THE EFFECTS OF SELF-MANAGEMENT STRATEGIES AS INITIAL TRAINING AND AS SECONDARY TRAINING FOR AFFECTIVE SELF-DISCLOSURE OF UNDERGRADUATE MALES Dissertation for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY NANCY LEE VOIGHT 1975 ‘:9}\6 Date 0-7 639 This is to certify that the thesis entitled Effects of Self-Management Strategies as Initial Training and Secondary Training for Affective Self-Disclosure of Undergraduate Males presented by Nancy Lee Vo ight has been accepted towards fulfillment of the requirements for Ph . D . degree in Education Major professor July 23, 1975 ABSTRACT THE EFFECTS OF SELF-MANAGEMENT STRATEGIES AS INITIAL TRAINING AND AS SECONDARY TRAINING FOR AFFECTIVE SELF-DISCLOSURE OF UNDERGRADUATE MALES BY Nancy Lee Voight Self-management strategies were used in the present study as primary or secondary treatment procedures for training 36 undergraduate male volunteers in self- disclosure of affect. Three training conditions were used-- goal—directed practice, self-reinforcement, and no—treatment control. The two treatment groups utilized programmed texts which defined the target behavior and outlined a three-week personal change project for which a person would develop a shaping plan, establish weekly goals, record daily progress on cards, and contract for change. The self—reinforcement group administered reinforcers to themselves when they met goals. Subjects were tested after three weeks on an audio response test and three pencil—and—paper measures. All subjects were previously exposed to the measures in a comparison study. The level of treatment in the comparison study was used as a blocking variable for the present study. Nancy Lee Voight Two~way multivariate analysis of variance techni— ques were used for hypothesis testing. The five dependent variables were amount of affect, quality of affect, anxiety, skill, and attitude. Amount and quality scores were derived from ratings made by two trained raters from audio tape recordings of simulated interactions with best friends. The criteria for quality had five subscales: reference, time orientation, appropriateness of affect, reason given, and specificity of reason. Hoyt inter-rater reliability was .99 for amount and .99 for quality. Anxiety was self-reported as the subjects responded to the audio stimulus situation. Skill was measured by a multiple- choice discrimination test with a Hoyt reliability of .87. Attitude was measured by responses on a one to seven Likert-type scale which was designed to assess reactions and opinions to disclosure of affect. Hoyt reliability was .84 for this test. All measures and criteria were developed especially for this study and the companion research. Three hypotheses were tested at alpha level .05. Only two were supported. As predicted, no significant interaction was found between the blocking variable, prior treatment, and the present treatments. Also as predicted, the effects of self-reinforcement and goal-directed practice were equivalent. However, the third hypothesis was not supported. This hypothesis predicted significant differ- ences between no-treatment control and self-reinforcement or goal-directed practice. Although no differences were Nancy Lee Voight significant, observation of group means showed that the goal-directed practice group scored best on all measures, and self-reinforcement was second on all but anxiety. Supplemental tests were run to explore other dimen- sions not included in formal hypothesis testing. Multi- variate analysis of variance and alpha level of .05 were used when significance testing was done. Two analyses tested the three hypotheses using reduced or transformed variables. The three most highly correlated variables-- amount, quality and skill-dwere used in one analysis. In the other, a mean quality score was used in place of quality and amount. These two analyses yielded results consistent with the results of hypothesis testing. In addition, repeated measures multivariate analysis of variance was used to assess differences among groups on eight subscale dimensions of amount, quality, skill, and anxiety. These dimensions covered all combinations of initiator-respondent, positive-negative affect, and male— female partner. A significant multivariate and univariate interaction was found for quality on the feeling dimension. For amount, univariate significance was found for the same dimension. On another supplemental test, the independent variable was the subject's total training experience including prior treatment. In this onedway multivariate test, the both treatments group was significantly better than other groups. No difference was found between prior Nancy Lee Voight treatment and self—management or between these two and the neither treatment group. The latter finding was contrary to prediction. Finally, Open-ended responses to a debriefing ques— tionnaire were categorized. Subjects reported gains in awareness and improvement of skills. However, only 80% reported that they always did their change project, and only 54% of self-reinforcement and 29% of the goal-directed group met their goals most or all of the time. Results of the study suggested need for revision and further study. The combination treatment seemed promising, but the self—management packages alone needed strengthening. Also subject attrition caused a reduced sample size, and increased size would have produced a more powerful design. In retrospect, it also seemed that amount and quality were the best measures and that the other measures could have been omitted. In conclusion, it appeared that self-management could be useful in combina- tion with other treatments and that either goal setting or self-reinforcement would be useful. However, stronger treatment and methodology should be used to determine fully if self-management itself is valuable in training for self- disclosure of affect. THE EFFECTS OF SELF-MANAGEMENT STRATEGIES AS INITIAL TRAINING AND AS SECONDARY TRAINING FOR AFFECTIVE SELF-DISCLOSURE OF UNDERGRADUATE MALES BY Nancy Lee Voight 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 1975 ACKNOWLEDGEMENTS In the preparation of this dissertation and my three years at Michigan State University, several people have provided stimulation, assistance, and support. I would like to acknowledge and thank the following persons for their contributions to my growth and education: Norman R. Stewart, major professor, for consistent concern and attention and for stimulation to think in terms of systems and interdependency of parts of a whole. Also, Norm has always had faith in my abilities, even when I was doubtful. Pamela Highlen, colleague and friend, for inspria- tion and challenge when discussing ideas and preparing projects. Richard Johnson, Robert Craig, and Thomas Gunnings, committee members, for thoughtful sharing of their experi- ence and knowledge. Dick, also, for his open-door and quick mind to answer questions and give advice. Robert Carr, Nancy Martin, and William Schmidt for excellent guidance in developing computer programs and giving statistical input. ii My spring term practicum students for their flexi- bility and understanding of my time demands. Jan Skinner especially for volunteering to type tape transcripts. Pat Kay and Pam McCrary for excellent rating Pilot test subjects for giving of their time. Ron Riggs also for assistance in reliability computation. Finally, Jay Van Hoven, whom I married between Chapters III and IV, for his support and understanding. iii TABLE OF CONTENTS Page ACKNOWLEDGEMENTS . . . . . . . . . . . . . ii LIST OF TABLES . . . . . . . . . . . . . . vi LIST OF FIGURES . . . . . . . . . . . . . xi Chapter I. INTRODUCTION AND REVIEW OF LITERATURE . . . . 1 Introduction . . . . . . . . . . . . 1 Purpose . . . . . . . . . . . . 1 Importance . . . . . . . . . . . 2 Generalizability . . . . . . . . . 4 Summary . . . . . . . . . . . . 4 Review of Related Literature . . . . . . 5 Self-Disclosure . . . . . . . . . 5 Self-Management Strategies . . . . . . 14 Summary . . . . . . . . . . . . 21 II. EXPERIMENTAL DESIGN AND PROCEDURES . . . . . 23 Overview . . . . . . . . . . . . 23 Research Hypotheses . . . . . . . . . 24 Definition of Terms . . . . . . . . . 25 Research Design Over Time . . . . . . . 28 Design Over Variables . . . . . . . . . 29 Procedures . . . . . . . . . . . 31 Population and Sample . . . . . . . 31 Treatment . . . . . . . . . 32 Instrumentation and Data Collection . . . . 37 Data Analysis . . . . . . . . . . . 43 Summary . . . . . . . . . . . . . 45 iv Chapter Page III. RESULTS . . . . . . . . . . . . . . 47 Overview . . . . . . . . . . . . 47 Hypothesis Testing . . . . . . . . . . 50 Test of Hypothesis 1 . . . . . . . . 50 Test of Hypothesis 2 . . . . . . . . 50 Test of Hypothesis 3 . . . . . . . . 52 Additional Data . . . . . . . . . 53 Summary . . . . . . . . . . 53 Supplemental Testing . . . . . . . . . 55 Overview . . . . . . . . . 55 Test Across Total Training Experience . . 57 Multivariate Testing with Three Dependent Variables . . . . . . . 61 Repeated Measures . . . . . . . . . 63 Test with Mean Quality . . . . . . . 80 Debriefing Questionnaire . . . . . . 84 Summary . . . . . . . . . . . . . 89 IV. SUMMARY, DISCUSSION AND IMPLICATIONS . . . . 91 Summary . . . . . . . . . . . . . 91 Limitations . . . . . . . . . . . 95 Discussion of Results . . . . . . . . . 98 Implications of the Study . . . . . . . 102 Suggestions for Future Research . . . . . 103 In Retrospect . . . . . . . . . . 104 BIBLIOGRAPHY . . . . . . . . . . . . . . 106 APPENDICES A. Advertisement for Subjects . . . . . . 113 B. Poster Advertisement . . . . . . . . 114 C. Letter to Subjects . . . . . . . . 115 D. Self-Management Projects . . . . . . 116 E. Sample Goal Cards . . . . . . . . . 158 F. Audio Performance Test . . . . . . . 162 G. Anxiety Scale . . . . . . 166 H. Amount and Quality Rating Criteria . . . 169 I. Skill Test . . . . . . . . . . . 179 J. Attitude Scale . . . . . . . . . . 185 K. Debriefing Questionnaire . . . . . . 188 Table 1. LIST OF TABLES Page Multivariate and Univariate Analysis of Variance and Step-Down §_Statistics, for Quality, Skill, Amount, Anxiety, and Attitude--Prior Treatment x Follow-up Treatment Interaction Cell Size N_= 3 . . . 51 Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Quality, Skill, Amount, Anxiety, and Attitude—~Se1f-Reinforcement (n = 12) - Goal-Directed Practice (n_= IZI‘= 0 . . . . 51 Multivariate and Univariate Analysis of Variance and Step-Down §_Statistics for Quality, Skill, Amount, Anxiety, and Attitude--No—Treatment Control (3 = 12) — Self-Reinforcement (n = 12) + Goal- Directed Practice (2 = 12) = 0 . . . . . . 52 Observed Cell Means and Standard Deviations for Goal-Directed Behavior, Self-Reinforcement, and No-Treatment Control on Quality, Skill, Amount, Anxiety, and Attitude . . . . . . 54 Correlation Matrix for Quality, Amount, Anxiety, Attitude and Skill Dependent variables 0 O O O O O O O O O O O O 55 Multivariate and Univariate Analysis of Variance and Step-Down 3 Statistics for Quality, Skill, Amount, Anxiety, and Attitude--Both Treatments (g_= 12) - Neither Treatment (n_= 6) + Prior Treatment Self-Management (g = 6) (a = 12) vi Table Page 7. Multivariate and Univariate Analysis of Variance and Step-Down E Statistics for Quality, Skill, Amount, Anxiety, and Attitude--Prior Treatment (n = 6) - Self-Management (n = 12) = D . . . . . . 58 8. Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Quality, Skill, Amount, Anxiety, and Attitude--Neither treatment (n_= 6) — Prior Treatment (né6)+Se1f—Reinforcement (§?12) 2 = 0 . . . . . . . . . . . . . . . 59 9. Observed Cell Means for Both Treatments, Neither Treatment, Prior Treatment, and Self— Management on Quality, Skill, Amount, Anxiety, and Attitude . . . . . . . . . . . . 60 10. Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Quality, Amount, and Ski11--Prior Treatment x Treatment Present Interaction Test size. n_= 3 . . . . . . . . . . . . . . 61 11. Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Quality, Amount, and Ski11--Goal-Directed Practice (3': 12) - Self-Reinforcement (g=12)=0............62 12. Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Quality, Amount, and Ski11--No Treatment Control (n_= 12) - Goal-Directed _ Self-Reinforcement Practice (né12) (n'= 12) 2 T ..63 13. Multivariate and Univariate Analysis of Variance and Step-Down E Statistics for Eight Subscale Dimensions of Quality-— Prior Treatment x Present Treatment Interaction Test Cell Size 2,: 3 . . . . . 65 vii Table Page 14. Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Eight Subscale Dimensions of Skill-- Prior Treatment x Present Treatment Inter- action Test Cell Size n_= 3 . . . . . . . 66 15. Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Eight Subscale Dimensions of Amount-— Prior Treatment x Present Treatment Interaction Test Cell Size n_= 3 . . . . . 67 16. Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Eight Subscale Dimensiofis of Anxiety-- Prior Treatment x Present Treatment Interaction Test Cell Size n_= 3 . . . . . 68 17. Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Eight Subscale Dimensions of Quality-- Goal-Directed Practice (n_= 12) - Self-Reinforcement (n = 12) = 0 . . . . . 72 18. Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Eight Sucscale Dimensiofis of Skill—— Goal-Directed Practice (n’= 12) — Self-Reinforcement (E_= 12) = 0 . . . . . 73 19. Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Eight Subscale Dimensions of Amount-- Goal-Directed Practice (3 = 12) - Self-Reinforcement (n_= 12) = 0 . . . . . 74 20. Multivariate and Univariate Analysis of Variance and Step-Down 2 Statistics for Eight Subscale Dimensions of Anxiety-- Goal-Directed Practice (2.: 12) - Self-Reinforcement (2’: 12) = 0 . . . . . 75 21. Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Eight Subscale Dimensiofis of Quality-— No-Treatment Control (n_= 12) - Goal-Directed + Self-Reinforcement Practice (2:12) (2.: 12) 2 viii Table 22. 23. 24. 25. 26. 27. 28. Page Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Eight Subscale Dimensions of Skill-- No-Treatment Control (a = 12) - Goal-Directed + Self—Reinforcement Practice (.17 12) (n_= 12) _ 0 2 — 0.77 Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Eight Subscale Dimensions of Amount-- No-Treatment Control (n_= 12) - Goal-Directed Self-Reinforcement Practice (né12) (2_= 12) _ 0 2 _ .078 Multivariate and Univariate Analysis of Variance and Step-Down‘g Statistics for Eight Subscale Dimensions of Anxiety- No-Treatment Control (2 = 12) — Goal-Directed Self-Reinforcement Practice (g_= 12) (2.: 12) 2 ‘0..79 Correlation Matrix for Quality Total, Amount, Anxiety, Attitude, Skill and Mean Quality Dependent Variables . . . . . . . . . . 81 Multivariate and Univariate Analysis of Variance and Step—Down F Statistics for Anxiety, Attitude, Skill and Mean Quality—- Prior Treatment x Present Treatment Inter- action Test 3 = 3 . . . . . . . . . . 82 Multivariate and Univariate Analysis of Variance and Step-Down 3 Statistics for Anxiety, Attitude, Skill, and Mean Quality--Goa1-Directed Practice (n_= 12) - Self-Reinforcement (2.: 12) = 0 . . . . . . 82 Multivariate and Univariate Analysis of Variance and Step-Down 5 Statistics for Anxiety, Attitude, Skill, and Mean Quality--No-Treatment Control (n'= 12) - Goal-Directed + Self-Reinforcement Practice (n=12) (n=12) 2 ix Table 29. 30. 31. 32. 33. Frequency Count of Categorized Responses to "What Was of Most Value to You in the Training?" . . . . . . . . . . . Frequency Count of Categorized Responses to "Did You Learn Anything New About Your“ self? If so, What?" Frequency Count of Categorized Responses to "How Do You Feel You've Benefited From the Experience?" . . . . . . . . . Frequency Counts of Subjects Who Did Their Self-Management Project and Met Goals Regularly . . . . . . . . . . . . Frequency Counts of Preference for the Role of Initiator or Respondent, the Expression of Positive or Negative Affect and Male or Female Partners . . . . . . . . . . Page . 85 88 89 LIST OF FIGURES Figure Page 1. Post-test only design over time with randomized blocks . . . . . . . . . . 28 2. Two-factor fully crossed design across variables. Equal n's per cell . . . . . . 30 3. Flow-chart diagram of self-reinforcement treatment program . . . . . . . . . . 34 4. Flow-chart diagram of goal—directed practice Program . . . . . . . . . . . . . 35 5. Interaction effect for amount: feeling .dimension 6. Interaction effect for quality: feeling dimension . . . . . . . . . . . . . 70 xi CHAP TER I INTRODUCTION AND REVIEW OF THE LITERATURE Introduction Males in this culture have been socialized in ways which inhibit their expression of emotion and emotional reactions (Lynn, 1964). This concern can be conceptualized as a need to increase verbalized affect and develop clearer affect statements. According to social learning theory and recent research, some of the ways to institute changes in verbal behavior include behavioral modeling, coqnitive modeling with self-statements, practice in goal-directed behavior, and positive reinforcement. The latter two, practice and reinforcement could be used as primary training methods or as secondary training or follow-up procedures for short- term modeling or cognitive treatments. Purpose The overall purpose of this study was to assess the results of using social learning change strategies and behavioral measurement techniques in self-disclosure research. The study investigated the effects of the self— management strategies of goal-directed practice and self- reinforcement both as treatments for changing verbal affective self—disclosure and as secondary treatments given a social learning treatment or a cognitive treat— ment. Differences in amount and quality of verbalized affective self-disclosure were primarily measured by behavioral tests rather than by traditional pencil-and- paper tests. Importance Interest has recently been expressed in the application of behavioral change strategies to humanistic goals (Thoresen, 1974; Thoresen & Mahoney, 1974). Self- management strategies have been suggested as methods well suited for this purpose; however, no reported study has used self-management for altering or maintaining affective self-disclosure statements. In Maslow's terms (1954), self-disclosure may be a component in satisfying one's need for understanding. It also aids in communication of risk-statements, a component in transcendent behavior (Jourard, 1968). However, much of the prior self—disclosure research has yielded ambiguous or contradictory results. In a recent review, Cozby (1973) cites inadequate pencil— and-paper measures and poor situational control as possible sources of research confusion. Thoresen (1974), speaking of humanistic goals in general, noted that they are often loosely defined and subject to ambiguity in interpretation. The present study introduced multiple variables (the major ones being behaviorally tested) as well as clear opera- tional definition, and control of situational variables. The study was focused on the affective component of self— disclosure. In addition to the desirability of combining human- istic and behavioral concepts, it is valuable to investigate methods of self-management for both treatment and follow-up. Treatment effects were important as the first application of self-management to this area. As a secondary treatment, self-management could help to alleviate a continued problem of short-term treatments, the loss of effect during follow-up. Finally, a researchable self—management issue was addressed. Self-reinforcement has been assumed to be the critical element in the chain of self-monitoring, self- evaluation, and self—reinforcement. However, the possibility remained that goal-directed practice, which leads to a positive self-evaluation, could be as effective as formal self-reinforcement contingencies applied to goal—directed practice (Kanfer, 1970). Generalizability The findings of this study might have impact beyond the present situation. Generalization would be limited to college students. However, as much research deals with university students, findings from this study could be generalized to other single male college students. Within this population, replications would be possible to explore: 1. use of self-management as a follow-up procedure for other short-term treatments 2. use of self-management as a treatment for other affective concerns 3. use of the complete self—management package and the social learning and cognitive structuring packages (Highlen, 1975) as adjuncts to other treatments designed to develop general social skills or desirable personal growth group or therapy behaviors Summary The present study utilized some new treatment and measurement methods with self—disclosure research and con- trolled for several variables which have had confounding effects in prior studies. Also, the efficacy of two self- management packages was tested in both primary and secondary treatment situations. As these were new areas of inquiry regarding affective self-disclosure, results may be of interest to future researchers. Review of Related Literature Self-Disclosure Self-disclosure, as defined by Pearce and Sharp (1973), encompasses verbal interactions in which one person voluntarily tells another any information about himself or herself which the other is unlikely to know or discover from other sources. These communications can include both past and present content and can consist of sharing events, values, or feelings (Benner, 1968). Several psychological theorists have considered self-disclosure important for full personality functioning (Fromm, 1947, 1956; Goffman, 1959; Horney, 1936, 1950; Jourard, 1959, 1964; Rogers, 1957). It would appear that development of self-disclosing capacities, then, would be an element in one's personal growth. In the past seventeen years, researchers have studied self-disclosure in order to understand its compo- nents and the characteristics of the self-disclosing person. Many of these studies have yielded ambiguous results which are difficult to interpret because of four consistent problems in self-disclosure studies: 1. content ambiguity-~Self-disclosure can be cogni- tive, attitudinal, or affective. It can also be past- or present-oriented. Many studies deal with these three elements as a set, and this may obscure certain differences. 2. situational context--Se1f—disclosure appears to be very situation—bound, and it is difficult to generalize from lab or test situations to actual behavior. 3. measurement validity--The majority of early studies used pencil-and—paper report measures such as the Jourard Self-Disclosure Questionnaire. There is, however, no strong evidence of the validity of this measure (Benner, 1968; Hurley and Hurley, 1969). 4. statistical analysis--It is not uncommon for correlational studies to be done relating certain variables to self-disclosure. Aside from the measurement difficulty cited above, sometimes very small and possibly meaningless differences have been reported as significant statistically and these are then cited as "proof" of a significant relationship. Contradictory findings have resulted. As this inconsistency is consistent, it is not the intention of this review to critique methodologically all prior self- disclosure research. Rather, the more stable prior find- ings will be discussed briefly to form a tentative des- cription of both the components of self—disclosure and the characteristics of a self-disclosing person. Concepts and findings will be related to a social learning framework. Theoretical input on self—disclosure. Writings of modern psychologists have covered two domains which are related to self-disclosure. One domain deals with the negative effects of alienation and estrangement from others. These conditions are hypothesized to be factors in modern neuroses (Fromm, 1956; Horney, 1936). The other domain represents the positive aspects of transparency and sharing which were once equated with mental health (Jourard, 1959). However, research studies failed to demonstrate positive correlations between self-disclosure and mental health Cozby, 1973). Jourard (1964) then hypothesized that a curvilinear relationship exists between self-disclosure and mental health, and that some high disclosers were indiscriminate and made others anxious with their instant intimacy. These high self-disclosers may well be having difficulty with situation discrimination and are sharing with others irregardless of the nature of their personal relationship. In learning theory terms, the response has been generalized to inappropriate situations, apparently because the person misinterpreted antecedent events. In relation to alienation and estrangement neuroses, the playing of roles (Goffman, 1959) and games (Berne, 1964) may be a factor. People conceal their most intimate feelings and thoughts, and this may be a factor in their recognition of their own separateness from others. Polansky and Brown (1969) indeed conceive of the desire for self- concealment as a defense against separation anxiety. They hypothesize that a low—discloser wants to be exactly the same as the other person, and that low-disclosers are reminded of their separateness when they do disclose. The response described here appears to be an anticipation of negative consequences which elicits an avoidance response. It may be that this avoidance response is a component of neuroses. In relation to positive mental health, self- disclosure has been linked to the self—actualization process. In order to be transcendent and commit oneself to experiencing, one should be Open, alert, and transparent (Jourard, 1964). Transparency implies a spontaneous fusion of feeling and action as well as a willingness to commit oneself to others through self—disclosure. In addition, the self-actualized person glories in his/her separateness or uniqueness because of self—acceptance. In learning terms, this relates to self-observation, self- evaluation, and self-reinforcement. Theoretical writings, then, focus on two dimensions-- avoiding neuroses and attaining self-actualization. Both of these dimensions can be related to self—disclosure. .Although the amount, intimacy level, and situational compo- nents of "normal" self—disclosure are not discussed, it seems that amount alone is not related linearly to positive mental health. Individuals who would increase their self- disclosure, then, could follow the following learning framework: 1. development of the response 2. attention to situational cues 3. extinction of irrational anxiety 4. self-observation 5. self-evaluation 6. self-reinforcement It is this framework which has been used in the present study as well as in the companion research by Highlen (1975). The components of self-disclosure. Self- disclosure has been related to situational and relationship components such as reciprocity, liking, intimacy, social approval, time, and personality of those who are inter- acting (Cozby, 1973). Reciprocity and social approval seem to be the best supported of these components. Both are relationship variables in that they deal with concepts which must be examined in the light of a relationship. They are also situation-bound in that relationships vary in intimacy and power. What is appropriate for an indi— vidual in one relationship may not be appropriate in another. Reciprocity has been examined by Jourard and associates (1959b, 1960, 1963), as well as Bundza and Simonson (1973), Ehrlich and Graever (1971), Levinter and Senn (1967), Levin and Gergin (1969), Tognoli (1969), and Sayicki (1972). Despite methodological differences, most of these report that one's self-disclosure in a dyad is to some degree a function of the other's self-disclosure. It also appears that high self-disclosing therapists and 10 others can elicit increased self-disclosure from low- disclosers. Trust, reinforcement, and modeling have all been hypothesized to account for the reciprocity effect; however, none have been adequately researched at this time. Social approval as a component of self-disclosure can be easily understood from a learning context. Behaviors which are socially reinforced, especially by significant others, are likely to be repeated. Thibaut and Kelley (1959) hypothesize that each individual has a repertoire of alternative responses and that he/she will choose among alternatives based on the expectation of rewards or punish— ment from others. Argyle and Kendon (1967) see this not so much as a striving for reward but as a desire for equili— brium so an individual can balance anxiety and rewards in making a self-disclosing statement. Support for social approval as a component of self- disclosure came from experimental studies--the majority of which were limited to self-referenced affect, which made them especially pertinent for the present study. The major variable has been a difference in reinforcement schedule or reinforcing response. Taylor et a1. (1969), in a study not limited to affect, found that continuous positive reinforce— ment and delayed continuous positive reinforcement were related significantly to longer duration and depth of self- disclosure than either continuous negative or later negative conditions. In studies dealing with affect in the counsel- ing interview, Crowley (1970), Barnabei (1972), and ll Auerswald (1972) all utilized high-level simulation situa- tions and behavioral measures. Crowley conditioned client positive and negative self-referenced affect using counselor restatement as a social reinforcer. Crowley found that positive affect was conditioned while negative affect was not. Barnabei (1972) used non-contigent counselor reflec- tions, probes, confrontation, and free-style responding and noted the effect on self-referenced affect. As could be expected, non-contingent reinforcement did not produce differences in affect emission. Auerswald (1972) used restatement and interpretations as reinforcers for produc- tion of affect and found that interpretation served as a more powerful reinforcer for her population. These studies, then, indicate that: 1. social approval can reinforce self«disclosure and increase it, and, 2. in therapy, counselor statements can be contingent reinforcers to modify self- disclosure within the interview. In regard to research dealing with components of self-disclosure, then, several variables may be important. At present, both reciprocity and social approval appear to be significant elements. Within social approval, specific affect research has demonstrated that reinforcement from others can influence affective responses. 12 Characteristics of the self-disclosing person. Cozby (1973) in a review of self-disclosure literature has cited studies relating to family patterns, sex, race, cultural factors, mental health, and personality charac- teristics in describing the disclosing person. Although no characteristic has been clearly defined, the self— discloser is more likely to be female and middle class in this culture. The general tendency to do research on college students may bias this description, however. Cozby contended that family patterns appear to be more influential in determining the object of one's self- disclosure than the extent of self—disclosure. Social class rather than race may be influential in determining self- disclosure (Jaffee & Polansky, 1962; Tulkin, 1970). Cozby believed that this may be a function of social class differences in reinforcement of verbalizations during a child's development. A modeling effect may also be operat- ing. Findings regarding mental health have been discussed previously. Currently, research on personality correlates has proved inconclusive. The one characteristic which is most consistently found is that females tend to disclose more than males (Pederson & Breglio, 1968). Although a number of studies have reported no sex differences (Certner, 1973; Mulcahy, 1973; P109, 1965; Vondracek & Marshall, 1971) no study has reported male disclosure greater than female. Considering 13 the present state of self-disclosure research, Cozby (1973) concluded that this may actually support a sex difference. If so, this would be consistent with sociological theory. Balswick and Peek (1973),for instance, stated that the learning of male sex role results in a male's thinking that he should not be emotionally expressive. Lynn (1964) noted that boys are taught their sex role by punishment for girl- 1ike behaviors in the absence of a positive male model and without clear direction. This results in inconsistent socialization for males, hostility towards females, and anxiety regarding sex-typed behavior. Studies which have found females to be more self-disclosing have frequently included intimacy level as a dependent variable in addition to amount (Davis, 1973; Green, 1973; Jourard & Lasakow, 1958; Pederson & Higbee, 1969; and Sousa-Poza et al., 1973). Kopfstein and Kopfstein (1973) reported no sex differences in intimacy level, but their study used intimacy questions asked of another, which may represent a somewhat separate construct. The use of a social learning treatment and self- disclosing feeling statements as the dependent variable make the Green study (1973) of special interest. The information on characteristics of self- disclosers is contradictory and vague. Although sex may be a variable (and it is the most consistently supported one), contradictory evidence can be cited regarding it, also. 14 In conclusion, self-disclosure cannot be fully described from the literature. Theorists believe self- disclosure can work against alienation and, thus, neurosis. Other theorists believe it is a component in self- actualization. Apparently self-disclosure is reciprocal and sensitive to rewards—costs expectancies in a given relationship. It is difficult to support any particular personal characteristics of self-disclosers, although sex and sex-role learning may be variables. Many self- disclosure studies have research flaws stemming from content ambiguity, situational context, measurement validity, and statistical analysis. Input from this review of self-disclosure literature has been incorporated into the present research in the following ways: 1. situations and relationship factors will be controlled by limiting to best male and female friend 2. sex differences will be controlled by limiting the study to males 3. behavioral measures and inferential statistics will be used 4. only the affective component will be measured Self-Management Strategies Behavioral humanism. Thoresen (1974) has noted that self-management techniques can be used to reach 15 humanistic ends by behavioral means. Humanistic tenets include a focus on awareness of self and feelings, communi- cation, transcendence and growth, freedom, individual self-determination, and responsibility for one's own behavior (Thoresen, 1974; Thoresen & Mahoney, 1974). In May's terms (1969) one can develop Will by setting goals and engaging in behaviors to reach that goal. Learning theory, as well as self-disclosure research, would indi- cate that positive reinforcement contingencies increase the probability that behaviors will be increased. The essence of self—management, according to Mahoney (1972), is that an organism makes a response to modify the probability of another response. In other words, the person is his/her own change agent and controls the target behavior. In terms of the current study, which sought to utilize self-management strategies to modify or maintain affective self-disclosure responses, aspects of behavioral humanism which are relevant are: l. a focus on present, feeling state responses 2. individual standards for performance and self-evaluation 3. contracting for change and assuming respon- sibility for change 4. self-reinforcement 5. personal growth through self-disclosure 16 The Specific treatment packages used in this study are outlined in Chapter II. Experimental clinical studies and theoretical formulations which relate to the content of the packages will be reviewed here. Goal—directed practice. As noted above, May (1969) cites goal setting and goal-directed behavior as charac- teristic of Will. Stieper and Wiener (1965) have identified goal setting as an important factor in traditional counseling processes. In two theoretical models, self-management is described as a three-component process: (a) self-monitoring or self-awareness, (b) self-evaluation or standard reaching, and (c) self-reinforcement (Bandura, 1971; Kanfer, 1971). It is the self-evaluation component which is equivalent to the goal-directed treatment. In order to selfvevaluate, a person must know precisely what behaviors to evaluate, must commit him/herself in light of that personal standard. The importance of standard setting and realistic self—evaluation has been stressed by Bandura (1971). The establishment of a poor self-concept (Rogers, 1957) may well be related to expectations of others which are too high or too vague. This leads to internalized high standards which make positive self-evaluation impossible. By changing this process in terms of a person's standards of progress in practicing self-disclosing, the goal-directed practice provides experiences in positive self-evaluation. 17 Research in goal-directed practice has been scanty, although one component, the self-monitoring process, has been acknowledged both as a selfeobservational or awareness- recording device and as a possible change technique itself (Goldfried & Merbaum, 1972; Kazdin, 1974; Mahoney & Thoresen, 1974; Tharp & Wetzel, 1968; Thoresen & Mahoney, 1974). Contracting for increasing self-disclosure is reported to be successful in group counseling by Ribner (1974). Thoresen and Mahoney (1974) note that comparing performance to a personal standard constitutes self- evaluation. Kanfer and Phillips (1970) note that perhaps standards are traditionally vague, and a person may easily slip into negative evaluations or external evaluations in the absence of a clear criteria. Daily goal-setting can help facilitate positive self—evaluation by providing a clear-cut standard. If an individual develops objective internal standards, then,these standards should be resistant to change by external forces (Festinger, 1954). Therefore, goal-directed practice may lead to internalized positive self-evaluation in regard to affective self- disclosure. Three reported research studies deal directly with self—evaluation or goal setting. When compared to self— monitoring in the management of obesity, goal-setting did not significantly improve performance (Mahoney, 1972). Social reinforcement appeared to be superior to a combination 18 of self-observation and goal-setting in a group study by Kolb, et al. (1968). Stuhr (1972) used a training pro— cedure for eliciting personal feeling questions and insti- tuted training and practice in self—management as follow-up. Results indicated that performance only slightly decreased after two weeks, but the exact role of selfuevaluation processes was unclear. Also, other follow-up treatments were not compared for relative efficacy. DeSpite disappoint- ing results, it can hardly be said that the four studies constitute a condemnation of self-evaluation treatments. Additional research is clearly called for. Goal-directed practice has been traditionally considered important for change. Standard setting and self-evaluation are connected in goal-directed practice because a goal constitutes a clear standard against which to measure performance. If the goal were met, a positive self-evaluation could occur. Research results on goal- setting as a treatment technique have been negative. As these studies are few in number and do not deal with goal- setting as a follow-up, further research seems justified. In addition, Kanfer (1970) notes that at least partial independence exists between self—evaluation and self— reinforcement, one of the treatments to which it will be compared. Self-reinforcement. In the context of changing behavior in vivo, one has to realize that external l9 reinforcement may not automatically occur. In fact, because of the homeostatic mechanisms in interpersonal systems, initial responses to change may be negative, and the person may have difficulty implementing the change. In Kanfer and Karoly's terms (1972), the individual is in a situation in which it is possible to engage in a high- probability behavior (non-self—disclosure),but wishes to make a low-probability response (an affective self- disclosure statement). Self-reinforcement may provide the person with a reward expectancy sufficient to elicit the low-probability response. Self-reinforcement is closely related to personal growth. In order to be responsible for oneself, one must be sensitive to feedback from others, but must rely on oneself for self-reinforcement in conflict situations. For this to occur, a person may need to value self—reinforcement over external reinforcement (Dollard & Miller, 1950). Self-reinforcement can also provide a means by which one can modify or learn divergent behaviors which.are not commonly reinforced socially, but have value in expressing uniqueness. Two aspects of self-reinforcement will be considered here--self—reinforcement as treatment and as follow-up techniques. Examples of behaviors studied will also be reported. In a study by Rehm and Marston (1968), self- reinforcement was compared to a control group and to a 20 non-specific therapy group. The self-reinforcement group demonstrated significantly greater reduction in anxiety. Several well-designed studies have compared self- reinforcement with external reinforcement in a variety of settings. It would appear that the two reinforcement processes are equivalent. Bandura (1969) believes that self—reinforcement may be more powerful than external means because it eases transfer of training and hence maintenance and generalization. Self-reinforcement has shown slight superiority over external techniques for controlling dis- ruptive children (Bolstad & Johnson, 1972). Equivalence has been demonstrated in dealing with college study habits (Jackson & Van Zoost, 1972), academic performance (Glynn, 1970), motor tasks (Hall, 1973), verbal discrimination performance (Unmacht & Obitz, 1974), and memory (Bellack & Tillman, 1974). Thus, self-reinforcement appears to be supported as a treatment strategy. It would be preferred over external reinforcement in cases involving humanistic goals or the desirability of establishing maintenance effects. As a secondary technique, self-reinforcement has not been fully studied. Bandura (1969) and Kanfer (1970) write that self-reinforcement should facilitate both persistence of behavior change and generalization. In fact, it seems to be a natural way to demonstrate changes in vivo which have been established in counseling situations. Too often such a transition is not made and treatment effects quickly 21 fade. Two studies which deal with self—reinforcement as a follow-up technique have been reported. In 1962, Marston and Kanfer trained subjects in a nonsense syllable discrim- ination task to a 60 percent accuracy criterion. On brief follow-up trials, an external reinforcement group improved performance, a self-reinforcement group maintained per- formance levels, and an extinction group declined in accuracy. In a non-comparative follow-up, Kaufman and O'Leary (1972) used self rewards as an auxiliary device to encourage maintenance effects. Unfortunately, no other treatment was run, so it is difficult to use results of this study. Summary In reviewing research and theoretical writings on self-disclosure and self-management, it appears that both. the topic and the technique need further study and clari- fication. Although self-disclosure seems to be a component in positive mental health and striving for humanistic goals, the role of affective self-disclosure is unclear. Affect, however, seems to be controlled by social learning variables and would be responsive to reinforcement and positive self-evaluation. Therefore, primary and secondary treatments for affective self-disclosure which utilize goal- directed practice as a method for creating positive self- evaluation or self-reinforcement for making appropriate 22 affect statements could be effective. Their relative efficacy is difficult to predict. Since both treatments utilize a self-management framework, they are consistent with humanistic goals for the self-directed person. CHAPTER II EXPERIMENTAL DESIGN AND PROCEDURES Overview Experimental procedures were used in this study to assess the effects of goal-directed practice and self- reinforcement strategies on the affective self-disclosure of 36 single undergraduate males. Eighteen of the subjects had previously received training in affective self- disclosure through the training modes of social modeling and cognitive structuring (Highlen, 1975). For these subjects, the self-management strategies were considered to be secondary procedures. The remaining 18 subjects had received attention placebo or no-treatment conditions from Highlen (1975). For these subjects, self-management was considered to be the primary treatment for affective self- disclosure. The self-management strategies were presented in programmed text format, and each subject developed an individualized change program and change contract before the finished training. A no-treatment-control condition (traditional follow-up) was also utilized. All subjects were tested after three weeks' time, and scores were 23 24 computed on five dependent measures. All measures were devised especially for this study and the companion study by Highlen (1975). Three measures were pencil-and-paper measures testing attitudes, skill in discrimination, and self-reported anxiety in simulated dyadic interactions. The remaining two measures, amount of affect and quality of affective self-disclosure,involved the recording of the subjects' actual responses made in simulated dyadic interactions with their best male or female friends. Typed transcripts were made from the recorded responses and these transcripts were rated by trained raters. This chapter will outline the experimental design and procedures used to administer treatment, gather data, and analyze results. Research Hypotheses Because of the exploratory nature of this research, it was difficult to develop expectations of results since such expectations involved broad extrapolations from vaguely related studies. However, it was predicted that no interaction effect would be demonstrated between goal- directed practice, self-reinforcement and no-treatment control, and levels of the prior treatment, the blocking variable. In addition it was predicted that the main effects of goal-directed practice and self-reinforcement would be equivalent and either would be superior to the no-treatment control condition. 25 Hypotheses, then, were: 1. No interaction effect will be demonstrated between levels of prior treatment and levels of present treatment. 2. The main effects of goal-directed practice and self-reinforcement will be equivalent. 3. The main effect of no-treatment control will be less than the main effects of either goal-directed practice or self- reinforcement. Definition of Terms The following terms were used consistently through- out this study. To facilitate understanding, they were defined as follows: 1. affective self-disclosure--voluntary verbal statement made as an initiator or respondent in a dyadic interaction which expresses the speaker's emotions in feeling terms, is present-oriented, and self-referenced (using the pronoun I). 2. amount of affect-—the number of responses containing at least one verb, adverb, or adjective affect word which expresses emotion in the context of the reSponse. 3. quality of affect--the quality of a statement previously classified as affect under the definition for amount of affect. Quality is determined by ratings on the following five categories: reference, time orientation, appropriateness of affect, reason, and specificity of reason (see Appendix H). 26 4. concomitant anxietyé—anxiety level self—reported on a 1-7 scale as a person engages in affective self- disclosure in simulated dyadic interactions. 5. self-disclosure skills--the ability to discrim- inate appropriate affective self-disclosure as measured by responses to a multiple choice paper-and-pencil test. 6. attitudes towards self—disclosure--evaluative responses regarding affective self-disclosure as measured by a pencil-and-paper instrument. 7. prior treatment—-any one of the four treatment conditions and tests which are already completed. These treatments defined by Highlen (1975) are: a. social modeling package-~a one—hour video tape presentation with instruction on affect self-disclosure and modeling examples of appropriate and inappropriate interactions with best friends b. cognitive structuring package--a one-hour video tape presentation identical to the social modeling package except for the addition of covert rehearsal c. attention p1acebo--a one-hour audio relaxation tape d. no-treatment control—~other than during testing,no contact with the subjects 8. present treatment--any one of these treatment packages in this study used as primary or secondary train- ing eXperiences. 27 a. self—reinforcement b. goal-directed practice 0. no-treatment control 9. self-reinforcement--a self—management package including operationalized definition, description of pro- cedure, goal setting, principles of shaping, inventorying of reinforcers, choice of reinforcers, principles of self- reinforcement, a model of practice, self-managed goal- directed practice with monitoring, and self-reinforcement. 10. goal-directed practice--a self—management treatment package including operationalized definition, description of procedure, goal setting, principles of shaping, a model of practice, self-managed goal-directed practice with monitoring. 11. no—treatment control--no contact with the subjects except testing. 12. self—managementf—any response made by an organism to modify the probability of another response (Cautela, 1969). 13. self-control--the process through which an individual becomes the principal agent in guiding, direct- ing, and regulating those features of his own behavior that might eventually lead to desired positive consequences (Goldfried & Merbaum, 1972). 28 Research Design Over Time Subjects volunteered and received treatment on a staggered schedule encompassing nearly three weeks. However, each subject had a three-week waiting period before testing. As subjects volunteered, they were assigned to treatment groups by a previously determined random schedule. This insured both that random assignments to prior treatment occurred and that subjects were randomly assigned to present treatment from prior treatment blocks. The study took the form of a post-test only experi- mental design with randomized blocking, as represented in Figure l. X1 0 X2 0 X3 0 T1 WWW X1 X2 X3 T2 sows: orbc> X1 0 X2 0 X3 0 T3 WWW X1 0 X2 0 X3 0 T4 WWW Figure l. Post-test only design over time with randomized blocks. 29 The st represent levels of present treatments. T's represent levels of prior treatment; 5 is random assign- ment; QDis post-testing. The solid line between groups is indicative of random assignment. This design, as described by Campbell and Stanley (1963), presents no threats to internal validity. A no-treatment control group was included in both prior and present treatment. Generalization is limited to similar groups which have received prior treatment and multi-component testing but no other threats to external validity were inherent in the design. Design Over Variables The variable matrix was a two-factor fully crossed design with multiple dependent measures and equal numbers of observations per cell. The independent variables were T = prior treatment, a blocking variable, and x = present treatment. The design is illustrated in Figure 2. The dependent variables were: 1. amount of affect-—measured by ratings of the subject's verbal responses in 16 simulated dyadic inter- actions with best male and female friends. The criteria for measuring this variable are in Appendix H. 2. quality of affective self-disclosure--measured by ratings of the subjects verbal responses which had previously been rated as affect. The criteria for measur- ing the five categories of this variable are in Appendix H. 30 51 54 57 T1 . . . s3 56 59 510 513 816 T2 - - - 512 515 518 519 522 525 T3 . . . 521 524 527 S28 531 534 T4 0 o o 530 533 536 Figure 2. Two-factor fully crossed design across variables. Equal n's per cell. 3. concomitant anxiety--measured by the subject's self report of anxiety on a 1-7 scale. The anxiety was reported for each situation on the verbal test. The anxiety scale is located in Appendix G. 4. affective self-disclosure skills--measured by multiple choice responses to a pencil-and-paper measure designed to test discrimination of appropriate affective self-disclosure in situational contexts. This measure is located in Appendix I. 5. attitudes toward affective self—disclosure-- measured by evaluations given on a seven point scale regarding reactions to one's own self-disclosure. This measure is located in Appendix J. 31 Procedures Population and Sample Subjects were drawn from a pool of single under- graduate male volunteers who were enrolled at Michigan State University in the Spring term, 1975. They ranged from freshman to senior in class rank and from 18 to 26 years old. The average age was 21.5 years. Prior to the study advertisements were placed in the Michigan State News, a daily campus newspaper, to solicit volunteers (see Appendix A). In addition, posters were displayed in numberous dormitories (see Appendix B). As volunteers called, they were randomly assigned to a prior treatment group and to a present treatment group. The random assign- ment procedure was determined prior to assignment of subjects so that early volunteers would not be clustered in any one group. When newspaper advertisements and dormitory posters failed to produce sufficient numbers of volunteers, sub- jects were solicited by a paid recruiter who went to various dormitories and contacted people with information about the project. After three weeks, recruiting was halted when the total number of subjects equaled 52. All subjects in this study had been previously tested by Highlen (1975) using the five dependent measures. In addition, 18 had received short-term videotape training on affective self-disclosure. 32 Subjects were aware of the time commitment of the study and that it required two sessions separated by a three-week period. They were also fully aware of the topic of the training, and all had expressed an interest in developing their ability to express feelings more effectively. Prior to their final testing appointments, all subjects were contacted by mail and informed that a choice of three reinforcers (money, a plant, or a bottle of wine) would be available after completing the testing and a debriefing questionnaire. This letter is included in Appendix C. The value of the reinforcers averaged $1.00 each. Because of the attrition of five subjects and the need for equal p's per cell, subjects were randomly dis- carded from treatment groups and usable p_reduced to 36. At the final appointment, subjects who had not received a self-management treatment package were given one. Also each was offered the opportunity to list his summer address and receive a copy of the treatment results. Treatment Subjects received one of three treatments. The first was the self-reinforcement group. In Phase 1, a programmed text was presented to each subject individually. This programmed text is in Appendix D. Given variations in reading time, instruction varied in length from 45 to 75 minutes. Instruction was divided into seven units, with 33 tests for understanding following the first six units. Provision was made for recycling if the subject's under- standing did not meet criteria of 90% for each unit. A flowchart diagram of the Phase 1 programmed text is presented in Figure 3. At the end of Phase 1, subjects showed their change contracts to an experimental assistant, reviewed the content of the programmed test,and discussed their own self-reinforcement projects with the assistant. Assistants were instructed only to give suggestions if a subject misunderstood content or the nature of the change project. No alterations were made in change contracts or self- reinforcement project plans. Subjects then returned their programmed texts and handed in their contracts. During Phase 2, subjects were instructed to use the packet of cards they prepared while working on the pro- grammed text to record when they meet their goals received reinforcement. In addition they were to record the number of feeling statements made to males and females and whether their experience with expressing affect that day had been good, bad, or neutral. COpies of sample cards are included in Appendix E. The goal-directed practice group also received a two-phase treatment. In Phase 1, diagrammed in Figure 4, subjects completed a programmed text on goal-directed behavior. Reading time varied from 30 to 60 minutes. As 134 Figure 3. program. Flow-chart diagram of self-reinforcement treatment INTRODUCE PRESENT GIVE CHECK UNDERSTANDING DESCRIBE TREATMENT OPERATIONAL EXAMPLES SELF- I“. DEFINITION OF DOES NOT MEETS F. REINFORCEMENT DEFINITION MEET CRITERIA CRITERIA PROCEDURE 1.0 2.0 3.0 4.1 4.2 5.0 4.0 ECYCLE_7 CHECK UNDERSTANDING MEETS [DOES NOT MEET I PRACTICE GIVE PRESENT GOAL CRITERIA CRITERIA GOAL EXAMPLES SETTING 9.2 9.1 WRITING 0F GOALS 6.0 ' 9.0 7.0 #ECYCLE J PRESENT GIVE T DEVELOP CHECK UNDERSTANDING ,PRESENT A PRINCIPLES EXAMPLE SHAPING A 1. , INVENTORY OF DE PLAN .. DOES NOT I MEETS 1. OF SHAPING SHAPING 12.0 MEET CRITERIA CRITERIA REINFORCERS 10.0 11.0 13.1 13.2 14.0 ‘t 13.0 IRECYCLETE —*RECYCLE—r CHECK UNDERSTANDING CHOOSE DIRECT To OF GOOD REINFORCER REINFORCER COMPLETE 18 . 0 INVENTORY MEETS DOES NOT ‘ 1 15. CRITERIA MEET CRITERIA 19.1 19.2 GIVE PRESENT PRINCIPLES OF EXAMPLES REINFORCEMENT 19.0 OF REINFORCEMENT 17.0 GIVE EXAMPLES OF DEVELOP REINFORCEMENT CHECK UNDERSTANDING REINFORCEMENT PLAN - PLAN *---4D 20.0 21.0 MEETS DOES NOT CRITERIA MEET CRITERIA . 22.1 22.2 I 22.0 CHECK UNDERSTANDING MODEL IMPLEMENTATION SUMMARIzE TOTAL PLAN 24.0 23.0 MEETS DOES NOT CRITERIA MEET CRITERIA ‘I 25.1 25.2 RECYCLE 1 25.0 MAKE CONTRACT REINFORCE END TREATMENT PROGRAM 26.0 27.0 28.0 35 .Eouwoum OOHDOOMQ pouomufltiamom mo Eoummflp DHGSOISOHM .v Guzman o.mH Zmo m>HO Bzmwmmm o.m mmbomoomm DmBUmmHD IAO -o.m _ , III N.m m.m 0.5 o.w dHMMBHmU 9mm: HO WUHBUHU Bzmwmmm MUDaomBZH 36 in self-reinforcement, provision was made for recylcing and mastery learning. Each subject also discussed his goal-directed practice project, change contract, and the instruction content with an assistant. Conditions were the same as they were with the self-reinforcement group. Through the instruction, subjects developed a self-management plan as well as a concept of the definition of verbal affective self-disclosure; but no instruction in self-reinforcement principles was given. Each subject had a recording booklet to note daily goals achieved, their reactions, and the number of statements to males and females. This group also contracted to meet goals and record in their booklets for three weeks. The goal- directed group had all the components to make accurate self-evaluations and a self—management procedure designed to yield positive self-evaluations, but no provision had been made for them to self-reinforce. The essential components of both treatment packages followed the Bandura (1971) and Kanfer (1971) models for self-management. Self-observation was facilitated by a clear operational definition. Graduated self-imposed goals and contracting for change were seen as elements in self-evaluation. In the self—reinforcement condition, choice of reinforcers and self-determined reinforcement plans were used to plan for self-reinforcement. 37 The no-treatment control group was dismissed following their prior treatment posttest and advised to return for follow-up in three weeks. At the final testing time, four procedures were used: 1. All subjects were posttested. 2. Subjects who have received both prior treatment- social learning/cognitive structuring and present treatment-self-reinforcement/goal-directed practice were debriefed using a questionnaire and feedback sheet. 3. Subjects from prior treatment-attention placebo/ no treatment control were randomly assigned to a group treatment of either social learning or cognitive structuring. 4. Subjects from present treatment—no treatment control were administered either self-reinforcement or goal-directed practice. Assignment was random. Instrumentation and Data Collection Two measurement procedures were used to yield data for the five dependent variables. One procedure involved a behavioral test which measured amount of affect and quality of verbal affective self-disclosure as well as self-reported concomitant anxiety. The second procedure, pencil-and-paper testing, included a skill test and an attitude test. Both procedures had been previously administered to the subjects as part of their prior treat- ment. 38 The two pencil-and-paper tests were pretested on a population of 20 male undergraduate volunteers from a psychology class at Olivet College, a small liberal arts college 30 miles from Michigan State University. Volunteers were offered extra class credit for taking the two tests. Hoyt reliability coefficients were computed for the pre-test (Guilford, 1954). The reliability for the skill test was .31, and the reliability for the attitude test was .83. All measures were then administered to a pilot test group of 20 male college graduates who had received prior treatment. Reliability was again computed for the skills and attitude tests using Hoyt's formula. Reliabilities were .91 for skills and .86 for attitude. Both of these measures were then revised following item analysis with procedures described in Mehrens and Lehmann (1973). Some poor items were omitted and many foils on multiple choice items were rewritten. The audiotape recordings of the pilot test subjects‘ responses for amount and quality were transcribed and used to train the raters used in the actual study. The raters, two female high school English teachers, held bachelor's degrees and both had over 30 hours towards a master‘s degree. Both were employed in the same school. They had been close friends for six years. They were paid for two training sessions and were paid $1.00 per transcript during the actual test. For training of raters, the following teaching methods were used: 39 1. instruction on criteria 2. discussion of criteria 3. modeling of use of criteris with transcripts 4. guided practice using criteris with.transcripts 5. discussion and clarification 6. reliability check 7. discussion of reliability check 8. revision of criteria 9. final reliability check for training The first reliability check, on two transcripts, yielded Hoyt coefficients of 1.00 for amount, and .09 for quality. Within quality subscales, the coefficients were: 1. reference = .51 2. time orientation = .00 3. appropriateness of affect = .00 4. reason present = .84 5. specificity of reason = 1.00 Following revision of criteria and further training of the raters, a second reliability check was run on two trans- cripts. This yielded a Hoyt coefficient of 1.00 for amount and .99 for quality. At this rating, quality subscale coefficients were: 1. reference = 1.00 2. time orientation = 1.00 3. appropriateness of affect = 1.00 4. reason present = .85 5. specificity of reason = .99' 40 Copies of the final measures as well as the rating criteria are in the appendices. The behavioral test con- sisted of 16 stimulus audio tape situations (see Appendix F). Eight involved male partners and eight females. Situa- tions equally represented the subject in the role of initia— tor and as respondent. The situations presented positive and negative emotions equally. Subjects were instructed to consider the taped voices those of their best male friend or best female friend to control for situation intimacy level. After each stimulus, subjects responded into a second recorder. A lS-second time limit was imposed. On a paper scale, subjects circled the number (one through seven) which represented their anxiety regarding self—disclosure on that stimulus and in that situation. The one-seven scale was in a Likert-type scale format (see Appendix G). Following the behavioral test, typescripts were made of the subjects' responses and sentences were unitized using a modification of Auld and White's (1956) rules. In the unitizing for this study, all "thank you" phrases were omitted because they were classified as incomplete units. For measuring amount of affect, a list of feeling words developed by Crowley (1970) and expanded by Auerswald (1972) was used (see Appendix H). The raters rated the typescripts by giving a score of one to a response which included at least one affect unit and a score of zero to responses which had no affect units. These scores were totaled across responses for each subject. 41 Quality ratings were made according to the criteria outlined in Appendix H. Five separate ratings were made as follows: 1. reference - scoring range, 0—12 2. time orientation - scoring range, 0-6 3. appropriateness of affect - score 0 or 3 4. reason present - score 0 or 5 5. specificity of reason - score 0 or 8 Using the criteria for amount and quality, the trained raters scored subject typescripts. Eight randomly selected type- scripts were chosen to be rated for the reliability check for inter-rater reliability for the actual study. Hoyt reliability coefficients were .99 for amount and .99 for quality. Quality subscale reliabilities were: 1. reference = .99 2. time orientation = .98 3. appropriateness of affect = .97 4. reason present = .98‘ 5. specificity of reason = .99 Scores on the quality variable were totaled for each indi— vidual. A total was also computed for the anxiety measure. The behavioral measure, then, yielded three numerical totals which correspond to three dependent variables: amount, quality, and anxiety. The mean score for quality was 346.36 with a standard deviation of 102.39. Scores ranged from 0 to 582. A perfect score would have been 624. 42 The mean score for amount was 10.28 with a standard devia— tion of 2.83. Scores ranged from 1-16, which was a perfect score. For anxiety, the mean score was 46.67 with a standard deviation of 12.13. The range of scores was from the minimum possible, 16, to 64. It was possible to score 112. The paper-and-pencil tests involved skill and atti- tudes. The skill test was a four—foil multiple choice measure with correct reSponses scored as one point and incorrect responses scored as zero (see Appendix I). Some distractors were cognitively oriented, and some involved blaming or other-referenced affect. Question stems were equally divided between positive and negative stimuli, respondent and initiator situations, and male and female best friends as targets. Scores on the skill test were totaled to develop one score for each individual. The Hoyt reliability coefficient computed for the skill test in this experiment was .87. Skill mean score was 10.78 with a stan— dard deviation of 5.42. Scores ranged from 1-23, with 24 representing a perfect score. The attitude test consisted of items measuring reactions or evaluations of statements about self-disclosure of affect (see Appendix J). Items on the attitude measure were rated on a seven-point Likert—type scale on a range from strongly agree to strongly disagree. All items were brief sentences. Scores were totaled following scale adjustment 43 to account for positive-cued questions (with "strongly agree" expected) and negative—cued questions (with "strongly disagree" expected). A Hoyt reliability coefficient of .84 was computed for the attitude scale. The mean score on the attitude scale was 145.94 with a standard deviation of 17.48. The actual range of scores was 99—198, and the possible range was 30-210. Totals were used as data from all five measures. Totals were considered preferable to averages because less information was lost. A debriefing questionnaire was developed to accom- plish two purposes: 1. to gather additional anecdotal data from the subjects regarding the self-management experience 2. to assess reactions to the experiment The debriefing questionnaire is included in Appendix K. All questions were open-ended and allowed the subject con- siderable freedom to respond. Data Analysis Data was analyzed using a two-way multivariate analysis of variance to test both interaction and main effects. The variables were arranged for the step-down F procedure in the following order: 1. amount 2. quality 44 3. skills 4. attitude 5. anxiety Although not a part of hypothesis testing, other procedures were used to analyze additional data: 1. A one-way multivariate analysis of variance technique was used to explore differences on the five dependent variables. A new independent variable representing the subject's total experimental experience was formed for this analysis. Four amounts of training were compared. These were: a. both treatments--subjects who received cognitive structuring or social modeling from Highlen (1975) as well as either goal-directed practice or self- reinforcement in the present study. b. neither treatment--subjects who received either attention placebo or no treatment control from Highlen (1975) and no-treatment control in the present study c. prior treatment--subjects who received either cognitive structuring or social modeling from Highlen (1975) and no-treatment control d. self-management--subjects who received attention placebo or no treatment control from Highlen and goal-directed practice or self-reinforcement 2. A two-way multivariate analysis of variance was utilized to examine the most discriminating three 45 variables-—quality, amount and Skill apart from anxiety and attitude. 3. A repeated measures multivariate analysis of variance was used to examine the male-female, initiator- respondent, and positive-negative dimensions of the quality, skill, and anxiety measures. 4. A two-way multivariate analysis of variance was run using the subjects' mean quality score (quality total divided by amount total) instead of total quality score. Four dependent variables were used in this analysis. They were mean quality, skill, anxiety and'attitude. 5. Frequency counts were computed on responses to the debriefing questionnaire. Summary The procedures outlined above can be eXpected to provide a valid test for interactional and treatment effect hypotheses on each dependent variable. With the pilot testing and behavioral measures, instrumentation in the present study should be both reasonably valid and reliable. Prior reviews of self-disclosure measures have called for such checks as well as for increased use of behavioral measures (Benner, 1968; Hurley and Hurley, 1969; Vondracek, 1969a; 1969b). The data analysis techniques should be strong enough to detect large differences if they exist although the number of subjects was small. Finally, the treatment packages were based on instructional theory and 46 behavior change theory. Mastery learning with provision for modeling, practice and recycling allowed all individuals to grasp essential components fully. These components are elements of self-management theory as described by Bandura (1971) and Kanfer (1971). It is expected that replication could be accomplished with relative ease. CHAPTER III RESULTS Overview Hypothesis test results are presented in this chapter along with the results of supplemental tests. The experiment was designed so that posttest measurement provided scores on five dependent variables for each of 36 single male undergraduates. The five dependent variables measured dif- ferent aspects of self-disclosure of affect. These aspects were: 1. amount of affect 2. quality of affect 3. anxiety 4. skill 5. attitude Scores for amount and quality of affect were totaled from ratings made by two trained raters on set criteria. Raters judged typed transcripts made from the subjects' recorded responses to sixteen simulated dyadic interactions. Con- comitant anxiety was self-reported on a paper-and-pencil 47 48 measure as subjects responded to the situations. The anxiety score was totaled across situations. The skill and attitude tests were pencil-and-paper measures assessing skill in discriminating quality affect statements and personal attitudes towards self-disclosure of affect. Totals were used as scores for these measures. TWo-way multivariate analysis of variance was used for hypothesis testing. A blocking variable, level of prior treatment was used. The main independent variable involved three levels of treatment. These were: 1. goal-directed practice 2. self-reinforcement 3. no-treatment control Three hypotheses tests were run. All tests were done with alpha level set at .05. It was predicted that there would be no interaction effect. It was also predicted that goal- directed practice and self-reinforcement would be equivalent and that either would be superior to no-treatment control. Supplemental data was gathered and analyzed to explore certain dimensions and comparisons not formally tested by hypotheses. A.one-way multivariate analysis of variance was run comparing four levels of a transformed independent variable which corresponded to the subjects' complete training experience. The training amounts compared were: 1. both treatments--training in both experiments 2. neither treatment--no training 49 3. prior treatment--training from previous study only 4. self-management--training from present experiment only This test also used an alpha level of .05 and all five dependent variables. It was predicted that Group One onld be superior to Groups Two, Three; and Four. It was also predicted that Groups Three and Four would be equivalent and that either would be superior to Group Two. The second supplemental test involved a duplication of hypothesis testing using only three dependent variables-- amount of affect, quality of affect, and skill. Third, the comparisons used in hypothesis testing were made across repeated measures formed on three dimen- sions in the amount of affect, quality of affect, anxiety, and skill tests. These dimensions were: initiator- respondent, positive-negative affect, and male-female interactant partner. Fourth, hypotheses tests were duplicated substituting a mean quality of affect score for the total quality score and using all dependent measures except amount and quality. Finally, descriptive statistics and frequencies were computed on subjects' open-ended responses to questions of relevance on a debriefing questionnaire. All results from hypothesis testing and supplemental testing are contained in this chapter. Results are discussed and further explained in Chapter IV. 50 Hypothesis Testing Test of Hypothesis 1 The first hypothesis predicted that there would be no interaction effect. The probability of a Type I error was set at .05 for the multivariate test. In univariate significance testing, alpha level was partitioned for the five dependent measures and each was tested at the .01 level. The results of this analysis are presented in Table 1. No differences were significant, so Hypothesis 1 was supported (§_= 1.14 with 15.55.6 df). It stated: H - No interaction effect will be demonstrated 1 between levels of prior treatment and levels of present treatment. Test of Hypothesis 2 The second hypothesis predicted that effects of the two self-management procedures would be the same. The hypothesis stated: H2: The main effects of goal-directed practice and self-reinforcement will be equivalent. This hypothesis was tested by multivariate (E.< .05) and univariate (p.< .01) analyses of variance. The hypothesis of no difference was supported by the failure to reject the null form hypothesis at p_< .70 (E = .6030 with 6,20 9f). Results are presented in Table 2. 51 Table l Multivariate and Univeriate Analysis of Variance and ’Step-Down F Statistics, for Quality, Skill, Amount, Anxiety, and Attitude Prior Treatment x . Present Treatment Interaction Cell Size N.= 3 Multivariate §_= 1.14 with 15,55.6 df, p_< .35 Dependent Mean Squares Univariate F Step-Down Measures Between 6,24 df p F_ p Quality 4606.86 .44 .73 .44 .73 Skill ‘32.85 1.12 .36 1.58 .22 Amount 2.61 .33 .81 2.11 .13 Anxiety 178.02 1.21 .33 .91 .45 Attitude 281.69 .92 .45 .86 .48 Table 2 Multivariate and Univariate Analysis of Variance and Step-Down §_Statistics for Quality, Skill, Amount, Anxiety, and Attitude Self—Reinforcement (p_= 12) — Goal-Directed Practice (p'= 12) = 0 Multivariate §'= .60 with 5,20 df, p_< .70 Dependent Mean Square Univariate F Step-Down Measures Between 1 , 24 d_f_ ‘ p g p Quality 3765.03 .36 .55 .36 .55 Skill 4.17 .14 .71 .01 .93 Amount 4.17 .52 .48‘ .52 .48 Anxiety 337.50 2.29 .14 2.18 .15 Attitude 32.67 .11 .75 .06 .81 52 Test of Hypothesis 3 The third hypothesis was structured to test the effects of either treatment compared to no treatment. The hypothesis stated: H3: The main effect of no-treatment control will be less than the main effect of either goal-directed practice or self- reinforcement. This hypothesis was not supported by the multivariate analysis, and no significant differences were found between treatment and no treatment (F = 1.00. with 5,20 df; p < .44). Univariate analysis of variance using alpha level of .01 also demonstrated no significance. These results are presented in Table 3. Table 3 Multivariate and Univariate Analysis of Variance and Step-Down 3 Statistics for Quality, Skill, Amount, Anxiety, and Attitude No-Treatment Control (3’: 12) — Self-Reinforcement (p_= 12) + Goal-Directed Practice (2.: 12) = 0 Multivariate §'= 1.00 with 5,20 pg; p < .44 Dependent Mean Square Univariate §_ Step-Down Measures Between 1,24 d£_ p_ F. p_ Quality 51392.18 4.90 .04 4.90 .04 Skill 88.89 3.02 .10 .43 .52 Amount 40.50 5.06 .03 .39 .54 Anxiety 4.50 .03 .86 .00 .98 Attitude 382.72 1.25 .27 .02 .88 53 Additional Data Further information regarding the results of hypo— . theses tests can be derived from observing the cell means and for goal-directed practice, self-reinforcement, and standard deviations no-treatment control. These data are presented in Table 4. On all dependent measures except anxiety, the goal—directed practice group means were the highest of the three groups. The low score on anxiety represented the best mean on this measure. 'Therefore the goal-directed practice group scored best on all five measures; self—reinforcement ranked second on all measures except anxiety. In addition to the cell means, the correlation matrix for the dependent variables has been reported. This matrix is in Table 5. It‘appears that while quality and amount are highly correlated, both are only slightly corre— lated with attitude and skill, which are unrelated to each other. The anxiety measure is not correlated with any of the other four dependent variables. Summagy \ Hypothesis tests supported two hypotheses and failed to support one. Multivariate analysis of variance techniques were used for the tests of all hypotheses and alpha level was set at .05. In addition to reSults of hypotheses tests, relevant observed cell means and the variable correlation matrix were reported. 54 .muOchm Boa MODMOMUCM ouoom 309R Hm.ma wm.ma mH.m mm.m OH.NHH I m.HvH 5H.mv mh.m mm.m m.mmm ANH n CV Honucou psmauomueloz om.mH No.0 Hm.a vh.m mh.mh I H.h¢H hm.om mm.oa om.HH m.omm ANA u CV usmfim0H0msflom Imaom mm.m~ 5H.OH wm.m mo.m Hm.mm I v.mvH hH.mv Nv.HH mm.NH m.mmm ANH H CV H0H>m£mm CODOOHHQIHMOO .Q.m coo: .Q.m CD02 .o.m com: .o.m com: .n.m coo: moonw OCSDHDD< «mumsxcfi pcooE< Haflxm wpflaoso Opsufluum Cam .mumflxsm .ucsoam .Hawxm .mpwamso co Houucou usoaumoualoz CEO .pcofimOH0mswomlmaom .Hoa>mnmm Couoonwolamoo How msOMDOH>OQ pumpsoum Cam memos HHOU OOPHOmoo s EHEES 55 Table 5 Correlation Matrix for Quality, Amount, Anxiety, Attitude and Skill Dependent Variables Quality Amount Anxiety Attitude Skill Quality 1.00 Amount .98 1.00 Anxiety .08 .10 1.00 Attitude .45 .49 .06 1.00 Skill .50 .42 .04 .05 1.00 Sppplemental Testing Overview Supplemental analyses were designed to assess addi- tional data which was available from the experimental pro- cedures employed, the dependent measures used, and the debriefing questionnaire completed by the subjects. Because the present study was conducted in conjunction with one by Highlen (l975).it was possible to assess the differences at final posttesting (a three-week time differential from the posttesting by Highlen) between subjects who had received training in both studies, subjects who had received no training, subjects who received prior treatment only, and those who had received self-management only. The nature of the dependent measures provided three additional tests which utilized group comparisons identical to those in hypothesis testing. In one of the three 56 analyses, the three most highly correlated measures--amount of affect, quality of affect, and skill--were used as the only three variables. In another, a mean score for quality of affect was computed and used instead of the total quality score. This new variable was used with three other variables-—skill, attitude, and anxiety. The skill, anxiety, amount of affect, and quality of affect measures each con- tained items which could be analysed across dimensions of initiator-respondent, positive-negative affect, and male- female interactant. Because of this, repeated measures two- way multivariate analysis of variance techniques were used to assess differences among the testing contrasts for each of those four dependent measures. Finally, responses to the open-ended debriefing questionnaire were categorized and are reported in this chapter. The debriefing questionnaire provided data regarding subjects' impressions of the value and benefit of their total training experience (including treatment by Highlen, 1975), their estimates of their relative facility across the initiator-reSpondent, positive-negative affect, and male-female interactant dimensions, and a report of the degree to which they actually fulfilled their self— management contracts. These supplemental analyses, excluding debriefing questionnaire material, were tested with alpha level set at .05. Although formal hypotheses were not made, 57 comparisons were tested and are reported in a manner similar to hypothesis testing. Test Across Total Training Experience Test for both treatments. This analysis predicted that the final test results across five dependent variables would demonstrate that the combination of training from the Highlen (1975) experiment and the present study would be superior to either training experience alone or to neither training experience. One-way multivariate and univariate analysis of variance tests were run with multivariate alpha at .05 and univariate alpha at .01 each. The multivariate test was significant (3.: 6.50 with 5,28 df; p_< .0004). Two univariate tests were significant. These were skill (§_= 8.03 . with 1,32 d:; E.< .0079) and attitude (§_= 11.2764 with 1,32 df; p < .0021). gestatistics for the univariate test and all other analysis results are pre— sented in Table 6. Test for equality of single treatments. This analysis predicted that the effects of prior treatment alone (tested three weeks after training) and self-management alone (tested at the end of the three-week practice period) would be equivalent. The prediction of no difference was supported, and the null form test statement was not rejected. Results are presented in Table 7. Table 6 58 Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Quality, Skill, Amount, Anxiety and Attitude Both Treatments (3.: 12) - Neither Treatment (p_= 6) + Prior Treatment ( = 6) + Self-Management (p_= l2) 0 3 Multivariate F_= 6.50 with 5,28 df; p < .0004 Dependent Mean Square Univariate F Step-Down Measure Between 1,32 d_f p ' E ‘ 3 Quality 43719.25 2.50 .12 2.50 .12 Skill 245.68 8.03 <.01 5.00 .03 Amount 55.13 4.26 .05 14.81 <.01 Anxiety .50 .00 .95 .10 .75 Attitude 3726.72 11.28 <.01 4.24 .05 Table 7 Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Quality, Skill, Amount, Anxiety, and Attitude Prior Treatment (3 = 6) - Self-Management (n_= 12) = 0 Multivariate §_= .79 with 5,28 gg Dependent Mean Square Univariate F Step-Down Measure Between 1,32 df_ p“ F. 2 Quality 128.07 .01 .93 .01 .93 Skill .44 .01 .90 .05 .83 Amount .11 .01 .93 1.62 .21 Anxiety 205.44 1.42 .24 1.09 .30 Attitude 361.00 1.09 .30 1.17 .29 59 Test for superiority of single treatments over no treatment. This analysis predicted that the groups which received a single treatment would receive significantly higher scores than members of the neither treatment, prior treatment or self-management. This hypothesis was not supported (§_= 1.39 with 5,28 df; p_< .26). Results of the multivariate and univariate analyses are in Table 8. Table 8 Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Quality, Skill, Amount, Anxiety and Attitude Neither treatment (n_= 6) — Prior Treatment (n = 6) + Self-Reinforcement (n_= 12) 2 0 Multivariate §.= 1.39 with 5,28 95! E.< .26 Dependent Mean Square UnivariaEe F Step-Down Measures Between 1,32 df p F p Quality 52121.44 2.98 .09 2.98 .09 Skill 17.01 .56 .46 .13 .72 Amount 33.35 2.58 .12 .62 .44 Anxiety 280.06 1.94 .17 2.17 .15 Attitude 24.50 .07 .79 1.02 .32 Observed cell means were examined to further support directionality of differences from the three comparisons. When considering that low anxiety scores are better than 60 high ones, the both treatments group scored better on all measures except anxiety, on which the prior treatment mean was superior. Prior_treatmen; and self-management scored better than the neither treatment group on skill, but only the prior treatment group was superior to neither treatment on attitude. Observed cell means are presented in Table 9. Two predictions made regarding the comparison of effects among both treatments, neither treatment, prior treatmen , and self—management were supported. Bpgh tneatments.proved better than either treatment alone or no treatment on two variables, attitude and skill. The single treatments were equivalent, but they were not better than the neither treatment group. Table 9 Observed Cell Means for Both Treatments, Neither Treatment, Prior Treatment, and Self-Management on Quality, Skill, Amount, Anxiety, and Attitude Group Quality Skill Amount Anxiety* Attitude Both Treatments (p_= 12) 395.64 14.50 12.00 .46.83 160.33 Neither Treatment (n'= 6) 241.00 7.50 7.33 52.50 137.00 Prior Treatment (p.= 6) 344.85 9.67 10.17 39.83 145.67 Self- Management I§_= 12) 350.56 9.33 10.0 47.00 136.17 *Low scores indicate low anxiety. 61 Multivariate Testing with Three Dependent VarIabIes The three most highly correlated measures were used in this analysis to explore effects of the reduced number of variables on the contrasts used for hypothesis testing. Results were similar to those for hypothesis testing with all five dependent variables. The two-way multivariate and univariate analysis of variance procedure was used with alpha level at .05 for multivariate and .016 for univariate tests. In analysis of the interaction between prior treatment and present treatment (goal- directed practice, self—reinforcement and no-treatment control), no significant interaction was found (§~= 1.34 with 9,53.69 df; p_< .24). This supported the original hypothesis. Results are presented in Table 10. Table 10 Multivariate and Univariate Analysis of Variance and Step-Down §_Statistics for Quality, Amount, and Skill Prior Treatment x Treatment Present Interaction Test size n = 3 Multivariate §_= 1.45 with 9,53.69 df; p_< .24 Dependent Mean Square Univariate §_ Step-Down Measure Between 1,32 d; p. F' p Quality 4606.86 .44 .73 .44 .73 Alnount 2.61 .33 .81 1.64 021 Skill 32.85 1.12 .36 2.06 .14 62 The analysis followed the hypothesis that the main effects of selfvreinforcement and goal-directed prac- tice were equivalent. This prediction was also supported (§.= .29 with 3,22 gf; p_< .83). Results of both multi- variate and univariate tests are in Table 11. Analysis was made to assess differences between goal-directed practice, self-reinforcement, and no—treatment control. Contrary to prediction, no significant differences were found. It has been hypothesized that either self- reinforcement or goal—directed practice would be superior to traditional follow-up, but this was not supported (F_= 1.82 with 3,22 df; p_< .17). Results are in Table 12. Table 11 Multivariate and Univariate Analysis of Variance and Step-Down F_Statistics for Quality, Amount, and Skill Goal Directed Practice (n’= 12) - Self-Reinforcement (p_= 12) = 0 Multivariate §_= .29 with 3,22 df; p_< .83 Dependent Mean Square Univariate F. Step-Down Measure Between 1,24 d_f_ p_ ' F p_ Quality 3765.02 .36 .55 .36 .55 Amount 4.17 .52 .48 .43 .52 Skill 4.17 .14 .71 .12 .74 63 Table 12 Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Quality, Amount, and Skill No-Treatment Control (n|= 12) - Goal-Directed Practice (ps12) - Self-Reinforcement (n=12) = 0 2 Multivariate §_= 1.82 with 3,22 df; p'< .17 Dependent Mean Square Univariate §_ Step-Down Measure Between 1 , 24 d_f p F p Quality 51392.18 4.90 .04 4.90 .04 Amount 40.50 5.06 .03 .13 .72 Skill 88.89 3.02 .10 .69 .42 Hypothesis tests were duplicated using only the three dependent variables quality, amount, and skill. Consistent with results of testing using all five measures, Hypotheses l and 2 were supported in their null form, but Hypothesis 3 was not. Repeated Measures The questions and situations used for the quality of affect, amount of affect, skill, and anxiety tests were designed to yield subscale scores across three dimensions: initiator-respondent, positive-negative affect, and male female interactant. The combinations of these dimensions yielded eight new dependent variables. Four separate two- way multivariate and univariate analysis of variance tests were run. Each test corresponded to one of the original 64 measures--quality, amount, skill, or anxiety. The same comparisons used for hypothesis testing were made. In the analysis for interaction, it was again pre- dicted that no significant interaction would occur. This hypothesis was supported for multivariate tests of three of the four variables. These were quality (F.= 1.36 with 48,87.71 df; p < .11), skill (§_= .66 with 48,87.7l d:; p < .94), and anxiety (F_= .86 with 48,87.71 9:; p_< .71). For amount, however, an interaction, significant at the p < .04 1eVe1, was found (F.= 1.54 with 48,87.71 df). For the univariate tests, only one dependent variable, initiator—negative feeling, was significant for amount (§_= 6.30 with 6,24 df; p‘< .005) and also for quality <1: 5.83 with 6,24 SEF.E < .0008). Alpha levels were .05 for the multivariate test and .005 for the univariate test. Results of the interaction test are presented in Tables 13, 14, 15, and 16. Graphs were used to explore the significant univariate interactions for amount and quality on the posi- tive-negative feeling dimension. The amount interaction is presented in Figure 5; the quality interaction is presented in Figure 6. The test between training groups predicted that goal-directed practice and self—reinforcement would have equal effects. For the repeated measures two-way multi- variate test, this prediction was supported for all four measures. With d£.8 and 17 in each case, §_ratios were: 65 .mosmowmwsmfim monocmaA mswaoom x xom x OHOH he. mm. mm. we. ma.¢H me. mo.H ma. hm.H mw.vmm msfiHoom x OHOH mm. vm.a mm. NH.H mo.mma mswamom x xmm no. em. as. me. h¢.woa xom x OHOH oo. mm.e .mooo. mm.m «h.mmve mafiammm mm. on. hm. hm. ow.mvm xom mp. mm. mm. Hm. 5H.Hb~ DHOH mo. mo.~ mo. mo.m H>.¢mm~m usmumsoo m M .m I .mm vm.m somsumm sowmsmefla mamomnsm csonlmoum m ODMHHM>HGD mumsom com: AH. v_m 1mm HS.sm.me EDA; 6m.a u.m mumNAm>ADH52 m N lam TNHm HHOU Dmmfi COHHUQHTHCH HGTEMGHB #CTWTHQ X HGQEMTHB HOflHm l MDHHMSO mo msowmcofiwn mamomnsm unmwm Mom moanmaumum m czonlmoum cam moamwum> mo mammflmcm muowhm>wso can oumfiHMPHuHsz ma OHQCB 66 mafiamom x xom x macs we. mm. om. om.H Ho. em. m~.H mm. em.H mo. mstoom x OHOH mm. mm. mu. mm. Ho. msfiHmom x xom mm. ma. hm. ov. Ho. xmm x OHOH em. as. as. mm. mo. mcwamow pm. Hm. em. mv. mo. Mom hm. Hm. mm. em. Ha. OHOH as. so.H ms. so.H em.m Demumeoo m _M .m I. mm v~.m cmmzuom sowmsoawn anomnsm csoolmmum m mumwum>wso mnmsvm coo: em. v.m “no as.sm.me EDA; so. u m ouoflum>wuasz m N m ONHm HHTU HWOH. GOfiHUMHUUH—H UCGEMOHB flfimwmhm X #QOEWOHH. HOHHm I Hflflxm no mcosmcmEAn manomnsm Deana How moflumwumum m czooimoum Usm mosofinm> mo mamhaosd Tameo>wsD can oumHHMPMDHsz vH wanna 67 .ODGMOMMMGOMM mmuosoou om. mm.H vm. mv. Ho. mswammm x xmm x macs om. mm.H ms. es.H AN. manammm x macs om. HS.H am. ms.a ma. maflammm x Mom mm. om. am. on. Ha. xom x THOR oo. ma.m .moo.v om.e om.m SEAHmom mm. mm. mm. om. mm. Mom mm. Hm. em. me. mm. THOR ma. mm.H ma. mm.a oo.ma usmumcoo m .M m I .mw v~.m :003umm sowmsmfiwo mamomnsm ssoanmoum m ODMHHOPMCD ouwsvm coo: so. v.m «mm as.sm.ms EDA: sm.a H.M mumHHOPwuflsz m U M". TNHm HHTU #mTB COHHUMHTDGH “COEHMTHB “Gmmem N #COEWOHB HOHHQ I undead mo MGOHmcmEHQ mamomnSm unmflm How mowumwumum m szonlmoum Dam mosmaHm> mo mammamsm ODDHHM>HGD paw ouDMHDPMDHDZ ma OHQMB 68 am. we. we. mm. «A. mafiamou x xmm x macs on. me.H SN. ms.H km. mcsammm x maou as. mm. em. as. om. mafiamom x New mm. mm.a ma. mm.fl mm.a xmm x maou om.. «m. cm. mm. m~.m maflaomu om. as. me. oo.H «H.m xmm Hm. om. cm. as. mm.m macs mm. SH.H mm. SH.H Hm.ama Seahmeoo m M m I. mm va.m soospom scamsmefla mamomnsm csooimoum m Gunfium>flsa mumsvm coo: as. v m «mm as.sm.me EDA; 6m. u m EDRAHESADHSE m u m ONHm HHOU umoa sofluomumusH ucmfiummus usommum x usmsummua Howum sumsxaa mo mcoflmcmEMo manomnsm Demflm How mosumflumum M ssoalmoum Cam OUGMHHO> mo mamhaoss mumwum>flsb cam oumaum>fluasz 0H OHDOB 69 Mean Score 3.00 2.75 2.50 2.25 2.00 Goal-Directed 1.75 Behavior 1.50 No-Treatment Control 1.25 1.00 .75 .50 .25 Self— 0.00 Reinforcement Social Cognitive Attention No-Treatment Modeling Structuring Placebo Control Figure 5. Interaction effect for amount: feeling dimension. Mean Score 95 90 85 80 75 7O 65 60 55 50 45 40 35 30 25 20 15 10 Figure 6. 70 Goal-Directed Practice No-Treatment Control Self- Reinforcement Social Cognitive Attention No-Treatment Modeling Structuring Placebo Control Interaction effect for quality: feeling dimension. 71 quality = 1.42 p‘< .26 skill = .26 py< .97 amount = 1.48 p‘< .23 anxiety = 1.09 p < .42 Alpha level for the multivariate test was .05. Alpha was partitioned for the univariate test and the level of .005 was used. Multivariate and univariate results for the test of no-difference across subscale dimensions between the goal-directed practice group and the self-reinforcement group are presented in Tables 17, 18, 19, and 20. The final repeated measures analysis tested for differences across dimensions among no-treatment control, goal-directed practice, and self—reinforcement. If differ— ences were found, it was predicted that either goal-directed behavior or self-reinforcement would be superior to no- treatment control. Multivariate differences were found for quality (F_= 2.83 with d: 16 and 34; py< .005) and amount (F = 2.66 with df‘lG and 34; p < .008). However, because of the interaction effect reported previously, no univariate tests were significant at alpha level .005. Non—significant multivariate §_ratios were reported for skill (§_= .64 with 16 and 34 df; p, <.82), and anxiety (§_= .85 with 16 and 34 df; p'< .62). Results are presented in Tables 21, 22, 23, and 24. The interaction effect complicated inter» pretation of significance testing for this comparison. Where differences did occur, they were interactional in form and did not rely solely on main effects. 72 mm. vo. hm. mo. vm. msHHTTM x me x THou om. no. mm. mo. vv.m msHHTTw x THOH mo. mv.m NH. mm.m b>.mmm msHHTTM x me mm. mm. mm. Ho. mH.H XTT x THOH mm. mm. mm. oo.H «H.mmh mGHHTTm mm. om. eh. HH. . mm.hm me mo. H~.v 0H. mm.~ mm.mmvm THOH mm. Hv. mm. Hv. mm.ommw usmumsoo m .M m .I .mm ¢N.H sTT3uTm sonsTEHn THTOTOWW szoolmTum m TDTHHT>HGD Tumswm :TTS mm. v m «wm SH.m E653 NS.H u M TDTHHT>HHHSZ o u Ame u my Demamouomcflmmuneom . Ame u I wuHHTso mo TGOHTETEHQ THTOmnsm uanm m czoolmTum OCT TOCTHHT> mo MHTMHTG4 TDTHHTPHED OCT TuTHHTPHuHsz 5H THQTE mv TUHDUTHm CTHOTHHQ HTOO How TOHumHuTum 73 mm. no. em. mH. co. msHHTTm x me x THOH we. om. Hm. SH.H mo. maaeomn A Teen ms. HH. am. be. cc. mcflemmm A Row om. vH.H Hm. vm. Ho. xTT x THOH . mm. mo. mm. mm. 60. SEHHTTM mm. mo. mm. Ho. 00. me mm. mm. mm. mm. HH. THO.H we. NH. vb. NH. ow. usmumsoo m M m I .mm vm.H sTTzuTm consTEHQ THTOmnsm :3OQImTum m TDTHHTPHGD Tumsvm GETS hm. v.m «mm SH.m EDHS em. u.m mumnuMSHuHsz o u ANH u mv usTETUH0msHTmIMHTm I ANH ".mv TOHuOTHm CTDOTHHQIHTOO I HHHMM no mcosmcmEHo Temomnsm DESHM Mom HOHDHHDEDM m GBOQImTum Cam TOGTHHT> mo THmmHTsé TDTHHT>HCD OCT TuTHHT>HuH52 mH THDTB 74 mm. mo.e ms. as. oo. mcHHmmm x Men x 6H69 mm. mo. He. Am. so. chHmmm x THOR mo. Hm.m me. mm.~ mm. mcMHmmm x xmm mm. 5H. oo.H oo. oo. xwm N THOH am. To. an. om. he. mcwammm om. me. he. mo. vo. me Ho. mm.h mo. mm.m hH.¢ THOH mv. mm. me. am. 5H.v unnumsoo m M m mm «N.H sTT3UTm COHmsTEHQ THTomnsm s3ooImTum m TDTHHTPHGD Tumswm sTTz mm. v.m 2mm_se.m EDAS w¢.H M TDTHHT>HDHDE o u ANH u my DETETOHOMQHTmIMHTm I ANH ".mv TOHDTTHT OTDUTHHQIHTOO l unsoad mo TGOHTGTEHQ THTOmnsm ucmHm How MOHumHuTum m EBOQImTum OCT TOGTHHT> mo THmmHTsd TDTHHT>HCD Cam TuTHHT>HuH52 mH THQTB 75 an. no. em. . so. Ho. msHHTTm x xTT x THOH mo. sm.m mo. mm.m Ho.~ SEHHTTH x «Hon km. mm. mm. Ho. Ho. meHHmmH x xmm mm. Hm. . mm. we. no. me x THOH mH. Hm.H mm. mm. Ho.~ meHHmmm mm. 00. oo.H oo. oo. me mm. mo. mm. Ho. vo. THOH vH. mm.~ «H. m~.~ om.bmm ucmpmsoo m M m .l .mm v~.H sTTsme GOHmsTEHo THTomasm ssoaImTum m TuTHHT>HsD Tumswm GETS we. v.m «mm 5H.m EHHS mo.H u m THEHHE>HHH62 o n ANH u my usTETOH0msHTmIMHTm I ANH u my TOHuOTHm pTuOTHHQIHTOO l muTHxsd mo msonsTEHo THTOmndm Danm How TOHDMHuTum m :3OQImTum Dam TOCTHHT> no THthmsd TuTHHT>HsD CGT THTHHTPHHHDS om THQTB 76 m5. Hm. me. mm. mm.w~ mGHHTTH x XTm x THOH Ho. mm.m mH. om.H Ho.mHm mGHHTTm x THOH mo. mn.m mm. m¢.H mm.mmH mGHHTTM x xTT me. me. NH. hm.m mm.Hov xTT x THo..H so. mo.m so. SH.m mm.HoH~ meHHmmm OH. mo.~ om. n~.H mh.wmm me mo. mm.m «N. om.H Nh.mm~H THOH 0H. nm.~ 0H. hm.m mm.~thm DGTHMGOU m M m. I. MM vm.m GTTsuTm GOHTGTEHQ THTomnGm GsoQImTum m TDTHHT>HGD THTva GTTS mo. v.m “mm Hm.SH EHHS . Hm.~ u.m TDRHA6>HDH52 N o n ANH ".mv uGTETOHOMGHTMIMHTm + “NH u my TOHDOTHT pTuoTHHQIHTOO I ANH u GV HOHuGOO uGTEuTTHBIoz I SDHHTGO mo MGOHTGTEHQ THTomnGm quHm How MOHumHuTum m GBOQImTum OGT TOGTHHT> mo MHTSHTGG TuTHHT>HGD OGT TumHHTPHuHGS HN THQTB 77 mm. om.H mv. mm. Ho. mGHHTTm x me x THOH om. HH. hm. hm. Ho. mGHHTTm x THOH mm. mH. mp. mm. Ho. mGHHTTm x me mm. Hm. mm. mH. Ho. me x THOH mm. NH. mm. mq.H hm. mGHHTTw em. mm. mv. on. «H. me om. Hm. mm. . Hm. mo. THOH mH. mo.m mH. mo.m Hm.m uGTuTGoo m _M .m I. mm ¢~.m GTTsnTm GonGTEHn THTomnsm GsooImTum m THTHHTPHGD THTva GTTS . . . S. v m “mm HASH 5H3 «6. n M 33353:: m o u “NH n my DGTETOHOMGHTSIMHTm + ANH u my TOHDOTHG CTDOTHHQIHTOO I ANH u GV HOHHGOU pGTEuTTHBIoz I HHme mo mGOHmGTEHQ THTomon uSmHm How TOHHMHuTuw m GsoOImTum OGT TOGTHHT> mo MHTSHTGG THTHHT>HGD OGT TuTHHT>HuHGS NN THQTB 78 .TUGTOHMHGTHT mTpOGTQR om. «m. mH. oo.m mo. mGHHTTm x xTM x THOH Ho. oo.m no. mo.m he. mGHHTTm x THOH HH.. NS.H mm. mS.H mH. maHHmmm x New mm. mm. HH. mm.m vm. xTT x THOH mm. SH.H SH. om.m SS.H SSHHTTT «H. mH.m Hm. v~.H mm. xTM mo. mo.v mo. hm.~ HH.~ THOH mo. mn.m mo. mn.m mm.mm uGTuMGOO m M m I .mm vm.~ GTTsuTm GonGTEHQ THTOTQGm GsoQImTum m TDTHHT>HGD THTGOm GTTS mo. v m @m 36H 21qu SEN u M 3353352 0 n I N I “NH n GV HGTETOHOMGHTGIOHTm + ANH n GV TOHHOTHT OTMOTHHQIHTOO ANH ".mv HOHDGOU uGTEuTTHBIOZ I uGGOEG mo mGoHMGTEHD THTOmon quHm How MUHumHuTum m GBOQImTum UGT TUGTHHT> mo THMSHTGG TuTHHT>HGD CGT TDTHHT>HDHGS mm THQTB 79 mm. mo. mm. mo. No. mGHHTTm x me x THOH Hm. Hn.H om. vn.H mo.H mGHHTTm x THOH em. mH.H me. mm. mo.H mGHHTTm x me mv. om. om. mm. we. me x THO“H HH. mm.m Hm. Hm.H . mm.m SEHHTTH mm. Ho. mm. Ho. mo. me mm. so. no. mo. mo. THOH mm. mH.H mm. 0H.H oo.HnH HGTDTGOU m M m I mm vm.m GTTBDTm GOHMGTEHQ THTomnsm GBOQImTum m TDTHHT>HGD THTGOm GTTS mm. v m “mm em.eH EDHS mm. u.m TDMHHE>HDH52 N AmH u my pGTETOHomGHTmIMHTm + ANH u my TOHHOTHG CTDOTHHQIHTOO I ANH u my HOHDGOU HGTEDTTHBIOZ I muTHxGG mo MGOHMGTEHQ THTomon quHm How MOHDMHuTum m GBOQImTum CGT TOGTHHT> mo THmmHTGG TDTHHT>HGD UGT TuTHHT>HuHGS vm THQTB 80 The four two-way repeated measures multivariate analysis of variance tests using subscale dimension scores demonstrated a significant multivariate and univariate interaction relating to quality: feeling dimension, and a significant univariate interaction on amount: feeling. Thus, the no-interaction hypothesis was not supported. Because of the interaction, main effects were not consis- tent. No significant differences were found between goal- directed practice and self-reinforcement. Multivariage significance was found in the comparison of no«treatment control to goal-directed practice and self-reinforcement. However, no univariate significance was found due to the interaction. Test with Mean Quality By dividing the total quality score for each.sub— ject by the amount score, a new variable, mean quality, was formed which correlated .49 with quality (total) and .42 with amount. The complete correlation matrix is presented in Table 25. Two-way multivariate and univariate analysis of variance techniques were used to analyze data from four dependent measures: anxiety, attitude, skill, and mean quality. Tests were made for interaction and compari- sons of treatment effects. Multivariate alpha level was established as .05; univariate alpha was .008. 81 Table 25 Correlation Matrix for Quality Total, Amount, Anxiety, Attitude, Skill and Mean Quality Dependent Variables Mean Quality Amount Anxiety Attitude Skill Quality Quality 1.00 Amount .98 Anxiety .08 Attitude .45 Skill .50 Mean Quality .49 1.00 .09 1.00 .47 - .06 1.00 .42 - .03 .05 1.00 .42_ .07 .38 .27 1.00 The test for interaction yielded an‘F ratio of 1.64 with 12 and 55.85 g; (p < .11). None of the univar- iate tests were significant. All multivariate and univar- iate results are presented in Table 26. As no interaction was predicted, these results support the hypothesis state- ment. The prediction of no difference between goal- directed practice and self-reinforcement was supported by results using anxiety, attitude, skill, and mean quality as dependent variables (§.= .53 with 4,21'df; E.< .71). Results are presented in Table 27. 82 Table 26 Multivariate and Univariate Analysis of Variance and Step-Down E Statistics for Anxiety, Attitude, Skill, and Mean Quality Prior Treatment x Present Treatment Interaction Test 3 = 3 Multivariate F = 1.64 with 12,55.85 df; p < .11 Dependent Mean Square Univariate F Step-Down Measure Between 3,24 df p 5| p Anxiety 178.02 1.21 .33 1.21 .32 Attitude 281.69 .92 .45 .92 .45 Skill 32.86 1.12 .36 .99 .42 Mean Quality 78.15 2.35 .10 3.64 .03 Table 27 Multivariate and Univariate Analysis of Variance and Step-Down F Statistics for Anxiety, Attitude, Skill, and Mean Quality Goal-Directed Practice (ps12) - Self-Reinforcement (n612) = 0 Multivariate §.= .53 with 4,21 df; E,< .71 Dependent Mean Square Univariate F Step-Down Measure Between 1,24 df p F 8 Anxiety 337.50 2.29 .14 2.29 .14 Attitude 32.67 .11 .15 .05 .83 Skill 4.17 .14 .71 .08 .79 Mean Quality .00 .00 .99 .00 .97 83 In comparison of the no-treatment control group with the goal-directed practice group and self—reinforcement group, no significant differences were found (§_= 1.04 . with 4,21 df; p.< .41). This was contrary to prediction and did not support the hypothesis. Results of the multi- variate and univariate tests are in Table 28. Duplication of hypothesis testing using three of the dependent variables with a new variable, mean quality, resulted in no significant interaction and no differences among main effects. These findings supported two hypo- theses but failed to support the hypothesis of difference between no-treatment control and the average effects of goal-directed practice and self—reinforcement. Table 28 Multivariate and Univariate Analysis of Variance and Step-Down F_Statistics for Anxiety, Attitude, Skill, and Mean Quality No—Treatment Control (§\= 12) - Goal-Directed Self-Reinforcement Practice (n=12) (n=12) " = 0 2 Multivariate §_= 1.04 with 4,21 df; p_< .41 Dependent Mean Square Univariate E, Step-Down Measures Between 1,24 SE. p. F_ p Anxiety 4.50 .03 .86 .03 .86 Attitude 382.72 1.25 .27 1.23 .28 Skill 88.89 3.02 .10 2.49 .13 Mean Quality 79.07 2.38 .14 .42 .52 84 Debriefing Questionnaire The debriefing questionnaire, completed by subjects after final testing, asked for comments regarding the total training experience and experience in practicing the use of feeling statements. Responses were Open-ended and were categorized for reporting purposes. Results reported are for those who received either goal-directed practice or self-reinforcement unless specifically stated. In responding to an initial set of questions designed to assess self-reported outcomes,.subjects reported that they gained in awareness and skill,and that they felt their relationships with others in their natural environment had improved. Frequency counts of categorized responses to these questions are presented in Tables 29, 30, and 31. Subjects were asked two questions regarding their self-management projects. One asked if they actually did the project. Eighty % (10 of 13) of the self- reinforcement group responded in the positive. One person reported that he did it "most of the time," one "sometimes," and one did not do the project. Eighty-two % (14 of 17) of the goal-directed practice group reported completing the project and 18% (three subjects) did not. However, Frequency Count of Categorized Responses to "What 85 Table 29 Was of Most Value to You in the Training?" Both Treatments F Self-Management 5 Audio test and anxiety Awareness 2 test 1 Awareness 4 Clear explanations 1 Definition of feeling Daily appraisal 1 statement 1 Evaluation of goal Daily practice 4 progress 1 Examples on videotape 1 Definition and method 1 Motivation to work on Goals achieved 1 feelings 1 Results achieved 2 Improved affect expression 1 Skill improvement 2 Programmed text 2 Training in openness 1 Reinforcement l Videotape 3 2,: l7 n.= l4 86 Table 30 Frequency Count of Categorized Responses to "Did You Learn Anything New About Yourself? If so, What?" Both Treatments §_ Self-Management F Awareness of easy-going Awareness of blaming 2 nature 1 Can express feelings withe Awareness of own quiet- out being unmanly 1 ness and nervousness 1 Communicate feelings less Awareness of resistance often than believed 8 to technique in affec- tive domain 1 Competition from other men 1 Express positive affect to females; negative to males 1 Easier to talk to females 1 Feeling statements some- times more effective 1 Nothing new 1 Hard to express feelings l Openness has positive Moods influence relation- results 2 ship 1 Positive self-image 1 Nothing new 3 Owning feelings valuable l Retreat to empathy from self-disclosure 1 Value of expressing feelings 1 When he pushes himself, he gets better results 1 16 n_= 15 II: II 87 Table 31 Frequency Count of Categorized Responses to "How Do You Feel You've Benefited from the Experience?" Both Treatments F. Self-Management F Awareness 7 Awareness 1 Easier to make friends 1 Express self better 1 Improved communication 4 Expression of feelings; not repression 1 Know people better 4 Improved communication 2 Learned about self 1 Improved relationships 2 Method constantly in Owning of feelings 1 his mind 1 Personal reactions more Practice 1 precise 1 Trust others more 1 Reflection and action on openness 2 Skill improvement 1 n“: 20 n‘= 12 when asked if they regularly met their goals, only 8% (one) of the self-reinforcement group and 29% (five) of the goal- directed practice group responded that they always met their goals. Forty-six % (6 of 13) of the self-reinforcement group met their goals most of the time. Eight % (l of 13) of self-reinforcement met goals "sometimes" as did 12% (2 of 17) of the goal-directed practice group. Thirty-eight % of the self-reinforcement group (5 of 13) and 59% (10 of 17) 88 of the goal-directed practice group did not regularly meet their goals. The frequency counts are presented in Table 32. Table 32 Frequency Counts of Subjects Who Did Their Self-Management Project and Met Goals Regularly Self-Reinforcement 2,: 13 Goal—Directed Practice 2': 17 Did project §_ Did project ‘yg yes 10 yes 14 mostly l mostly 0 sometimes 1 sometimes 0 no 1 no 3 Regularly met goals Regularly met goals yes 1 yes 5 mostly 6 mostly 0 sometimes 1 sometimes 2 no 5 no 10 Finally subjects were asked to judge whether it was easier for them to initiate or reSpond, express positive or negative feelings, and share feelings with male or female partners. Results for all four groups are presented in Table 33. Across groups the consensus was that respondent roles and positive emotions were easier, but that they could express affect to members of both sexes with equal ease. 89 Table 33 Frequency Counts of Preference for the Role of Initiator or Respondent, the Expression of Positive or Negative Affect and Male or Female Partners Both Neither Prior Self- Treatments Treatment Treatment Management n_= 15 n]: 4 n'= 5 n_= 15 Initiator 5 l l 3 Respondent 10 3 1 9 Neither 0 0 3 1 Positive 13 3 4 9 Negative 0 O 0 0 No preference 2 0 l 3 Female 8 l l 3 Male 4 0 1 5 No preference 3 3 3 7 Results from the debriefing questionnaire provided information from the subjects about their reactions to the entire training experience. Subjects reported increases in awareness and skills. Most completed their self-management projects, but many failed to meet their goals regularly. Across groups, subjects reported a preference for respondent role and positive affect, but no preference between male and female interactants. Summary Hypotheses tests supported two hypotheses and failed to support one. There was no interaction effect, and the main effects of goal-directed practice and self- reinforcement were equivalent. However, no significant 90 differences were found in comparing goal-directed practice and self-reinforcement with no—treatment control. Supplemental tests indicated that these findings held when the dependent variables were reduced to three (quality, amount, and skill) or to four (anxiety, attitude, skill and mean quality). When repeated measures subscales dimensions were used as dependent variables, however, a significant multivariate and univariate interaction occurred around the feeling dimension for quality. On the same dimension, a univariate significance was found for amount. No differences were detected between the two self- management groups on the repeated measures. Although multivariate differences were noted between no—treatment control and the self management groups, the interaction with prior treatment made this effect inconsistent. Results were reported for four groups formed on the basis of a subject's total training experience. Significant differences were found between both treatments and the other groups on attitude and skill. Other comparisons between neither treatment, prior treatment, and gel;- management revealed no significant differences. Finally, results of the debriefing questionnaire yielded impressions and judgments which may help with the interpretation of results. The interpretation, discussion, and implications follow in Chapter IV. CHAPTER IV SUMMARY, DISCUSSION AND IMPLICATIONS SummaEy The purpose of this research was to assess the effects of two self-management strategies used as primary training or as secondary- procedures for training of males in self-disclosure of affect. The subjects were 36 under- graduate males enrolled at Michigan State University during the Spring quarter, 1975. All had participated in a prior study by Highlen (1975) which immediately preceded the present research. All subjects were tested in the prior study and half received one-hour videotape training in affect self-disclosure. Prior treatment involved four possible training conditions. These conditions were utilized as blocking variables. A related purpose was to assess the application of behavioral change techniques to the achievement of humanis- tic goals. The target research behavior was conceptualized in humanistic terms as owning and sharing personal feelings with significant others. However the Operational definition and self-management change techniques were taken from 91 92 behavioral theory and change methodology. As self~ management essentially involves a person controlling his/ her own change contingencies, it was considered especially appropriate for us in attaining humanistic goals because of the personal responsibility, freedom, and commitment inherent in self-change methods. Finally, the research was designed to correct several flaws noted in prior self-disclosure studies which were reviewed in Chapter I. 1. content ambiguity--This research was limited to disclosure of affect in the present. Other studies had tended to treat cognitive and affective self-disclosure or past and present timing of disclosures as equivalent. 2. situational context--Testing for the present research was conducted in part with simulated interactions with best friends. Although this method was not as valid as measurement in actual situations, it was an improvement over pencil-and-paper measures. Also subjects practiced disclosing affect in their own environment with their best friends. 3. measurement validity-~The simulated interaction test was considered superior to self-report questionnaires which have frequently been used in this field. 4. statistical analysis--Parametric inferential statistics were used in this study to compare treatment effects. Using a pre—set alpha level (.05) greatly reduced the chance of reporting small differences as significant ones. 93 Therefore, this study attempted to introduce a stronger methodology and new change techniques for training in affect self-disclosure. Differences among three treat- ments and interaction with prior treatment were studied in data analysis. The study was designed as a posttest only with three levels of treatment and four levels of prior treat- ment, a blocking variable. Subjects were randomly assigned to blocks and randomly assigned from blocks to treatments. All subjects were volunteers. As volunteers were recruited over several days, training was staggered over three weeks, and all subjects received individual training. Each sub— ject had a three-week waiting period before final testing. For the two self-management groups, this time was designated as the span of their individual change projects. At their final appointment, subjects were tested and given small reinforcers for participating. After test- ing was completed, the control subjects were given treatment. Testing covered five dependent variables. These were amount of affect, quality of affect, anxiety, attitude, and skill. Amount and quality were measured by ratings of typescripts prepared verbatim from audiotape recordings. Anxiety was a concomitant self—report measure. Attitude and skill were measured from pencil-and—paper tests. Hoyt reliability for ratings was .99 for amount and .99 for quality. Reliability for the attitude test was .84; skill reliability was .87. 94 Two-way multivariate analysis of variance was used for hypothesis testing. Two hypotheses were supported in that no interaction was found and no significant difference was found between goal-directed practice and self- reinforcement. However, the third major hypothesis, which predicted that goal-directed practice and self—reinforcement were superior to the no-treatment control condition, was not supported. Supplemental testing resulted in the findings of hypothesis tests remaining unchanged when only quality, amount, and skill were tested as well as when anxiety, attitude, skill and mean quality were used. On repeated measures multivariate and univariate analysis of variance using eight subscale dimensions of amount, quality, skill, and attitude as dependent variables, a significant multi- variate and univariate interaction was found for quality on the feeling dimension. For amount, univariate signifi- cance was found for the same dimension. These interactions indicated quite a variance within present treatment groups on this dimension. Therefore, when significant differences were found between present treatment groups, the differ- ences were largely attributed to interaction effects. On another supplemental test, the independent variable was a subject's total training experience. In this one-way multivariate test, the both treatments group was significantly better than other groups. No difference 95 was found between prior treatment and self—management or between these two and the neither treatment group. The latter finding was contrary to prediction. Finally, results of a debriefing questionnaire were described. It seemed from these responses that the sub- jects gained in awareness and in skills. However, subjects also reported that only 80% always did their change pro- jects. Also, only 29% of the goal-directed group met their goals most or all of the time; in the self-reinforcement group, this was 54%. The purpose, methodology, and results of this research were presented as summary and background for further consideration of the results and what they mean. Limitations Three major limitations were considered relevant to understanding this research. These limitations concern validity and soundness of measures, nature of the subjects and subject attrition, and the design and methodology of the study. Validity. Three matters of validity need to be discussed. The first is a general one which relates to any study of self-disclosure. How can a researcher or practi- tioner be sure that the verbalization or responses of an individual represent disclosures of self and not parroting of the phrases of others, a script, or other artificial 96 communication? It is most difficult to tell. This pre- sented a problem in the present research. In the audiotape test from which amount and quality were measured, standard situations were used for all subjects. Interactions were simulated and some subjects did report that they would have preferred interacting with another person. To some extent, the artificial nature of the simulation must have affected the authenticity of the subjects' self-disclosure. A related validity question is one which relates to all the dependent measures. Although tested for relia- bility, none of the measures was tested for concurrent or predictive validity. With this question unsettled, it is hard to interpret results completely. The third possible threat to validity is that two scores were derived from ratings. With raters, although reliability was high, the possibility of consistent ratings which were too high or too low exists. Raters were care- fully trained, however, which lessens the probability of this as a serious limitation. Nevertheless, it should be considered. Subjects. Subjects in this experiment were under- graduate male volunteers from a large state university. Generalization of results must necessarily be limited to such populations. It is important to note that results cannot validly be extended to non-volunteer populations. Also subjects received tangible reinforcers for their 97 attendance at final testing. Although they did not know reinforcement would be offered until shortly before their appointment day, it is entirely possible that the offer of reinforcement influenced their decision to attend that session. In addition, all subjects in this research.were previously tested with.the same measures and half of them had been previously trained. The effect of testing may be a limitation, as well as the fact that the subjects were quite aware of their involvement in special experimental conditions. Another aspect of limitations related to subjects in subject attrition. The loss of subjects resulted in the loss of statistical power because cell size for hypothesis testing was reduced. Also, subjects were randomly drOpped from all groups until equal cell sizes were achieved, the subjects who did not return belonged overwhelming to the no-training conditions of prior treatment or the traditional follow-up groups. It is difficult to account for differ- ences between those who drOpped out and those who stayed except in terms of possible differences in interest and motivation. Therefore, it is conceivable that the subjects in either of the no-treatment groups were more interested and motivated than other subjects. This certainly would constitute a limitation to the study. 98 Design. Some limitations are inherent in the posttest only design. The primary one, according to Campbell and Stanley (1963) is a limit or generalization to other groups which participate in special experimental conditions. Because of previous testing, there could be a reactive effect of testing. A final limitation was the choice of five multiple measures. As quality and anxiety were the variables of major interest, it might have been better to use only these two measures. This would have altered the univariate alpha level and increased power. The limitations of this study include possible threats to validity as well as problems connected with subjects and design. These 1imitations,which may have decreased power and precision, may be partially responsible for the failure to achieve statistical significance. Discussion of Results Results of hypothesis testing were very interesting. All seemed consistent with predictions and observed cell means except the contrast between traditional follow-up and the average of goal-directed practice and self; reinforcement (see Table 4). Except for anxiety and atti- tude, it seems possible that the trend of change would favor detection of differences if the power and precision of the test were both higher or if treatments were made more potent. The smaller differences between goal-directed 99 practice and self—reinforcement make it seem unlikely that such changes would produce significant differences in terms of this contrast. Therefore, it appears that the two treatments do have equivalent effects. If this is true, economy is gained through use of the goal—directed practice package as it is the shorter one to complete and the cards require somewhat less recording. A factor to be considered in discussing hypothesis test results is the fact that not all self-management subjects fulfilled their change contracts. Insufficient motivation and unrealistic goals may account for many of the difficulties. It is worth noting that several more of the self-reinforcement groups actually met their goals, which was a condition for reinforcement. This seems to indicate a rationale for possible use of self-reinforcement, despite the fact that treatment effects were equivalent. Therefore, with college males the addition of reinforcement to a goal-directed change prOgram for affective self- disclosure does not significantly change the treatment effect, but the reinforcement may serve as additional incentive to meet goals. As for the supplemental tests, the one testing effects of the total training eXperience was the one of most interest to this researcher. The groups receiving bpip; treatments scored better than all other groups, and significant differences were found. This seems to support the contention underlying this research that short-term 100 treatments are best combined with change-oriented secondary treatment programs if results are to be maintained. The failure to detect differences between either single treatment and neither treatment is difficult to understand. Observed cell means indicate that the neither treatment group scored lower. The lack of significance may be related to possible attrition bias in the neither treat— meg; group. Low power and precision should also be con- sidered as possible causes. The other supplemental tests provided little new information. The interaction on the quality and amount repeated measures tests are difficult to explain. On the feeling dimension, positive feelings are preferred by the sample, as indicated by their responses to the debriefing questionnaire. However, this preference was uniform, and the interaction effect denotes lack of uniformity. Possibly, individual differences are overly represented because of small cell size. The possibility also remains that of those who practiced self-disclosure of affect in their own environments, some may have had very unpleasant experiences with only positive or only negative emotions. The debriefing questionnaire lends some final light on results. As noted in Chapter I, Bandura (1971) and Kanfer (1971) both conceptualize a chain of (a) self— monitoring or self-awareness, (b) self-evaluation, and (c) self-reinforcement. It was one of the intentions of 101 this study to compare the last two. However, responses on the debriefing questionnaire most clearly indicate that self-awareness was a major result. Other gains for subjects included improved skills and better interpersonal relation- ships. In summary,it seems possible that some differences were not detected because of insufficient power, small cell size, and differential subject attrition. The equivalence of goal-directed practice and self-reinforcement would probably hold even if these limitations were corrected. This is the first experimental evidence to suggest that the achievement of goals is inherently reinforcing for college students. This reinforcing effect is probably related to positive self-evaluations which come from reaching goals. On the other hand, although the groups scored similarly on all measures, many more of the self-reinforcement group regularly met their goals during their change projects. As there is no reason to believe that self-reinforcement goals were more realistic than those set by the goal-directed group, it seems likely that reinforcement served as moti- vation for meeting goals. Therefore it appears possible that self-reinforcement in this study proved to be a motivator only and that overt self-reinforcement was not a necessary condition for change. In this context, the positive feedback gained from a self-evaluation of goal achievement could be considered a covert self—reinforcer which was at least as effective as overt self—reinforcement. 102 Of the supplemental tests, the most important and interesting is the significant differences between bggh treatments and all other conditions. This combination of modeling plus self-management and cognitive structuring plus self-management is similar to many two-phase behavioral strategies used in counseling practice. As the subjects' change projects were totally unsupervised, this combined treatment cost in experimenter time was limited to produc- tion of the learning packages. Although this researcher chose to supervise the experiment personally, an assistant or clerk could easily supervise up to six subjects at a time. Thus, both treatments do not require large amounts of professional time to administer. Audiotape and videotape equipment are necessary, however. Still, both treatments are effective and relatively inexpensive. Implications of the Study Few clear implications can be drawn from this research. Many results were not statistically significant and others were contradictory. It is apparent that both treatments proved most effective, that no interaction occurred between prior treatment_and no-contact control treatments, and that goal-directed practice and self— reinforcement produced equivalent results. Implications from these results are primarily: 103 1. support for the contention that multi-phase change strategies are more effective than one-shot short term treatments 2. support for the benefit of further research into the relationship between goal—achievement and self- reinforcement In terms of the stated purposes of this study, there are further implications. The assessment of the effects of self—management is, in this researcher's opinion, incomplete due to research limitations and merits further study. In terms of the application of behavioral methods to humanistic research and the correction of earlier research flaws, much has been accomplished. Behavioral strategies, though not yet proven effective, have been used to train subjects in self-disclosure of affect. New measures, although currently untested for validity, were able to provide medium level simulation tests with controlled content and intimacy level. It would seem that these accomplishments support the use of behavioral techniques and measurement in self- disclosure research. Suggestions for Future Research As this research was conducted without many concrete prior findings, much of the preceding section has indicated directions for future research. In general, the desir- ability of continued use and development of behavioral 104 measures and behavioral techniques for accomplishing humanistic goals has been given some support by this research. Specifically, however, suggestions for future research include: 1. comparison of the effects of goal achievement and self-reinforcement as primary treatment in analogue studies 2. further study of self—management strategies as secondary treatment procedures with short term interven- tions 3. study of affect self—disclosure patterns among women (calls frompotential female subjects indicated that they believed they were better than men in amount but not in quality). Future researchers would be wise to reduce to only two dependent variables, amount and quality. In addition, a cell size of at least five should be considered minimum. In Retrospect A dissertation is meant to be a learning experience, and this researcher would alter several aSpects of her study if it were repeated. These alterations include: 1. subject selection--It would be preferable to have a large pool of quasi-voluntary subjects, such.as introductory psychology students offered extra credit. This would have saved time and money. In addition, the attrition of subjects would have been controlled. 105 2. size of sample--Sample size would be increased to 60. 3. dependent measures-—Only quality of affect and amount of affect would be used. 4. treatment packages--More explanation and development of the rationale and validity of the target behaviors would be given. 5. change project cards--On each weekly goal card, the definition of the target would be repeated. 6. supplemental tests--All supplemental tests except the test for total training experience and the debriefing questionnaire would be omitted. These changes would provide a sounder, more economical research program. Results of this research are not conclusive. As the research was in new areas, it is perhaps not surprising that it generated more questions than it answered. This researcher believes that both affect self-disclosure and self-management are areas where further research is merited. She trusts that her study will prove valuable to others in designing future studies. BIBLIOGRAPHY BIBLIOGRAPHY Argyle, M. & Kendon, A. The experimental analysis of social performance. In L. Berkowitz (Ed.), Advances in experimental socialypsychology. Vol. 3. New York: Academic Press, 1967. Auerswald, M. ‘The differential reinforcing power of two counseling techniques, restatement and enterpre- tation, on client_production of affect. Unpub- lished doctoral dissertation, Purdue University, 1972. Auld, F. & White, A. M. Rules for dividing interviews into sentences. The Journal of Psyghology, 1956, 53, 273-281. Balswick, J. O. & Peek, C. W. The inexpressive male: a tragedy of American society. Family Coordinator, 1971, 29, 363-368. Bandura, A. Principles of behavior modification. New York: Holt, Rinehart and Winston, 1969. Bandura, A. Vicarious and self-reinforcement processes. In R. Glaser (Ed.), The natureiof reinfpicement. New York: Academic Press, 1971, pp. 228—278. Barnabie, F. Determining the effects of three counselor verbal responses on client verbal behavior. Unpub- lished doctoral dissertation, University of Tennessee, 1972. Bellack, A. S. & Tillman, W. Effects of task and experi- menter feedback on the self-reinforcement behavior of internals and externals. Journal of Consulting and Clinical Psychology, 1974, 1;, 330-336. Benner, H. J. Self-disclosure as a construct. Unpublished doctoral dissertation, Michigan State University, 1968. Berne, E. Games_people play. New York: Grove Press, 1964. 106 107 Bolstad, O. D. & Johnson, S. M. Self-regulation in the modification of disruptive classroom behavior. Journal of Applied Behavior Analysis, 1972, i, 443-454. Bundza, K. A. & Simonson, N. R. Therapist self-disclosure: Its effect on impressions of therapist and willing- ness to disclose. Psychotherapy: Theopy, Research, and Practice, 1973, i9, 215-217. Campbell, D. T. & Stanley, J. C. Experimental and quasi- experimental designs for research. Chicago: Rand McNally, 1963. Cautela, J. R. Behavior therapy and self-control: Techniques and implications. In C. M. Franks (Ed.), Behavior therapy: Assessment and status. New York: McGraw-Hill, 1969. Certner, B. C. Exchange of self-disclosures in same-sexed groups of strangers. Journal of Consulting and Clinical Psychology, 1973, 52, 292—297. Cozby, P. C. Self-disclosure: A literature review. Psychological Bulletin, 1973, 19, 73-91. Crowley, T. J. The conditionability of positive and nega- tive self-reference emotional affect statements in a counseling :ype interview. Unpublished doctoral dissertation, UnIVersity of Massachusetts, 1970. Davis, T. S. Relationship of personality_and task demands to self-disclosing behavior and psychological health. Unpublished doctoral dissertation, Univer- sity of Southern California, 1973. Dollard, J. S Miller, N. E. Persopglity and psychotherapy. New York: McGraw-Hill, 1950. Ehrlick, H. S. & Graever, D. B. Reciprocal self-disclosure in a dyad. Journal of Experimental Social ngchology, 1971, 1, pp. 389-400. Festinger, L. A theory of social comparison processes. Human Relations, 1954, 1, 117-140. Fromm, E. The art of lovipg. New York: Harper Bros., 1956. Fromm, E. Man for himself. New York: Rinehart, 1947. Glynn, E. L. Classroom applications of self-determined reinforcement. Journal of Applied Behavior Analysis, 1970, 3, 123-132. 108 Goffman, E. The presentation of self in everyday life. New York: Doubleday, 1959. Goldfried, M. R. & Merbaum, M. (Eds.) Behavior change through self-control. New York: Holt, Rinehart, and Winston, Inc., 1973. Green, A. H. The effect of various forms of modeling and instructional procedures upon the modification of self-disclosing verbal behavior. Dissertation Abstracts, 1972, 33, 3938B. Guilford, J. P. Psychometric methods. New York: McGraw Hill, 1954, pp. 263-301. Hall, H. V. Effects of direct and self-reinforcement as a function of internal-external control. Perceptual and Motor Skills, 1973, 31, 753-754. Highlen, P. S. Effects of social modeling and cpgnitive structuring strategies on affective self disclosure of single undergraduate college males. Unpublished doctoral dissertation, Michigan State University, 1975. Horney, K. Neurosis and human growth. New York: Norton, 1950. Horney, K. The neuroticgpersonality of our time. New York: Norton, 1936. Hurley, J. R. and Hurley, S. J. Toward authenticity in measuring self-disclosure. Journal of Counseling Psychology, 1969, 16, 271-274. Jackson, B. and Van Zoost, B. Changing study behaviors through reinforcement contingencies. Journal of Counseling Psychology, 1972, $2! 192—195. Jaffee, L. D. & Polansky, N. A. Verbal inaccessibility in young adolescents showing delinquent trends. Journal of Health and Human Behavior, 1962, 3, Jourard, S. M. Healthy personality and self-disclosure. Mental Hygiene, 1959, 13, 499-507. (a) Jourard, S. M. Self-disclosure and other-cathexis. Journal of Abnormal and Social Psychology, 1959, .EZ! 428-431. (b77 109 Jourard, S. M. The transparent self: Self-disclosure and well-bein . Princeton, New Jersey: D. VanNostrand Co., 1964. Jourard, S. M. Disclosing man to himself. Princeton, New Jersey: D. VanNostrand Co., 1968. Jourard, S. M. & Lasakow, P. Some factors in self- disclosure. Journal of Abnormal and Social Psychology, 1958, 56, 91-98. Jourard, S. M. & Richman, P. Factors in the self-disclosure inputs of college students. Merrill-Palmer Quarterly, 1967, g, 166~172. Jourard, S. M. & Landsman, M. J. Cognition, cathexis, and the 'dyadic effect' in men's self-disclosing behavior. Merrill-Palmergparterly, 1960, 6, 178-186. Kanfer, F. H. Self-regulation: Research, issues, and speculations. In C. Neuringer and J. L. Michael (Eds.), Behavior modification in clinical psychology. New York: Appleton-Century-Crofts, 1970, pp. 178- 220. Kanfer, F. H. The maintenance of behavior by self-generated stimuli and reinforcement. In A. Jacobs and L. B. Sachs (Eds.), The_psychologygf_priy§te events. New York: Academic Press, 1971, 39-59. Kanfer, F. H. & Daroly, P. Self-control: A behavioristic excursion into the lion's den. Behavior Therapy, 1972, 3, 398-416. Kanfer, F. H. 8 Phillips, J. S. Learning foundations of behavior therapy. New York: John Wiley and Sons, 1970. Kaufman, K. F. & O'Leary, K. D. Reward, cost, and self- evaluation procedures for disruptive adolescents in a psychiatric hospital. Journal of Applied Behavior Analysis, 1972, g, 293-309. Kazdin, A. E. Self-monitoring and behavior change. In M. J. Mahoney and C. E. Thoresen (Eds.), Self- control: Power to the erson. Monterey, Calif- ornia: Brooks-Cole, 1973. Kolb, D. A., Winter, S. K., and Berlew, D. E. Self- directed change: Two studies. Journal of Applied Behavior Science, 1968, 2, 453-471. 110 Kopfstein, J. H. & Kopfstein, D. Correlates of self- disclosure in college students. Journal of Consult- ing and Clinical Psychology, 1973, $1, 163. Levin, F. M. & Gergin, K. J. Revealingness, ingratiation and the disclosure of self. Proceedings of the 77th Annual Convention of the American Psychological Association, 1969, 4, 447-448. Levinger, G. & Senn, D. J. Disclosure of feelings in marriage. Merrill-Palmer Quarterly, 1967, 13, 237-249. Lynn, D. B. Divergent feedback and sex-role identification in boys and men. Merrill-Palmer'Quarterly, 1964, E, 17-23. Mahoney, M. J. Self-reward and self-monitoring techniques for weight control. Unpublished doctoral disserta- tion, Stanford University, 1972. Mahoney, J. J. & Thoresen, C. E. Self—control: Power to the person. Monterey, California, 1974. Marston, A. R. & Kanfer, T. H. Human reinforcement: Experimenter and subject controlled. Journal of Experimental Psychology, 1963, 66, 91-94. Maslow, A. Motivation and personality. New York: Harper, 1954. May, R. Love and will. New York: Norton, 1969. Mehrens, W. A. & Lehmann, I. J. Measurement and evaluation in education and psycholggy. New York: Holt, Rinehart & Winston, 1973. Mulcary, G. A. Sex differences in patterns of self- disclosure among adolexcents: A developmental perspective. qurnal of Youth and Adolescence, 1973, 2, 343-356. Pearce, W. B. & Sharp, S. M. Self—disclosing communication. The Journal of Communication, 1973, 23, 409-425. Pederson, D. M. & Bueglio, V. J. Personality correlates of actual self-disclosure. Psychological Reports, 1968' a, 495-501. Pederson, D. M. & Higbee, K. L. Personality correlates of self-disclosure. Journal of Social Psychology, 1969, lg, 81-89. 111 Plog, S. C. The disclosure of self in the United States and Germany. Journal of Social Psychology, 1965, Q, 193-2030 Rehm, L. P. & Marston, A. R. Reduction of social anxiety through modification of self-reinforcement: An instigation therapy technique. Journal of Consult- ing and Clinical Psychology, 1968, 32, 565-574. Ribner, N. G. Effects of an explicit group contract on self-disclosure and group cohesiveness. Journal of Counseling Psychology, 1974, 21, 116-120. Rogers, C. R. Client-centered therapy; Its current practice, implicationsy_and theory. Boston: Houghton Mifflin, 1951. Rogers, C. R. On becoming a psrson. Boston: Houghton Mifflin, 1961. Sayicki, V. & Oregon College of Education. Outcomes of nonreciprocal self—disclosure strategies. Journal of Personality and Social Psychology, 1972, 23, 271-276. '__ Sousa—Poza, J. F., Rohrberg, R. & Shulman, E. Field dependence and self-disclosure. Perceppual and Motor Skills, 1973, 36, 735-738. Stieper, D. R. & Weiner, D. V. Dimensions of Psychotherapy. Chicago: Aldine, 1965. Stuhr, D. E. The effects of social model characteristics in eliciting personal feeling questions. Disser- tation Abstracts, 1972, 13, 4105-4106A. Taylor, D. A., Altman, I. & Sorrentino, R. Interpersonal exchange as a function of rewards and costs and situational factors: Expectancy confirmation- disconfirmation. Journal of Experimental Social Psychology, 1969, 5, 324-339. Thibaut, J. W. & Kelley, H. H. The sgcial psychology of groups. New York: Wiley, 1959? Thoresen, C. E. Behavioral means and humanistic ends. In M. J. Mahoney and C. E. Thoresen (Eds.), Self- control: Power to the person. Monterey, Calif.: Brooks/Cole Publishing Co., 1974, pp. 308-322. 112 Thoresen, C. E. & Mahoney, M. J. Behavioral self-control. New York: Holt, Rinehart, and Winston, 1974. Tognoli, J. Response matching in interpersonal information exchange. British Journal of Social and Clinical Psychology, 1969, s, 116-123. Tulkin, S. R. Author's reply: Environmental influences on intellectual achievement. Representative Research in Social Psychology, 1970, 1, 29—32. Urmacht, J. J. & Obitz, F. W. Verbal discrimination task performance as a function of self vs. external reinforcement, I-E scale performance, and subject vs. experimenter determination of response. Journal of Psychology, 1974, §§, 139-148. Vondracek, F. W. Behavioral measurement of self—disclosure. Psychological Reports, 1969, 25, 914. (a) Vondracek, F. W. The study of self-disclosure in experi- mental interviews. Journal of Psychology, 1969, la, 55-59. (b) Vondracek, F. W. & Marshall, M. J. Self-disclosure and interpersonal trust: An exploratory study. Psychological Reports, 1971, 2Q, 235-240. Watson, D. L. & Thorp, R. G. Self-directed behavior: Self- modification for personal adigstment. Monterey, California: Brooks/Cole, 1972. APPENDICES APPENDIX A ADVERTISEMENTS FOR SUBJECTS ADVERTISEMENTS FOR SUBJECTS First Advertisement: IMPROVE YOIIR -- COMMUNICA'IIOI SKILLS A short term research and training program is being offered during the reminder of this term. The focus of the training is on the expression of feelings to others. If you are a single undergraduate male you qualify for the program. There will be two sessions and total time commitment will be approximately 4 hours. Sessions will be arranged according to free time in your class schedule. All materials will be provided, and there is no charge for this training. ll you an Inhnstul, call 353-3198. J Text of Second Advertisement: COMMUNICATIONS SKILLS TRAINING. Learn to express personal feelings more effectively. Free 4 hour training. Call 353-3798 or 355-1755. Volunteers. 113 APPENDIX B POSTER ADVERTISEMENT COMMUNlCAT lON SKILLS A 5lnor-l- “her-m research/braking Program will be concluded ducking hl'eremainder 04 “we. ‘l'er I'm-FM Program Wm {tows on Hm «PVCSSion o