a lO-minute period to determine base rates of SRA responses, an experimental period in which the various treatments were given, and a 15-minute post experimental period. Brody found the modeling reinforcement procedure to be the most effective in increasing the SRA rate 26 during both the experimental period and the post experi- mental period. The model tape was effective only in maintaining the base rate of responding compared to the control group which decreased their SRA responding over the interview. The model interaction procedure produced nonsignificant decrements over the three periods of the interview. Methodologically, this study is flawed by the experimenter being both the model and the "counselor" in her study. It would have been difficult for her not to have given out subtle cues as to which treatment con— dition was her favorite. Myrick (1969) developed a model in which a male counselor talks to a male student. Both audio and video tapes were made. The tape begins with an introduction by a narrator who discusses what counseling is, the role of the counselor, and development of counseling (1 min. 45 sec.). The client and counselor interact. The client begins by focusing on others, then shifts his focus to self-reference. The counselor reinforces self-reference (4 min. 10 sec.). The narrator interrupts the inter- action. He points out how the client began by blaming others, but then shifted to talking about himself. He suggests what to look for in the next segment (1 min. 10 sec.). Client engages in more self-references. The counselor reflects and clarifies. At the 27 conclusion, the counselor reinforces the client for talking about self. The client reports the experience as beneficial (5 min. 15 sec.). The narrator reinforces the client's use of self-references. He emphasizes the role of the client. He concludes with discussion of confidentiality and the role of the counselor. Thirty subjects watched the model on video tape, 30 subjects listened to the model on audio tape, and 30 subjects served as controls. Exposure to the model increased self-reference. The audio taped model gave the best results; however, the video taped model worked best for the girls and the audio taped model worked best for the boys. Krumboltz et a1. (1967) found that exposure to a video taped model produced more information seeking behavior than the control treatment. Different models incorporating different levels of counselor attentiveness and prestige failed to produce the differences that the authors had hypothesized. Pre—organizers. A pre-organizer is some device which points out what will be modeled. A pre-organizer tells the subject what to look for in the model. In a sense a pre-organizer is a model, or a model for a model. Truax and Carkhuff (1967) found that pre- organizers enhance the effects of models. Myrick (1969) interspersed pre-organizers within his model. Goi (1971) 28 included instructions about self-disclosure. The Role Induction Interview (Hoehn—Saric et al., 1964) could be considered a pre-organizer only condition. Whalen (1969) found instruction a necessary antecedent for modeling to be effective. Eskedal (1975) used modeling with pre-organizers to train counselors. He used a split screen 30-minute video tape of an experienced counselor establishing a positive relationship with a client. He assigned master's level counseling students to one of four groups: (a) symbolic role model with attentional cues, (b) symbolic role modeling only, (c) control with attentional cues, and (d) control only. Measuring acquisition of counseling skills with a true false instrument, he found both modeling treatments statisti- cally significant over the two control conditions. The modeling with attentional cues produced significantly more counseling skill acquisition than modeling only. Summary The beginning client is faced with the unique setting of counseling for which he may or may not have the appropriate behavioral skills. Since counseling can be conceptualized as a learning process, any method that would enhance his learning is likely to accelerate counseling progress. 29 Theoretically, modeling or imitation learning promises to teach beginning clients behaviors which would increase the probability of satisfactory counsel- ing results. A review of the literature indicates that exposure to a model has effects on counseling behavior. However, most research has been directed toward college students. A few studies have been executed using models with group counseling clients from the general popu- lation. No study was found using subjects from rural mental health centers. No research appears to have been done on clients from noncollege settings who primarily sought help for personal problems and only secondarily volunteered to participate in imitation learning research. Consequently, it is proposed to expose beginning clients at a rural, Small town community mental health center to an audio model of appropriate client interview behavior. The purpose is to determine if exposure to a model will have a positive effect on new clients' subsequent counseling session behavior. CHAPTER II DESIGN OF THE STUDY Rating Scale for Client InterviewfiBehavior The author along with three master's level coun- selors, each with five years experience in counseling, developed a rating scale to measure client interview behavior. The items designated with an asterisk were taken from Rogers (1967). (SELF-EXPLORATION:) looks to self for answers rather than to external sources expresses angry feelings expresses positive feelings about self expresses negative feelings about self expresses anxiety is able to express ambivalent feelings expresses feelings of dependency talks about frustrations discusses and admits own psychological needs *talks about his own reactions in or to the problem situation 30 31 *talks about the contribution of his own reaction to the problem *talks about his own understanding of his feelings, experiences, and attitudes (ASSUMES RESPONSIBILITY:) takes initiative, responsibility for talking sees self as one who will make primary changes does not blame others for his situation *talks about his direct involvement in a problem situation or event (TRUST COUNSELOR:) is comfortable in counseling situation expects or assumes that the counselor likes her appears to like counselor receptive when counselor gives "negative" feedback looks to counselor anticipating help has confidence in counselor (MOTIVATED:) talked about anxiety provoking content realizes that solutions may be difficult willing to incur anxiety in order to change is not overly dependent upon a counselor (HONEST:) tries to keep problem in perspective neither minimalizes nor maximizes problem 32 does not try to gain sympathy does not try to impress counselor communicates the affective aspects of the problem in addition to the cognitive aspects insights include emotional as well as cognitive components Each item on the scales is scored on a five-point scale. A score of 1 indicates very low frequency of the characteristic. A score of 2 indicates low frequency of the characteristic. A score of 3 indicates a moderate frequency of the characteristic. A score of 4 indicates a high frequency of the characteristic. A score of 5 indicates a very high frequency of the characteristic. Weighting the Rating Scale To obtain some measure of validity and to deter- mine if the items on the rating scale should be weighted, 10 counselors with a minimum of a masters degree in a counseling field plus two years experience were given the following instructions: Each of the following sheets contains a factor for measuring client interview behavior. Please rate each item under that factor independently. To rate independently means that the items should not be compared to each other. Use the following scale: 1 - not important 2 - of low importance 3 - of moderate importance 4 - of high importance 33 The mean of the ratings was 3.3 with a standard deviation of 0.4. Seven of the 31 items of the rating scale fell i one standard deviation from the mean (Appendix A). The range was 2.4 to 3.9. Considering the above statistics and the difficulty in finding a rationale for assigning weights, it was decided to leave the items of the rating scale unweighted. Given a mean rating of 3.3 (above moderate importance), the items comprising the Client Interview Behavior Scale appear to have content validity. Reliabilitygof the Rating Scale To provide estimates of interjudge reliability using the Client Interview Behavior Rating Scale, a pilot study was performed. Five counselors from a rural mental health center were asked to submit four audio recordings of client interviews ranging from poor to excellent in interview behavior. Each counselor held a masters degree and had worked at the center at least two years. Each counselor used his own criteria in selecting the four clients that he audio recorded. The 20 tapes were rated on the Client Interview Behavior Scale by the two judges used in the research. One tape was inaudible. Hoyt's analysis of variance to estimate reliability was applied to the results (Winer, 34 1962). The preferred estimate is that estimate removing systematic differences due to judges: = 1 _ MS residual k MS between people r The reliability coefficients are presented in Table 2.1. Table 2.1 Hoyt's Estimated Reliability Scale Interjudge Reliability Self-Exploration .67 Assumes Responsibility .52 Trusts Counselor .72 Motivated .64 Honest .88 Total Scale .78 n = 19 The raw data and computations can be found in Appendix B . From the reliability coefficients (Table 2.1) it ‘was concluded that the judges could obtain good relia- IJility using the Client Interview Behavior Scale. These data serve as an estimate of the interjudge reliability liY the same judges using the same rating scale but on a different sample of people. 35 Intercorrelations Within the Client Ifiterview Behavior Scale While the rating scale was being developed, it was apparent that the five sections of the rating scale could be overlapping. Also the items of each section could have been placed in more than one section. The placement of items was usually subjective and sometimes arbitrary. The Pearson product moment correlation coefficient was applied to the pilot study data to determine the intercorrelations among the five sections of the rating scale (Hayes, 1963). On the basis of the high intercorrelations shown in Table 2.2, the five subtests were collapsed into one scale. Table 2.2 Client Interview Behavior Scale Subtest Intercorrelation Matrix Trusts . Self- Honest Counselor Motivated Exploration Assumes Responsibility 0.84 0.86 0.87 0.81 Honest 0.78 0.90 0.83 Trusts Counselor 0.77 0.72 Motivated 0.79 36 Sample Power Test Using the pilot study data (Appendix B) a deter- mination of statistical power was made. Not only did this investigator want the experimental treatment effect to be statistically significant, but he also wanted it to be large. If the treatment had an effect, it should have a great enough effect to make its use practical in a counseling session. With practical application a primary concern, it was decided that the experimental treatment should yield a standard deviation difference between the experimental and control means. Applying the formula: 2 92 = n(u1-uz) 2J02 (see Appendix B) it was determined that 15 subjects each in the experimental and the control group would give a power of 0.80 for detecting differences of l g at the 0.05 alpha level. Power is defined as the probability that the decision rule rejects the null hypothesis when an alternate hypothesis is true. Subjects All female clients making an appointment for their initial counseling session at a rural mental health center were asked by the secretary if they had 37 had any type of counseling previously. Those who had not experienced counseling were assigned randomly to one of three female counselors and then randomly assigned to either the experimental or the control treatment. Ran- domization for counselor was made by rolling a die: 1 or 4 = counselor one; 2 or 5 = counselor two; 3 or 6 = counselor three. Randomization for treatment was made by flipping a coin. When the client appeared for her first interview she was given by the secretary a client participation form to read (Appendix C). Between December 1974 and August 1975 30 out of 50 new female clients agreed to participate in the research. Demographic data are presented in Appendix D. Treatments Preorganizer Three master's level counselors with a minimum three years professional experience were given a page elaborating good client interview behavior (Appendix E) and were asked to write a narrative, appropriate for clients, describing these characteristics. Three master's level counselors with a minimum two years experience then ranked the resulting three narratives. The narrative with the highest average ranking was designated as the preorganizer (Appendix F). The preorganizer was read by a secondary school teacher and an elementary school teacher to determine reading 38 level. Each had over 10 years experience. They agreed that a ninth grade reading level would be sufficient to comprehend the preorganizer. The Model Three women who had completed counseling at the mental health center to their own satisfaction and then had participated in empathy training in conjunction with their volunteer activities at the mental health center's telephone crisis line were coached in good client behavior. They were told the purpose of this research, given the article on good client characteristics (Appen- dix E), and given an opportunity to ask questions. No questions were asked. They were not instructed as to what problems to present. It was more important that they exhibited the good client characteristics in a spontaneous, natural manner, rather than covering the many problems of a composite client in an unnatural, artificial manner. An auditory model was chosen because the tasks or skills to be demonstrated were verbal. A video model would have shown posturing and eye contact, cues that could have been distracting from the verbal cues to be imitated. Each of the women then entered an audio recorded counseling session with a female counseling intern from the mental health center. The counseling intern was ”1' ..le 1. .n ‘l’ (I) 39 selected because of her excellent professional repu- tation and her nondirective style of counseling. She was instructed to be as nondirective as possible. The rationale was to have the counselor on the audio tape model to engage only in behaviors all counselors usually have in common and thereby keep to a minimum the variable of the model's counselor. The resulting three tapes were then rated on the Client Interview Behavior Rating Scale by three master's degree counselors with a minimum of two years experience. The tape with the highest rating was selected as the Model Tape (Appendix G). The Control Treatment Finding a control treatment that would not have a negative effect or positive effect on the control sub- jects was very difficult. Many approaches were con- sidered but all seemed to be biased in favor of either the control group or the experimental group. The most neutral was selected. The control equivalent to the experimental preorganizer was a one-page article relating the history and function of the mental health center. The control equivalent of the experimental audio model was a recording of equal length (11 minutes) in which one woman asked questions about counseling and a second woman gave the answers. The questions and answers were taken from Wolberg (1967). See Appendix I. 40 The Counselors Three female counselors working at a rural mental health center served as counselors for the subject's initial interview. Counselor one was in her middle forties. She was finishing her master's degree in social work. She described her counseling orientation as psy- choanalytic. Counselor two was 28. She had five years experience beyond her master's degree in social work. She described her counseling orientation as eclectic. Counselor three was 46. She recently completed her master's degree in rehabilitation counseling. She described herself as a transactional analysis adherent. All three counselors were unaware of the nature of the research in which they had participated. They were satisfied with the explanation that the research focused on improving client interview behavior. The Independent Raters Rater one had four years experience as a volun- teer and empathy trainer at a college town walk-in and telephone crisis center. She completed her master's degree in social work midway through this study. Rater two had four years experience as a volun- teer and empathy trainer at the same crisis center as rater one. She also finished her master's degree, in rehabilitation counseling, midway through this study. 41 Both women obtained employment as mental health therapists at a community mental health center upon receipt of their degrees. They were paid four dollars an hour for rating the tapes of this study. They were chosen because of the quantity of counseling interactions they had had with a broad spec- trum of crisis problems, and because of the further counseling process awareness they had to have in order to train others. Training the Raters Before the raters listened to the pilot study tapes, they spent one hour with this investigator dis- cussing the Client Interview Behavior Rating Scale. Their rating of the pilot study tapes was in effect further training. Before rating the experimental tapes they spent one hour with this investigator reviewing the Client Interview Behavior Rating Scale. Then they independently rated three audio recorded counseling sessions. After rating each of these training tapes they compared ratings and discussed their rating differences. By the third training tape, they had less than 10% difference in their total scores. 42 Because of an unexpected delay of six weeks in receiving the second half of the experimental tapes, the raters met by themselves to review the rating scale. Criteria Measures Four dependent variables were measured to deter- mine treatment effects: the judges' ratings of the counseling sessions on the Client Interview Behavior Scale, the judges' determination of the time required by the clients to first state a problem, the counselors' ratings of the counseling session on the Client Interview Behavior Scale, and the number of appointments each client kept. Judges' Ratings on Client Interview Behavior Scale The two independent judges rated the audio recordings of the subjects' counseling sessions using the Client Interview Behavior Scale. The items on the scale were grouped in varied order to reduce generali- zation effects. The judges' ratings were averaged. The range of possible scores was 31 to 155. Problem Statement Time The two judges tallied the time required for each subject to first state a problem. The criterion was that the client stated any personal problem or that 43 the counselor stated a client problem and the client agreed. Problem statement time was measured in minutes. Counselor Ratings on the Client Interview Behavior Scale The counselors rated each client on the Client Interview Behavior Scale. To insure no rater contami- nation, the counselors were instructed not to rate a client if she revealed anything at all above her experi- mental treatment. Four counselors ratings were not obtained because the clients mentioned to her counselor something about their treatment either before or after the recorded interview. No client mentioned her treat- ment during the recorded interview. Appointments Kept The number of counseling appointments each sub- ject kept was tallied (l, 2, or 3). Each client was given another appointment after the first and second interview. Procedure After the client signed the participation form, she was taken to a private room where she was given the appropriate experimental (Appendix H) or control package (Appendix I) by the secretary. The secretary told her to open the door when she finished listening to the tape and her counselor would come for her. 44 The first 30 minutes of the counseling session were audio recorded. After the initial interview the client was scheduled with a counselor for her future sessions. This counselor was not necessarily the counselor who did the initial interview, but the counselor next on the Center's rotation list. To assess the effects over time of exposure to a model, the clients that came for their third counseling session were audio recorded. When approximately one-half of the initial inter- views were completed, the recordings of these interviews along with any recordings of third interviews were given to the two independent judges for ratings. Six weeks later the judges rated the remaining data. The judges did not know whether the subject had been in the experi- mental group or the control group nor did they know whether the interview was the first or the third. Research Hypotheses Hypothesis 1: The experimental group will score significantly higher than the control group on the criteria measured at the initial interview. Hypothesis 2: The experimental group will score significantly higher than the control group on the criteria measured at the third interview. 45 Hypothesis 3: There will be no interaction among counselors and treatments. Statistical Model Multivariate Analysis of Variance Since this study uses multiple variables which are at least intuitively related, Multivariate Analysis of Variance (Whitla, 1968) was used. The multivariate F ratio indicates whether or not all the dependent variables when considered together as one dependent variable are significant at a given alpha level for a hypothesis being tested. The univariate F ratio can also be obtained for each of the dependent variables associated with the hypothesis. However, the univariate F ratio assumes that each dependent variable is indepen- dent of every other dependent variable of this study. Violation of this assumption raises the alpha level to an unknown degree. Because this study cannot assume that the dependent variables are independent, the step down F ratio was used. The step down F ratios for each dependent variable can have their exact alpha level determined prior to the statistical analysis of the data. Thus, the probability of type 1 error can be «adjusted for the effect of the other dependent variables. If four dependent variables were used in a study, the ill '1 I“ I ' '1! "J! F? 46 appropriate alpha level for each would be the alpha level of the multivariable F ratio divided by the number of dependent variables or .05/4 = .0125. Thus, the null hypothesis would not be rejected for any dependent variable unless the value of the obtained step down F ratio exceeded the tabled value of F at alpha equal to .0125. The order that the dependent variables are entered in multivariate analysis of variance is important. Once a variable is found to be significant the variables that follow are not testable because they violate the assumption of independence. Thus, it is important that the dependent variables be ordered regarding their respective contributions. Order of Presentation in Terms of Potentingignificance for the F Step Down Ratios 1. judges' ratings on Client Interview Behavior Scale 2. counselors' ratings on Client Interview Behavior Scale 3. time client requires to definitively state problem 4. number of appointments kept When analyzing the third interview Client Inter- View Behavior Scale ratings, the score is the difference Fifi-qt viii-l.“ ans 1.1! (II (I) ’2 (I) ' 1 CEI‘. (I) r) 47 between the rating on the third interview and the first interview for the judges' ratings. A difference score is not used for the counselor's rating because the coun- selors are not necessarily the same for both interviews. Summary Thirty female clients who had never had previous counseling experiences were assigned randomly to either an experimental treatment or control treatment and also assigned to one of three female counselors. The subjects were from a rural, small town community mental health center. The experimental treatment consisted of a written statement describing effective client counseling behavior (preorganizer) and an audio recording of a female client engaging in good client counseling behavior with a female counselor. The control treatment consisted of a written statement describing the history of the mental health center when the counseling took place and an audio recording of a woman asking questions about mental health which were answered by another woman. Audio recordings were made of the initial coun- seling interviews. To measure longitudinal effects, the third counseling session was recorded. Four dependent variables measured the relative éaffectiveness of the treatments. Two independent judges 'u’ere l Zlier.‘ rated a If." 1" JVAHA h :f ti 48 were trained to rate the recorded interviews on the Client Interview Behavior Scale. The counselors also rated the counseling sessions on the Client Interview Behavior Scale. The independent judges rated the amount of time each subject required to first state a personal problem. The total number of appointments kept was tallied for each subject. CHAPTER III ANALYSIS OF RESULTS First Interview The variable matrix and cell means are shown in Table 3.1. Each of the three counselors interviewed five clients exposed to the experimental treatment and five clients exposed to the control treatment. Four measures were taken on each client: Ml - the average judges' rating; M2 - the counselors' rating; M3 - the problem statement time, determined by the judges; M4 - the number of appointments kept. Because data were missing for three subjects on one or more of the dependent variables, these subjects were deleted. Thus, the design is unbalanced. A multivariate analysis of variance1 was per- formed with the probability of rejecting the null hypo- thesis when it is true set at 0.05. For the F step down ratios an equal partition of the alpha level was made lFinn program titled "Univariate and Multivariate Analysis of Variance, Covariance, and Regression, Version 4, June, 1968." 49 50 Table 3.1 Variable Matrix and Cell Means Independent Dependent Variables Variables Treatment Counselor M1 M2 M3 M4 E 1 80.7 94.0 3 7 2.0 E 2 110.4 113.4 1.7 2.6 E 3 94.6 90.0 2.2 2.6 C 1 101.3 127.3 2.0 2.3 C 2 112.8 120.4 1.6 2.6 C 3 92.8 90.4 2.7 2.2 N 27 E Experimental C Control M1 Judges' Ratings Averaged M2 Counselor Ratings M3 Problem Statement Time (minutes) — Number of Appointments Kept 51 for each of the four dependent variables. A step down F was considered significant for p < 0.0125 (0.05/4). The hypotheses were considered by the Finn pro— gram in the following order: (1) interaction between treatment and counselor, (2) main effects treatment, and (3) main effects counselor. The interaction hypothesis was tested first because if the model is to have practical significance, clients exposed to a model must do better than clients exposed to the control no matter who coun- sels them. Of crucial concern is that a counselor does not interact negatively with model treated clients. The hypotheses are presented in the anticipated form even though analysis of variance is nondirectional. Appropriate post hoc tests will be performed to infer direction for significant results. Hypothesis 1: The experimental group will score significantly higher than the control group on the criteria measured at the initial interview. The means for treatment effects are in the oppo- site direction predicted for the first three dependent variables: average judges' rating on the Client Inter- view Behavior Scale, the counselor's rating on the Client Interview Behavior Scale, and the average judges' (ietermination of problem statement time. The means for Ilumber of appointments kept was in the direction pre- ciicted (Table 3.2). 52 Table 3.2 Treatment Combined Means Dependent Variables Independent Variables M1 M2 M3 M4 Experimental 97.5 99.9 2.4 2.5 n = 13 Control 102.4 111.6 2.1 2.4 n = 14 N = 27 M1 = Judges' Ratings Averaged M2 = Counselor Ratings M = Problem Statement Time (minutes) M4 = Number of Appointments Kept 53 The multivariate F ratio for treatment main effects was equal to 1.1701. This F ratio is not sig— nificant at the 0.05 alpha level. None of the step down F ratios are significant at the 0.0125 alpha level. The univariate F ratios are presented, but are ignored because the four dependent variables cannot be assumed to be independent of each other (Table 3.3). To get some idea of the effects of the unbalanced design, cell means were used as estimators of the missing scores. The multivariate F ratio and the four step down F ratios did not yield significant differences (Table 3.4). Thus, it appears that the unbalanced design did not influence the failure to find significant statistical differences. The null hypothesis for treatment means was not rejected. Hypothesis 3: There will be no interaction among counselors and treatment. The multivariate F ratio for interaction effects of counselor and treatment was 0.6721. It is not sig- nificant at the 0.05 alpha level. None of the step down F ratios were significant at the 0.0125 alpha level. The univariate F ratios are presented but are ignored because they cannot be assumed to be independent (Table 3.5) . 54 Table 3.3 F Ratio for Multivariate Test of Equality of Mean Vectors for Main Effects Associated with Treatment F ratio for multivariate test of equality of mean vectors = 1.1701 d.f. = 4 and 18 p < 0.3569 g = 0.05 . Hypothesis Univariate Step Variable Mean Sq. F p < Down F < Judges' Rating 254.303 0.836 0.37 0.836 0.371 Counselor Rating 1292.915 5.205 0.033 4.163 0.055 Problem Statement Time 0.880 0.202 0.66 0.031 0.863 Number Appointments Kept 0.038 0.061 0.81 0.026 0.874 Note. d.f. for hypothesis = 1; d.f. for error = 21; N = 27; a = 0.0125. 55 Table 3.4 F Ratio for Multivariate Test of Equality of Mean Vectors . for Main Effects Associated with Treatment (missing scores estimated with cell means) F ratio for multivariate test of equality of mean vectors = 0.8786 d.f. = 4 and 21 p < 0.4935 g = 0.05 . Hypothesis Univariate Step Variable Mean Sq. F p < Down F p < Judges' Rating 163.333 0.578 0.455 0.578 0.455 Counselor Ratings 853.333 3.391 0.078 2.750 0.111 Problem Statement Time 0.0403 0.009 0.923 0.017 0.899 Number Appointments Kept 0.300 0.474 0.497 0.366 0.552 Note. d.f. for hypothesis = l; d.f. for error = 24; N = 30; a = 0.0125. 56 Table 3.5 F Ratio for Multivariate Test of Equality of Mean Vectors for Interaction Effects Among Counselors and Treatment F ratio for multivariate test of equality of mean vectors = 0.7207 d.f. = 8 and 36 p < 0.6721 g = 0.05 . Hypothesis Univariate Step Variable Mean Sq. F p < Down F p < Judges' Rating 169.107 0.556 0.582 0.556 0.582 Counselor Ratings 466.341 1.877 0.778 1.381 0.274 Problem Statement Time 1.843 0.422 0.661 0.182 0.835 Number Appointments Kept 0.206 0.329 0.723 0.919 0.417 Note. d.f. for hypothesis = 2; d.f. for error = 21; N = 27; a = 0.0125. 57 The multivariate F ratio and the four step down F ratios did not yield significant differences when the data were balanced by using cell means as estimators of missing scores (Table 3.6). The null hypothesis for interaction effects was not rejected. It was concluded that any combination of counselor and treatment did not significantly affect treatment outcome. Reliability of Judges To determine if the failure to find significant differences between treatment means could have been due to low reliability of the measures, Hoyt's analysis of variance was applied to the two independent judges' ratings on the Client Interview Behavior Scale to deter- mine reliability (Winer, 1962). The preferred statistic is that estimate removing systematic frame of reference for the individual judges: _ l _ MS residual MS between people = 1 _ 3840 18726 = 0.795 The reliability is sufficiently high to decrease likeli- hood of low reliability between judges as a reason for not 1Finding significant differences between treatment means. 58 Table 3.6 F Ratio for Multivariate Test of Equality of Mean Vectors for Interaction Effects Among Counselors and Treatment (missing scores estimated with cell means) F ratio for multivariate test of equality of mean vectors = 0.4535 d.f. = 8 and 42 p < 0.8813 g = 0.05 . Hypothesis Univariate Step Variable Mean Sq. F p < Down F p < Judges' Rating 154.033 0.545 0.557 0.545 0.587 Counselor Ratings 391.233 1.555 0.232 1.127 0.341 Problem Statement Time 0.785 0.188 0.830 0.054 0.947 Number Appointments Kept 0.100 0.158 0.855 0.243 0.786 Note. d.f. for hypothesis = 2; d.f. for error = 24; N = 30; q = 0.0125. 59 The Hoyt's reliability coefficient for the jIJdges' rating of problem statement time was: = 1 MS residual MS between people Trhis reliability is relatively low. Thus, it is possible ‘that there were differences between treatment means of problem statement time that the judges were unable to reliably measure. It should be noted that separate judge reliability ‘was calculated for the first and third interview ratings even though the judges rated the interviews concurrently and were unaware of their order. Counselor - Judge Reliability The judges' averaged ratings on the Client Inter- view Behavior Scale were compared to each of the three counselors' ratings on that scale using Hoyt's reliability coefficient (Winer, 1962). The results are presented in Table 3.7. Thus, when a counselor's ratings of her clients ‘was compared to the judges' ratings of those clients, strong positive reliabilities were not found. Counselor one did have a fairly high reliability coefficient, but 60 the other two counselors obtained relatively low relia- bilities. However, the counselors were not trained in the use of the rating scale and were participants in the interviews which they rated. Also, the counselors did their ratings over a nine—month period, while the judges did their ratings in two weekends about a month apart. Table 3.7 Counselor - Judges Reliability Counselor 1 Counselor 2 Counselor 3 0.72 -0.04 0.46 Counselor Main Effects No hypotheses were made regarding counselor effects, even though the counselors selected had dif- ferent levels of experience and presumably different levels of effectiveness. Since counselor effects might be of interest to some, a multivariate test of equality of mean vectors for main effects of counselors was per- formed (see Table 3.8, 3.9, and 3.10). The multivariate F ratio for counselor effects was 3.2596, which had a probability < 0.0057. Thus, taking all the dependent variables together simul- taneously, the hypothesis that the counselor main effects means are equal is rejected. 61 Table 3.8 Counselor Combined Means Dependent Variables Independent Variables M1 M2 M3 M4 Counselor 1 92.4 113.0 2.7 2.1 n = 7 Counselor 2 111.6 116.9 1.7 2.6 n = 10 Counselor 3 93.7 90.2 2.5 2.4 n = 10 Judges' Ratings Averaged Counselor Ratings Problem Statement Time (minutes) Number of Appointments Kept 62 Table 3.9 F Ratio for Multivariate Test of Equality of Mean Vectors for Main Effects Associated with Counselors F ratio for multivariate test of equality of mean vectors = 3.2596 d.f. = 8 and 42 p < 0.0057 g = 0.05 . Hypothesis Univariate Step Var1able Mean Sq. F < Down F P < Judges' Rating 1150.034 3.780 0.0397 3.780 0.0397 Counselor Ratings 1910.027 7.689 0.0032 5.772 0.0105 Problem Statement Time 3.167 0.725 0.4963 0.274 0.7637 Number Appointments Kept 0.459 0.733 0.4924 1.866 0.1835 Note. d.f. for hypothesis = 2; d.f. for error = 21; N = 27; a = 0.0125. 63 Table 3.10 F Ratio for Multivariate Test of Equality of Mean Vectors for Main Effects Associated with Counselors (missing scores estimated with cell means) F ratio for Multivariate test of equality of mean vectors = 3.2596 d.f. = 8 and 42 p < 0.0057 g = 0.05 . Hypothesis Univariate Step Variable Mean Sq. F p < Down F p < Judges' Rating 1144.433 4.049 0.031 4.049 0.031 Counselor Ratings 2226.900 8.849 0.001 7.492 0.003 Problem Statement Time 3.525 0.844 0.443 0.297 0.746 Number Appointments Kept 0.633 1.000 0.383 1.661 0.214 Note. d.f. for hypothesis = 2; d.f. for error = 24; N = 30; a = 0.0125. 64 Looking at the step down F ratios, the judges' rating failed to reach the 0.0125 alpha level. The next dependent variable, counselor ratings, was significant beyond the 0.0125 alpha level. The third and fourth dependent variables cannot be considered because once a significant step down ratio is obtained, making decisions on the lower ranked step down F ratios would violate the assumption of independence. No post hoc tests were done to determine the direction of counselor differences because they are of little consequence to this study. Furthermore, it cannot be determined whether a counselor's ratings represent the effectiveness of her counseling or the severity of her evaluating her clients. Third Interview To assess the effects of exposure to a model at the initial counseling session, the same dependent variables were measured at the third counseling inter- view. Only 15 of the 30 subjects remained in counseling for at least three sessions. In analyzing the data, the counselor factor was collapsed to prevent cell sizes from being very small and very unequal. Because data were missing on some of the depen- dent variables, use of multivariate analysis of variance would have resulted in further loss of subjects (n=13). 65 Thus, univariate analysis of variance was used (Winer, 1962). Since the dependent variables in the study cannot be assumed independent each from the other, use of the univariate F ratio does raise the level of alpha to an unknown degree. Any significant results will have to be interpreted with caution. Hypothesis 2: The experimental group will score significantly higher than the control group on the criteria measured at the third interview. The dependent variable, averaged judges' rating on the Client Interview Behavior Scale, for the third interview was the difference of that rating between the first and third interview for each subject (Table 3.11). The mean for the experimental treatment was +4. The mean for the control treatment was +0.143. The univariate F ratio for equality of treatment effects was not significant at the 0.05 alpha level (Table 3.12). The null hypothesis of equal treatment means as measured by judges' ratings was not rejected. Following agency policy, after the initial interview each subject was assigned to a counselor for the remainder of their sessions. This counselor could have been the initial interview counselor or one of three male counselors. 66 Table 3.11 Variable Matrix and Cell Means Independent Dependent Variables Variables M1 M2. M3 Experimental +4.0 101.63 4.0 Treatment n = 8 n = 8 n = 8 Control +0.14 92.5 2.29 Treatment n = 7 n = 6 n = 7 M1 = Difference Average Judges' Ratings between Interview 3 and Interview 1 3 ll 2 Counselor Ratings 3 ll 3 Problem Statement Time (minutes) 67 Table 3.12 Analysis of Variance - Judges' Rating Source of Variation SS d.f. MS F Treatments 55.5 1 55.5 0.25 Experimental Error 3062.9 14 218.8 F.95 (1,14) = 4.6 The mean counselor rating for the experimental treatment was 101.6. The mean counselor rating for the control treatment was 92.5 (Table 3.11). The univariate F ratio for equality of treatment effects was not significant at the 0.05 alpha level (Table 3.13). The null hypothesis of equal treatment means as measured by counselor ratings was not rejected. The judges determined the time each subject took to state a problem (Table 3.11). The mean time for the experimental treatment was 4 minutes. The mean time for the experimental for the control treatment was 2.3 minutes. The univariate F ratio for equality of treat- ment effects was not significant at the 0.05 alpha level (Table 3.14). The null hypothesis of equal treatment means as measured by problem statement time was not rejected. 68 Table 3.13 Analysis of Variance - Counselor Ratings Source of Variation SS d.f. MS F Treatment 285.5 1 28 0.14 Experimental Error 2695.4 13 207.3 F.95 (1,13) = 4.67 Table 3.14 Analysis of Variance - Problem Statement Time Source of Variation SS d.f. MS F Treatment 11 1 11 2.6 Experimental Error 59.9 14 4.3 F.95 (1,14) = 4.6 69 Reliability_of Judges The Hoyt reliability coefficient for the judges' ratings on the Client Interview Behavior Scale was: = l _ MS residual MS between people _ 172.29 —649".—‘25 ll H = 0.74 This reliability is sufficiently high to decrease the likelihood of low reliability between judges as a reason for not finding significant differences between treatment means. The reliability coefficient for the judges' ratings of third interview problem statement time was: _ 1 _ MS residual MS between people 7.59 10.13 = 0.25 This reliability is relatively low. Thus it is possible that there were differences between treatment means of problem statement time but the judges were unable to reliably measure that difference. 70 Counselor Judge Reliability The independent judges' averaged ratings on the Client Interview Behavior Scale and the six counselors' ratings on that scale were compared using the Pearson product moment correlation coefficient (Hayes, 1963). The product moment correlation was 0.11. Thus, the judges and the counselors do not correlate well on their ratings of the same subjects. However, the counselors were not trained and they were rating interviews in which they were participants. Also the counselors rated their interview over a period of nine months, while the judges rated the interviews in two weekends separated by about a month. Summary Multivariate analysis of variance was performed on the data of the initial interview. Significant dif- ferences were not found between treatment means on any of the dependent variables measured: judges' ratings of the Client Interview Behavior Scale, counselors' ratings on the Client Interview Behavior Scale, the number of appointments kept, and the amount of time required to first state a personal problem. No inter- action among counselors and treatments was found. Because of high attrition between the first and third interviews, the independent variable of counselor was collapsed and analysis of variance was performed on 71 each of the dependent variables measured at the third interview. No significant differences between treatment means were found. Hoyt's analysis of variance to estimate relia- bility was performed on the judges'ratings of the Client Interview Behavior Scale. The reliability was suffi- ciently high to make it unlikely that differences were not found because the judges could not reliably measure client behavior. Hoyt's estimates of reliability were applied to the judges' and counselors' ratings on the Client Inter- view Behavior Scale for the initial product moment cor- relations were performed between the judges' ratings and counselors' ratings on the Client Interview Behavior Scale for the third interview. These reliabilities were low. Hoyt's estimate of reliability was performed on the judges' ratings of the time the subjects required to first state a personal problem. These reliabilities were relatively low. CHAPTER IV SUMMARY OF RESULTS, DISCUSSION OF RESULTS, CONCLUSIONS AND IMPLICATIONS Summary of Results Exposure to a model of good client behavior as opposed to exposure to questions and answers regarding mental health did not produce significant differences in beginning female clients' interview behaviors as measured by two independent judges' ratings on the Client Inter- view Behavior Scale, the counselors' ratings on the Client Interview Behavior Scale, the amount of time required to first state a problem, and the total number of appoint- ments kept. Significant differences were not found at the first interview or the third interview. Discussion of Results The Sample The results of this study do not support a hypothesis that modeling is effective for improving the interview behavior of small town, rural female clients who come for counseling by their own choice. 72 73 The subjects of this study had made prior decisions to enter counseling and only secondarily agreed to partici- pate in this research. This research is unique in that the setting is a community mental health center. Most studies research- ing the effect of modeling in counseling are set in uni- versities. University students might well be better able to copy model behavior because of their more sophisticated intelligence, familiarity with psychology, and their greater experimental exposure. University students may be better skilled at deciphering what is expected of them. On the other hand the client of the community mental health center has come for her own pressing con- cerns and may not be interested in or able to puzzle out what the agency expects of her. She is more of a consumer than the university student. Of 50 new clients at the rural mental health center who were asked to participate in the research, only 30 agreed to participate. Thus, 40% of new client population were not studied. How the inclusion of these clients would have affected the results is not known. It is possible that the clients who declined to participate are the clients who would benefit most by exposure to a model. It may have been that these clients were more overwhelmed by anxiety than the volunteer clients. Thus the prospect of participating in research 74 only increased their anxiety even though if they had participated in the research their anxiety would have been lowered. It is possible that those who agreed to partici— pate in the research had less severe problems and were able to engage in good interview behavior and thus improved very little by exposure to a model. The range of possible scores on the Client Interview Behavior Scale is from 31 to 155 with 93 being the middle score. The means (Table 4.1) are somewhat higher than the middle score indicating that the average item rating on the Client Interview Behavior Scale was "engages in a moderate degree of the given behavior." The population from which the subjects of this study were drawn is different from the populations of other studies using modeling as an adjunctive to indi— vidual counseling. The population served by the mental health center of this study is predominately rural. The major industry is dairy farming. The area has several small cities with populations of less than 7,000. The per capita income as determined by the 1970 census was approximately 7,000 dollars. Although the county contains a large metropolitan city, the residents of the county very seldom enter that city. The focus of their geographical territory is the rural areas close to their own homes. 75 Table 4.1 Interview Means Dependent Variables Time at Measure M1 M2 M3 First Interview 100.07 106.8 2.8 n = 29 n = 28 n = 29 Third Interview 99.27 97.71 3.2 n = 15 n = 14 n = 15 M1 = Judges' Average Rating M2 = Counselor's Rating M3 = Problem Statement Time (minutes) 76 The community mental health center where this study took place had its main office in a town of 7,000 people. Because of the population's preference to identify only with areas close to the residences, the center had branch offices in four small towns within 15 miles of the main center. The subjects of this study could be characterized as rural oriented, low income, high school educated, farm and unskilled laborers, homemakers, big city shy, and politically conservative. The subjects of this study, as well as other clients at the mental health center, were primarily self-referred probably at the advice of a friend who had had contact with the mental health center. A few were referred by their physician, clergyman, or school. Why 50% of the subjects did not remain for three interviews is difficult to answer. The demographic data (Appendix D) does not indicate any differences among those who kept three appointments and those who did not keep four appointments. The author was a counselor at the agency for four years. Periodic analysis of client contacts have shown a mean of four counseling sessions per client. However, several modes must be considered to interpret that mean. Many clients had only one or two sessions and many clients had more than 10 sessions. Thus it 77 appears that this population has two groups of clients. The greatest number have only one or two contacts. A smaller number enter counseling on a long-term basis. The implications for this study may be that the subjects of this study were probably those that had concrete, specific problems for which they sought limited counseling solutions and very few of the sub- jects sought involved counseling to change their general personality and lifestyles. This study probably con- tained few subjects equivalent to the subjects of modeling research using college students. The college student prototype would be interested in the character- istics of good client interview behavior whereas the non- college prototype would be focused on his specific problem only. Procedure The subjects, when they arrived for their first appointment, were handed a paper requesting their par- ticipation in research. If they agreed they were escorted to a room where they read the preorganizer and listened to the model. When they finished listening, they opened the door. Then the counselor entered the room, introduced herself, and brought the client to her office for the counseling session. It is reasonable to assume that new clients have anxiety about entering counseling. This anxiety probably 78 begins to peak at the time of meeting the counselor. The presentation of the treatments occurred at the time the clients expected to meet the counselor. Thus, the clients had to wait another 10 to 15 minutes longer than expected. This unexpected chain of events could have increased the subjects' anxiety level such that any effects of either the experimental treatment or the control treatment were blocked or cancelled. These anxiety effects could have been reduced by presenting the treatments in the counselors' offices. Presenting the treatments after the initial interview might also have yielded significant treatment differences. The Model Many models other than the model used were con- sidered. One was a composite model in which several women would make brief statements. Another was a com- posite model in which excerpts from counseling sessions would be collated and then acted. Another was a model interview and a preorganizer integrated. The preorganizer would be incorporated into the audio tape. The pre- organizer would delineate the appropriate behavior and then the model would exemplify the behavior. The same sequence would occur for each desirable counseling behavior. The model chosen was the best of three coached counseling sessions. This type of model was chosen 79 because it seemed the most natural and easy with which to identify. However, it is possible that a model with a variety of problems covered would be more effective. Possibly the preorganizer would have been more effective if it stated the desired behaviors more spe- cifically, if it gave written examples along with audio taped examples. Also the subjects who came for their own problems, rather than volunteering to talk about problems for research, could have reacted negatively at having to listen to another person's problem. Perhaps a pre- organizer only treatment simply stating good client behavior would allow the client to personalize the con- cepts of good client behavior rather than having to translate from a model client's counseling session to her own counseling session. Another possibility could be that average intel- ligence clients would benefit by repeated exposure to a model whereas intelligent university students benefit from a single exposure to a model. Part of this investi- gator's impetus for this research was the observation that many clients entering group counseling from indi- vidual counseling changed their counseling behavior in the direction of the group members who were making the greatest progress in counseling. 80 In group counseling modeling occurs and is often followed by reinforcement. The model used in this study received no reinforcement. Perhaps the inclusion of reinforcement to the model would facilitate imitation learning for the population studied. Interaction The psychosocial, behavioral skills of the client subjects, the effects of the model treatment, and the shaping or counseling methods of the counselors could have interacted. The modeling treatment could not have been powerful enough to offset the skills or lack of skills of the client interacting with the reinforce- ment pressures of a given counselor. The Control Treatment The greatest problem in designing this research was finding a neutral control treatment. The control treatment probably was not neutral. It did answer questions about mental health and counseling. To the extent that receiving information about an ambiguous situation reduces anxiety, the control treatment would alleviate anxiety and thereby aid the clients in per- forming effective interview behaviors. Given no assurance that the control treatment was neutral, it may be that the model worked better than no treatment at all but the control treatment worked 81 as well as the model or that neither the model nor the control treatment had any effect. The treatment means were in the direction favor— ing the control treatment. Although the difference was not statistically significant for the decision rule alpha level, the step down F for counselor rating had a probability of 0.055 and a univariate F ratio had a probability of 0.033. These statistics raise the question of whether the control treatment might be studied further as a method of improving client interview behavior. Client Interview Behavior Scale Reliability and Validity The two independent judges previously trained in using the Client Interview Behavior Scale obtained good reliability coefficients of approximately 0.80 when rating the pilot data and the research data. When compared to these judges, the counselors generally yielded low reliability coefficients. The Client Inter- view Behavior Scale thus appears to be a reliable measure for those trained in its use. However, there is no assurance that even if the counselors had been trained that their ratings would agree with the judges. That a measuring instrument is reliable is a necessary criterion Of validity, but not sufficient con- dition to claim that it measures what it purports to measure. The fact that 10 counselors gave an overall 82 average weight to all the scale items of greater than moderate importance suggests content validity. However, the question whether or not the Client Interview Behavior Scale actually measures effective interview behavior is unanswered. The same question of validity begs an answer for each of the dependent variables. If the counseling received by the subjects was effective and the dependent variables of this study were able to measure that effectiveness then each measure should have been higher at the third interview than the initial interview. The means are presented in Table 4.1. Visual inspection of the means reveals very small dif- ferences between the first interview and the third inter- view. Thus it is possible that the Client Interview Behavior Scale does not measure client behavior relevant to counseling progress. However, it is also possible that the counseling received was ineffective or that the effectiveness of counseling cannot be measured by the scale until later interviews. Problem Statement Time Significant differences were not found for problem statement time. The reliabilities of judges were low; however, the means were two to three minutes. Improving these means would have very little practical 83 significance, unless it could be assumed that small increments of first problem statement time are cor- related with later interview behavior. An examination of the two judges' ratings of problem statement time reveals that one judge rated with very little variance compared to the second judge. Therefore, it must be assumed that each used different criteria for "stated a personal problem." Greater reliability might have been obtained by having more specific criteria other than "any personal problem." More specific criteria such as "time required to state a feeling about a personal problem" might have made this dependent variable a more sensitive measure. Another difficulty with using time required to first state a personal problem is that the first problem stated is not always the real problem. Clients often test the counselor by presenting less severe problems first and then when they feel that they can trust the counselor with their real concern do they reveal the problem for which they ultimately entered counseling. Number of Appointments Kept The rationale for including the number of appointments kept was that clients who were most satis- fied with their counseling and who were making the most progress would stay in counseling the longest. However, some clients do obtain all they want from one or two 84 sessions. Other clients drop out of counseling without receiving any benefit whether they stop at the first interview or the twentieth interview. Thus, the relation- ship between number of interviews kept and the counseling benefits is not strictly linear. Third Interview The first interview means were in the direction favoring the control treatment. At the third interview the means of judges' rating and counselor ratings were favoring the model treatment. Only 15 subjects remained at the third interview. Such few subjects does not yield much power for the statistical tests performed. A larger sample size might have yielded significant differences. The shift of means from the first to the third interview also raises the question of whether or not exposure to a model has a delayed effect. Another specu- lative explanation could be that those keeping three appointments had more severe problems and therefore benefited most from the model treatment. Replication of the study with a larger sample would be necessary to answer these questions. It was necessary because of agency policy for some of the subjects to have a different counselor for her second, third, and subsequent interviews. Those subjects that were transferred had male counselors for 85 later interviews. The effect of forming a new relation- ship with another counselor could have had a deleterious effect. Also, the model was female client, female coun- selor. Perhaps the model effect did not generalize to a male counselor. Unfortunately, the number of subjects is too small to meaningfully compare those subjects who had the same counselor for three sessions against those who changed counselors after the initial interview. Dependent Variable A perennial problem in the research of counseling has been finding dependent variables to measure the pro- cess and its effects. The dependent variables of this study appear to have content validity, but further validity is uncertain. Another dependent variable that might have been used could be a client satisfaction scale. Conclusions No evidence was obtained to support the hypothesis that exposure to a good client audio model increases good client interview behavior for female clients entering first-time counseling at a rural, small town mental health center. Evidence was found that the Client Interview Behavior Scale is a reliable measuring instrument when used by trained raters; however, only minimal evidence was found for its validity. 86 Implications for Further Research Another study using modeling with rural mental health center clients certainly would be worthwhile. The effectiveness of imitation learning in improving client interview behavior has been demonstrated with clients and pseudo-clients associated with universities. This author was unable to find research on modeling executed at community mental health centers. The question remains, "Does model exposure improve the process of counseling for clients with average intel- ligence and education?" The use of different models might yield positive results. Repetitive exposure to good client models might be more powerful in changing client problem-solving behavior. The complexity of the model may be an impor- tant independent variable to consider. The variety of models for such a nebulous process called effective client counseling behavior are numerous. The problem may not be whether models are effective but rather which models are most effective, in what situations, and how should they be administered. APPENDICES APPENDIX A DATA FOR WEIGHTING CLIENT INTERVIEW BEHAVIOR SCALE MEAN 10 APPENDIX A DATA FOR WEIGHTING CLIENT INTERVIEW BEHAVIOR SCALE Iten Self 5.4.4 1.4:“.15688 33333333333 «Jajrqndajnjqjanqunumq uhhhhhhhuhuu In». 33333333333 33:“. 3.4 231431464.“ ‘ (4)4 33hhhhhhuu hi43331u. 3‘414hn41u. 33333223314‘414 .u. 3333333333h. hhhhhh Hannah.” 33333333hhhh EXploration 012 123.456789111 Responsibility Assumes 3666 333a) 3:433 huh.” 3333 3.4 3.7.4. 14:414.“ 3H3OJ 3333 (414.414 .u..u.h..u. 22hlu. 123.“. Counselor lhmsts 322333 332.414.” 332223 313.423 3hhuuiu. 322:4 3|...» nu. 33323 2 33.4:464 hhlllh 1223314 1231456 Motivated 7L>A%g/ 3332 314 3.“. 141414. 3 3332 33u1 huh.“ hh3h. (”L433 .4314?) h3h2 443.43 Honest O r)...” 78 32333 Id. 33:41.4 30¢.2 33 322.46“. “huh.“ 232-4.“. in» 2214.“. 3323b... 222k.” 2226“.“ 123.45 87 3 - moderate importance 4 - high importance 1 - no importance 2 - low importance APPENDIX B PILOT STUDY DATA APPENDIX B PILOT STUDY DATA Subject Self Assumes Trusts Motivated HOnest Total Exploration Responsibility Counselor R1 R2 R1 R2 R1 R2 R1 R2 R1 R2 R1 R2 92 37 36 17 11 22 2h 16 13 19 18 111 102 129 89 h2 19 1h 27 ‘25 19 1h 25 23 139 118 22 39 81 18 1h 20 22 12 1h 15 20 100 111 69 30 h1 12 1h 21 22 13 18 16 17 92 108 60 20 33 11 1h 2h 21 11 18 15 18 81 100 70 17 26 5 11 1h 20 u 12 7 12 87 81 101 23 26 11 12 1h 20 13 12 13 15 7h 85 179 h3 29 17 16 23 23 13 16 12 17 108 101 h92 hl 35 18 18 29 23 18 15 20 22 126 113 33 2h 29 11 13 2h 28 12 13 1h 16 85 95 39 32 36 15 12 17 21 13 12 16 18 93 99 78 5h Mo 16 12 20 25 16 12 1h 22 20 111 11 h2 h1 20 17 29 27 15 1h 20 22 126 121 21 2h 31 10 12 16 20 7 9 8 13 65 85 10 h9 uh 18 16 25 28 17 17 22 21 131 126 hh 2h h3 9 18 17 25 9 15 12 18 71 119 86 31 36 1h 18 2h 26 11 12 12 15 ' 92 103 as 39 33 15 12 25 23 11 12 17 17 107 97 52 16 29 8 11 13 19 6 9 8 18 51 82 R1 = rater one R2 = rater two 88 89 Hoyt's Analysis of Variance to Extimated Reliability r = l - MS residual MS between people Self Exploration r = l - 768é18 233 . l = 0.67 Assumes Responsibility r = l - 128.7 18 271.3 1 = 0.52 Trusts Counselor r = l - 123.5 18 1.1 l = 0.72 Motivated r = l - 97 18 2 l = 0.6h Honest r = l - 66.6/18 536.6713 = 0.88 90 Total Scale r = 1 _ 29§2.8/18 13174.9718 = 0.76 Variance 2 2 _ a__)_(E_—Xp2 n = 776779.98 _ 39496.33 19 = 1386.8 a = 37.24 Power 2 2 n 2(u1-u2) ¢= 2 J 02 Where J - number of treatments n - number in each treatment <1)2_ 15(37.24)2 ‘ 2(2)(1386.99) = 3.75 ¢ = 1.94 APPENDIX C PARTICIPATION REQUEST FORM APPENDIX C PARTICIPATION REQUEST FORM Mental Health Center always has been interested in improving the effectiveness of its services. we would appreciate your cooperation in helping us find out what type of things improve the counseling process. If you agree to participate in our present research you will be asked to read a one-page article and listen to a lO-minute tape recording. After you listen to the tape recording then you will talk to your counselor. Then we will record your first counseling session and your third counseling session. These tapes then will be listened to by two experienced counselors from a mental health center in another city. You will not be identified and the record- ings of your sessions are strictly confidential. After the tapes are listened to, they will be erased. Unfortunately we cannot tell you the exact nature of the research until the results are all in; however, after June 1975, c0pies of our findings will be available from our office. Again you can be assured that nothing you say or anything about you 91 92 will be known by anyone besides the two people who will listen to the tapes and they will not know your name a ***** [:::] I agree to participate in the above research understanding that recordings of my sessions will be kept strictly confidential. signed [:::l I prefer not to participate. APPENDIX D DEMOGRAPHIC DATA * SUBJECT AGE 23 21 22 50 20 33 27 19 30 31 18 26 33 42 32 27 29 28 21 29 20 24 25 35 20 42 27 25 22 24 APPENDIX D DEMOGRAPHIC DATA OCCUPATION Unemployed Secretary Salesclerk Housewife Student Unemployed Salesclerk Unemployed Unemployed Housewife Student Clerk Typist Housewife Secretary Unemployed Student Secretary Billing Clerk Computer Operator Housewife Unemployed Housewife Hostess Housewife Cashier Housewife Homemaker Homemaker Secretary Housewife of selection for appointment. 93 High High High High High High EDUCATION School School School School - 10 grade School - 10 grade School 1 year of college High High High High High High High High School School - 11 grade School - 11 grade School - 11 grade School School School School In High School High High School School 1 year of college High High High High High School School School School School - 10 grade Junior at M.S.U. High High School School 2 years of college High High ** Kept all three appointments. School School * Subject number has no relationship to order APPENDIX E CHARACTERISTICS OF THE GOOD CLIENT APPENDIX E CHARACTERISTICS OF THE GOOD CLIENT Every counselor has, however vaguely, a mental image that defines a good client. Although there are individual differences among counselors in their defi- nitions of the good client characteristics, five factors seem crucial. The good client engages in self-explor- ation. He talks, states the problem, and discusses his attitudes, beliefs, feelings, and motives. The good client assumes responsibility. He sees himself as the one who will make the primary changes and does not blame others for his present predicament or expect changes in others to effect solutions for his problem. The good client has trust in his counselor, and perceives his counselor as one who will help. The good client is motivated. He is willing to incur anxiety in order to affect change, does not demand easy solutions, and is not overly dependent upon his counselor. The good client is honest. He tries neither to minimize nor maximize his problems, does not try to gain sympathy by exacerbating his symptoms nor try to impress the counselor with his psychological assets. 94 APPENDIX F PREORGANIZERS AND THEIR RANKINGS APPENDIX F PREORGANIZERS AND THEIR RANKINGS No. l Now that you have decided to talk to a counselor about some of your problems, you probably are wondering what counseling is all about. To help you get the most out of counseling in the shortest length of time, we offer the following suggestions. Obviously talking is important. Try to state your problem(s) the best you can. Your counselor is interested in your thoughts, feelings, and attitudes concerning your problem(s). Exploring your own feelings will be of good help to understand yourself. Although our problems have many causes outside of ourselves, try as much as possible to assume responsi- bility for yourself. Try to see yourself as the one who can change. Of course others influence our lives, but it's pretty hard to change others and our surroundings; however, we can change ourselves. You are no doubt wondering what kind of person your counselor is. In brief, he is concerned about your best interests and is specially trained to help you. Often because of past experiences with other people who 95 96 have let them down, persons are reluctant to trust their counselor. Your counselor has had much education and experience. Without a doubt, he has talked with others with problems like yours and will understand. Counseling does not provide only easy solutions. Making psychological changes can be difficult. Learning new, more effective ways of problem solving can be temporarily more painful than our familiar, but ineffec- tive, methods of handling our life situations. A last suggestion is to be honest about yourself. There is no need to impress your counselor with either how much you have going for you or how much you have going against you. You are important to your counselor, he wants to get to know the real you better so that he can better help you. If your counselor could not accept people as they are, he would not have made it as a coun- selor. 11.9.1.2. Now that you've taken the first step in seeking some help for your problems, you're probably wondering how counseling works and what your part is in counseling. In coming to this decision about seeking help you've probably thought that you needed to talk with someone. Your problem seems confusing to you and you have many different feelings about it. In talking 97 about your problem and these feelings to a counselor you'll be able to explore your feelings and come to a clearer understanding of the problem. Talking then, openly, about all of your feelings is an important part of counseling. In choosing to come here, you make a decision to do something about your problem. You made a choice and accepted responsibility for that choice. During the sessions that follow you will be faced with more decisions and additional changes. Changes in therapy are decisions only you as a client can make--no one can make you change or keep you from changing. In order to effect these changes you must accept the responsibility for your present problem and acknowledge yourself as the only one who can change it. The person you'll be working with in these sessions is a skilled counselor who knows how to listen. It's probably been difficult for you to talk with anyone about your problem due to your uncomfortable feelings about it. It is important to know, then, that your counselor is a person who is interested in you and wants to understand how you're feeling. The counselor listens to many problems each day and accepts and wants peeple to trust him. Your counselor feels this way about you. At times you will feel pretty overwhelmed by your problem and how to accomplish the changes you want. 98 You'll probably wonder why change is so painful. It is painful because we've allowed our problems to reach so severe a point. In order to change the problem it is necessary to go back through the problem and readjust our attitudes and feelings. This is very hard but in order to make changes it's necessary to accept this as part of counseling. Lastly, in dealing with your problem you have most likely rationalized and not looked at it clearly. To accomplish change in therapy, you will have to be very honest with yourself about the situation. Your honesty and realistic way of looking at the difficulty will help you in the future as well as how to avoid choosing another similar problem-situation. 191.3. In order for you to receive the maximum benefit from counseling you will have to learn to talk about yourself, your ideas, thoughts, feelings, fears and hopes without holding back. This is difficult at first, but as you get to know your counselor you will find yourself being able to open up more and more. As you become able to express whatever is on your mind, no matter how uncomfortable you may feel doing so, you will be making progress in the development of what can be called "the counseling relationship." In order for counseling to be helpful to you this relationship must 99 be one where you have made a commitment to yourself to be as straightforward as possible. Often it is tempting to tell your counselor only part of what is on your mind, holding back things that you are not able to say without feeling self-conscious or uneasy. However, in the long run it will help you if you say those things regardless of how reluctant you are to do so. You will find that as you get to know your counselor you will be able to say things to him (her) that at the beginning of counseling you kept to yourself, perhaps because you were afraid your counselor would think less of you if you revealed aspects of yourself that other people might think were unacceptable. But as time goes on you will discover that your counselor is not like most other people you have known because he or she will listen attentively to you without judging you or criticizing you. Even if now, at the start of counseling, you feel that because of your problems or symptoms you are weak or inferior to other people you will discover that as you and your counselor work together on solving your problems you are actually doing something that takes a great deal of courage. Looking at yourself in coun— seling is often painful and many other pe0ple who could benefit from counseling do not seek it because they would rather suffer than go through the often difficult process of change. 100 RANKINGS OF PREORGANIZERS Average Rater 1 Rater 2 Rater 3 Ranking No. 1 No. 2 No. 3 APPENDIX G RATINGS OF POTENTIAL MODEL TAPES APPENDIX G RATINGS OF POTENTIAL MODEL TAPES E: z E S ’53 .2 c1 9' ma 8 E13 '5‘ we as s a a 5% 8% es s a a TAPE 53:9 53 E! §3E3 §E £3 £3 R1 R2 R3 R1 R2 R3 R1 R2 R3 R1 R2 R3 R1 R2 R3 £Rl+iR2+iR3 1 56 52 53 17 18 20 26 25 30 18 18 20 25 22 25 M25 2 57 h8 53 19 15 15 30 25 26 18 17 18 25 19 23 h03 3 58 an 53 20 17 16 30 21 23 16 13 1h 25 22 2h 396 101 APPENDIX H EXPERIMENTAL TREATMENT APPENDIX H EXPERIMENTAL TREATMENT Thank you for participating in our research. Please read the next page and then follow the instructions on the last page. 102 103 Now that you have decided to talk to a counselor about some of your problems, you probably are wondering what counseling is all about. To help you get the most out of counseling in the shortest length of time, we offer the following suggestions. Obviously talking is important. Try to state your problem(s) the best you can. Your counselor is interested in your thoughts, feelings, and attitudes concerning your problem(s). Exploring your own feelings will be of good help to understand yourself. Although our problems have many causes outside of ourselves, try as much as possible to assume responsi- bility for yourself. Try to see yourself as the one who can change. Of course others influence our lives, but it's pretty hard to change others and our surroundings; however, we can change ourselves. You are no doubt wondering what kind of person your counselor is. In brief, he is concerned about your best interests and is specially trained to help you. Often because of past experiences with other people who have let them down, persons are reluctant to trust their counselor. Your counselor has had much education and experience. Without a doubt, he has talked with others with problems like yours and will understand. Counseling does not provide only easy solutions. Making psychological changes can be difficult. Learning new, more effective ways of problem solving can be temporarily more painful than our familiar, but ineffec- tive, methods of handling our life situations. A last suggestion is to be honest about yourself. There is no need to impress your counselor with either how much you have going for you or how much you have going against you. You are important to your counselor, he wants to get to know the real you better so that he can better help you. If your counselor could not accept people as they are, he would not have made it as a counselor. 104 Now that you have finished reading, please push the green PLAY button on the tape recorder. You will hear part of an actual counseling session in which the client does the things that you have just read. Even though her problems may be somewhat different than yours, it is hoped that you will find it helpful in getting a a feel for what happens in counseling. When the tape is finished push the red STOP button. cl CO: cl: 105 But it was really a . . . you know . . . a really strange predicament last night . . . (nervous laugh) . . . because all of a sudden we were there and . . you know . . . he had certain expectations that I really wasn't . . . you know -- he wanted me to . . . I guess . . . express, you know, that I really still cared for him or something. And at the very end he walked me out to the car and wondered if he could hug me and I said, "No" . . . you know . . . it's just that there was nothing the whole night . . . that had changed me from just feeling . . . that we were just there together for four hours. But I guess I really felt like I had to grab . . . um . . . the fact that . . . yes, you know . . . it's suddenly whenever you break up with somebody you think -- well, what was it . . . (nervous laugh) that you even cared about, you know, you're hurt, you know . . . like I felt hurt and angry and I wondered well . . . why was I so hung up on him, why did I really care . . . and yet, you know . . . last night I just felt like yah (giggle) . . . if things were different . . . if he was different, . . . I don't know in certain . . Last night you could identify reasons why you cared. Why I still, well, you know, why I liked the guy in the first place and why I did really like him, C0: C1: C0: C1: 106 you know. There was something there . . . I mean . . . you know this is making me feel like maybe I didn't have any reason to like him. Where did that leave you when you felt like you didn't have any reason to like him? Well, it really gets me hung up on . . . well . . (nervous laugh) . . . what am I looking for . . . you know . . . in a man, but . . . you know I get really . . . uh, just totally lost like . . . qualities . . . um . . . you know . . . personali- ties, um . . . (laugh) . . . I just don't know, I just don't know where to start. Somebody will ask me what it is I am looking for and I just go . . . ah, . . . you know . . . kinda blank. I mean, you know, there are . . . You feel like you should be able to come up with an answer . . . Well, yeh, there are certain things that I really should remember . . . you know, . . . like warm- ness . . . a person who tries to relate to me, you know . . . , things, . . . (sigh) . . . oh kinda of interactions that I am looking for . . . but it just really scares me when I keep thinking I don't know why, you know, I had all these really intense feelings . . . and now they're really crashed and it just really scared me that maybe I didn't have C0: C1: 107 any basis for even . . . (laugh) . . . having the feelings in the first place. And I felt at least a little bit more comfortable with . . . yeh, you know, I still like the guy . . . but like is where it's at (laugh) and really I want to be left out of where he's at, because it's not going to do me any good. All this sounds too that if you aren't left out you are going to feel pretty uncomfortable. Yeh, it was really funny when things ended . . . um . . . or at least when I finally said you know, . . . they're ended and then I felt like I was almost turning over a new leaf and almost being free from the burden of . . . uh . . . trying to cope with my own emotions. I felt like my emotions were almost . . . um . . . you know . . . in a better spot -- they were more complete. I wasn't struggling with this idea of you know -- I like him . . . but he is way off into dreamland with somebody else . . . you know, I mean . . . I just didn't know what his reaction was to me and when I decided, you know, apparently that was just no where to go . . . I felt more whole, more free . . and yet it was kinda you know . . . it shocked me in a way that man . . . I really feel I had to get out -- I was out, you know really . . . CO: cl: CO: Cl: C0: C1: 108 You felt better out by yourself than you had with it. Yeh, only I couldn't convince myself when I was in it, that I would feel better out, you know? mm uh You have that fear of, you know, that something being gone that you really wanted . . . you know . . . and if it's gone, you know . . . that emptiness. Then when you were out you didn't really miss it (interrupting) I didn't really have an emptiness, you know . . . this is kind of a shocking (nervous laugh) experience. Like it was a good experience . . . because I feel like I wasn't empty when it was gone. I was more me because I didn't have to, you know, kinda over exert myself in a way that there wasn't, I mean, . . . you know, I couldn't present myself anymore completely to him than I had . . . I mean, you know, I was showing him that I cared . . . telling him that I cared and he wasn't giving me what I needed. So, I was kind of over exerting myself and when I quit doing that . . . I felt like I became (nervous laugh) a full person again . . ., where I was just losing myself before . . . you know, I was just grasping at (nervous laugh) cer- tain things that'll tell me that I had good exper- iences . . . you know, I had certain things that really were o.k., but since he didn't come through CO: 01: CO: CO: 109 there's no sense in, you know . . . in really hang- ing on to something that isn't there . . . So last night . . . you're wondering if — o.k. if I sit in the same room with this guy is it going to be absolutely nothing or am I going to be able to pitch into what was there, you know and be friends on that level or am I going to be suckered into the whole thing. Yah, I went through just about you know every one of those levels . . . and almost more you know, usually it was like a whole month of telling myself things are over between us . . . and yet I've got to sit . . . four hours . . . four hours can be a very long time, but I just . . . you know, I felt pretty comfortable and I felt that he was actually more uncomfortable from where I was at last night. That's a change? Yah, I mean, you know I feel, you know I am feeling this way about a lot of men, really . . . that I don't really, you know . . . that I just don't have anything to offer them because I haven't gotten what you know, I needed in return and I'm not going to . . . you know, over exert myself to some things that are just not worth it. I am just kind of waiting for the . . . What do you mean by that -- you don't have anything to offer? CO: Cl: 110 Well . . . (sigh) apparently where . . . you know . . . they're at . . . I mean, . . . what they need isn't what . . . you know, what I'm you know, the kind of things that I can give 'em, . . . you know, the companionship or relationship . . . what I'm looking for, I guess. Is it that you feel that you don't have anything or that you are not willing to keep looking unless you can start getting something back? Well, sometimes it gets confusing like . . . you know, you wonder if what I've got is what anybody is looking for . . . you know and then . . . um . . . I just kinda tell myself well you're probably . . . you know . . ., there should be someone that can share with me . . . you know, more on an equal basis than what has been. You know, as far as both of us feeling that we're giving and we're receiving and not giving just (laugh) or taking, you know. And . . . you know, just kinda relaxing to say . . . you know, I'm not going to rush around looking for that. You know, it's just one of those things you just can't do . . . I mean I can't. You know, you stumble into a person that somehow eventually works out to be maybe the right person . . . but I'm almost nill on the idea of . . . you know, rushing out to even meet guys. Really, you know, CO: Cl: 111 it's kinda a lull of really caring about getting even . . . well, you know a lot of crazy relation- ships started. It's . . . you know, - I don't know . . . I'm grouping men into . . . you know, they generally want sex and then just kinda nothing else and I'm not going to put myself out there to even be mixed into it. You sound like you are angry about what has been happening. (embarrassed laugh) Yes - like even last night I got a telephone call that was from a guy that I used to work with . . . and the whole thing was . . . well could I come out and have a drink because . . . you know there was . . . I don't know . . . how many guys . . . two or three . . . at least there and they thought they would like to have me out there to have a drink. I'm not . . . you know, I have not been interested in drinking in . . . you know, I have not been drinking much and I've got a shift to do and I can't drink before the shift - and I'm just not coming (laugh) . . . and of course they say why not . . . don't you think you can . . . you know, just for an hour . . . you know, trying to tear you apart thing . . . come on, come on, . . . egg you on. CO: C0: C1: 112 How is that tearing you apart? Well they're trying to make me feel like . . . um . . . like you know I'm hurting them in some way . . . you know, because I'm not showing up. It's just a big fictitious thing to . . . To get you to feeling like the guys are asking you because you owe it to them? Well . . . yah . . . kinda in that way, you know, if I'm their friend I owe it to them to come out and have a good time because most of them want more than just a drink and . . . you know, it's just kinda (laugh) . . . that's not where I'm at. That's what I try to tell them, you know, that I'm really not interested in going out. I haven't been going out for that reason. APPENDIX I CONTROL TREATMENT APPENDIX I CONTROL TREATMENT Thank you for participating in our research. Please read the next page and then follow the instructions on the last page. 113 114 The Mental Health Center came into being in May, 1970. It was operated by one part-time psychologist. In the early autumn of 1971, three full- time professional mental health workers joined the staff. The process of developing the Center was begun with full force. By talking to ministers, teachers, doctors, nurses and other helping people, the community became aware of our counseling services. Since our beginning over 3,000 persons have come to our Center. At the present time well over 100 different people have regular counseling sessions at the Center. They come from all age groups; they have many different types of problems; and they come from all walks of life. The present staff consists of five full-time counselors, four interns, and two secretary receptionists. The Center is open in , , , and for those who are unable to come to the office. The Mental Health Center operates under a federal grant from the National Institute of Mental Health. It also receives funds from the state of Michigan and County tax revenue. Another portion of our operating cost comes from client fees. 115 Now that you have finished reading, please push the green PLAY button on the tape recorder. On the tape recording are answers to general questions that many people have about coming to a mental health center. When the tape is finished please push the red STOP button. 116 Transcription of Control Tape We put together a series of questions that many people ask about a Mental Health Center such as this one. The fellowing questions are those we have fbund are most frequently asked by be- ginning clients. The questions will be asked by one person and . answered by the other person. What is the real cause of emotional problems? Emotional problems are the product of bad learning experiences. These create conflicts which interfere with relationships with life and with people. Some of these conflicts are conscious, that is, the person is only partially aware of them. ther of these conflicts are conscious, that is, a person knows what the struggle is. Doesn't my needing counseling mean that I have a weak will or am.on my way to a mental breakdown? No. More and more people, even those whose problems are essentially normal ones are seeking help these days because they realize they can better themselves with counseling. The fact that you desire counseling is a compliment to your judgement and it is in no way an indication that you are approaching a.mental breakdown. 117 If I had an emotional problem doesn't that make me different from other people? All people have some neurotic problems. Their ability to live with these problems and to fulfill themselves creatively, in spite of these problems, constitutes the difference between what is called normality and emotional mal-adjustments. Most persons whom we call normal actually have neurotic problems, but they do not interfere with their life adjustments or happiness to any great degree. When such an interferance does occur the individual will want to do some- thing about himself and his problem. Why do sonmany'people have neurosis? It is impossible to excape conflicts in the process of living. The very act of growing up involves sacrifices and giving up of individual pleasures fbr the group interest. There are many deprivations and disciplines we all have to adjust to. The price of such an adjustment is often frustration and tension. The mild neurotic symptoms are present in all of us. Indeed they are more common than the common cold. 118 Can physical problems be caused by emotion? Many physical symptoms are psycho-somatic in nature which means they have an emotional or nervous basis. When you come to think of it, it is not really so strange that emotional strains or worries should produce physical symptoms. After all every organ in your body is connected to your brain by nerve channels so it is logical that when your nervous system is upset by some crisis or conflicts you may feel the effects in different parts of your body. Can counseling make a person happy no matter how bad his situation may be? Counseling will not stop a person from reacting to a bad situation. It should however make it easier for him to deal with the situation in a more constructive manner. Am I going to be forced during counseling to give up the pleasures in my life situation that I find valuable now? Nebody is going to force you to do anything. If any of the pleasures that you enjoy at the present time are destructive to you, you have a right to know this. Once you become aware of all the 119 facts you will then be in a better position to judge for yourself if you want to continue or discontinue certain pleasures. Many people have the mis-conception that counseling will make them terribly serious and prevent them from enjoying themselves. 0n the contrary, counseling relases the inhibitions to enjoyment and pleasure. What happens to the information about me? In scientific work, records are necessary since they permit a more thorough dealing with ones problems. And it is understandable that you.might become concerned about what happens to the information about you because much or all of this information is personal. Case records are confidential. No outsiders, not even your closest, relative or family physician is permitted to see your file without your written permission. Will I be required to take psychological tests? Quite often the counselor will request their client to take several psychological tests. The purpose of testing is to give your counselor a rough overview of your personality structure and functioning. 120 This further aids your counselor to understand you. Psychological testing also give your counselor something against which to check the impressions he has obtained in talking to you. This completes the reading and tape material. Please push the stop button on the tape recorder. Open the door in the office to let your counselor know you are through. She will come and get you in a few seconds. APPENDIX J EXPERIMENTAL DATA APPENDIX J EXPERIMENTAL DATA Interview One Counselor Subject Treatment 13 15 20 18 CO 19 11 12 14 16 17 WNNNN OO‘QNH CO2 NNNNN WKOQU'ID CO 1 U'lkDCDQO‘OthH Mean for S's keeping 3 appointments Mean for S's not keeping 3 appointments OOOOOMMMMM OOOOOML‘UI‘JD’JM OOOOONWMMM 121 M1 R1 R2 78 68 92 80 95 71 123 78 * * 111 96 134 106 95 87 92 71 133 64 110 103 112 80 109 79 153 127 135 95 124 107 114 69 124 116 134 101 131 105 112 109 105 93 83 66 9o 77 115 92 9o 54 130 77 112 58 140 126 73 67 96.5 102.8 5U $0 110 65 116 * 127 140 134 93 142 131 103 102 103 128 110 125 116 129 122 ldFJHPJFJHP‘FJHFH IHFJHIOP‘*FJHP4F‘ 94 93 70 103 90 105 85 82 101 79 I—‘l—‘l—‘Nl—‘O‘NWHW 104.5 2.8 110.1 1.7 N NNCDU'IN *Nl—‘QH QPJP‘HPJh‘HPJF‘w NNI—‘WNCDNU'Il-‘H WNWUJNUNNWW NWHI—‘WWNNWH HNwNwUJUWI-‘U Counselor CO CO CO 122 Interview Three Subject Treatment HI—‘l-‘I-‘l-‘l-JI-‘NI-‘H \nmANt-uoooomw WNNNN 00‘qu NNN \O\IU| N N (D uh OOOOONMHMM OOOOOMMMME’J OOOOOME’JMML‘U H mxomxlmoanE—I experimental treatment control treatment counselor judge's rating on Client Interview Behavior Scale 126 96 110 * 103 137 1: 113 108 * 95 108 145 117 * 97 * 'k M l 115 * 79 119 * 73 64 139 61 83 * 'k 102 104 111 102 * 83 W Nfl-I-‘Hfl-mfl-I'l-‘H II-Nfl-fl-l—‘fl-fl-fl-l—‘fi- I- *H *NNDH 39H '2'" N U! usun—Ia-ooa-x-Ne. l-uba-a-Hu-at-a-l—Ix- 2' *H X-UJOUJl-J *b counselor's rating on Client Interview Behavior Scale judge's rating of time to first state a problem number of appointments kept judge one judge two REFERENCES I! REFERENCES Allport, F. H. Social psychology. Cambridge, Mass.: Riverside Press, 1924. Bandura, A. Principles of behavior modification. New York: Holt, Rhinehart, and Winston, 1969. Bordin, E. S. Implications of client expectation for the counseling process. Journal of Counseling Psychology, 1955,‘2, 17-21. Brody, H. A. The effect of three modeling procedures on the frequency of emission of self-referent affect statements. Dissertation Abstracts, science & engineering, Aug. 1968, 22, 2, 767-8} Carnes, E. F., & Robinson, F. P. The role of client talk in the counseling interview. In A. Brayfield (Ed.), Readiness in modern msthods of counseling. New York: Appleton-Century Crofts, 1950. Carkhuff, R. R. Training as a preferred mode of treatment. Journal of Counseling Psychology, 1971, 18, 123-131. Corsini, R. Roleplaying in psychotherapy. Chicago: Aldine, 1966. Drakeford, J. W. Integrity therapy. Nashville, Tenn.: Broadman, 1967. Eskedal, G. A. Symbolic role modeling and cognitive learning in the training of counselors. Journal of Counseling Psychology, 1975, 22, 152-155. Gerwirtz, J. L., & Stingle, K. C. The learning of gener- alized imitation as the basis for identification. Psychological Review, 1968, 1:, 374-397. Goi, F. J. The effect of actual therapy pretraining upon the depth of self-exploration of university coun— selinggcenter students during the initiaI'inter- view. Unpublishedgdoctoralfdissertation, Temple —I— o Univer51ty, 1971. 123 124 Goodstein, L. D., & Grigg, A. E. Client satisfaction, counselors, and the counseling process. Personnel and Guidance Journal, 1959, 38, 19-24. Grigg, A. E., & Goodstein, L. D. The use of clients as Hays, W. Heller, Higgins, judges of the counselor's performance. Journal of Counseling Psychology, 1957, 4(1), 31-36. L. Statistics for psychologists. New York: Holt, Rinehart, & Winston, 1963. K. Effects of modeling procedures in helping relationships. Journal of Consulting and Clinical Psycholong 1969, 32, 522-526} W. H., Ivery, A. E., & Uhleman, M. R. Media therapy: A programmed approach to teaching behavioral skills. Journal of Counseling Psyf chology, 1970, 17(1), 20-26. Hoehn-Saric, R., Frank, J. D., Imber, S. D., Nash, E. H., Holt, E. Kell, B. Stone, R. G., & Battle, C. C. Systematic prepar- ation of clients for psychotherapy. 1. Effects on therapy behavior and outcome. Journal of Psychiatric Research, 1964, 2, 267-281. B. Animal drive and the learning process (Vol. 1). New York: Holt, 1931. L., & Mueller, W. I. Impact and change: A study of counseling_re1ationships. New York: Appleton- Century-Crofts, 1966. Kelly, G. A. Theppsychology of personal constructs. Volume—II Clinical diagnosis and psychotherapy. New York: Norton, 1955. Krumboltz, J. D., & Thorenson, C. E. The effect of behavioral counseling in group and individual settings on information seeking behavior. Journal of Counseling Psychology. 1964, 11(4), 324-332. Krumboltz, J. D., Varenhorts, B., & Thorenson, C. E. Lazarus, Non-verbal factors in the effectiveness of models in counseling. Journal of Counseling Psychology, 1967, 14(3), 412-418. A. A. The results of behavior therapy in 126 cases of severe neurosis. Behavior Research and Psychotherapy, 1963-1964, I, 72. 125 Lazarus, A. A. Behavioral rehearsal vs. non-directive therapy vs. advice in effecting behavior change. Benavior Research and Therapy. 1966,‘4, 209-212. Lennard, H., & Bernstein, A., et al. The anatomy of psychotherapy. New York: Columbia University Press, 1960. Lennard, H., & Bernstein, A. Role learning in psycho- therapy. Psychotherapy; Theory, Research and Maccoby, E. E. Role-taking in childhood and its conse- quences for social learning. Child Development, 1959, 39, 239-252. Marlatt, G. A., Jacobson, E. A., Johnson, D. L., & Morrice, D. J. Effect of exposure to a model receiving evaluative feedback upon subsequent behavior in an interview. Journal of Consulting and Clinical PsycholOQY. 1970, 34(1), 104-112. Marlatt, G. A. Exposure to a model and task ambiguity as determinants of verbal behavior in an interview. Journal of Consulting and Clinical Psychology, 1971, 3§72), 268-276. Miller, N. E., & Dollard, J. Social learning and imi- tation. New Haven: Yale University Press, 1941. Mowrer, O. H. Learning theory and the symbolicyprocess. New York: Wiley, 1960. Mowrer, O. H. The new group therapy. Princeton: Van- Nostrand, 1964: Myrick, R. K. Effect of a model on verbal behavior in counseling. Journal of Counseling Psychology, 1969, 16(3), TBS-190. Orne, M. T., & Wender, P. H. Anticipatory socialization for psychotherapy: Method and rationale. American Journal of Psychiatry, 1968, 124, 1202-1212. Parloff, M. Some factors affecting the quality of thera- peutic relationships. Journal of Abnormal and Social Psychology, 1956, 52,5-10. Patterson, C. H. Client expectations and social con- ditioning. Personnel and Guidance Journal, 1958, 31, 136-138. 126 Piaget, J. Playy dreams, and imitation in childhood. New York: Norton, 1952. Richard, H. C., & Willard, D. L. Note on recorded vs. actual model behavior and observed performance. Psychological Reports, 1968, 33(3), Pt. 1, 839- 840. Rogers, C. R., & Dymond, R. F. Psychotherapy and per- sonality change. Chicago: University of Chicago Press, 1954. Rogers, C. R. (Ed.). The therapeutic relationship and its impact: A study of psyCHotherapy with schizo- hrinics. Madison, Wisconsin: University of Wiscon31n Press, 1967. Skinner, B. F. Science and human behavior. New York: Macmillan, 1953. Snyder, W. U. The psychotherapy_relationship. New York: Macmillan, 1961. Stotland, E., Zander, A., & Natsoulas, T. Generalization of interpersonal similarity. Journal of Abnormal and Social Psychology, 1961, 33, 250-256. Swartz, A. N., & Hawkins, H. L. Patient models and affect statements in group therapy. In Proceed- ings of the 73rd Annual Convention of the American Psychological Association. WaEhington, D.C.: American Psychologicai Association, 1965. Truax, C. B., & Carkhuff, R. R. Personality change in hospitalized mental patients during group psy- chotherapy as a function of the use of alternate sessions and vicarious therapy pre-training. Journal of Clinical Psychology. 1965, 33, 225-228. Truax, C. B., Carkhuff, R., Wago, D., Kodman, F., & Moles, E. A. Changes in self-concepts during group psychotherapy as a function alternate sessions and vicarious therapy pre—training in institutionalized mental patients and juvenile delinquents. Journal of Consulting Psychology, 1966, 32, 309-314. Truax, C. B., & Carkhuff, R. R. Toward effective coun- seling and psychothergpy. ChiCago: Aldine, 1967. 127 Truax, C. B., & Wargo, D. G. Effects of vicarious therapy pre-training and alternative sessions on outcome in group psychotherapy with out- patients. Journal of Consulting and Clinical Psychology, 1969, 33, 440-447. Warren, N. C., & Rice, L. N. Structuring and stabilizing of psychotherapy for low-prognosis clients. Journal of Consulting and Clinical Psychology, 1972, 33, 173—181. , Whalen, C. Effects of a model and instructions on group verbal behaviors. Journal of Consulting and Clinical Psychology, 1969, 33J 509-521. Whitla, D. K. (Ed.). Handbook of measurement and assess- ment in behavioral sciences. Reading, Mass.: AddiSOn-Wesley, 1968. Winer, B. J. Statisticalyprinciples in experimental design. New York: McGraw-Hill, 1962. Wolberg, L. R. The technique of psychotherapy. New York: Grune and Stratton, 1967. rl. Ital \i taxman—ESL