FHE, EFFECT GFFHF COVERT BFHAvLoRs ’OF VISUAL "IMAGERY; SELFFMONLTORINGL “ AND SELF- EVALUATION upon , THE: 0va EXPRESSION OF . EMOTIONAL worms e ' Brssertatlon foF the Degree of Ph D l?-;3172‘~"-;;.{.;.;i MECHIGAN STATE UNIVERSITY 1973 ' ' T: " 9- w 1' ’32.? has been accepted towards fulfillment of the requirements for _Eh..n_ degree in m Major ptofessor Date July 24, 1973 Y II in! a IY ‘5” "ME 8 SDNS' _ BOOK BMDERY INCA Ian vemuzns L gunfighmllig H ~ ”1. W.'-(H.n ‘l - MIL ABSTRACT THE EFFECT OF THE COVERT BEHAVIORS OF VISUAL IMAGERY, SELF-MONITORING, AND SELF-EVALUATION UPON THE OVERT EXPRESSION OF EMOTIONAL WORDS BY Geoffrey Gilbert Yager The purpose of this study was to investigate the effects of covert behavior upon overt action. More speci- fically, three distinct types of cognitive behavior were experimentally manipulated in order to examine their differential effects on a counseling-related performance. The three covert behaviors included were: (a) self- monitoring of behavior, (b) self-evaluation of monitored behavior, and (c) visual imagery of positive, neutral, or aversive scenes. The criterion variable, an overtly observable task, consisted of the number of emotional words verbalized in a TAT-like story generated in response to a collection of cartoon pictures. The experiment was designed to include the three covert behaviors mentioned above within two independent variables. The first independent variable, "self- regulation," consisted of three levels: (a) a :o-self-moni condition, 1 rendition. ‘ th model 0 Kanfer's mo his or her t followed. self-monito If this mod Sh’pass th'. dependent . "'L ‘Aa'. in'GIVed ti 5 Va imager Geoffrey Gilbert Yager no-self-monitoring condition, (b) a self-monitoring condition, and (c) a self-monitoring with self-evaluation condition. These three groups paralleled, to some extent, the model of self-regulation which Kanfer (1971) proposed. Kanfer's model asserts that for a human being to regulate his or her own behavior, three necessary steps must be followed. In brief, the steps of self-regulation include self-monitoring, self-evaluation, and self-reinforcement. If this model is appropriate, subjects in the self- monitoring with self-evaluation condition should exceed subjects in the self-monitoring group who, in turn, should surpass those in the no-self-monitoring condition on the dependent variable. The second independent variable, visual imagery, involved the administration of positive, neutral, or aver- sive images contingent upon an actual or perceived increase in the number of emotional words expressed during one TAT- like story as compared to the immediately preceding story. The satisfaction of this contingency was decided by an experimenter tally in both the no-self-monitoring and self- monitoring conditions. In the self-monitoring with self- evaluation group, however, the subject's self-evaluation determined whether the contingency had been met. It was expected that the contingent administration of the covert visual images would differentially effect the emotional word output on subsequent stories. The positive imagery should have served very much as an external reinforcer and hould have 1:. the neat aversive itai emotional we either the A These hyper? of Cautela Sub level and g the College latter of 'w'ere assigq {3 levels d 3“in? the “a“ up f: StimuluS C §eq“‘est' wi Geoffrey Gilbert Yager should have led to a larger number of emotional words than in the neutral imagery condition. For similar reasons, the aversive images should have suppressed responding in emotional words and resulted in fewer emotional words than either the neutral or the positive imagery treatments. These hypotheses are consistent with the recent theorizing of Cautela (1972). Subjects in the study consisted of sixty-three upper level and graduate students who had enrolled in courses in the Colleges of Education or Human Ecology in the Spring quarter of 1973 at Michigan State University. Subjects were assigned randomly to one of nine treatment conditions (3 levels of self-regulation x 3 levels of visual imagery). During the experimental session, each subject was asked to make up fifteen stories, each in response to a cartoon stimulus card. An experimenter was present in the room to request, when appropriate to the contingencies or to the experimental condition, that the subject (a) read and imagine the scene presented on an imagery card, (b) stOp imagining a scene after 30 seconds, (c) take and complete an imagery rating form, or (d) take and fill in a self- monitoring or self-evaluation form. There were two major analyses performed on these data. The first, a multivariate analysis of covariance, employed the number of emotional words in the first story and a self-perception of empathic ability questionnaire as covariates with the ratings of emotional words in the remaining f second anal exployed t’”. enotional to measures d;| differences ables or or variables a Sit-Pom £0: Si Variable i in a Subs: lC‘Zated b of image of the“ Geoffrey Gilbert Yager remaining fourteen stories as the dependent variables. The second analysis, a repeated measures analysis of covariance, employed the basic design (3 X 3) with the addition of the emotional word ratings for stories 2-15 as the repeated measures dimension. Neither analysis yielded significant differences on the self-regulation and visual imagery vari- ables or on the interaction between these independent variables and the repeated measures variable. Hence, no support for the major hypotheses of the study was found. Significant differences were found on the measures variable in the repeated measures analysis. Additionally, in a subsidiary analysis, significant differences were located between imagery conditions in the mean self-ratings of image vividness and speed of imagery attainment. Neither of these findings, however, was of direct relevance to the primary hypotheses of the study. Despite the absence of significant findings relative to the major dimensions of the study, the present investi- gation has generated a number of hypotheses which may serve to further the exploration of the area of covert behavior. Among the specific suggestions offered to future researchers are (a) select a dependent variable which is not likely to be so easily influenced as to produce a "ceiling effect," (b) make the definition of the dependent variable oper- ationally clear to the subjects, (c) attempt to make sure that subject motivation and concentration remain high during the study, (d) in the self-regulation dimension, tell the subjects w: to actually iragery var; iniividuali process of References ‘ Cautela, J_ din Yo: Kahfer I F . 98? ‘ I evJ ‘ Geoffrey Gilbert Yager subjects what self-monitoring involves and influence them to actually record their behavior, and (e) relative to the imagery variable, incorporate a procedure to allow for individualization of imaginal scenes and practice of the process of forming visual images. References Cautela, J. R. Rationale and procedures for covert con- Kanfer, ditioning. Advances in behavior therapy. New York: Academic Press, 1 , PP- - . F. H. The maintenance of behavior by self- generated stimuli and reinforcement. In A. Jacobs and L. B. Sachs (Eds.) Theypsychology of private events. New York: Academic Press, I§71, pp. 55-60. THE EFFECT OF THE COVERT BEHAVIORS OF VISUAL IMAGERY, SELF-MONITORING, AND SELF-EVALUATION UPON THE OVERT EXPRESSION OF EMOTIONAL WORDS BY Geoffrey Gilbert Yager 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 1973 a: )fi- DEDICATION To my wife, Margie-- There's more to say, but any acknowledgment would pale in comparison to her actual overt and covert contributions. ii Ea myself. I experience Wished t Statistics in GXpressI 35 logica; expanding 3011.1st I fatiVe as and their is E‘s-1C7!) a. intel“est K given the a“ skill Profoundl thanks 9: all °f LI and “Kirk ACKNOWLEDGMENTS Early in my doctoral program, I set six goals for myself. (1) I wanted to gain all of the counseling skill, experience, and personal understanding of Bob Winborn. (2) I wished to not only learn all that Andy Porter knew about statistics and research design but also to be as effective in expressing that knowledge to others. (3) I hoped to be as logical, as open, and as helpful in exploring and expanding upon graduate students' research ideas as Dick Johnson. (4) I wanted to become as personable and infor- mative as Norm Stewart and to know as much about systems and their application to counseling. (5) I desired to know as much about learning theory and motivation as Mark Rilling. (6) I hoped to develop the flair, excitement, and genuine interest with which Steve Yelon teaches. Well, I've failed in all six goals, but to have been given the chance to share a mere fraction of the knowledge and skill implied in the above goals has been the most profoundly valuable educational experience of my life. My thanks go to each of these six men and, additionally, to all of the many other men and women with whom I have studied and worked during the past four years. iii TABLE OF CONTENTS Page LIST OF TABLES . . . . . . . . . . . . . Vii LIST OF FIGURES. . . . . . . . . . . . . ix Chapter I. THE PROBLEM, RATIONALE, AND RELATED RESEARCH. 1 Need . . . . . . . . . . . . . 1 Purpose . . . . . . . . . . . . 2 Theory and Related Research. . . . . . 5 Covert Behavior . . . . . . . . . S The Influence of Skinner . . . . . 6 A Possible Reconciliation .‘ . . . . 7 The Emphasis of the Literature Review . 7 Development of Covert Behavior . . . 9 Self-Control . . . . . . . . . . 10 Introduction to Self-Control . . . . 10 Stimulus Control Techniques . . . . 12 Self-Monitoring . . . . . . . . 13 Self-Reinforcement and Self- ‘ Punishment . . . . . . . . . 15 Clinical Applications of Self-Contro . 20 Additional Questions. . . . . . . 32 Related Studies Outside of Self-ContrOl . 33 Overview . . . . . . . . . . . . 35 iv Chapter E II. EXPE S .r v 111, 1 IV, Chapter Page II. EXPERIMENTAL DESIGN AND METHODOLOGY . . . . 36 Subjects . . . . . . . . . . . . 36 Measures . . . . . . . . . . . . 36 Self-Evaluation of Empathic Ability. . . 38 Emotional Word Measurement. . . . . . 39 Experimental Manipulations . . . . . . 43 Self-Regulation Dimension . . . . . . 46 No Self-Monitoring Treatment . . . . 46 Self-Monitoring Treatment . . . . . 47 Self-Monitoring and Self-Evaluation Treatment . . . . . . . . . . 47 Imagery Dimension. . . . . . . . . 48 Positive Imagery . . . . . . . . 51 Neutral Imagery. . . . . . . . . 51 Aversive Imagery . . . . . . . . 51 Hypotheses. . . . . . . . . . . . 52 Hypothesis One. . . . . . . . . . 53 Hypothesis Two. . . . . . . . . . 54 Hypothesis Three . . . . . . . . . 55 Experimental Design. . . . . . . . . 55 Analysis 0 I O O O O O O O O O O 57 III. ANALYSIS OF RESULTS . . . . . . . . . 59 The Primary Analyses . . . . . . . . 59 Subsidiary Analyses. . . . . . . . . 73 Descriptive Analyses . . . . . . . . 77 8mm. 0 O O O I O O O O O O O 86 IV. SUMMARY, DISCUSSION, AND CONCLUSIONS. . . . 89 Summary. . . . . . . . . . . . . 89 Discussion. . . . . . . . . . . . 92 Chapter ezeziocaA; APPENDICES APpendi A. h’; G B. C. C X Se‘. SeJ Chapter Theory . . . . . . . . . . . Self-Regulation . . . . . . . . Imagery O O O O O O O 0 O O 0 Sample . . . . . . . . . Design and Statistical Treatment Instruments . . . . . . . Treatments. . . . . . . . Self-Regulation . . . . . . . . Imagery . . . . . . . . . . . Conclusion . . . . . . . . . . . BIBLIOGRAPHY O O O O O O O O O O O O O O APPENDICES Appendix A. Typical Presentation Employed to Solicit Subjects . . . . . . . . . . . . B. Self-Evaluation Questionnaire. . . . . . C. Self-Evaluation Questionnaire Weightings for Each Item and Lie Scale Indicated. . . . D. Emotional Word Guide. . . . . . . . . E. Three Sample Stories With Emotional Words Underlined . . . . . . . . . . . F. Subject Consent Form. . . . . . . . . G. Demographic Data Sheet . . . . . . . . H. Cartoon Stimulus Pictures . . . . . . . I. Story Recall Checklist . . . . . . . . J. Story Comparison Form . . . . . . . . K. Imagery Rating Form . . . . . . . . . L. Complete List of Imagery Suggestions . . . M. Final Evaluation Form . . . . . . . . vi Page 93 94 96 96 97 98 99 100 104 106 108 118 120 123 126 127 132 133 134 149 150 151 152 158 3.2. 3.3. 3.4. 3.5. 3.5. no Table 2.1. 2.2. 3.1. 3.2. 3.3. 3.4. 3.5. 3. 6. LIST OF TABLES Personal and Demographic Characteristics . Experimental Design . . . . . . . . . Regression Analysis Portion of the Design: The Relationship Between the 14 Ratings of Emotional Words and the 2 Covariates (Self-Perceived Empathy and EW in the First Story). . . . . . . . . . . Summary of the Multivariate Analysis of Covariance and the Repeated Measures Analysis of Covariance on Fourteen Observations of the Number of Emitted Emotional Words. . . . . . . . . . Means and Pooled Standard Deviations of Emotional Words Emitted in Each Story for Each Self-Regulation Condition . . . Means and Pooled Standard Deviations of Emotional Words on Each Story for Each Imagery Condition . . . . . . . . . Mean Number of Emotional Words Emitted for. Each Story Number and Mean Differences for Story Number Contrasts of Possible Signi- ficance Using Duncan's New Multiple Range Test . . . . . . . . . ._ . . . Multivariate and Univariate Tests for Imagery Effects Upon Mean Vividness of Image and Mean Speed of Image Attainment. . . . . vii Page 37 56 60 63 66 67 71 74 3.9. 3.10. 3.11. 3.12. 3.13. 3.14. Table 3.7. 3. 8. 3.9. 3.11. 3.12. 3.13. 3.14. Cell Mean and Standard Deviations on Vividness and Speed of Image Attainment Variables . . Duncan's New Multiple Range Post Hoc Com- parisons of Differences Between Imagery Conditions on Vividness of Image and Speed of Image Attainment. . . . . . . . . Means and Standard Deviations of Number of Emotional Words Emitted Response to Particular Cartoon Pictures . . . . . . Means and Standard Deviations of Image Vividness and Speed of Imagery Attainment for Each of the Possible Positive Images. . Means and Standard Deviations of Image Vividness and Speed of Imagery Attainment for Each of the Possible Neutral Images . . Means and Standard Deviations of Image Vividness and Speed of Imagery Attainment for Each of the Possible Aversive Images. . Frequency and Percentage of Subjects Falling Within Each Category of Response on the Question: ”What, in Your Opinion, Was the Probable Purpose of This Study?” . . . Frequencies, Percentages, and Chi-Squared Tests for Last Three Questions of the Final- Evaluation Form on the Self-Regulation Condition . . . . . . . . . . . . viii Page 74 76 78 79 80 81 83 85 Figure 2.1. 3.1. 3.2. 3.3. The Gr LIST OF FIGURES Figure Page 2.1. The layout of the experimental room during the SMSE condition . . . . . . . . . 44 3.1. Graphical presentation of mean numbers of emotional words for each story for each self—regulation condition. . . . . . . 68 3.2. Graphical presentation of mean numbers of emotional words for each story for each imagery condition . . . .. . . . . . 69 3.3. Graphical presentation of the mean number of emotional words emitted for each story position . . . . . . . . . . . . 72 ix 33. C' have becc: SBttings 1972) , TI involves his or he PIOCESQ C the neceg The Cone. Perhaps Such st: impro'v'e: self‘CQ deSCI-l; have re the“ CHAPTER I THE PROBLEM, RATIONALE, AND RELATED RESEARCH N_e_§_d_ Over the past several years, self-control strategies have become increasingly popular in clinical and applied settings (Cautela, 1969; Kanfer & Phillips, 1970; Mahoney, 1972). The self-control approach, in its simplest form, involves nothing more than allowing the client to direct his or her own change. The counselor may instigate the process of change, but it is the client who must carry out the necessary steps to achieve the desired counseling goal. The concept of self-control has rapidly been accepted, perhaps to some extent, as a result of its association with such strongly-favored societal ideals as freedom and self- improvement. From the beginning of recorded time, the concept of self-control has beem employed by mankind as a trait descriptive of heroes and leaders. Labels for this trait have ranged widely from "guts” to "will power“ to ”the power of positive thinking." Only recently, however, have investigators begun to move in the direction of a better definition control or 1970a). De ness, and I there is a of. the den“ Ir. one cent: are those others. ”‘5 decid. Phenomena not, as S litératu. definition of the parameters of the phenomenon of self- control or self-regulation (Kanfer, 1971: Cautela, 1967, 1970a). Despite the compelling logic, theoretical sound- ness, and clinical effectiveness of such recent formulations, there is a great need for well-designed empirical validation of the developing theories of self-control. In examining the self-control strategies, there is one central component: covert behaviors. Covert behaviors are those behaviors which are not directly observable by others. Such behaviors as thinking, feeling, imagining, and deciding fall within the covert category. These covert phenomena and their role in self-control strategies have not, as yet, received much direct attention in the research literature. Purpose The purpose of this study was to investigate the effects of covert behavior upon overt action. More speci- fically, three distinct types of cognitive behavior were experimentally manipulated in order to examine their differential effects on a counseling-related performance. The three covert behaviors included were: (a) self- monitoring of behavior, (b) self-evaluation of monitored behavior, and (c) visual imagery of positive, neutral, or aversive scenes. The criterion variable, an oVertly observable task, consisted of the number of emotional words (U1 story gen pictures . Observati folled behavior dependi: Positiv d9 liVeI pre'senI future s““*"‘-Ii‘-F.ai SE1I~I 3nd ( E'Jal Se‘! . '\ words (Ullmann s McFarland, 1957) verbalized in a TAT-like story generated in response to a collection of cartoon pictures. Each of these covert variables may be interpreted as components of the model of self-regulation that has been developed by Kanfer (1971). Kanfer's model asserts that for a human being to regulate his or her own behavior, three necessary steps must be followed. First, a behavioral observation or gglgfmonitoring phase occurs. This is followed by a gglgfevaluation stage in which the observed behavior is compared to a predetermined criteria. Finally, depending on the judgment made in the self-evaluation, a positive or negative gglgfreinforcement (which could be delivered in the form of a positive or negative image) is presented. The type of self-reinforcement will effect the future probabilities of the self-regulated response. In summary, self-regulation is a process of self-monitoring, self-evaluating, and self-reinforcing. In assessing the effectiveness of self-monitoring and of the combination of self-monitoring with self- evaluation, the present experiment provided research evidence directly relevant to Kanfer's (1971) model of self-regulation. If the model was appropriate, the group receiving instructions to both self-monitor (SM) and self- evaluate (SE) should have emitted more emotional words in telling their stories than those who were asked only to self-monitor. Likewise, the SM-only group should have outperforr (Note: I: carry out to a no-St the lack ’. should ha- condition Presented in the n TAT-like story. In the s herSelf analogc 83113 1t u outperformed the no-self-monitoring (NSM) control condition. (Note: It was expected that a number of subjects would carry out self-initiated SM and SE despite their assignment to a no-SM control. However, it was also anticipated that the lack of specificity in these self-initiated efforts should have made them somewhat less powerful than in those conditions where SM and SE were directly requested). The reinforcement variable, visual imagery, was presented contingent upon an actual or perceived increase in the number of emotional words expressed during a given TAT-like story as compared to the immediately preceding story. The satisfaction of this contingency was decided by an experimenter tally in both the NSM and SM conditions. In the SM and SE group, however, the subject himself/ herself determined (based upon a self-evaluation) whether the contingency had been met. Use of the visual images in this study was directly analogous to the covert positive reinforcement and covert - sensitization procedures of Cautela (1970a, 1967). Cautela's counseling treatment programs employ imagery to increase or decrease the likelihood of certain desired or undesired behaviors. Cautela (1972) indicated that the covert procedures which he employs are founded on the assumption that a reinforcing or punishing stimulus, when presented in imagination, will function in a manner. essentially similar to an externally applied stimulus. Thus, the positive imagery condition should have served to increase _ words) wh. decreased theories eXiStencl w£WEd a environ: Princip] Sth a ‘ eXPErim. tCtally which i I radlcé Prospe: latiQAI increase the occurrence of the target behavior (emotional words) while the aversive imagery condition should have decreased the target behavior's frequency. Those subjects exposed to a neutral imagery condition were expected to be little influenced by the contingent administration of that particular covert behavior. Thus, the purpose of the study was to test Kanfer's model (1971) of self-regulation and Cautela's (1971a) theories of covert reinforcement. Theory and Related Research Covert Behavior At its outset, behaviorism rigidly denied the existence of any and all unobservable phenomena. Man was viewed as a very direct product of his social and biological environment. Watson (1919) and his colleagues applied the principles of scientific methodology to human behavior in such a way that only behavior externally visible to the experimenter was deemed worthy of consideration as either a cause or an effect of human action. This "monistic" view totally denied the importance of mental activity, a behavior which is clearly unobservable to anyone other than the individual doing the thinking. Despite the many criticisms engendered by such "radical" views, behaviorism has survived and, in fact, prospered during the years following Watson's early formu- lations. A broad base of laboratory research has now establishe of learni: fotmdaticfl oriented J corp lex a ' which we: as a beha‘fior1 established a science of human behavior on a firm foundation of learning theory. In recent years, it appears, this foundation has been sufficiently firm for behaviorally- oriented psychologists to begin to venture into the more complex and less easily examined areas of mental functioning which were prohibited by Watson's original assertions. The influence of Skinner. Skinner (1953) took the first steps in examining covert (nonobservable) mental behaviors by hypothesizing that thoughts or images are nothing more than a specific response class which happens to be observable only by inference. ”A purely private event would have no place in a study of behavior, or perhaps in any science; but events which are . . . accessible only to the individual himself often occur as links in chains of otherwise public events and they must be con- sidered" (Skinner, 1953, p. 229). Skinner chose to regard ”public" and "private" events as subject to exactly the same rules of acquisition, maintenance, and extinction. Skinner's operant view of "private events” has recently initiated a remarkable expansion of knowledge in the area of covert behavior (Bandura, 1969, 1970: Cautela, 1969; Kanfer & Phillips, 1970; Kanfer, 1970a: Mahoney, 1972; Kanfer & Karoly, 1972, and Elson, 1972). Most of this research has been related, directly or indirectly, to the process of "self-control." Since much human behavior occurs in the absence of immediate and obvious external y control, contempor. between t: Change . I U (D I control, it is assumed that covert and private events must provide for self-control of these behaviors. A possible reconciliation. Thoresen (1973) argues in a recent article that the emphasis on self-control by contemporary behaviorists may suggest a possible link between the humanistic and behavioral approaches to behavior change. The ideals of self-direction, independence and self-improvement appear to be present in some form in the concepts expressed by both groups. Clearly, however, total agreement is not imminent, for there are many who continue to argue with modern behavioristic theorizing from a cognitive (Beck, 1970), "humanistic” (Portes, 1971), and a strict behavioral (Gewirtz, 1971) point of view. Perhaps the only single agreement concerning covert behavior to which all of these approaches might subscribe would be that covert occurrences do have some influence. As human beings, all of those associated with the "helping professions" are fully aware of the effects of thoughts, beliefs, and images upon behavior. The primary emphasis, then, of contemporary behavioral research on covert behavior is not to answer the question, "Does covert action have effects?" but rather, "How do covert actions affect particular behaviors?” The emphasis of the literature review. As has been indicated, a substantial literature has developed which is indirectly related to this last question. However, behavior ' ature mus contribut as Ellis their wo: pectic a; little 5: primari l} 1970). l experimeg the Proc separate C0gniu, Activit RIEQ of “Quid t treatir increa: However, since the research evidence relevant to covert behavior has advanced so rapidly, the review of the liter— ature must be greatly restricted. For example, significant contributions to therapy have been made by such theorists as Ellis (1962) and Ellis and Harper (1961). However, their work on a cognitive (and, therefore, covert) thera- peutic approach will not be covered since, by stimulating little scientific research, it continues to remain, primarily, at the speculative level (Bergin, 1970: Nawas, 1970). The review of the literature, then, will emphasize experimental research findings on the influence of covert behavior in clinical settings and, more specifically, in the process of behavior modification. The behavioral counselor need not hypothesize a separate cognitive system (as Ellis has) to explain cognitions: he or she need only define certain human activities as cognitive and treat difficulties in this area of response in a parallel manner to the way he or she would treat any overt difficulty. The ultimate goal in treating either overt or covert behavior would be to increase the capabilities and potentialities of individual clients (Ullmann, 1970). The notion that the contemporary behaviorist neglects thoughts and feelings in his or her work is unsubstantiated. It would, therefore, be in- structive at this point to examine some of what has been written within the behavioral area which relates to covert behavior and its significance. 2 born with such as t ceptive s essential of covert paper. verbally thinking fore, m 1 behavior learning (1968, 1 mental p irwestig inclhides brought general: .Speech~l “to: bc verbali zI ”Shin, relatim 3. «Eve 1 ' 17y Pothe Development of covert behavior. All humans are born with certain immediately operative covert behaviors such as those arising from the interoceptive and proprio- ceptive systems. These innate covert behaviors are essentially physiological in nature and are not the types of covert behavior that have provided the focus of this paper. Alternatively, this investigation focused upon the verbally- or imaginally-mediated covert behaviors, such as thinking and reasoning which are not innate, and, there- fore, must be learned over time. During development, these behaviors gain power over overt behavior as part of the learning process. Bem (1967) and Meichenbaum and Goodman (1968, 1971) have experimentally examined this develop- mental process. Their findings from three well designed investigations clearly suggest that the sequence of growth includes: (a) Initially, the child's overt behavior is brought under the control of an adult's overt reinforcement, generally of a verbal nature; (b) the child develops overt ”speech-for-self" which gains control over his own overt motor behavior, and (c) the child's overt motor behavior comes under the control of the implicit (covert) self- verbalization such as are commonly called thoughts or cognitions. Research by Kanfer (1966), which showed a positive relationship between appropriate self-reinforcement and age level, adds additional support to the developmental hypothesis. More recently, Mahoney, Thoresen, and Danaher (1972) , l provided behavior develope a learni monitori Process reinforfi n 10 (1972), in a very cleverly designed investigation, have provided evidence indicating that internalized (covert) behavior can be modified by traditional operant methods. Covert verbal behaviors, then, are apparently developed in progressive stages through an operant learning method. Self-control of behavior appears to evolve out of a learning context in which extensive external parental monitoring is gradually shifted to self-monitoring. This process is accelerated by the continuing guidance and reinforcement provided in the child's early environment. Self-Control The self-control research constitutes the major contribution to the literature on covert behavior. This paper presents (a) a general introduction to the self- control area, (b) a mention of stimulus-control techniques, (c) a discussion of self-monitoring, (d) a very brief review of the literature on self-reinforcement and self— X punishment, and (e) an extensive look at the clinical applications of self-control. Introduction to self-control. Self-control has been defined in a number of ways (Skinner, 1953, p. 228; Goldimond, 1965, p. 853; Cautela, 1969, p. 324; Kanfer & Phillips, 1970, p. 415; Mahoney & Thoresen, 1972, p. 5). Essential to each definition is the individual acting in such a way as to direct and manage his or her own internal and external behaviors. This process may include the suppressi attemptinl or it may: are unlit-cl under fir lines) . by the " forming 1 3Ct3 as 1 Object C1 Self-coal SEIf-con bEhavior 11 suppression of behavior that is undesirable (e.g., attempting to eliminate cigarette smoking or alcoholism), or it may involve the development of desired behaviors that are unlikely in the present environment (e.g., courage under fire or writing term papers by self-imposed dead- lines). The central focus of self-control, as is implied by the "self-" designation, is that the individual per- forming the behavior is identical to the individual who acts as the agent of change. ”He is both subject and object of his own behavior" (Elson, 1972, p. 6). Since the self-controller must determine his or her own actions, all self-control techniques must necessarily involve covert behaviors. Decisions of this nature are made covertly. In a review of self-control research, Cautela (1969) stresses the advantages of teaching self—control techniques to counseling clients. Included in his list of benefits are: (a) The likelihood of enhancing the client's perceived effectiveness since he/she has such an important part in the therapeutic process (This idea is also stressed in Kanfer & Martson, 1963a.); (b) self-control can be taught in groups rather than the traditional one-to-one format of counseling; (c) if a relapse of the distressing behavior occurs, the client can readminister his/her treatment; and (d) new behavior is likely to be maintained longer when the client perceives that he/she is responsible for the change. (n self-cont first of provide f_ 1953; Fell 1965). § 0f discri 0f learn:J serve to {0: Thus 12 Stimulus control techniques. The employment of self-control includes several possible techniques. The first of these requires alteration of the environment to provide for stimulus control over the behavior (Skinner, 1953; Ferster, Nurnberger & Levitt, 1962; Goldiamond, 1965). Stimulus control is a term descriptive of the effect of discriminative stimuli on behavior. If, in the course of learning, an organism has been reinforced for a behavior in the presence of certain stimuli, those stimuli will serve to "set the stage" for further responding in the future. For example, a pigeon which has been trained to peck a key for food only in the presence of a red light will not peck the key to any great extent when the red light is off. When lighted, however, the red bulb appears to "control" the pigeon's behavior: whenever it is lit, the pigeon peeks! Applying the same principle to self-control, a student with poor study behavior may improve his grades by choosing to study consistently at one location in the library. This location would then "set the stage” for his study behavior. In addition, the student would be well- advised not to attempt to study in his room‘which has already been highly associated with (become a discrimin- ative stimulus for) such nonproductive behaviors as bull- sessions, television viewing, daydreaming, and sleeping. Thus, altering the environment or controlling the stimuli associate) self-mama S self—cont behavior. the subja In the pi feeiback the use d on the c ‘_ 13 associated with a behavior can prove helpful in implementing self-managed changes in that behavior. Self-monitoring. The second possible method of self-control is self-monitoring or self-observation of behavior. This approach involves a careful recording of the subject's actions by the subject himself or herself. In the process of self-observation, the individual receives feedback on his/her immediate behavior and, often, through the use of checklists or graphs, he/she is given feedback on the cummulative progress of the self-modification process. Self-monitoring, operating alone, will normally produce some changes in behavior in the desired direction (Mahoney & Thoresen, 1972). Experimental evidence for the effectiveness of self-observation has recently been presented. McFall (1970) requested subjects to either monitor (a) the number of times they smoked in class, or (b) the number of times they wanted to smoke but did not. Through external validation by inconspicuous observers, McFall's groups showed differential increases in the behavior which they were asked to record: Those who self-monitored the smoking, increased, while those who observed their restraint from smoking reduced their cigarette intake. McFall's objective for the study was to demonstrate that self- monitoring is a reactive technique when applied to behavioral research. McFall explained his results in terms of a CO\ is crit: hypothe: HcFall': purely : He argu4 desired perform investiq Offered (1972), re(Enlat; reSPORd: perfOrm' Priate : argue t. reinfOr Clinica 1971; E 1972) Pl ‘— JChfiSCr y-CFall prOVJid‘ CbSe rv'j 14 of a covert "self-conscious" reaction about a behavior that is critically observed. Orne (1970) disagreed with McFall's hypothesis and very cogently explained certain flaws in McFall's design. Orne chose to explain the McFall study purely in terms of experimenter "demand characteristics." He argued that the subjects had likely guessed their desired behavior and were, very simply, attempting to perform in that manner. More relevant to the research carried out by this investigation is the explanation of McFall's (1970) data offered by Kanfer (1970b, 1971) and Kanfer and Karoly (1972). Kanfer's explanation involved his model of self- regulation: the subject self-monitors his/her behavior and responds to the collected information by evaluation his/her performance and following that evaluation with an appro- priate self-reward or punishment. Kanfer would, of course, argue that self-monitoring without subsequent self- reinforcement would not increase the frequency of an observed behavior. In the short time since McFall's article, four clinical case applications (Rutner & Bugle, 1969: Mahoney, 1971; Broden, Hall, & Mitts, 1971; and Bolstad & Johnson, 1972) and five experimental studies (Lawson & May, 1970: JOhnson & White, 1971; Thomas, Abrams, & Johnson, 1971; McFall & Hammen, 1971; and Herbert & Baer, 1972) have provided additional support for the belief that self- observation can, by itself, promote behavior change. Each of the a': directicr (the case native h; cigarettél decrease' classroc: increase behavior flier me rESP-arc): 15 of the above studies demonstrated changes in the desired direction. Among the changes attributed to self-monitoring (the case studies could be explained by a number of alter- native hypotheses) were increased study behavior, decreased cigarette smoking, decreased negative self-thoughts, decreased hallucinations, increased dating, decreased classroom disruption, decreased aversive vocalizations, and increased parental attention to children's appropriate behavior. Self-reinforcement and selfepunishment. A third major method of self-control stressed in the behavioral research is the contingent implementation of various types of self-rewards or self-punishments. Before considering the treatment methods arising from this approach, the literature on self-administration of tangible rewards and punishments must be examined. The overt self-rewarding and punishing behaviors are of particular interest since they provide the most direct observable information about the internal cognitions which serve to generate them. In other words, although the self-reinforcement studies allow subjects to receive tangible rewards, subjects must make covert decisions to evaluate whether they deserve or do not deserve the self-reward. In a general sense, self- reinforcement (SR) and self-punishment (SP) must be viewed as the principal motivating factors in self-regulation and self-control. investig paradi Ll) research is free or an a' in truc adequa, Studies associa ation) Which m' eXPerim reinfoq' IEarnin Vidual' TendEUC to the Variab: the 13s} initia. Condit which 16 One of the designs frequently employed in the investigation of SR and SP is the "directed learning paradigm" (Kanfer, 1970a; Kanfer & Phillips, 1970). This research model involves a situation in which the subject is free to administer a desired reward (self-reinforcement) or an aversive stimulus (self-punishment) SQTliE.but is instructed to do so contingent on a self-evaluation of the adequacy of the preceding task response. In many of these studies, an ambiguous criterion behavior (e.g., a word association task for which there is no ”correct" associ- ation) is included to control for possible confounding which might result from differential learning between experimental conditions. Thus, differences in self- reinforcement or self-punishment are not the result of learning: they instead reflect differences in the indi- vidual's tendency to self-reinforce or self-punish. Tendencies to self-reinforce or self-punish are important to the area of covert behavior: An understanding of what variables will effect the amount of SR or SP would provide the behavior modifier significant tools to help a client initiate or perpetuate a self-change project by manipulating conditions which change the amount of SR or SP (both of which are commonly covert in nature). ' Using the "directed learning paradigm," a number of excellent research investigations have been carried out by Kanfer, Marston, and their colleagues. The effects of a number of variables upon the acquisition and modification ofSR an AddithI‘. motivati behavior and Else literat; POSitive reinforr fiarston) Duerfel, learned 5 MarSt (Harstc gm ex; 17 of SR and SP responses have been examined in some detail. Additionally, studies have been performed to assess the motivating effects of SR and SP upon contingently-associated behavior. Bandura (1971), Kanfer (1970a), Mahoney (1972), and Elson (1972) have offered comprehensive reviews of this literature. Some of the conditions which have been shown to be positively related to the amount of SR include external reinforcement of the self-reinforcement response (Kanfer & Marston, 1963b; Dorsey, Kanfer, & Duerfeldt, 1971; Kanfer & Duerfeldt, 1967b, 1968; and Marston, 1969), the amount of learned discrimination on a target task (Kanfer, Bradley, & Marston, 1963), the ambiguity of the performance task (Marston, 1964) and the amount of nurturance provided by the experimenter (Bandura, Grusec, & Menlove, 1967). The amount of self-reinforcement, on the other hand, is inversely related to the age of the subject (Kanfer, 1966) and the difficulty of the task (Dorsey, Kanfer, & Duerfeldt, 1971). Another variable which has demonstrated consistent effects on self-reinforcement is that of modeling. Bandura and Kupers (1964) and Bandura and Whalen (1966) both demonstrate that a subject is likely to follow the self- reinforcement practices of the model to which he/she has been exposed. A general hypothesis can be drawn from the set of SR findings cited above: the higher the subject's con- fidence in the appropriateness of an answer, the more likely t confiden arbiguit previous the covd (SR) whe indicate the beh- that Sp that Sp 1968; K Marston Subjeczt 18 likely the subject will be to self-reward. A subject's confidence may be the result of better learning, reduced ambiguity, earlier reinforcement, increased age, or previous vicarious rewarding experience. In each case, the covert thought appears to be directing overt action (SR) when increased assurance of self-evaluation is indicated. Self-punishment has not yet been fully explored by the behavioral researchers. There are clear indications that SP and external punishment are directly related but that SP is not directly related to SR (Kanfer & Duerfeldt, 1968: Kanfer, Duerfeldt, & LePage, 1969). Additionally, Marston and Cohen (1966) demonstrated that frustration of subjects created high levels of self-punishment. Relative to the extensive literature on the vari— ables important to the development and maintainance of a self-reinforcing response, the research on the effect of SR upon behavior has been surprisingly sparce. A question of importance is: Does the application of self-reinforcement affect the strength of the behavior which it follows? Marston (1967) is among the few who have experimentally examined this question. He found that not only was SR effective in facilitating learning, but also it was more beneficial than an external reinforcement. Rehm and Marston (1968) and Benke and Harris (1972) were able to effect positive change in more complex behavior, dating concerns and study behavior, by training their subjects to use appr Peterson trate th genciesi applicai some ve SR‘s ef group w hat Che S task, E Portia: for con 'COrrefi Study, meat w. traine aSSUmel have b 0f the 19 use appropriate self-reinforcers. Finally, Epstein and Peterson (1973a) describe three case studies which illus- trate the clinical effectiveness of self-imposed contin- gencies. (More attention to the general area of clinical applications has been deferred to later in this paper.) A study by Kanfer and Duerfeldt (1967a) presents some very interesting findings relevant to the question of SR's effect on behavior. The investigation involved a group which was encouraged to self-reward for correct matches to tachistoscopically presented stimuli (a learning task, 22; an ambiguous situation). During the first portion of the experiment, SR subjects rewarded themselves for correct matchings by turning on a green light labeled "correct." During a final "extinction" period of the study, the green light was disconnected and no reinforce- ment was delivered. Despite this fact, those subjects trained in SR, continued to improve performance! If we assume the learning was of an operant nature, there must have been reinforcement occurring to increase the accuracy of the matches. This type of self-reinforcement (no overt rewards) must have been.motivated by purely covert factors. This indication of the effect of covert self-reinforcement is of direct relevance to the present study. An investigation by Johnson (1969) on the extinction of a self-reinforcement condition indicated an extinction curve for SR subjects which did not differ from that associated with the removal of external reinforcement. However, p.21) m subjects the nece for cove Goodman , and fir emPIOYe respon: Stratet tuere apprca. N‘nil 20 However, in criticizing Johnson's study, Elson (1972, p. 21) makes the justifiable point that since Johnson's subjects were first graders, they may not yet have achieved the necessary internalization of verbal behaviors required for covert self-reinforcement (Bem, 1967: Meichenbaum & Goodman, 1968, 1971). No direct tests of the effect of self-punishment upon learning were found during the preparation of the review of the literature. To summarize the relevance of the SR and SP research to the present study, (a) self-reinforcement and self- punishment can be generated within laboratory settings, (b) these reinforcing (and punishing?) conditions may be overt or covert in nature (Kanfer & Duerfeldt, 1967a), and (c) the self-reinforcing of overt behavior can promote change 'in the overt response. Clinical applications of self-control. The third and final technique of self-control is the one most often employed in clinical settings: the alteration of the response repertoire by the self-managed implementation of strategies which may, and usually does, involve self- monitoring and self-reinforcing. Within this general area there are two major subheadings: classical conditioning approaches and operant conditioning orientations. Cautela (1969) reviewed possible self-control techniques and mentioned five which were based upon the responder desensiti zation, q approaché actions behavior of incon has been example , anxiety . anxiety- control% I analOgo" PaVlov. 21 respondent conditioning model: (a) relaxation, (b) desensitization, (c) thought stopping (d) covert sensiti- zation, and (e) assertive training. Each of these approaches employs the explicit use of incompatible covert actions to reduce the probability of other specific covert behaviors (often emotions such as anxiety). This matching of incompatible behavior with the behavior to be reduced has been termed counter-conditioning (Bandura, 1969). For example, the behavior of relaxation is incompatible with anxiety. If a person can learn to relax in potentially anxiety-producing situations, the anxiety response can be controlled. The reduction of anxiety in this manner is analogous to the early classical conditioning studies of Pavlov. Just as the dog learns to associate a bell with meat powder, the individual who learns to relax in formerly anxiety-provoking situations will learn to associate relaxation with those situations. When the strength of association is greater for the relaxation response than for the anxiety reaction, the anxiety will drop out of the individual's response repertoire. Similarly, in desensitization (Wolpe, 1958), a heirarchy of anxiety-provoking events is imagined with relaxation, and the client finds the originally fear- provoking stimuli are no longer immobilizing. Although desensitization is not normally a self-control technique, Cautela (1969) and Watson and Tharp (1972) do advocate its use given certain precautions. The wolves th 's:op.' Th behaviors a pulsive be? Cox iragery tec In this pr: series of . ‘a w an unde the Client and: an a I“ u ‘39 liager §'~_us . Salmu‘us 1 22 The third respondent technique, thought stopping, involves the introduction of an imagined shout of the word "stop." This behavior is incompatible with such covert behaviors as obsessive thoughts, hallucinations, or com- pulsive behavior. Covert sensitization (Cautela, 1966, 1967) is an imagery technique which has generated considerable interest. In this procedure, the client is encouraged to draw up a series of cognitive images of varying stages of approach to an undesired behavior (e.g., cigarette smoking). As the client pictures each step very clearly in his or her mind, an aversive covert response is slowly introduced to the imagery. Normally the behavior employed as the noxious stimulus is nausea. As the client's imagination comes closer and closer to the undesired behavior (e.g., taking out the cigarette; beginning to light it), he/she also imagines the process of beginning to vomit all over the undesired object (the package of cigarettes). From a respondent viewpoint, the imagery establishes a classically- learned association of aversive stimuli and the undesired target behavior. The client is sensitized to any and all occurrences of the behavior he/she wishes to reduce. The aversive response (feeling nauseous) is incompatible with the enjoyment of the target behavior. Covert sensitization, which is viewed as a respondent or classical conditioning model, has many components which would indicate that its effects might well be resul For inst nauseous thoughts the aver thoughts cons eque easily ‘ rather Cautela asked a 23 be resultant from operant as well as respondent conditioning. For instance, a client may be instructed to practice the nauseous scene immediately following occurrence of any thoughts about the undesired target response. In effect, the aversive images become consequences of the covert thoughts related to the undesired target behavior. As a consequence of a behavior, these aversive covert images can easily be interpreted as self-punishment which is an operant rather than a respondent technique. Furthermore, in Cautela's covert sensitization training, the subject is asked at the end of each scene to imagine that he/she is moving away from the undesired behavior (e.g., cigarette smoking). The client, then, is covertly reinforced (negatively reinforced) by the removal of the aversive visual image and is positively reinforced by imagining himself/herself feeling better and better. It matters little, however, whether the effect results from respondent or operant conditioning: a client who has been "sensitized" in this way is much more likely to be able to exercise self-control and modify his/her undesirable behavior. Demonstrations of the effectiveness of the technique in altering the overt behavior of clients are found in several recent studies. Seven case studies have indicated successful application of covert sensiti- zation to the reduction of alcoholism and obesity (Cautela, 1966), sadistic sexual fantasies (Davison, 1968), smoking (Tooley & Pratt, 1967), fetishism and addiction (Kolvin, 1967) , s 1969), r drug ad: studies : of the i of alte sensiti the tec the thr QIOUps. Self~zq Support altl'loud CO‘v'ert later 1 tends Sens 1t 24 1967), sexual deviation (Barlow, Leitenberg, & Agras, 1969), homosexual behavior (Curtis & Presly, 1972), and drug addiction (Wisocki, 1973b). Since none of these case studies employed a design allowing for direct assessment of the hypothesized treatment (e.g., an "A-B-A" design), the results obtained could be open to an unlimited number of alternative explanations. The only direct experimental assessment of covert sensitization (Lawson & May, 1970) failed to indicate that - the technique was more effective in reducing smoking than the three alternative treatments presented to the control groups. Since one of these controls involved only a simple self-monitoring procedure, the study clearly does not support the effectiveness of covert sensitization. But, although there is little solid evidence to support the covert sensitization treatment directly, this paper will later present evidence from the operant literature which tends to bolster arguments for the use of Cautela's covert sensitization. The final classical conditioning self-control treatment, assertive training, has been examined to some extent by Wolpe (1968). An assertive behavior can be used as an incompatible response to social anxiety. Nothing, however, has yet appeared in the self-control literature which uses assertive training. In summarizing the applications of classical con- ditioning to the area of self-control, it must be emphasized that al zation, asserti naiada; make a flexib; technic OPeran direct Proces coined t0 cov 13139134 (CCT) desirJ The EA °an b; diffe: “eigh- °°ver on e1 25 that all five techniques examined (relaxation, desensiti- zation, thought stopping, covert sensitization, and assertive training) are employed to reduce or eliminate maladaptive behaviors. None employs a model that would make a desired behavior more likely. It is this additional flexibilty which is allowed through the use of the operant techniques. Perhaps the earliest self-control technique of an operant nature was proposed by Homme (1965) who focused directly upon covert behaviors and their role in the change process. Since Homme's approach is operant in nature, he coined a new term in the literature, "coverant," to refer to covert operants which include such behaviors as thinking, imagining, and feeling. Briefly, Homme's coverant conditioning therapy (CCT) involves systematic attempts to increase certain desired coverants while reducing other undesired coverants. The underlying assumption of treatment is that if a person can be led to think differently, that person will act differently. For example, a client who wishes to lose weight may very infrequently perform the coverent (i.e., covert operant) of thinking about how his/her stomach bulges on either side of the belt buckle. Such coverants are not pleasant and, therefore, do not occur often. Homme argues that were such thoughts to increase substantially, they would begin to interfere with such overt behaviors as eating a third piece of pie or canceling a tennis match. The solutio likely I as poss of the of a hi plishnr matter client he tak her dw drinks 10W pr Certai FrESeJ 'A ticulé 26 solution, then, is simple: attempt to increase the un- likely coverants (thinking about the belt buckle) as much as possible. To accomplish this goal, Homme suggests use of the Premack Principle (1965, 1971): Make performance of a highly probably behavior contingent upon the accom- plishment of the low probability coverant. It does not matter what high probability behavior is chosen. Thus, one client may think about how terribly fat he is just before he takes one of his frequent naps; another may contemplate her dwindling wardrobe before every cup of coffee she drinks. Both, however, would be increasing a coverant of low probability which would tend to be incompatible with certain fat-producing behaviors. Homme's original paper (1965) is a fine theoretical presentation of his ideas. A more recent publication by Mahoney (1970) provides additional insight into the par- ticulars of the coverant control methodology. Neither article, however, is as directly relevant to the present study as is the underlying assumption of the Homme approach: covert behaviors can and do effect overt behavior in predictable ways. There are several investigations which have appeared in the literature to support Homme's technique. Keutzer (1968) compared Homme's coverant control therapy (CCT) in an experimental design to reduce smoking with a breath- holding technique and a treatment which resembled an ”in vivo" covert sensitization (smoking was made highly aversii ments contro 1971) inform cover contr effec a bra resul COVEr and a diffej for a Self- all 0 Todd tn. 27 aversive). Keutzer found no difference among the treat- ments but all were significantly more effective than the controls. A study on weight reduction (Horan & Johnson, 1971) showed similar results in that Homme's CCT, an information group, and a group receiving non-contingent coverants had all lost significantly more weight than a control, but again, treatments were not differentially effective. Tyler and Straughan's (1970) test of CCT and a breath-holding technique demonstrated no significant results at all. Finally, Lawson and May (1970) compared covert sensitization, CCT, a contractual management group, and a self-monitoring control to find no treatment group differences, but significant decremental treads over time for all treatments. This again points to the fact that the self-monitoring groups seem to be equal in performance to all other treatments. Mahoney (1971), Johnson (1971), Flannery (1972), Todd (1972), and Tooley and Pratt (1967) report successful case study applications of the Homme coverant control methodology. Several of these studies, however, employed additional techniques which tend to obscure a direct test of any one method.‘ Echoing the earlier statement made in discussing Homme's techniques, Cautela, while referring to his own operant self-control procedures, explicitly states that "the manipulation of covert processes can influence overt processes in a predictable manner" (Cautela, 1970a, responsa reinforc reinfori diately imagine hehavic, increas covert The C021 overt <1 Ekamp 1% rewardj fa"Grit maybe the de the GI behaVi C «3110'! 28 p. 34). In the therapeutic implementation of his ideas, Cautela (1970a) employs imaginal manipulation of both a response and certain reinforcing stimuli (covert positive reinforcement or CPR). In what is called covert negative reinforcement (CNR), an aversive image is removed imme- diately as the behavior which is to be increased is imagined (Cautela, 1970b). In both CPR and CNR, the behavior that is pictured is one that the client wished to increase in frequency. Extensive client practice of the covert linkages is necessary in either form of treatment. The covert behaviors of imagery in a short time lead to overt changes in the client's behavioral repertoire. In applying his techniques Cautela would, for example, carefully decide with his client what would be a rewarding covert behavior, perhaps the imagining of'a favorite fishing spot, or a truly aversive covert behavior, maybe a severe disappointment or fear. Then, depending on the desired outcome of treatment, Cautela would encourage the client to imagine specific aspects of a desired behavior (e.g., appropriate social relationships) and follow it with the covert reinforcement; or alternatively, the client would imagine the covert aversive stimulus (e.g., snakes and spiders) and remove it contingent upon the imagination of the desired behavior. Once again, both of these techniques would serve to increase the likelihood of a desired overt behavior. Cautela, projectJ estimate contingt gation i Support an expe 1971) w are Coal exPlair Characg attitu: POSsib queasy on de: “(AQUA 29 There is increasing research support for each of Cautela's techniques. In a laboratory study of CPR, Cautela, Steffen, and Wish (1972) found that subjects shown projected circles could be influenced to over- or under- estimate their baseline estimates of circle sizes by contingent presentation of reinforcing images. The investi- gation was well designed and did provide substantial support for the effectiveness of CPR. Another attempt at an experimental investigation (Cautela, Walsh, & Wish, 1971) was rather sloppily designed, and although its results are consistent with the CPR hypothesis, it is more easily explained in terms of quite blatently direct demand 'characteristics. The criterion variable was a score on an attitude questionnaire which was administered twice (a possible testing effect). Since the overt behavior of questionnaire-taking is particularly open to biasing based on demand characteristics, its choice for the CPR study was undoubtedly inappropriate. In the area of CNR there are again two studies which have been completed. As in the CPR investigations, there is one well designed laboratory research study (Asher & Cautela, 1972) and one poorly designed attitude questionnaire investigation (Cautela & Wisocki, 1969). Although the second of these could be safely disregarded, the first does present solid evidence (again, in a circle estimation task) that CNR will effect overt behavior. knowledii and covi the teri equival pleasur zation this is covert There a COP-“Part (sensi 1972) contrC to in! ‘ 30 The present study has added to the accumulating knowledge in the field of covert positive reinforcement and covert punishment (CP). The literature does not use the term "covert punishment," but it is essentially equivalent to Cautela's covert sensitization. (When a pleasurable but undesirable stimulus in.a covert sensiti- zation session is followed by imagining noxious sensations, this is not different from contingently administering a covert punishment following the undesirable stimulus.) There are two studies, both well planned and executed, which compare covert positive reinforcement with covert punishment (sensitization). The first of these (Manno a Marston, 1972) used the two treatments plus a minimal treatment control in a weight reduction program. The CPR was employed to increase incompatible responses to eating, while CP was intended to suppress eating behavior. Both experimental groups had greater weight loss than the control. The second study (Epstein & Peterson, 1973b) was carried out in a laboratory setting. The criterion behavior was the choice of numbers from 1 to 100. For one group, numbers ending in 1-3 were reinforced by covert positive images and numbers ending in 7-9 were punished with covert aversive images. The consequences for the second group were reversed. The results demonstrated that the contingent covert stimuli significantly effected the mean rates of the numbers selected in the designated category (1-3 or 7-9). As would theoretically be expected, the CP suppressed the ca: indicat The sar perforr in a w? (altho Punish; Harstc abilit rates that Stlnu 31 the response class that it followed, while CPR stimulated responding in the preceding class. The covert reinforce- ment number categories were changed somewhat more than were the categories influenced by covert punishment. This indicates to some extent that SR is more powerful than SP. The same contention has recently been made in a study performed by Mahoney, Moura, and Wade (1973) who found that in a weight reduction experiment, self-reinforcement (although 222 covert in this study) was superior to self- punishment, self-monitoring or control conditions. Both sets of data previously mentioned (Manno & Marston, 1972; Epstein & Peterson, 1973b) demonstrate the ability to self-presented covert imagery to control the rates of overt responses. This supports the contention that covert stimuli follow the same principles as overt stimuli. Cautela's hypotheses gain considerable support from these two well-designed studies. As a final note, Cautela has very recently postu- lated two additional covert analogues to traditional operant techniques: covert extinction (Cautela, 1971b) and covert modeling (referred to in Flannery, 1972). The Flannery (1972) article reports a successful case study employing covert modeling and other techniques with a drug-dependent college dropout. Wisocki (1973b) also presents a case study using a number of Cautela's tech- niques in aiding a client to alter his behavior in three areas: nent oi behavic questi' just p nearly study, admini group in the Kanfeq in say r6f1e( a Cen( If th hE/shl 32 areas: (a) elimination of heroin addiction, (b) improve- ment of "self—concept," and (c) establishment of pro-social behaviors. Additional questions. There are several additional questions to be examined based on the experimental research just presented. Why do differing treatments tend to yield nearly equal results? Why in the Horan and Johnson (1971) study, for example, did the group receiving a non-contingent administration of coverants perform equally as well as the group employing the contingencies? The answer perhaps lies in the unobservable area of covert behavior. It may be that Kanfer and Karoly (1972) correctly identified the problem in saying that "the contract occupies a special position, reflecting the intention statement or performance promise, a central component of the [self—lcontrol process" (p. 410). If the client is sufficiently committed to the contract that he/she has made with his/her counselor (or, it could be a self-contract), the program will probably be successful; if not, the contingency—fulfillment behavior (Mahoney, 1970) will become increasingly difficult. Within Horan and Johnson's (1971) study, there is evidence of a decreasing commitment to the contract, since, on each successive week, the subjects displayed a decreasing implementation of the requested anti-target behavior coverants. The decreased implementation of coverants or the lack of contingency- fulfillment behavior are, in essence, a cessation of self- monitoring. When self-monitoring stops, self-directed behavic i would q regulai clusion enp loyi investi resear} contro how do lying the 1: eXamix Seem ' n 0 r1 0 £1: ’1. f” Pfl 33 behavior change stops! This interpretation of the data would clearly fit Kanfer's (1971) formulation of self- regulation, self-control, or self-management. This con- clusion, of course, was particularly related to the employment of a self-monitoring condition in the present investigation. In summarizing the entire area of self-control research, one might say that despite the host of self- control studies, very few relate directly to the question: How do covert behaviors effect overt actions? The under- lying aspects of each self-control research design employ the initiation and generation of covert behavior, but the examination of the direct effects of coverants does not seem to be a major research thrust at present. Related Studies Outside of SEIT:Eontrol A few studies outside the self-control area must be briefly cited because of their direct relevance to the present investigation. First, there have been a number of inquiries (e.g., Shaw, Edson, & Bell, 1960; Brookover, Paterson, & Thomas, 1962: and Brookover, LePere, Hamachek, Thomas, & Erickson, 1966) which report strong relationships between expectancy measures (covert behaviors), in the form of self-concept scales, and academic achievement. Thus, the covert behaviors proportedly arising from "self- concept" occur in conjunction with achievement. The difficulty, of course, with a correlational study is that behavic to the technic has co: Murray built counte 34 there is no indication of causation. The self—concept expectancy may effect achievement or vice versa. Better evidence of expectancy sets in the area of behavior change is found in the recent controversy relating to the effectiveness of the systematic desensitization technique (Wolpe, 1958)., For instance, synopsizing what has come to be a rather large collection of articles, Murray and Jacobson (1969) and Jacobs and Wolpin (1971) built strong cases for an expectancy effect (unrelated to counter-conditioning hypotheses) in behavior modification. Among the studies indicating such expectancy effects are Oliveau, Agras, Leitenberg, Moore, and Wright (1969) and Leitenberg, Agras, Barlow, and Oliveau (1969). Both studies indicated that the systematic desensitization treatment was significantly more effective when it was introduced as a highly reliable treatment of fear than when it was introduced as a physiological experiment (with no apparent therapeutic effects). Davison and Wilson (1973) indicated their belief that there is undoubtedly some enhancement of success in desensitization that is attributable to a positive ex- pectancy, but expectancy is not the total explanation of desensitization's effect on behavior. Cited among the articles supporting this position is a very fine study by Woy and Efran (1972). In their investigation, high ex- pectancy and low expectancy desensitization groups differed in on 13' the e a lie Chapfi resu 35 in only one outcome measure and that was a self-perception of improvement rating scale. The effect of this review of the expectancy literature on this study was that subjects were given a questionnaire to assess their expectancies of empathic performance before the experimental treatment. The scores on a "self-concept of empathic ability test" should be correlated with eventual scores on the outcome criteria (the number of emotional words). Making the covert behavior of expectancy overt through employment of a written test, an adaptation of Coopersmith's self-esteem scale (1959), was expected to provide an excellent covariate to employ in the analysis of the data. Overview Chapter II contains a description of the subjects, the experimental treatments, the measurement instruments, a list of specific procedures, testable hypotheses, the research design, and the statistical analysis procedures. Chapter III involves a reporting of the results of the data analysis. Chapter IV includes a discussion of these results and a summary of the study. CHAPTER II EXPERIMENTAL DESIGN AND METHODOLOGY Subjects Sixty-three volunteer subjects were solicited in the Spring term, 1973, from upper level and graduate courses in the Colleges of Education and Human Ecology at Michigan State University. A total of sixteen Education and Human Ecology classes were visited in seeking student volunteers. Appendix A provides an illustration of a typical presentation. Since the subjects for this study were not drawn at random from a defined population, a summarization of the personal and demographic data collected is included in Table 2.1. This more careful description of the "sample" of 63 will allow the reader to better determine the nature of the hypothetical population of all similar samples to which the results can legitimately be generalized (Corn- field & Tukey, 1956). Measures The study employed two types of dependent measures. First, the concomitant variable (covariate) was a 36 EU. «a «hCuU-t nrafiuv H. no a 2. ~Ch33522 CC... I N1 ' H. fiH‘XrH- tlto'nil . O I ~QCOELOL 37 AmH.HHV> .nowmm .om w ..>umm HmcCOmuom .mcflaomcsou va.ovv Awm.mmvva Hwnuo announce AwH.HHV> AwH.HHV> mearuo .>eo tango Awm.mavoa .om xumocooom Awm.ovo Awh.mmvmm .Ed .oumd cowmmosmu Awm.mvvam Aw~.~mvva .¢.z .¢.m Aw~.me~ Awm.nvm cmflmoo .>oo .Emm .om cuss: Awm.mmvva Awm.mvm .om HMfloomm mcflommm w om humanoEmHm 000m Emumoum ooumwo HOnmz no mmu< coaumuaaowoomm AwH.HHVh Awm.~mvva Awh.movmv ham: caboose Honuo >moHoom steam cofiumooom mmoaaou Awo.ava o Awm.m.m Awm.H.H Amm.ove AmH.HHV> lav.mmvoa Awo.nvvom Aw~.mv~ +wm mmnam omloe molav ovuom mmnam omlmm mmIHm om moon: was Awm.ohvmv Awm.mmvma mamsom was: xom no u Houoa Ammomucwonom ocm wocosoouwv :3ooxooum ucocomaoo moflumwuouomnmnu mowumfinouomumno ownmmumOEmo can Hmcomuom H .N wands Self-E '-:"‘_—."C -h a naire Coope: study Of err; fOrm _ Who 1 de 0r 38 self-evaluation of empathic ability questionnaire (Appendix B). Second, a set of fifteen measurements was collected reflecting the number of emotional words emitted for each subject on the fifteen cartoon-stimulated stories. Self-Evaluation of Empathic Ability» The self-evaluation of empathic ability question- naire was based on a self-esteem inventory developed by C00persmith (1959). The modified version employed by this study contained thirty-eight items covering a wide variety of empathy-related characteristics. Items appeared in the form of self-descriptive sentences, such as, "I am a person who is comfortable with emotions," and "It's difficult for me to speak about feelings." The respondent checked his or her agreement or disagreement with each statement on a 6-point Likert scale, ranging from Yvery strongly agree" to "very strongly disagree." The scoring of the question- naire simply involved a summation of the weights assigned to each of the subject's responses. The scale also con- tained eight questions constituting a lie scale which were not employed in the analysis of this study. These "lie" questions and the weightings assigned to each of the responses for all items have been included in Appendix C. Since the self-evaluation questionnaire had been developed for this study, there is no existing predictive or concurrent validity evidence available. Content validity, however, was checked through collaboration with 39 a group of doctoral students who examined the proposed empathy-related questions prior to its printing. An internal-consistency reliability of the instrument for the subjects in this study was estimated by Hoyt's (1941) Analysis of Variance procedure yielding a reliability of .9071 and a standard error of measurement equal to 4.9805. Emotional Word Measurement The ratings of the number of emotional words emitted during the stories produced the major dependent variables of the investigation. The definition of the words which would be counted as "emotional" words followed the outline of Ullmann and McFarland (1957): General definition: Words with a special ”punch" to them, which convey tension, action, or feeling, which breathe life into communication. Specific definition: Nouns which deal with inter- personal relationships of a tensional nature, such as competition, hope, approval, trouble, strength, sanity, argument, decision, problem. Verbs which deal with human tension or motivations, such as strive, plead, hang, restore, try, wonder, love, lose, regret, endure, must, want, Stare, frustrate. Modifiers, either single words, or groups of words counted as one emotional word which tell of the human condition beyond the overtly descriptive. Such words as extra kick, reached the end, cheer up, wrong, bewildered, dazed, strained, willful, rash, impulsive, cool, going too far, tense, depressed and decisive are emotional words. Words which are descriptive of the stimulus such as young, old, male, female, mother and son . . ., graveyard . . . are not emotional words. Words which are not in any of the above categories, but which communicate emotion. Exclamations such as "heck with her," "this is hard," or "like me fixing to leave home" are examples. Unusual or unexpected com- binations of words which are expressive and are not due to the subject's inattention to stimuli, such as holy protector, side of sympathy, but it has been done, are examples (Ullman & McFarland, 1957, p. 82). the ty word ( train words, conta emoti l‘VaS __au The expe: throJ The subs expe rat: in—S 40 Comments and exclamations to the experimenter during the time alloted for telling the story were included in the word count along with purely story-oriented verbalizations. An additional word list, developed during the rater- training for this study and used to identify emotional words, is included in Appendix D. Additionally, Appendix E contains three transcripts of sample stories in which the emotional words have been underlined. Every story told by each of the sixty-three subjects was rated three times for its number of emotional words. The first rating, in every case, was made during the experimental session. Subsequent ratings were performed through the use of audiotaped recordings of every story. The two out-of-session ratings were averaged for use in the subsequent analyses. Since the two raters were also the experimenters in the study, one of the two out-of—session ratings was performed by the individual who tallied the in-session rating. Since it was necessary to request the subjects to control the recording of their own stories, the experi- menter attempted to observe carefully whether the microphone was turned on or off at appropriate times. Despite this attempted surveillance, certain stories were not recorded for seven subjects. For three subjects, only one story was absent; for two subjects, two stories were missing, and there was one subject each for whom three and five stories were not recorded. In each case of missing storie 41 stories, a regression equation was calculated to predict the missing post-session rating using the in-session ratings for that subject. These regression equations were felt to provide the best available estimate of the missing data scores. Correlations between in-session and out-of- session ratings ranged from .45 to .92 with only two corre- lations falling below .70. The predicted out-of-session ratings were used in all subsequent analyses for those stories which were not recorded. The question of validity is essentially irrelevant to this variable since no argument was made that the vari- able, in and of itself, is meaningful. In some ways, emotional words can clearly be viewed as analogous to counseling related behaviors, but this was not much more than an interesting coincidence. The behavior of emitting emotional words was chosen primarily on the basis that it was relatively easily observable and that it should have been relatively easily influenced by experimental manipu- lations. It was also a particularly good variable to employ since, from the subject's point of view, it made intuitive sense that emotional words and empathy were related. This link was solidified by an experimental instruction which related empathy with the number of emotional words included in the stories. This manipulation, it was believed, would motivate subjects to attempt to increase emotional words if they could. 42 Reliability of the raters is of considerable interest, and analysis of variance estimates of reliability (Ebel, 1951) were calculated for both the set of three ratings (one in-session and two out-of-session) and for the set of two post-experimental ratings (out-of-session). The inter rater reliability (for individual ratings) for all three ratings was .702 with a standard error of measurement equal to 14.68; for the two post-experiment ratings, a reliability for individual ratings of .910 was found (rxx = .953 for average ratings) with a standard error of measurement of 2.140. These reliabilities are not un- expectedly high since the rating of emotional words, particularly in the less pressured post-experimental setting, was relatively straight forward. A second form of reliability for these ratings of emotional words was also determined. This reliability estimate was, however, more directly related to the treatment manipulations, and, therefore, the discussion of this second type of reliability will be delayed until later in this chapter. The several additional measures which are collected as part of the treatment procedures will be explained later in this chapter. These include self-estimates of the number of emotional words emitted and imagery self-ratings. the w and Cr Michi' hinse two e serve EXpla eXper exce. play 43 Experimental Manipulations Subjects, who had signed up for specified one and three quarter hour time blocks, arrived one at a time at the waiting room area adjacent to a group of small counseling and consultation rooms in the 250 complex of Erickson Hall, Michigan State University. Each subject was invited to fix himself or herself coffee and then was taken by one of the two experimenters directly into one of the cubicles which served as the experimental setting. The experimenter explained that nearly all of the instructions for the experiment had been recorded on an audiotape. The following excerpt of the experimental instruction tape was then played for each subject: You have volunteered for participation in a research study related to the question of empathy. You will be asked to make stories which will serve to help us evaluate our conceptions of the term empathy. In addition, you may also be asked during the next hour to generate mental images. You will be requested to make these images as clear and as realistic as possible. Before we start, however, there are a few forms that we'd like you to fill out. The subject was then left alone to fill out the release form (Appendix F), the demographic data sheet (Appendix G), and the modified Coopersmith (1959) self- esteem scale (Appendix B). After approximately five to ten minutes, the experimenter returned and sat inconspicuously in the corner of the experimental room (see Figure 2.1) until the subject had finished each of the forms. Upon completion of the forms, the subject was informed that the experimenter 44 ”—u-II I u -.' n 0 .. x R. I ( . ._.. C G l 3 Ir - - -1..- (1i h=-T=> f ‘@;>—_.Q*_~Dmfi i g ' "‘77 "l \ I . . , . n 3 L t J “:1 ) L_—l L—-"-- .\ j r””””'% 7”, (”W ’” i 1 g‘ m i n % Ii] L.-_MJ ! ! L J l. 3 \ (W J; ‘\ .---_.‘!_'f_| \ ii.-- -. -- r K __.__J Fig. 2.1. The layout of the experimental room during the SMSE condition. a. One-way mirror b. Recorder #1--records subject's stories c. Subject consent form, demographic data form, self-perceived empathy form d. Lamp e. Final evaluation form f. Microphone--used by Ss to turn on and off Recorder 1 g. Recorder #2--plays experimental instructions h. Pencils , i. Subject's chair j. Stack of 15 cartoon pictures k. Imagery index cards 1. Story Recall Checklist (self-monitoring card) m. Story Comparison Form (self-evaluation card) n. Imagery Rating Form 0. Control to turn of and on experimental instructions p. E's chair q. Clicker to signal and of imagery r. Door 45 would remain in the room, but would be attempting to remain as unobtrusive as possible. This then, indicated to the subjects that their attention should be forward toward the table rather than backward toward the experimenter. The experimenter then turned on recorder #2 and the following message was given: All the materials you will need are presented here on the table. Instructions will be provided both by this recorded tape and by the experimenter directly. In front of you on the table you will find a set of fifteen cartoons. When you are requested, take the top card and examine the picture on that card. You will be asked to make an imaginative story about this picture. You will be asked to tell this story out loud. In telling your story, mention something about the past, that is what led up to or otherwise explains the picture: the present, that is, what's going on now in the picture, and also the future, that is, the outcome of the story. Feel free to tell any kind of story that you wish. Try to make it interesting to yourself’ (adapted from Ekman, Krasner, & Ullmann, 1963, p. 388). The only restriction that you'll have on your story is time. You must begin when the first tone sounds and you must end two and one quarter minutes later when the tone goes off again. Before the first tone, you will have thirty seconds in which to prepare some story ideas. When the second tone is given you will still be talking. When that happens, however, just stop and wait for your next instruction. If you should finish a story early, that is before the second tone is sounded, give more details about certain aspects of the story that you've already told or, alternatively, you may continue the story further into the future. The entire test will be recorded. However, in order to record your stories, you will have to help out by remembering to turn on your microphone before you begin to speak and to turn it off after you have finished each story. Thus, when the first tone sounds, you'll pull the small switch on the mike towards you: when the second tone sounds, you'll push the switch away from you. The story that you tell will indicate to some extent how empathic a person you are. The attribute of empathy can be expressed through one's use of appropri- ate feelings and emotion-laden words. hs y the room individu Please d O.K. card. The fifteen 5 i: and adults such as fis television. Krasner, a: instrument: treatment. according from one 1 U1 a quarter instructi inrespon secOnd Cc tehd to e 46 As you'll notice, the experimenter is present in the room. At some point in the next hour or so, this individual will ask you to follow certain instructions. Please do so when that request is made. O.K. Are you ready? Pick up the first picture card. The drawings involved in this task (Appendix H) were fifteen 5 inch X 7 inch simple cartoon pictures of children and adults engaging in relatively commonplace activities such as fishing, cutting a piece of cake, or watching television. Weiss, Krasner, and Ullmann (1960) and Ekman, Krasner, and Ullmann (1963) have employed very similar instruments. Before each subject was administered the treatment, the deck of picture cards was randomized according to a previously developed list of random numbers from one to fifteen. Up to the point of completion of the first two and a quarter minute story, each subject had received identical instructions. Thus, the number of emotional words expressed in response to the first picture provided an appropriate second covariate for the data analysis. Those who naturally tend to employ emotional words were expected to continue to O do so during the experiment- Self-Regulation Dimension No self-monitoring treatment. One third of all subjects were randomly assigned to a no self-monitoring (NSM) condition. This group received no further instructions from the experimenter concerning emotional words or their relevance t include emc from the ir instructio: take the n- tones toll .815 Subjects V Following thereafte menter tC This {on ObseI‘Ve . Study; Irv In t “Ural timg YOu' In this subject EValuat final c expel-iv recall 47 relevance to the study. Any attempt by these subjects to include emotional words in their stories must have resulted from the initial introduction since the further taped instructions contained no more than brief instructions to take the next picture card with the appropriately timed tones following. Self-monitoring treatment. Another third of the subjects were assigned to a self-monitoring (SM) treatment. Following the completion of the first story and every story thereafter, these subjects were requested by the experi- menter to fill out a "story recall checklist" (Appendix I). This form instructed the subject to self-monitor or self- observe the target behavior of primary interest in the study: Think back over the story you have just related. Try to remember all of the components that you included in that story, but, in particular, think about the number of feelings that you talk about. Indicate below your best estimate of the number of times that you expressed an emotional word in the story you've just completed. In this treatment there was no direct encouragement for the subject to attempt any comparisons between stories or to evaluate his/her performance on the task. Self-monitoring and self-evaluation treatment. The final group of subjects were given instructions by the experimenter to both self-monitor and self-evaluate (SMSE). As in the SM group, each story was followed by a "story recall checklist." Additionally, however, after every story fron fill in tr contained their per} story: W. feeli: compa this EXpre emoti '1 inVOlved desoribe, Ways On rece iVed anISiVe in eVer: the “Sn bath tn emetic: factor an ima, incl-ea frOm C In the thet‘s. 48 story from the second onward, the subject was instructed to fill in the "story comparison form" (Appendix J). This form contained directions for the subjects to self-evaluate their performance relative to the immediately preceding story: Were you able to incorporate more feelings or fewer feelings into the story you've just completed? When compared to the story immediately preceding it, was this story greater in the number of emotional words expressed, about the same, or smaller in the number of emotional words expressed? Imagery Dimension The second independent variable of the study involved visual imagery. Each of the three conditions described earlier (i.e., NSM, SM, and SMSE) was split three ways on the imagery variable. A third of each group received one type of image, either positive, neutral, or aversive. Although the presentation of images was contingent in every condition, the contingencies were not the same for the NSM and SM groups as they were for the SMSE group. For both the NSM and SM conditions, the experimenter's tally of emotional words expressed during a story was the deciding factor in whether the subject would or would not receive an image. If the tally indicated that the subject had increased in the number of emotional words (EW) expressed from one story to the next, an image card was presented. In the case that the number of emotional words remained the same, subjects were only requested to take an image card _i__f_ the the immediat involved the Plea card ov: reverse lma Close y and as actual] touch, bring 1 the cl; rate y. Th HOVlever, t would not SUbjeCt's If the £0 emOtional indicated had been the form earlie: 4 49 card ig the subject had received an image card following the immediately preceding story. The imagery instructions involved the following: Please take an index card from the table. Turn that card over and read the scene that is written on the reverse side. Imagine the scene that is written on your card. Close your eyes and try to make that image as clear and as vivid as possible. Make it seem as if you were actually there and part of that scene. If smell, touch, sound, or taste are appropriate to the scene, bring them clearly into your imagery. When you hear the clicker sound, open your eyes; stop the image; and rate your image on the green colored card (Appendix K). These instructions were the same in the SMSE group. However, the contingencies for whether the subject would or would not receive an image card were provided by the subject's self-evaluation on the "story comparison form." If the form indicated a rating of "greater" numbers of emotional words, an image was presented. If the card indicated "about the same," an image was given only if there had been an image given following the preceding story. If the form indicated "smaller," no image was presented. In discussing the ratings of emotional words earlier in this chapter, mention was made of a second type of reliability coefficient on these ratings. This second form of reliability involved nothing more than determining the proportion of appropriate decisions made by the experi- menter during the experimental treatment. For example, in the NSM and SM conditions, the experimenter's tally of emotional words decided whether the subject would or would not receive an image card. The reliability of these “on-the-spo decisions b and computi session and condition, decided by not includq resultant . number of difference (e.g., In- Decision n 1 = 24.5 I not deliVe considErm Pmportig1 that Were decisiOn ‘ fielded a total Of tIEatment indicated very near F 1' . Jal 117' 50 "on-the-spot" decisions was determined by making similar decisions based on the more reliable post-session ratings and computing the proportion of agreements between the in- session and out-of-session decisions. Since, in the SMSE condition, the determination of imagery presentation was decided by the subject, those in the SMSE condition were not included in the calculation of this percentage. The resultant exact prOportion of agreement was .813. In a number of the cases of a ”lack of agreement," the decisional difference was within only one or two emotional words. (e.g., In-session: story 1 = 22 EW; story 2 = 23 EW - Decision made to deliver image card. Out-of—session: story 1 = 24.5 EW; story 2 = 24 EW - Decision would have been to not deliver image card.) Since such differences could be considered to be inconsequentially differing, a second proportion was calculated where all "lack of agreements" that were within two emotional words of the alternative decision were included as "agreements." This reliability yielded a proportion of .912 or 535 "agreements" out of the total of 588 decisions made by the experimenters during the treatment sessions. .The high reliabilities reported here indicated that the treatment was, in fact, administered in very nearly the manner in which it was designed. Following a thirty second presentation of the visual imagery, each subject completed an imagery rating form (this is the "green card" referred to in the earlier instructions). The rating involved two dimensions: (a) the vividness oi to 'vivid") be clearly to 'immedia {g were rando imagined s lottery or Sixteen p( the deck . list of a It was it: be sample) images w Positive “Bi-QUE: subjecm 0‘ (b) ditiOn The im. 51 vividness of the image on a l to 7 scale ("almost no image" to "vivid"), and (b) the speed with which the image could be clearly attained, again, on a l to 7 scale ("no image" to "immediate”). Positive imagery. The third of the subjects who were randomly assigned to the positive imagery condition imagined such positive scenes as (a) winning the Michigan lottery or (b) soaking in a hot tub. There was a total of sixteen possible positive images. Between each subject, the deck of images was shuffled to insure randomness. (A list of all of the images employed is found in Appendix L.) It was important that a variety of visual imagery stimuli be employed so that satiation to one image would not occur (Cautela, 1972). Neutral imagery. As indicated earlier, the neutral images were presented on the same contingent basis as the positive images to the one third of the total group assigned to the neutral condition. This treatment asked subjects to imagine such scenes as (a) a pencil sharpener, or (b) a piece of chalk. Aversive imagery, The image cards in this con- dition describe such scenes as (a) biting into a maggot- filled hamburger, or (b) vomiting all over one's kitchen. The images were administered on the same basis as the previous two imagery conditions. Fc subject w (Appendix the probe had had 6 attempte: Addition Subjects exPerime the expe was not Stated, t<3 ten 1 the the the Sub 52 Following completion of the fifteen story, each subject was asked to fill out the "final evaluation form" (Appendix M). This form asked what the subject felt was the probable purpose of the study, what thoughts the Subject had had during the experiment, and whether he or she attempted to maximize the number of emotional words. Additionally, two questions were asked to determine if the subjects were self-monitoring or self-evaluating during the experiment. After the "final evaluation form" had been answered, the experimenter explained to the subject that the study was not actually related to empathy as the introduction has stated. Also, the experimenter usually spent another five to ten minutes explaining the experimental procedures and the theory behind the investigation. During this time, the subject was again asked to cooperate in not discussing the experiment with classmates. Hypotheses Although the literature on self-control of behavior makes implicit use of the assumption that covert behavior affects overt action, very few direct attempts to assess the effect of covert reinforcement and covert punishment have been performed (Mahoney, 1972). Additionally, few well controlled laboratory investigations have dealt with the crucial issues raised by Kanfer's model of self- regulation (1971). In light of these facts and in 53 consideration of the lack of consistent successful appli- cation of any of the covert self-control methodologies, this experiment assessed the effects of two distinct covert behaviors upon an overt performance task. The two independent variables included in this study were (a) the self-regulation dimension (NSM, SM, and SMSE conditions) and (b) the type of imagery (positive, aversive, or neutral imagery) which was contingently administered following an actual (in NSM and SM groups) or perceived (in SMSE) increase in emotional words from one story to the next. Hypothesis One Kanfer's (1971) model of self-regulation suggests a three step process: self-monitoring; self-eValuation, and self-reinforcement. For maximally effective self- change, all three components are needed. Thus, to remove any one of these components would produce a deficit in effectiveness. Therefore, the following hypothesis was developed relative to the self-regulation dimension. In a population of volunteer upper level and graduate students who are enrolled in courses in the College of Education and Human Ecology at Michigan State University, there will be differences in the number of emotional words emitted in response to-TAT-like cartoon pictures between (a) groups not asked to self-monitor, (b) groups requested to self-monitor, and (c) groups requested to self-monitor and self-evaluate. It is expected that the no self-monitoring group (NSM) will produce fewer emotional words than will the self- monitoring condition (SM) which will result in fewer such emotional words than the self-monitoring and self- evaluation (SMSE) condition. Sinc repeatedly ( effectivenes develop ove either a ma bY an inter variable a: analysis , I1 contingen “We! of anticipat Strength“ neutral negative suppreSs emot Volt Stud Cati 54 Since the criterion variable was collected repeatedly (during each of fifteen stories) and since the effectiveness of the differing levels of treatment may develop over time, this hypothesis could be supported by either a main effect for the self-regulation variable or by an interaction effect between the self-regulation variable and the measures variable in a repeated measures analysis. flypothesis Two It was expected that the visual imagery, presented contingent on behavior, would differentially effect the number of emotional words produced by subjects. It was anticipated that positive imagery would tend to reinforce, strengthen and increase the number of emotional words; neutral imagery was expected to have little effect, and negative imagery was theoretically expected to reduce or suppress subjects' responding through emotional words. The imagery condition will effect the number of emotional words expressed in TAT-like stories for volunteer subjects who are upper level or graduate students enrolled in classes in the Colleges of Edu- cation and Human Ecology at Michigan State University. The effect of the imagery conditions will be such that those subjects in the positive imagery group will produce more emotional words than those in the neutral imagery condition who will, in turn, produce more than subjects in the aversive imagery condition. As with the first hypothesis, the repeated nature of the criterion variable leads this hypothesis to two possible sources of support either in a main effect for the imagery variable, or more likely, in an interaction between 55 measures and imagery in a repeated measures analysis. (This is particularly true with the imagery condition, since it was expected that imagery would slowly effect the number of emotional words throughout the course of the fifteen story treatment.) Hypothesis Three The review of the literature provided no research or theoretical support for an expectation of an interaction between the self-regulation condition and the imagery variable. Thus, the third hypothesis takes the null form. There will be no interaction between the self- regulation and imagery variables as measured by the number of emotional words expressed in TAT-like stories. Experimental Design Consistent with the hypotheses given above, an experimental design with repeated measures was employed for this study. The 3 x 3 factorial design with multiple dependent variables (number of emotional words emitted during the fifteen stories) and the independent variables of self-regulation and visual imagery is shown in Table 2.2. It should be noted that an attempt was made to increase the precision of this general design by the inclusion of two covariates, both of which were expected to be highly related to the criterion variable. The first, ”self- concept of empathic ability" is a modification of the Coopersmith self-esteem inventory (1959). The second is Indepei Self-Regu N0 SEIfT Monitorin Self- MOnitorix Self~ MonitOI-i an Self- COVariat 56 Table 2.2 Experimental Design N = 63 Independent Variables Dependent Variables Self-Regulation Visual Imagery 14 Ratings of Emotional Words emitted during TAT-like stories (Stories 2 through 15) No Self— L— Positive images 7 Subjects per cell Monitoring Neutral images Aversive images Positive images Self- Monitoring Neutral images Aversive images Positive images Self- Monitoring Neutral images and Self— Aversive images Evaluation Covariate measures: 1. Self-concept of empathic ability questionnaire 2. The number of emotional words expressed during the first story dwinmberi first carto The of cartoons introducing felt to be gain in 9e: An introduced addition t Possible t to be £305: Opposite ; agains: s F itshould indicated Ship to ‘ gation. to be 11: ditions 57 the number of emotional words emitted in response to the first cartoon picture. The use of a randomization procedure on the order of cartoons probably served to decrease precision by introducing some unexplained error variance, but it was felt to be worth the precision loss for the association gain in generalization. Another experimental manipulation likely to have introduced unexplained variance was the imagery cards. In addition to the randomization of imagery cards, it was possible that for some individuals, certain scenes expected to be positive (or negative) may have elicited just the opposite reaction. Feasibility problems, however, mitigated against self-selection of reinforcing or aversive images. Finally, in a consideration of the precision issue,' it should be stated that the review of the literature indicated no other variable with a demonstrated relation- ship to the criterion variable studied in this investi- gation. No subject characteristics have been demonstrated to be likely to interact with any of the treatment con- ditions described for this study. Analysis The data was analyzed from two approaches. A multivariate analysis of covariance (MANCOVA) was performed employing the fourteen emotional word ratings and the two covariates (see Table 2.2). This analysis was run on the Finn Mult 1968). 'I program a (Jennricl covarianc to storiv It was t} qUeStion. Pendent ' would £01 and Poss. independ‘ reported marine r . 58 Finn Multivariate Analysis of Variance program (Finn, 1968). The second major analysis employed both the Finn program and the Jennrich Analysis of Variance program (Jennrich, 1961) for a repeated measures analysis of covariance with fourteen emotional word ratings (responses to stories two through fifteen) and the same two covariates. It was the intention of the first analysis to relate to questions concerning the effects for the two major inde- pendent variables and their interaction, while the second would focus primarily upon the repeated measures' effect and possible interactions between measures and the other independent variables. The rating scales and the self- reported data were analyzed and reported in a descriptive manner . in the C( IBEd in level of tistical whiCh WE invesns analyse: be reps. descrip in the variate anaIYSi determi and the CHAPTER III ANALYSIS OF RESULTS The Control Data Corporation 6500 computer system in the Computer Center at Michigan State University was used in calculating many of the statistical analyses. The level of significance for all tests was .05. Chapter III contains the reporting of the sta- tistical analyses of the results of the study. The analyses which were relevant to the primary hypotheses of the investigation are presented first. Then, several subsidiary analyses of data indirectly related to the hypotheses will be reported. Finally, attention will be given to several descriptive analyses of various measures that were included in the study. The Primary Analyses The first analysis to be considered.was the multi- variate analysis of covariance (MANCOVA). As with any analysis of covariance approach, the initial step was to determine the relationship between the covariate measures and the dependent variables. Table 3.1 presents the S9 60 Table 3.1 Regression Analysis Portion of the Design: The Relationship Between the 14 Ratings of Emotional Words and the 2 Covariates (Self-Perceived Empathy and EW in the First Story) Dependent Multiple Correlation Variable With 2 Covariates F p Less Than Time 2 .7005 25.0519 .0001 Time 3 .6174 16.0196 .0001 Time 4 .6749 21.7457 .0001 Time 5 .5354 10.4498 .0002 Time 6 .5328 10.3068 .0002 Time 7 .5587 11.7996 .0001 Time 8 .5657 12.2357 .0001 Time 9 .6222 16.4217 .0001 Time 10 .6158 15.8859 .0001 Time 11 .6369 17.7443 .0001 Time 12 .5415 10.7876 .0002 Time 13 .5288 10.0910 .0002 Time 14 .5655 12.2241 .0001 Time 15 .6391 17.9567 .0001 degrees of freedom for hypothesis = 2 degrees of freedom for error = 52 P Value for test of hypotheses of no association between dependent and independent variables = 2.8109 d.f. = 28,78 P less than .0002 evidence test of n (the emot covariate Story 1) 78. this by Chance high deg: dependen bee“ con 0f Varia two Sets of the c empathic was not .8193. Varialsi. EffeCti‘ first 5 predict 61 evidence relative to this question. The F-value for the test of no association between the 14 dependent variables (the emotional words in stories 2 through 15) and the two covariates (self-perceived empathy and emotional words on Story 1) was 2.81. With degrees of freedom equal to 28 and 78, this result indicated a relationship which could occur by chance alone only two times out of ten thousand. The high degree of relationship between the covariates and dependent variables indicated that the MANCOVA should have been considerably more precise than the straight analysis of variance would have been. More understanding of the relationship between the two sets of variables was found in the stepwise contribution of the covariates. The F-test for the contribution of the empathic ability questionnaire to the regression equation was not significant: F = .6349 with 14,39 D.F. and p < .8193. Thus, in terms of predicting the set of dependent variables, the empathy questionnaire was relatively in- effective. The second covariate, emotional words in the first story, was found to contribute significantly to the prediction of the fourteen dependent variables: F 6.5345 with 14,40 D.F. and p < .0001. An examination of the univariate Fs associated with the relationship between Story 1 and the remaining fourteen stories indicated each was significant beyond the .0001 level. It was clear that most of the overall relationship between the covariates and depender with the the canc depender equaled the depe the two lation y measure. the StO] .4284 a, depender Um Cove highly (‘.l427 is recs C18 ar 1y variabi relevam Table :3 Em thr “U32 not to 1 m0“: the 62 dependent variables stemmed from Story 1's common variance with the dependent measures. Further data relevant to this subject was fOund in the canonical correlation analysis between covariates and dependent measures. The first canonical correlation equaled .8346 and accounted for 4.98% of the variation in the dependent variables. The standardized coefficients for the two covariates in forming this maximal canonical corre- lation were -.9952 for Story 1 and -.0395 for the empathy measure. Clearly, the largest contribution was coming from the Story 1 variable. The second canonical correlation was .4284 and only accounted for 1.3% of the variation in the dependent measures. Here the standardized coefficients of the covariates were reversed in order with empathy weighted highly (1.0046) and Story 1 receiving very little weight (-.l427). The unique contribution of the empathy variable is recognized in the second canonical variate, but it clearly did not provide a very effective explanation of the variability of the dependent measures. The information on the analysis of the data relevant to Hypotheses One, Two, and Three is found in Table 3.2. Before considering the results of Table 3.2, the three hypotheses will be restated in summary form. Hypothesis One: There will be differences in the number of emotional words emitted between (a) groups not asked to self-monitor (NSM), (b) groups requested to self-monitor (SM), and (c) groups requested to self- monitor and self-evaluate (SMSE). It is expected that the NSM group will produce fewer emotional words than * A. ‘ So: \- MGil-e of Self- Merv (I) SR x I \ 1., Value for mdePendent \— 63 Table 3.2 Summary of the Multivariate Analysis of Covariance and the Repeated Measures Analysis of Covariance on Fourteen Observations of the Number of Emitted Emotional Words A. MANCOVA (3 levels of self-regulation x 3 levels of imagery with 2 covariates* and 14 measures) Multivariate Multivariate P less Than Source d.f. F Mode of Self-Regulation (SR) 28,78 .643 .9052 Imagery (I) 28,78 .853 .6745 SR x I 56,153.86 .846 .7614 F Value for test of hypothesis of no association between dependent and independent variables 2 2.8109 D.F. - 28,78 p < .0002 Y B. Repeated Measures ANCOVA (3 levels of self-regulation x ' 3 levels of imagery x 14 repeated measures with 2 covariates) Source d.f. d'f'adj. ssadj. MSadj. F P Less Than Mode of Self-Regulation (SR) 2 2 95.13 47.57 1.589 .2139 Imagery (I) 2 2 28.70 14.35 .479 .6219 SR x I 4 4 109.05 27.26 .911 .4648 Replications (R) Within SR x I 54 52 1556.88 29.94 Measures (M) 13 13 1371.59 7105.51 2.438 .0338** 53.x M 26 26 790.74 30.41 .787 .9213** I X M 26 26 1074.73 43.34 .952 .6273** SR.X I x M 52 52 1998.46 38.43 .887 .7850** Error (R x M: SR.X I) 702 700 30455.62 43.38 $2 for measures is .09 F Value for test of hypothesis of no association between dependent and independent variables - 32.84 D.F. - 2,52 p < .0001 *Covariates employed were: 1. the number of emotional words emitted in the first storv 2. the self-perception of empathic ability questionnaire **Probabilities based on an exact test of repeated measures effects. will th words i “Y: the nu jects emotio condit in the H1 betwe. measu E logically {Table 3 multiVaz thEn' i‘. remen-‘be H FIOVQG in elt] repeat indepe analYs neithe 64 will the SM condition which will result in fewer such words than the SMSE group. Hypothesis Two: The imagery condition will effect the number of emotional words expressed. Those sub- jects in the positive imagery group will produce more emotional words than those in the neutral imagery condition who will, in turn, produce more than subjects in the aversive imagery condition. Hypothesis Three: There will be no interaction between self-regulation and imagery variables as measured by the number of emotional words expressed. Hypothesis Three, the interaction hypothesis, is logically the first to be examined through the MANCOVA and the repeated measures ANCOVA. The results of both tests (Table 3.2) support this hypotheses since no significant multivariate (p < .7614) or repeated measures (p < .4648) interaction between SR and I was found. Hypothesis Three, then, in its null form is not rejected. It must be remembered, however, that the hypothesis has not been "proved," it merely has not been rejected. For Hypothesis One and Two, support might be found in either the main effects (in either the MANCOVA or in the repeated measures ANCOVA) or in an interaction between the independent variables and measures in the repeated measures analysis. Both analyses are reported in Table 3.2, and neither give support to the hypotheses as stated. There was no significant SR X I X M interaction (p < .7850), I X M interaction (p < .6273), or SR X M interaction (p < .9213). Additionally, no treatment effects were found for imagery (MANCOVA: p < .6745 and repeated measures A. p < .9052 De No, the : argues £02 3.3 and 3 (Figures although 3 indicate . For examp 0f the me. large as Also, the Where the group, I Significa '1 indicates the “are of eithe; only One for the j group 8x< 65 measures ANCOVA: p < .6219) or self—regulation (MANCOVA: p < .9052 and repeated measures ANCOVA: p < .2139). Despite the lack of support of hypotheses One and Two, the importance of these hypotheses to the research argues for the inclusion of the interaction means (Tables 3.3 and 3.4) and a graphical presentation of these means (Figures 3.1 and 3.2). The graphical presentations, although ggt_representing any significant differences, indicate some trends in the direction of the hypotheses. For example, no mean of the NSM group, with the exception of the mean at the time of the second story, was ever as large as the corresponding mean in the SMSE condition. Also, there were only two cases among the fourteen stories where the means of the SM condition exceeded the SMSE group. It must be stressed that these tendencies were not significant. The graphical presentation of the imagery conditions indicates similar non-significant tendencies. No mean of the aversive subjects was as high as the corresponding mean of either the neutral or of the positive group. Also, in only one case after the fifth story (allowing some time for the images to begin to have an effect) was the positive group exceeded by the neutral condition. Returning once again to Table 3.2, it must be noted that the only significant difference obtained.was on the measures variable in the repeated measures analysis of covariance (p < .0338). Omega squared (62), a measure of 66 sodassas>m va.oa vm.h mu.aa oH.HH no.o vo.- nh.a ~o.o~ ma.oa H¢.~H ov.au oo.aa o~.o« Hm.m vo.o~ ow inflow nus: ocauouusox oo.on vm.on no.u~ mv.a~ oH.h« vo.mu ov.m~ H~.on oa.nn ov.~m «0.0” on.o« hm.on H~.m~ rn.ma m inflow no.o mm.o ua.oa nn.m vm.¢ vm.oa qo.o mm.» am.h oa.m «h.m nw.¢ mm.a hu.o on.m om usuuouasoz oa.o~ oo.o~ v~.o~ an.r~ on.o~ om.m~ nv.m~ ~n.n~ H>.m~ mH.m~ Hm.h~ no.>~ oa.o~ w>.mu wo.h~ m nuaom nh.r oo.o no.a om.o an.¢ so.o~ oh.h no.o nm.o om.oa vm.s vv.oa m~.h mo.o no.0 om usauouusox ma.>m mo.o~ nn.m~ «H.o~ an.v~ mm.m~ ~m.o~ vu.m~ a~.s~ na.o~ ~m.m~ va.n~ vh.w~ ov.mu na.ou m awasm 02 ma «a ma NH Ha ca o o s o m v n N ass: noduwcsou undue an wonesz aboum aboum room as owuuasm mouoz ascofiuosm uo mCOwusw>oo ousocsum ouaoom one uses: coauaocoo coauoflamoauuaom room now m.m wanna 67 oa.p n¢.a ob.m hv.¢ vo.m oc.o m~.o mo.> n~.m o~.o vn.m on.» oo.m an.m hm.h om o>uauo>< nn.h~ vn.nu ah.m~ ho.n~ nm.n~ o¢.n~ vo.vm on.m~ ~m.vn nv.wu hn.n~ om.m~ oo.vu on.nu oo.m~ m ao.NH au.ua vo.na Hm.aa mn.oa an.aa mn.n~ oo.aa hm.na nm.na m~.m~ on.na oa.ad NA.HA on.~a om Huuusoz nn.ou ~m.ou mn.»~ ho.o~ mo.s~ ~o.o~ an.on pa.s~ ov.on oa.on ~u.on mm.b~ no.u« no.h« vo.o~ m Hm.o hn.o «u.o no.9 o>.o oo.~n nu.p so.o aa.aa ma.a~ «v.9 no.0 n«.o mn.m nm.o om o>wuuuom «o.cn mn.a~ no.¢~ va.cn va.o~ mv.on ov.on ~H.¢N uo.nn Na.nn vs.¢~ no.0“ om.m~ hm.mu vn.a~ m ma va .nH ad ad cu m o h w n v n a cool ceauavcou amass: auoum undue auooasn ceaufivcou abomQEH noun uOu muoum comm co unuoz aqueduOEm no uncauoa>oa cuoocaum ooaoom can undo! v.n vague acuunznnymuhnom ~n 5!. 53.3“. Eozuuuon a maiden «EOIIMHOD o \rNh :«HOOEO‘INHOD. OZ 0 WNH 68 50.3.3500 «5.31.268 «30m sumo new muoun scum now «who: “HS—0.3080 no 3085: 503 no coauuuconoum Hananmuuu in 6: ma 3." p nu Na — .3 — OH - m . chews—z human /° m a n N > _ _ :oauuagludom :3: gauoudcoxluaom - wndhoa «com-laden o 3.30» «coxludom oz 0 AN .NN VAN Ion .Hn fimn swan m0130‘ 3° Ion-m rrlll hows—q IbthObc‘ - REQUIEH wanna-.2 o E's-.H .bflOfiIOAN V 69 .:owu«vcoo bummed :25 new roam nuuo now mayo», Hun—039:0 no 938.5: .305 no defluuucouuum Hmownmuuu .~.m .mflm nuonlsz unoum 3 i S 2 3 3 o m s m m 4 .n N. 3 . < p _ . n _ - p . - .nu v.3“ I N . .mu m u N ‘ 1 IWN 'O' u n .S m. TL u 1? .mw m I- d on .QN w I x on / 0/ /o\ .3 -mn Eamon—H abanuo>< - (lull: buomaaH Hduusoz o .nn buowuan ohaudnom ¢ an the proport of the vari {For a disc presents U. comparisons "Jest (Kirk, Were found graphicalh shown were‘ between me‘ Primary hy' medsures v time effec any One CE time (Or . The highe Stories i words OCC one' Of < thefirefore analYSES h‘J’LDOthESe Thus' Hy; 70 the proportion of variance accounted for, suggests that 9% of the variance is accounted for by the measures effect. (For a discussion of 62, see Kirk, 1968, p. 127.) Table 3.5 presents the means for story order and a set of post hoc comparisons between means using Duncan's New Multiple Range Test (Kirk, 1968, pp. 93-94). Thirteen simple contrasts were found to be significant at the .05 level. Figure 3.3 graphically presents these same story means. In examining this figure, it should be remembered that the differences shown were across all groups. This significant difference between measures was, in fact, unrelated to any of the three primary hypotheses of the study. The differences on the measures variable must be attributable to some type of time effect since cartoon stimuli were randomly assigned and any one cartoon was unlikely to be presented at the same time (or story position) for one subject as for another. The highest mean for emotional words was expressed to the stories in time position number six; the fewest emotional words occurred in telling story number two. (Story number one, of course, was used as a covariate in the study and, therefore, does not appear in any of the repeated measures analyses.) In summary of the results of relevance to the major hypotheses, it was found that there was no significant interaction between self-regulation and imagery conditions. Thus, Hypothesis Three was not rejected. However, there were no significant effects for either the self-regulation .mo. n o as names cuosuon ouconomuwo ucuowmacmams 71 vm.pn n cam . O w Nu m mm nu u m . mo.~ . a max 3m w 3533.53 uoz on R .. . . . m «00 n ma N ans nuuunucou no noocausum ma ha a m .vo.m n~.~ ea.pu u mam .po.n .oo.~ m~.~ Hm.on a v- .m~.v .mv.n .hm.~ no.~ mm.m~ a Ham .-.v .Hm.n .om.n .mo.n .mm.~ ma." we.mu u a- moauoum mH.on u mm hm.m~ u pm os.m~ u mam Hm.m~ a cam mv.mu u mm vs.e~ u mm . . . «m sou «can: cognac puss cocoa oamauanz 302 m.csoc=a ocean consowmwcuam oanwnuom uo nunsuuucoo nonaoz muoum How moonsnoumaa can: one nonsuz huoum some now vouuwsm mono: accoauoam no nonsoz can: m.m manna. J) 7 3 31 '72 .cofifimom roam £000 new oouuafio 3503 HnGOHuOEo uo none—E 50E 0:» mo dehumuconoum Huownmauo .m.m .mam :35: been 3 i 9 3 S 3 m a s m n a n A P _ _ . _ _ I _ _ . _ t on r. . - . S 4a .on In rmn 5 and imagery measures va to both Hyp neither. Al ses of the carried 01 imagery C! for each attainmen '1 3‘6° The Vectors I degrees Thfl'efOI in the c ”of a1: uniVari‘ Similar Variabl p<.02 COndi ti The tre 73 and imagery conditions or for their interaction with the measures variable. These findings led to negative answers to both Hypotheses One and Two since support was found for neither. Subsidiary Analyses Although not directly related to the main hypothe- ses of the study, a multivariate analysis of variance was carried out to determine if there were differences between imagery conditions on the mean imagery rating form variables for each subject: vividness of imagery and speed of attainment. The results of this analysis are located in Table 3.6. The multivariate test of the differences between mean vectors yielded an P value of 2.4994 which with 4 and 118 degrees of freedom was significant at less than .0463. Therefore, differences did exist between imagery conditions in the combination of a subject's mean vividness rating (for all images) and mean speed of image attainment. The univariate tests, also reported in Table 3.6, indicate similar significant differences for the two imagery mean variables (vividness: p < .0302; speed of attainment: p < .0294). The means and standard deviations for each imagery condition on the two variables are indicated in Table 3.7. The trend in both measures is that the highest ratings are located in the neutral imagery condition with the lowest 74 Table 3.6 Multivariate and Univariate Tests for Imagery Effects Upon Mean Vividness of Image and Mean Speed of Image Attainment Multivariate r Ratio = 2.4884 p < .0463 D.F. = 4,113 Univariate Variable MS Univariate F P Less Than error Mshypothesis Vividness .5325 1.9784 3.7152 .0302 Speed of Attainment 1.5426 5.7755 3.7440 .0294 D.F. for Hypothesis - l D.F. error = 60 Table 3.7 Cell Mean and Standard Deviations on Vividness and Speed of Image Attainment Variables Vividness Speed Cell N Mean S.D. Mean S.D.* Positive Imagery 21 5.619 .675 5.222 1.382 Neutral Imagery 21 5.859 .606 5.823 .812 Aversive Imagery 21 5.250 .879 4.778 1.435 *Hartley's homogeneity of variance test on speed scores yields F max = 3.04 d.f. = 3,20 p < .05 that variances are homogeneous 75 in the aversive imagery group. It is also indicated in Table 3.7 that Hartley's homogeneity of variance test (Kirk, 1968, p. 62) yielded a significant difference between variances at the .05 level on the speed of imagery variable. Since the analysis of variance is particularly robust to a violation of the homoscedasticity assumption (i.e., equal variances) when cells are of equal size, this finding does not effect the validity of the analysis. From inspection of the standard deviations in Table 3.7, it is clear that the speed of imagery attainment variable varied less in the neutral imagery condition than in either the positive or aversive conditions. The Duncan's New Multiple Range Post Hoc comparisons for differences between levels of the imagery variable on the vividness and speed of attainment measures is presented in Table 3.8. On both variables, the only significant differences were found between the aversive imagery con- dition and the neutral imagery group. The neutral imagery subjects tended to have more vivid images which were attained with greater speed. An analysis of variance was performed on the raw number of images which were obtained by each of the nine treatment groups. There was no significant interaction between self-regulation and imagery variables, nor was there any main effect for either the self-regulation or imagery variable. Since no effects of this subsidiary analysis proved significant and since it was only 76 Table 3.8 Duncan's New Multiple Range Post Hoc Comparisons of Differences Between Imagery Conditions on Vividness of Image and Speed of Image Attainment Vividness of Imagery xneutral xpositive xaversive ineutral . . .240 .609* ipositive . . .369 iaversive ‘ ° Speed of Image Attainment 5Eneutral ipositive sEaversive ineutral . . .601 1.045* ipositive ’ ° '444 iaversive ' ' *Significant at .05. 77 tangentially related to the primary questions of the research, no tabular presentation of these results has been included. Descriptive Analyses The differences in the number of emotional words emitted in response to the various specific pictures (no matter what their order of presentation to any one subject) are descriptively illustrated in Table 3.9. The tennis picture stimulated the largest mean output of emotional words across all sixty-three subjects (32.67), while the cartoon picture of bird watching drew, on the average, the fewest emotional words in response (22.27). The table indicates that, indeed, there were inherent differences between pictures in their ability to stimulate emotional words. Descriptive data on the imagery ratings given in response to each of the possible image cards appear in Tables 3.10, 3.11, and 3.12. This descriptive information for the positive images is given in Table 3.10; for the neutral in Table 3.11, and for the aversive images in Table 3.12. ”Winning the Michigan lottery” appeared to be the most effective positive image, while ”lying on a warm, sunny beach” seemed to be the least effective. 0f the neutral images, ”the pencil sharpener" was most quickly imagined, and ”the banana peel” was imagined most vividly. The ”empty movie screen" was the 78 Table 3.9 Means and Standard Deviations of Number of Emotional Words Emitted in Response to Particular Cartoon Pictures Picture Title* Mean EW S.D. 1. Tennis 32.67 11.08 2. Fisherman 25.52 10.12 3. Cake 23.50 10.56 4. Chemist 28.08 10.54 5. Dog 28.32 9.29 6. Shiver 29.22 11.16 7. Baby 25.34 9.18 8. Report Card 28.56 9.77 9. Couch 28.52 11.05 10. Secretary 25.58 11.37 11. Sick 29.46 8.93 12. Teacher 28.94 10.10 13. Birds 22.27 10.19 14. Banjo 25.21 9.93 15. Grandmother 27.42 8.83 N = 63 for all pictures *Copies of these cartoon pictures are located in Appendix H . Table 3.10 79 Means and Standard Deviations of Image Vividness and Speed of Imagery Attainment for Each of the Possible Positive Images Number of Vividness Speed Abbreviated Title Subjects of Image* Receiving Mean S.D. Mean S.D. Image 1. Fireplace, soft music, a loved one 9 5.44 1.94 5.33 1.94 2. Worry-free weekend 9 5.78 1.56 5.56 1.51 3. Call from.an old friend 5 5.80 1.30 5.40 1.82 4. A job well done 8 5.00 1.20 5.25 1.49 5. Michigan Lottery winner 8 6.12 .99 6.00 .92 6. Ability to be invisible 7 5.14 1.77 4.71 2.06 7. Orange-red sunset 10 5.80 .79 5.50 1.27 8. Lively party 10 4.60 1.35 4.90 1.66 9. Free food 6 5.67 1.21 5.33 1.37 10. Mountain pathway 6 5.67 1.21 5.33 1.37 11. Warm, summy beach 5 ' 4.60 2.07 4.40 2.07 12. Outstanding test.performance 4 5.75 .50 5.75 1.50 13. Eating ice cream 9 5.89 1.36 5.78 1.48 14. Floating above the clouds 10 5.60 1.35 5.20 1.87 15. Bright, warm.summer vacation 6 5.83 .98 5.33 .82 16. Soaking in hot water 12 5.92 .67 5.92 .79 *Appendix L contains a complete list of the images presented. 80 Table 3.11 Means and Standard Deviations of Image Vividness and Speed of Imagery Attainment for Each of the Possible Neutral Images Number of Vividness Speed Abbreviated Title Subjects of Image* Receiving Mean S.D. Mean S.D. Image 1. Gray rock 7 6.00 1.15 6.43 1.13 2. Ice cube 9 6.22 .83 5.44 1.01 3. Pencil sharpener 7 6.43 ‘ .79 6.62 .74 4. Empty movie screen 6 5.50 2.07 5.33 1.97 5. Bowl of water 11 5.91 .94 5.73 1.01 6. Piece of chalk 9 5.44 1.74 5.67 1.73 7. Banana peel 9 6.56 .53 6.11 .78 8. Empty blackboard 7 6.14 1.07 6.00 1.15 9. Pile of sand 9 5.78 1.64 5.56 1.88 10. Dotted line in road 10 5.50 1.78 5.40 1.71 11. Empty white paper 8 5.75 1.04 5.75 1.04 12. Man walking 8 6.00 1.31 6.00 1.07 13. White t-shirt 10 5.60 1.90 5.80 1.40 14. Empty ash tray 6 6.50 .84 5.83 1.47 15. Electric fan 8 5.63 1.41 5.75 1.28 *Appendix L contains a complete list of the images presented. 81 Table 3.12 Means and Standard Deviations of Image Vividness and Speed of Imagery Attainment for Each of the Possible Aversive Images Number of Vividness Speed Abbreviated Title Subjects of Image* Receiving Mean S.D. Mean S.D. Image 1. Polluted lake 7 5.00 1.15 4.57 1.40 2. Starving baby 10 5.80 .92 5.50 1.27 3. Cigar-filled room 8 5.50 1.60 5.25 1.39 4. Unprepared for exam 11 4.91 1.44 5.09 1.64 5. Town dump 8 5.88 .83 5.62 1.30 6. Slimy, puss-filled rash 10 5.10 1.66 5.50 1.58 7. Pitch-black cave 9 5.33 1.58 5.00 1.73 8. Dog with diarrhea S 5.80 1.64 6.00 1.00 9. Obesity 11 5.45 1.63 5.36 1.91 10. Awakened by prank calls 9 5.44 1.13 4.89 1.17 11. Car accident 8 5.12 1.64 5.12 1.64 12. Maggots in hamburger 7 6.14 .69 6.28 .76 £13. vomiting 7 5.14 1.34 4.57 1.90 .14. 50 job rejections 10 5.50 .71 5.50 1.35 15. Sexual advance by disliked individual 6 5.00 1.10 4.83 1.47 *Appendix L contains a complete list of the images presented. 82 least effective neutral image in both the vividness and speed of attainment dimensions. For the aversive images, "the maggot-filled hamburger" was the most vivid and most rapidly attained image. While "being unprepared for an important examination" was the least vivid aversive image, ”the polluted lake” and ”vomiting" were both equally slow in their speed of attainment. The second major area of descriptive data was obtained through the analysis of the various questions on the final evaluation from. The first question on this form asked, "What, in your opinion, was the probable purpose of this study?“ The frequencies and percentages of the groupings into which all responses were categorized are presented in Table 3.13. The largest number of subjects (41.27%) responded that they felt the study was about one's ability to empathize with others. This is, of course, the expla- nation which was given to subjects in the experimental instructions. Another 20.63% felt, perhaps, that the similarity of the situation to a TAT test might indicate that the study was one of a projective testing technique. Only two subjects expressed explanations closely related to the actual purpose of the study: One indicated that it was a study of the effect of self-observation; the other felt.it was an attempt to use reinforcing images to alter the content of the stories. The subject was asked in the second question in the final evaluation form to ”Briefly outline the types of 83 Table 3.13 Frequency and Percentage of Subjects Falling Within Each Category of Response on the Question: ”What, in Your Opinion, Was the Probable Purpose of This Study?" Probable Purpose N Percentage 1. Test of one's ability to identify and empathize with others 26 41.27 2. Projective test to gain knowledge of subject's thoughts and feelings 13 20.63 3. Test of one's honesty (Was the self- perception of empathy questionnaire honestly completed?) 7 11.11 4. Find out how one thinks other people feel 5 7.94 5. Test of how images will be included in the story themes 5 7.94 6. Test of creativity or expressiveness 3 4.76 7. No idea of probable purpose 2 3.18 8. Test of the effectiveness of self- observation 1 1.59 9. The effects of reinforcing images upon the content of the stories 1 1.59 84 thoughts or ideas that occurred to you as you participated in the experiment.” Here, the responses were so diversified that no summary categories seemed to be appropriate. However, some common thoughts expressed included: "How am I doing on my stories?” "How little creativity I have!" "I wonder what this experiment is about?" "It is inter- esting to see how much of my own personal experiences I include in my stories!“ ”How sad these pictures are!" "I'm becoming very bored!” ”I'm tired to telling stories with lots of feelings: they all sound the same!" "As the experiment went along, I became more and more comfortable and less self-conscious!” Although these examples do not cover the range of ideas and thoughts which were written in response to this question, these are representative of the majority of the responses. The third question on the final evaluation form was, ”Did you attempt to maximize the number of feeling/ emotional words in the stories that you related during the experiment?" The responses to this question are parti- cularly relevant to the self-regulation dimension of the design. Thus, a frequency and percentage breakdown for each self-regulation condition is contained in Table 3.14. More subjects in the SMSE condition (61.9%) than in either the NSM (33.3%) or SM (42.8%) groups attempted to maximize their emotional words. The x2 test of homogeneity, however, on these data indicated that these differences were not significant. 85 Table 3.14 Frequencies, Percentages, and Chi-Squared Tests for Last Three Questions of the Final- Evaluation Form on the Self- Regulation Condition Question NSM SM SMSE 3. Did you attempt Yes 7 33% 9 42.8% 13 61.9% to max EW? No 14 67% 12 57.2% 8 28.1% X2 = 5.991, 2 degrees of freedom, p > .05 not significant 4. Did you SM? Yes 1 4.8% 4 19.1% 8 28.1% No 20 95.2% 17 80.9% 13 61.9% 2 x = 7.22, 2 degrees of freedom, p < .05 significant 5. Did you SE? Yes 4 19.1% 12 57.2% 18 No 17 80.9% 9 42.8% 3 2 85.8% 14.2% x - 18.89, 2 degrees of freedom, p < .001 significant 86 The fourth and fifth questions on the final evaluation form asked, in effect, if the subject had self- monitored ("Did you keep track of the number of feeling/ emotional words?') or self-evaluated ("Did you compare the performance of emotional words on one story with another story?"). A frequency and percentage breakdown on each of these two questions for each self-regulation condition is contained in Table 3.14. On the self-monitoring question (number 4), only 19.1% of the SM group indicated that they had self-monitored, while only 28.1% of the SMSE treatment indicated self-monitoring. There were, however, significant differences between the three conditions on this question as measured by a x2 test at a .05 level of significance. The NSM group showed the lowest amount of self-monitoring with only one individual answering yes to that question. The responses to question 5, the question on self- evaluation, indicated very clear differences in the x2 analysis. The NSM group had only 19.1% self-evaluations: the SM had 57.2%; while the SMSE condition contained 85.8%. The differences between these frequencies on the x2 test of homogeneity was significant beyond the .001 level. Summary Evidence was found to indicate a significant relationship between the set of fourteen dependent vari- ables (number of expressed emotional words) and the two covariates. This supported the decision to employ a 87 multivariate analysis of covariance as one of the means of analyzing the design. It was also found, however, that the self-perception of empathy covariate did not add signifi- cantly to the regression equation. Only the number of emotional words expressed during the first story was significantly contributing to the relationship with the fourteen dependent variables. The MANCOVA and the repeated measures analysis of covariance were then reported. Hypothesis Three was not rejected. It stated: There will be no interaction between self-regulation and imagery variables as measured by the number of emotional words expressed. The MANCOVA and repeated measures ANCOVA indicated no interaction between the two major independent variables employed in the study. There was no support for Hypotheses One or Two provided by either the MANCOVA or repeated measures analysis. Hypothesis One stated: There will be differ- ences in the number of emotional words emitted between NSM, SM, and SMSE self-regulation conditions. Hypothesis Two stated: The subjects in the positive, neutral, and aversive imagery conditions will perform differentially in their output of emotional words. Neither the main effects for these treatment nor the measures by treatment interactions (in the repeated measures ANCOVA) were found significant. There were significant differences found between repeated measures which reflected differences in emotional *word output associated with the amount of time that the 88 subjects had been in the experimental setting. These differences had occurred across all treatment conditions. Two analyses that were not directly related to the major hypotheses of the study were briefly reported. There were significant differences found only between the neutral imagery and the aversive imagery conditions in the subject's mean vividness of imagery and mean speed of image attainment variables. In each case the neutral condition was higher: the subjects in the neutral group tended to form more vivid images more quickly. No significant differences were found between imagery treatments or self-regulation treatments in the number of images attained during the experiment. Finally, descriptive tables were presented which gave (a) the means and standard deviations of emotional words elicited in response to each specific cartoon card, (b) the means and standard deviations of vividness and speed of image attainment for every possible image presented in the study, and (c) frequencies and percentages of responses to the five questions on the final evaluation form. CHAPTER IV SUMMARY, DISCUSSION, AND CONCLUSIONS Summary The purpose of this study was to investigate the effects of covert behavior upon overt action. One type of covert activity, self-regulation, was incorporated into the first of two independent variables of major importance. The levels of this variable consisted of a no-self- monitoring condition, a self-monitoring condition, and a self-monitoring with self-evaluation condition. These three groups paralleled, to some extent, the model of self— regulation which Kanfer (1971) has proposed. If the model is appropriate, subjects in the self-monitoring with self- evaluation condition should exceed subjects in the self- monitoring group who, in turn, should surpass those in the no-self-monitoring condition on the dependent variable (number of emotional words expressed during TAT-like stories). A second covert behavior, visual imagery, was also built into the design as an independent variable. The three levels of this variable involved the administration 89 90 of positive, neutral, or aversive images contingent upon an actual or perceived increase in the number of emotional words expressed during one TAT-like story as compared to the immediately preceding story. The satisfaction of this contingency was decided by an experimenter tally in both the no—self-monitoring and self-monitoring conditions. In the self-monitoring with self-evaluation group, however, the subject's self-evaluation determined whether the con- tingency had been met. It was expected that the contingent administration of these covert visual images would differ- entially effect the emotional word output on subsequent stories. The positive imagery should have served very much as an external reinforcer and should have led to a larger number of emotional words than in the neutral imagery condition. For similar reasons, the aversive images should have suppressed responding in emotional words and resulted in fewer emotional words than either the neutral or the positive imagery treatments. These hypotheses are con- sistent with the recent theorizing of Cautela (1970a). Subjects in the study consisted of a sixty-three upper level and graduate students who had enrolled in courses in the Colleges of Education or Human Ecology in the Spring quarter of 1973 at Michigan State University. Subjects were assigned randomly to one of nine treatment conditions (3 levels of self-regulation X 3 levels of visual imagery). During the experimental session, each subject was asked to make up fifteen stories, each in response 91 to a cartoon stimulus card. An experimenter was present in the room to request, when appropriate to the contingencies or to the experimental condition, that the subject (a) read and imagine the scene presented on an imagery card, (b) stop imagining a scene after 30 seconds, (c) take and complete an imagery rating form, or (d) take and fill in a self-monitoring or self-evaluation form. There were two major analyses performed on these data. The first, a multivariate analysis of covariance, employed the number of emotional words in the first story and a self-perception of empathic ability questionnaire as covariates with the ratings of emotional words in the. remaining fourteen stories as the dependent variables. The second analysis, a repeated measures analysis of covariance, employed the basic design (3 x 3) with the addition of the emotional word ratings for stories 2-15 as the repeated measures dimension. Neither analysis yielded significant differences on the self-regulation and imagery variables or on the interaction between these independent variables and the repeated measures variable. Hence, no support for the major hypotheses of the study was found. Significant differences were found on the measures variable in the repeated measures analysis. Additionally, in a subsidiary analysis, significant differences were located between imagery conditions in the mean self-ratings of image vividness and speed of imagery attainment. 92 Neither of these findings, however, was of direct relevance to the primary hypotheses of the study. Discussion Although invalidity in the theoretical basis of this study could have resulted in its demonstrated lack of significance, there are also a number of possible alter- native explanations. It is the purpose of this discussion section to focus attention upon each potential explanation of the data and to integrate the results from the subsidiary and descriptive analyses to help "make sense" out of the total package of data collected. To make this integration easier, the discussion has been divided into five possible areas of explanation: theory, sampling, design and sta- tistics, instrumentation, and treatment. However, before exploring these sources of infor- mation, the basic orientation of the discussion should be established. As was mentioned earlier in the paper, Figures 3.1 and 3.2 illustrate the measures by treatment interaction curves for the self-regulation condition and imagery variables, respectively. Neither graph pictures a significant effect for either treatment or for their interaction with measures, but both demonstrate remarkably consistent trends. In the graph referring to the self-regulation variable (Figure 3.1), the NSM condition did not exceed the SMSE condition in its respective means on any other 93 than the second story (which is the first reported on the graph). The SM group fell below the NSM condition at only one story time. Likewise, the SM condition's means exceeded all but one corresponding mean for the NSM group. A very similar situation with respect to the imagery variable is presented in Figure 3.2. Consistently across time, the positive and neutral imagery groups' means on emotion words exceeded the aversive imagery condition's means. Furthermore, after some early fluctuation in treatment means, the positive group remained consistently higher than the neutral group in emotional word output (from the fifth story to the end of treatment). Perhaps, even these remarkable consistencies could be disregarded, however, were they not in precisely the directions predicted by the hypotheses of this study. These non-significant tendencies in the data, then, provide interesting material for speculation and discussion. Despite the lack of statistical support for the basic hypotheses, these graphs present promising indications. Given this outlook on the data as a basic orientation, each of the possible general causes of non-significant findings, which were listed earlier, will be briefly examined. Theory A possible cause of the outcome of this study might be an error in, or a misinterpretation of, the literature on self-regulation and covert visual imagery. The lack of 94 significance may simply result from the lack of validity of theoretical formulations. Consideration of possible theoretical questions will focus first on the self- regulation variable and secondly upon the visual imagery dimension of the study. Self-regulation. Kanfer's model (1971) of self- regulation may not be appropriate. A recent article by Glynn, Thomas, and Shee (1973) provided a conceptual basis for self-regulation which may, in fact, argue that Kanfer has made a theoretical error in his model. The model elaborated by Glynn (1973) involved a four step process of self-regulation rather than a three step model such as Kanfer has advocated. The four components described were (a) self-assessment: the individual must examine his or her own behavior to determine if a specific behavior has occurred (essentially similar to self-monitoring); (b) self-recording: the individual must objectively record the frequency of performance of the observed behavior (assumed to be part of Kanfer's self-monitoring, but conceptually distinct from it); (c) self-determination of reinforce- ment: the individual must decide what type of reinforce- ment should be delivered upon performance of a given behavior (not included in Kanfer's model, but, perhaps, assumed to be part of the self-evaluation component): and (d) self-administration of reinforcement: the individual must dispense his own reinforcement contingent upon 95 performance of the behavior (combining both the self- evaluation and self-reinforcement aspects of Kanfer's model). The model of Glynn, Thomas, and Shee, then, would provide a number of possible hypotheses for further investigation. For instance, by not having the subjects in this study objectively record their frequency of emotional words, the effectiveness of attempted self- monitoring may have been lost. Another hypothesis would involve the lack of self-determination of reinforcement in this study.' If (as Glynn believed), this aspect is crucial to self-regulation, there would be little surprise in finding no significant differences between self-regulation treatments. The type of ”reinforcement" was clearly not self-determined: it was decided long in advance of the experimental session! Briefly, a second theoretical issue within Kanfer's model may be raised relevant to Thoresen and Mahoney's (1972) recent theoretical article. Thoresen and Mahoney's view of self-control stresses the crucial need for an implementation of strategy to change a behavior. Inherently, the subjects of this study may have attempted some strategies to increase emotional words, but such strategies may have been very poorly developed. Perhaps, without a plan or strategy, self-monitoring and self-evaluation are little more effective than no direct attention to the behavior at all. This idea, however, carries little weight in view of the numerous studies cited earlier that demonstrated that 96 self-monitoring alone was as effective as many specific types of treatment programs. Imagery. Since few studies have been reported that point to the effectiveness of covert reinforcement and punishment through visual imagery (Cautela, Steffen, & Wish, 1972; Asher & Cautela, 1972; Manno & Marston, 1972; and Epstein & Peterson, 1973b), it may be possible that Cautela's ideas are not theoretically sound. This seems unlikely, however, in that each of the investigations referred to above were apparently very well-designed and executed. Even these few studies, then, argue strongly against a rejection of Cautela's ideas on a theoretical basis. Sample Poor samples may yield results which will not generalize because of particular traits which are not indicative of the population as a whole. A specific sample might result in non-significant differences for variables such as self-regulation or imagery which would not gener- alize to other samples. Such a situation mgy be present in these data since the "sample" was not drawn from a given population. Subjects were volunteers from a large number (but not, by any means, all classes) in the Colleges of Education and Human Ecology, and it may be hypothesized that such subjects could be very unique in their responses. Although this is a possibility, the review of the literature 97 provided no evidence of any potential subject characteristic which may interact with either the self-regulation or imagery treatments. One possible sample-related hypothesis, however, which could explain non-significant results is the instruction that the test was an empathy test and that the number of emotional words would indicate one's empathic ability might have created a ”ceiling effect“ among volunteer students in education and human ecology. Since "empathy" is a highly valued attribute, all subjects might have immediately hit their ceiling in the number of emotional words expressed and may have stayed nearly at that level thereafter (no matter what their treatment condition). Such an effect may well be present in the data reported here. Those slight trends indicated in Figures 3.1 and 3.2 could, perhaps, signify the relatively weak treatment effects masked by the stronger phenomena of a ceiling effect. Design and Statistical Treatment Findings of no significant differences may often be the product of an imprecise analysis of the data, most often resulting from a poorly planned design. In the case of an imprecise analysis, differences between treatments may be obscured by unexplained error in the data. The design used in this study, however, employed no obvious flaws. Although some factors of importance may 98 remain hidden in the randomization procedures, no variables of importance were knowlingly disregarded. Also, although the regression analysis on the covariate-dependent variable relationship indicated that the MANCOVA gave up one more degree of freedom for error than may have been necessary, it is not possible that this "unnecessarily" lost degree of freedom would have made any difference in the signifi- cance tests. As a final note in considering the design question, it might be pointed out that significant differ- ences were obtained on the repeated measures dimension of analysis. Since the design was powerful enough to detect these differences which were on the order of three to five points, one might argue that the analyses should have picked up significant differences on the treatment dimensions if true differences were present. Instruments A lack of validity and reliability of the measures employed in a reseach study may produce sufficient error variance, in and of themselves, to hide significant differ‘ ences. The important instruments in this study were direct ratings of a relatively straight forward criterion: expressed emotional words. Content validity for the ratings may be examined by reading the definition of emotional words included in the instrument section of this paper and by reference to Appendix D. The training of the raters for 99 the experiment was carefully accomplished in brief sessions over the period of several days. The inter-rater reliability of the two raters was found to be very high (r = .95) on those estimates which were averaged and employed in the analysis. Based on these data, it is unlikely that a lack of validity and reliability in the measuring instrument was the direct cause of the non-significant results found in the study. Treatments When treatments are poorly developed or sloppily administered, unexplained variation is introduced into the data. Such unexplained variance, of course, makes signifi- cant differences more difficult to obtain. Also, of concern in the treatment area, is that the expected treatment conditions may not have been present for one reason or another. There are several treatment concerns related to this study. While treatment conditions were neither poorly designed nor sloppily administered, there may be concerns about whether treatments were appropriate and, in some instances, whether the assigned treatments were actually carried out. As in the theory area, the present section will be subdivided to consider questions relating to the self-regulation and imagery variables separately. 100 Self-regulation. The final evaluation form raised several questions as to whether subjects were receiving the treatment to which they were assigned. Specifically, it was indicated (in Table 3.14) that 80.9% of the subjects in the SM condition were 222 ”trying to keep track of the number of feeling/emotional words included in each story." This was true deepite the fact that each of these subjects filled out a "story recall checklist” (Appendix E) after each story. The percentage of 61.9 in the SMSE condition who indicated no self-monitoring reflects the same trend. Could it be that subjects merely guessed randomly at their estimates of emotional words for the "story recall checklist" without making any direct attempt to self- monitor? There are two pieces of data which speak to this question. First of all, an analysis of variance estimate of reliability (Ebel, 1951) was calculated between subjects' ratings (estimates based on their responses to the "story recall checklist") and the average scores of the two post- session ratings of emotional words in each story. Of course, only subjects in the SM or SMSE conditions were included in this reliability since NSM subjects did not receive self-monitoring instructions. The reliability for individual ratings estimate obtained was .028 with a standard error of measurement of 16.98. A reliability of .028 indicates very little consistency between the two sets of ratings; they are clearly considerably less 101 consistent than were the experimenter's in-session and out- of-session ratings (as reported earlier, rxx = .702). The second piece of date relevant to the subjects' self-monitoring considers only subjects in the SMSE group. In this condition, participants were asked to fill out a ”story comparison form” on which they evaluated whether their most recent story had contained more emotional words than the preceding story. A type of validity estimate for this self-evaluation form was calculated by comparing the decisions made on the form with those that would have been made based on the out-of-session tallies of emotional words. For example, how often did the subjects indicate an increase or decrease in emotional words when, in fact, the emotional words had increased or decreased correspondingly? The overall percentage of correct decisions on these evaluations was 60.5%. This percentage indicates that the appropriateness of these decisions was not too much over a chance level. Again, this can be compared to the experi- menter's percentage of correct in-session decisions (81.3%) which was reported in Chapter II. The relatively low reliability and validity of the subjects' self-monitoring and self-evaluation ratings may be due to two sources (a) a lack of attempted self- monitoring (as the answers to question 4 on the final evaluation form indicate), or (b) the subjects' lack of a working definition of an emotional word. Either reason, or the combination of both, would lead subjects to become 102 somewhat confused and unable to self-monitor effectively. Further studies in this area must more clearly define the self-monitoring operation and the variable to be monitored. Perhaps, the crucial deficit of this study was the lack of explicitness in these two areas, for it must be remembered that the review of the literature stressed that without self-monitoring, there appears to be no self-control or self-regulation effects! The few who did carefully self- monitor may be those whose scores are reflected in the non- significant trends found in Figure 3.1. Another interesting finding relevant to possible error in the treatment administration is related to both the self-regulation and imagery dimensions of the study. In comparing the graphs of the treatment X measures inter- actions (Figures 3.1 and 3.2) with the graph of the mean number of emotional words expressed across time (Figure 3.3), it is found that the largest differences between treatment conditions correspond to the largest means reported in the measures figure. Although this tendency seems to be present for both self-regulation and imagery variables, it seems particularly clear on the imagery dimension. For example, the highest means in the measures graph (Figure 3.3) were found at story positions 6 and 7; the maximal differences between imagery conditions (Figure 3.2) were found at story positions 6 and 7. Additionally, the lowest means on the measures graph were found at story positions 2, 4, and 11; at these same poiitions, the differences between imagery 103 conditions were either minimal (2 and 4) or in an unpredicted direction (at story positions 4 and 11, the neutral imagery mean exceeded the positive imagery mean). It must con- tinually be stressed, however, that what is being examined here is a trend in the data which is ESE significant. A hypothesis to explain this trend might start with the assumption that the differences between times of story presentations (Figure 3.3) reflect a "concentration" or "motivation" effect. When the subject is concentrating or is motivated, the number of emotional words becomes greater. Expanding on this assumption leads to the following sequence of possible occurrences: As subjects came into the experimental situation, they were anxious and self- conscious (This idea is supported by the responses to question 2 in the final evaluation form.) and could, therefore, not concentrate completely on the task at hand. As time moved on, however, the subjects became comfortable and began to enjoy the task and its challenges (again, an often expressed idea in response to question 2). By story 8, however, the concentration and motivation had peaked and boredom and other unrelated, distracting thoughts (once more, a very prevalently expressed idea) reduced subjects' efforts toward the task-related behavior. Finally, at about the eleventh story, the subjects began to "see the end" of the experiment and, again, became more motivated and purposeful in their efforts. 104 This rather drawn-out hypothesis is particularly relevant in that the "motivation" and "concentration" that are referred to above may be exactly what Kanfer and Karoly (1972) have called ”the contract” or the commitment that one has to do what they have agreed to do. If the subjects were motivated or committed to "the contract (i.e., not bored, distracted by self-consciousness, or otherwise distressed)," they would, perhaps, tend to respond more closely to theoretical expectations than if they were not similarly committed. Relative to this point, future similar studies should, perhaps, employ a concealment of the length of the session and the number of picture cards to be presented. Also, stories could be of a longer duration relative to a smaller number of cartoon cards. Third, the above hypothe- sized "boredom" effect may exist only during the stories in the 8-11 positions. If such is the case, an extension of the number of cartoon pictures (and stories) might yield the expected significant differences. In summary, the question of subject commitment or motivation needs to be considered in planning further studies. Imagery. The treatment presented in the imagery condition was not completely consistent with Cautela's ideas, and, for this reason, it may not have been as effective an approach as would have been true otherwise. The alterations of Cautela's clinical approach to fit the 105 laboratory setting included: (a) There was no relaxation training such as Cautela employs to help subjects better imagine the visual scenes, (b) there was no concentrated practice in the forming of visual imagery, and (c) specific images were randomly administered rather than individualized as Cautela would advocate. Each of these steps in Cautela's typical clinical application were excluded from this study based on time and feasibility considerations. The research literature does present evidence indicating significant effects without relaxation training (Manno & Marston, 1972; Epstein 8 Peterson, 1973b) and without individualization of images (Manno & Marston, 1972), but there has been no laboratory study reported that employed covert reinforcement or covert punishment without some prior practice in imagery. Thus, it may be that this study failed to find significance because of a divergence from Cautela's stated procedures. It would definitely be recommended that further attempts to validate covert reinforcement or punishment at least employ (a) some attempted individualization of images (making sure that an image elicits the expected positive or aversive reaction in a particular subject), and (b) specific training in imagining visual scenes. A final treatment question arises from the results of a recent study by Wisocki (1973a) which found a signifi- cant relationship between the vividness of subjects' images and the effectiveness of a covert reinforcement treatment program in the reduction of test anxiety. The subsidiary 106 analysis of the vividness and speed of attainment means for each subject in the present study is directly related to Wisocki's finding. Perhaps, the fact that the aversive group's images were less vivid than were the neutral group's images indicates that the aversive treatment was not as effective in influencing emotional words as it might have been. These differences on vividness and speed of attain- ment may have resulted from the lack of individualization and practice which was stressed in the preceding paragraph. Conclusion Despite an absence of significant findings relative to the major dimensions of the study, the present investi- gation has generated a number of interesting hypotheses which may serve to further the exploration of the area of covert behavior. Among the specific suggestions offered to future researchers are (a) select a dependent variable which is not likely to be so easily influenced as to produce a "ceiling effect," (b) make the definition of the dependent variable operationally clear to the subjects, (c) attempt to make sure that subject motivation and con- centration remain high during the study, (d) in the self- regulation dimension, tell the subjects what self-monitoring involves and influence them to actually record their behavior, and (e) relative to the imagery variable, incor- porate a procedure to allow for individualization of 107 imaginal scenes and practice of the process of forming visual images. BIBLIOGRAPHY BIBLIOGRAPHY Asher, L. M., & Cautela, J. R. Covert negative reinforce- ment: An experimental test. Journal of Behavior 222E221.and Experimental Psychiatry, 1972, 3, 1-5 Bandura, A. Princi 1es of behavior modification. New York: Holt, Rinehart, and Winston, Inc., I969. Bandura, A. Vicarious and self-reinforcement processes. In R. Glaser (Ed.), The nature of reinforcement. New York: Academic Press, , pp. - . Bandura, A., Grusec, J. E., & Menlcve, F. L. Some social determinants of self-monitoring reinforcement systems. Journal of Personality and Social PBXChOIOS! ' l I - ' Bandura, A., & Kupers, C. J. Transmission of patterns of self-reinforcement through modeling. Journal of Abnormal and Social Psychology, 1964, E2, [-9. Bandura, A., & Whalen, C. K. The influence of antecedent reinforcement and divergent modeling cues on patterns of self-reward. Journal of Personalit and Social Psychology, 1966,‘§, 373-382. Barlow, D. H., Leitenberg, H., & Agras, W. S. Experimental control of sexual deviation through manipulation of the noxious scene in covert sensitization. Journal of Abnormal Psychology, 1969, 2, 596-601. Beck, A. T. Cognitive theory: Nature and relation to behavior therapy. Behavior Therapy, 1970, l, 184-200. Ben, S. L. Verbal self-control: the establishment of effective self-instruction. Journal of Experimental Psycholggy, 1967, 131 485-491. 108 109 Beneke, W. M., & Harris, M. B. Teaching self-control of study behavior. Behaviour Research and Thergpy, 1972' 3:2, 35.41. Bergin, Allen E. Cognitive therapy and behavior therapy: Foci for a multidimensional approach to treatment. Behavior Therapy, 1970, 1, 205-212. Bolstad, O. D., a Johnson, S. M. Self-regulation in the modification of disruptive classroom behavior. Journal of Applied Behavior Analysis, 1972, 5, 443-454. Broden, M., Hall, R. V., & Mitts, B. The effect of self- recording on the classroom behavior of two eighth grade students. Journal of Applied Behavior “312818, 1971' A. - e Brookover, W. B., LePere, J. M., Hamachek, D. B., Thomas, 8., & Erickson, E. L. Self-conce t of ability and school achievement II: Im rovin academic achieve- ment throu 5 students' self-cance t enhancement. East LansIng, MiEHIgan: EducatIonaI PEEIIcatIon Services, 1966. Brookover, W. B., Paterson, A., 8 Thomas, S. Self-cance ts of ability and school achievement. East LansIng, 1 gan: O ce 0 Resear an Publications, Michigan State University, 1962. Cautela, J. R. Treatment of compulsive behavior by covert sensitization. Psychological Record, 1966, $3, 33-41e Cautela, J. R. Covert sensitization. Psychological Reports, 1967' 223 ‘59-‘68. Cautela, J. R. Behavior therapy and self-control: Techr niques and implications. In C. M. Franks (Ed.), Behavior thera : A raisal and status. New York: McGraw-HIII, 1559, pp. 323-345. Cautela, J. R. Covert reinforcement. Behavior TherapyI 1970a,'r, 33-50. Cautela, J. R. Covert negative reinforcement. Journal of Behavior Therapy and Erperimental Psychiatry, I I - e Cautela, J. R. Covert conditioning. In A. Jacobs, 8 L. B. Sachs (Eds.), The s cholo of rivate events. New York: Academic Press, I9gla, pp. We . 110 Cautela, J. R. Covert extinction. Behavior Therapy, 1971b, 2, 192-200. Cautela, J. R. Rationale and procedures for covert con- ditioning. Advances in behavior therapy. New York: Academic Press, I972, pp. 85-96. Cautela, J. R., Steffen, J., & Wish, P. An experimental test of covert reinforcement. Journal of Con- sulting and Clinical Psycholqu, 1972, in press. Cautela, J. R., Walsh, H., 8 Wish, P. The use of covert reinforcement in the modification of attitudes toward the mentally retarded. Journal of Psy- chology, 1971, 11, 257-260. Cautela, J. R., & Wisocki, P. A. The use of imagery in the modification of attitudes toward the elderly: A preliminary report. The Journal of Psychology, Coopersmith, S. A method for determining self-esteem. Journal of Abnormal and Social Psychology, 1959, p "' e Cornfield, J., & Tukey, J. W. Average values of mean squares in factorials. Annals of Mathematical Statistics, 1956, 31, 9077915. Curtis, R. H., & Presly, A. S. The extinction of homosexual behavior by covert sensitization: A case study. Behavior Research and Therapy, 1972, $2! 81-83. Davison, G. C. Elimination of a sadistic fantasy by a client-controlled counter-conditioning technique: A case study. Journal of Abnormal Psychology, 1968, 13, 84-90. Davison, G. C., 5 Wilson, G. T. Process of fear-reduction in systematic desensitization: Cognitive and social reinforcement factors in humans. Behavior Therapy, 1973, 3, 1-21. Dorsey, T. E., Kanfer, F. H., & Duerfeldt, P. R. Task difficulty and non-contingent reinforcement schedules as factors in self-reinforcement. Journal of General Psychology, 1971, 21, 323-334. Ebel, R. L. Estimation of the reliability of ratings. Psychometrika, 1951, lg, 407-424. lll Ekman, P., Krasner, L., & Ullmann, L. P. Interaction of set and awareness as determinants of response to verbal conditioning. Journal of Abnormal and Social Psycholggy, 1963, 55, 387-389. Ellis, A. Reason and emotion in psychotherapy, New York: Lyle Stuart, 1962. Ellis, A., 8 Harper, R. A. A guide to rational livin . Englewood Cliffs, N.J.: Prentice-Hall, 1961. Elson, S. E. Behavioral self-control: A survey of the literature. Unpublished manuscript, Michigan State University, 1972. Epstein, L. H., & Peterson, G. L. The control of undesired behavior by self-imposed contingencies. Behavior Theraezp 1973a, 5., 91-95. Epstein, L. H., & Peterson, G. L. Differential conditioning using covert stimuli. Behavior Therspy, 1973b, 4' 96-99e Ferster, C. B., Nurnberger, J. I., & Levitt, E. B. The control of eating, 1962. Cited in Cautela, J. R. Behavior therapy and self-control: Techniques and implications. In C. M. Franks (Ed.), Behavior therapy: Appraisal and status. New York: McGraw- H1 I I pp. - e Finn, J. D. Multivariance . . . Univariate and multivariate analysis of varianceLcovarianceyandre ression. §:fortran IVprogramI version 4. Buffalo, N.Y.: State university 0 New Yor at Buffalo, 1968. Mimeograph. Flannery, R. B. Use of covert conditioning in the be- havioral treatment of a drug-dependent college dropout. Journal of Counseling Psychology, 1972, 13, 547-550. Gewirtz, J. L. The roles of overt responding and extrinsic reinforcement in "self-" and "vicarious-reinforce- ment" phenomena and in ”observational learning” and imitation. In R. Glaser (Ed.), The nature of reinforcement. New York: Academic Press, I971, PP- 2’5’555- Glynn, E. L., Thomas J. D., & Shee, S. M. Behavioral self- control of on-task behavior in an elementary class- room. Journal of Applied Behavior Analysis, 1973, g, 105-‘1'1'3. 112 Goldiamond, I. Self-control procedures in personal behavior problems. Psychological Reports, 19654.11! 851-868. Herbert, E. W., & Baer, D. M. Training parents as behavior modifiers: Self—recording of contingent attention. Journal of Applied Behavior Analysis, 1972, 2, Homme, L. E. Perspectives in psychology: XXIV. Control of coverants, the operants of the mind. Psychologi- cal Record, 1965, 13! 501-511. Horan, J. J., a Johnson, R. G. Coverant conditioning through a self-management application of the Premack principle: Its effect on weight reduction. Journal of Behavior Thera and Erperimental Psy- Chfatry p I _I - ' Jacobs, A., & Wolpin, M. A second look at systematic desensitivation. In A. Jacobs, 8 L. B. Sachs (Eds.), The s cholo of rivate events. New York: AcademIc Press, I971, pp. 78-198. Jennrich, R. I. 1604 Analysis of variance. Madison, Wisconsin: University of Wisconsin Numerical Analysis Department Laboratory Bulletin, May 15, 1961. (Mimeographed.) Johnson, S. M. Effects of self-reinforcement and external reinforcement in behavior modification. Proceedingg of the 77th Annual Conventiony APA, 1969, 1, Johnson, 8. M., a White, G. Self-observation as an agent of behavior change. Behavior Therapy, 1971, 3' 488-497. Johnson, W. G. Some applications of Homme's coverant control therapy: Two case reports. Behavior Kanfer, F. H. Influence of age and incentive on children's self-rewards. Psychological Reports, 1966, 12! 263-274. ‘ Kanfer, F. H. Self-regulation: Research, issues, and speculations: In C. Neuringer, & J. L. Michaela (Eds.), Behavior modification in clinical s - cholo . New York: Appleton-Century-Crofts, I979a. Kanfer, Kanfer, Kanfer, Kanfer, Kanfer, Kanfer, Kanfer, Kanfer, Kanfer, Kanfer, Kanfer, Keutzer ' 113 F. H. Self-monitoring: Methodological limitations and clinical applications. Journal of Consultin and Clinical Psychology, 19705, 25, 148-152. F. H. The maintenance of behavior by self- generated stimuli and reinforcement. In A. Jacobs, 8 L. B. Sachs (Eds.), The s cholo of rivate events. New York: Academic Press, 1971, pp. 39-60. F. H., Bradley, M. M., & Marston, A. R. Self- reinforcement as a function of degree of learning. Psychological Reports, 19524.12! 885-886. F. H., & Duerfeldt, P. H. Motivational properties of self-reinforcement. Percoptual and Motor Skills, 1967a, 23, 237-246. F. H., & Duerfeldt, P. H. Effects of pre-training on self-evaluation and self-reinforcement. Journal of Personality and Social Psychology, 1967b, _, F. H., & Duerfeldt, P. H. Comparison of self- reward and self-criticism as a function of types of prior external reinforcement. Journal of Per- sonality and Social Psychology, 1968,'§, 261-268. F. H., Duerfeldt, P. H., & LePage, A. L. Stability of patterns of self-reinforcement. Psychological Reports, 1969, 31, 663-670. F. H., & Haroly, P. Self-control: A behavioristic excursion into the lion's den. Behavior Therapy, 1972, s, 398-416. F. H., & Marston, A. R. Conditioning of self- reinforcing responses: An analogue to self- confidence training. Psychological Reports, 1963a, 1_3_, 63-70. F. H., & Marston, A. R. Determinants of self- reinforcement in human learning. Journal of Experi- mental Psycholgy, 1963b, gs, 245-754. F. H., a Phillips, J. S. Learnin foundations of behavior therapy. New York: John WIIey and Sons, nc., a, pp. 407-454. C. S. Behavior modification of smoking: The experimental investigation of diverse techniques. Behavior Research and Therapy, 1968, 3, 137-157. 114 Kirk, R. E. Experimental design: Procedures for the behavioral sciences. Belmont, Ca1ifornia: roo s o e, . Kolvin, I. ”Aversive Imagery" treatment in adolescents. Behaviour Research and Therapy, 1967, 5, 245-248. Lawson, D. M., a May, R. B. Three procedures for the extinction of smoking behavior. Psyohological Record, 1970, 33, 151-157. Leitenberg, H., Agras, W., Barlow, D., & Oliveau, D. Con- tribution of selected positive reinforcement and therapeutic instructions to systematic desensiti- zation therapy. Journal of Abnormal Psychology, 1969, 11, 113-118. Mahoney, M. J. Toward an experimental analysis of coverant control. Behavior Thergpy, 1970, l, 510-521. Mahoney, M. J. The self-management of covert behavior: A case study. Behavior Therapy, 197l,'g, 575-578. Mahoney, M. J. Research issues in self-management. Behavior Therapy, 1972, 2, 45-63. Mahoney, M. J., Moura, N. G. M., & Wade, T. C. Relative efficacy of self-reward, self-punishment, and self- monitoring techniques for weight loss. Journal of Consulting and Clinical Psycholggy, 1973, 39, Mahoney, M. J., & Thoresen, C. E. Behavioral self-control: Power to the person. Educational Researcher, 1972, 1, 5-7. Mahoney, M. J., Thoresen, C. E., & Danaher, B. G. Covert behavior modification: An experimental analogue. Journal of Behavior Therapy and Experimental 531611131237, 9 _' " e Manno, B., & Marston, A. R. Weight reduction as a function of negative covert reinforcement (sensitization) versus positive covert reinforcement. Behaviour Research and Therapy, 1972, $2, 201-207. Marston, A. R. Variables affecting incidence of self- reinforcement. Psychological Reports, 1964, 11, 879-884. 115 Marston, A. R. Self-reinforcement and external reinforce- ment in visual-motor learning. Journal of Experi- mental Psychology, 1967, 11, 93-512 Marston, A. R. Effect of external feedback on the rate of positive self-reinforcement. Journal of E eri- mental Psychology, 1969, pp, 175-179. Marston, A. R., 8 Cohen, N. J. The relationship of negative self-reinforcement to frustration and intropunitive- ness. Journal of General Psychology, 1966, 11, 237-243. McPall, R. M. The effects of self-monitoring on normal smoking behavior. Journal of Consultin and Clinical Psycholggy, I975, :5, 155-115. McFall, R. M., & Hammen, C. L. Motivation, structure, and self-monitoring: Role of nonspecific factors in smoking reduction. Journal of Consultin and Clinical Psycholggy, I97I,‘§Z, 85-83. Meichenbaum, D., & Goodman, J. The developmental control of operant motor responding by verbal operants. Journal of Egpgrimental Child Psychology, l969,{1, Meichenbaum, D., & Goodman, J. Training impulsive children to talk to themselves: A means of developing self- control. Journal of Abnormal Psychology, 1971, 11, 115-126. Murray, E. J., & Jacobson, L. I. The nature of learning in traditional and behavioral psychotherapy. In A. E. Bergin, & S. L. Garfield (Eds.), Handbook of psychotherapy and behavior change. New York: By, I PP. - e Nawas, M. M. Wherefore cognitive therapy? A critical scrutiny of three papers by Beck, Bergin, and Ullmann. Behavior Thergpy, 1970,‘l, 359-370. Oliveau, D., Agras, W., Leitenberg, H., Moore, R., 3 Wright, D. Systematic desensitization, therapeuti- cally oriented instructions, and selective positive reinforcement. Behavior Research and Therapy, 1959!.10 27-34. Orne, M. T. From the subject's point of view, when is behavior private and when is it public: Problems of inference. Journal of Consultin and Clinical Psycholggy, 1975,‘§§, 145-147. 116 Premack, D. Reinforcement theory. In D. Levine (Ed.), Nebraska symposium on motivation. Lincoln: Uni- versity 0 Ne ras a Press, , pp. 123-180. Premack, D. Catching up with common sense or two sides of a generalization: Reinforcement and punishment. In R. Glaser (Ed.), The nature of reinforcement. New York: Academic Press, I971, pp. 121-155. Portes, A. On the emergence of behavior therapy in modern society. Journal of Consultin and Clinical Psychology, 1971, E, 115-147. Rehm, L. P., & Marston, A. R. Reduction of social anxiety through modification of self-reinforcement: An instigation therapy technique. Journal of Con- Rutner, I. T., & Bugle, C. An experimental procedure for the modification of psychotic behavior. Journal of Consulting and Clinical Psychology, 1969, 2!, Shaw, M. C., Edson, R., a Bell, H. M. The self-concept of bright underachieving high school students as revealed by an adjective check list. Personnel and Guidance Journal, 1960, 32, 193-196. Simkins, L. The reliability of self-recorded behaviors. Behavior Therapy, 1971,‘g, 83-87. Skinner, B. F. ‘Science and human behavior. New York: Macmillan, . Thomas, E. S., Abrams, K. S., a Johnson, J. B. Self- monitoring and reciprocal inhibition in the modifi- cation of multiple tics of Giles de la Tourette's syndrome. Journal of Behavior Therapy and Experi- mental Psyc a ry, , , - . Thoresen, C. E. Behavioral humanism. In C. E. Thoresen (Ed.), Behavior modification in education. Chicago, Illinois: The finiversity of Chicago Press, 1973, Todd, P. J. Coverant control of self-evaluative responses in the treatment of depression: A new use for an old principle. Behavior Therapy, 1972,'3, 91-94. Tooley, J. T., & Pratt, S. An experimental procedure for the extinction of smoking behavior. Psycholggical Record, 1967,‘11, 209-218. 117 Tyler, V. 0., & Straughan, J. H. Coverant control and breath holding as techniques for the treatment of obesity. Psychological Record, 1970, 32, 473-478. Ullmann, L. P. On cognitions and behavior therapy. ’ Behavior Therapy, 1970, 1, 201-204. Ullmann, L. P., & McFarland, R. L. Productivity as a ' variable in T A T protocols: A methodological study. Journal of Prcjective Techniques, 1975,‘gl, 80387. Watson, D. L., 6 Tharp, R. G. Self-directed behavior: Self-modification for ersonaI ad'ustment. Monterey, California: 5rooks7CoEe, 1972. Watson, J. B. Ps cholo from the stand oint of a behaviorist. Failaaelpfiia: Lippincott, 1919. Weiss, R. L., Krasner,iL., & Ullmann, L. P. Responsivity to verbal conditioning as a function of emotional atmosphere and pattern of reinforcement. Psy- chological Reports, 1960,'§, 415-426. Wisocki, P. A. A covert reinforcement program for the treatment of test anxiety: A brief report. Behavior Therapy, 1973a,'£, 264-266. Wisocki, P. A. The successful treatment of a heroin addict by covert conditioning techniques. Journal of Behavior Therapy and nggrimental PsyEfiiatEy, r o " o Wolpe, J. Ps chothera b reci rocal inhibition. Stan or , a ornia: tan or On vers ty Press, 1958. Woy, J. R., & Efran, J. S. Systematic desensitization and expectancy in the treatment of speaking anxiety. Behaviour Research and Therapy, 1972, 12, 43-49. APPENDICES APPENDIX A TYPICAL PRESENTATION EMPLOYED TO SOLICIT SUBJECTS APPENDIX A TYPICAL PRESENTATION EMPLOYED TO SOLICIT SUBJECTS When I spoke briefly to Dr. _____ on Monday morning, he indicated that possibly a few of you might be interested in volunteering to participate in a research study that I am presently running for my dissertation. Unfortunately, I cannot tell you too much about the study in advance. This limitation results from the fact that a prior knowledge of the study would undoubtably bias your scores during the experiment. What I can tell you is: (l) the task involved is of a verbal nature; (2) it is designed to be non-threatening in nature; (3) it is not dangerous; (4) it is completely ethical; (5) the task would only take one and one-half hour of your time: and, finally, (6) your results will be handled with anonymity. Following your participation in the study, I be glad to explain, in detail, the purpose of the experi- ment and what I had hoped to find in terms of results. In addition, I'd be pleased to send you a copy of an abstract of the final results sometime late in the summer. 118 119 I do need subjects badly, but please do not feel that you are in anyway obligated to sign up. Dr. has allowed me to circulate this note, but I'm sure he wishes to put no pressure on you to sign. All appointments will be in Room 250 Erickson Hall. The sign-up sheet is attached. As you'll notice, there are possible times available during most hours of the day for the remainder of this week and all of next week. For those of you who do participate, I will send you a post card reminder about three days prior to your appointment. If you wish to participate and cannot find a convenient time, please sign your name on the back of the last page (include your phone number). Later in in the week, I'll call to arrange a mutually satisfactory time. Thank you very much! APPENDIX B SELF-EVALUATION QUESTIONNAIRE APPENDIX B SELF-EVALUATION QUESTIONNAIRE Subject Number Please mark each statement in the column which most closely refers to the extent of your agreement with the statement given. There are no right or wrong answers, and you are encouraged to respond with your initial impressions. Please be as honest as possible in your answers and make sure that you mark a response for each statement. The column abbreviations refer to the following key: VSD--Very Strongly Disagree SD--Strongly Disagree D--Disagree A--Agree SA--Strongly Agree VSA--Very Strongly Agree VSD SD D A SA VSA l. I spend a lot of time day-dreaming. 2. I'm pretty sure of myself. 3. I often wish I were someone else. 4. I'm easy to like. 5. I never worry about anything. 6. I find it very hard to talk in front of a class. 7. I am a person who is comfortable with emotions. 8. There are lots of things about myself I'd change if I could. 120 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 121 I'm a good friend to have when someone's in trouble. VSD SD SAT vsa I get upset easily at home. I always do the right thing. I am proud of my ability to communicate with others. I know what to do without being told. I'm a lot of fun to be with. I'm often sorry for the things I do. I'm.popular with my peers. Other people will often attend to my feelings. I'm never unhappy. I hear what others say through their actions as well through their words. I can feel my body change when I'm emotional. I'm pretty happy. People expect too much of me. I like everyone I know. I argue a lot. I understand myself. Iris pretty tough to be me. Things are all mixed up in my life. I never make major errors in judgment. It's difficult for me to speak about feelings. I have a low opinion of myself. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 122 I don't like to be with other people. SD SA I'm never shy. When people are worried, I can usually tell it. I often feel ashamed of myself. If someone I knew was depressed, I'd ask what was wrong. If I have something to say, I usually say it. I always tell the truth. I try to be helpful to others. I don't care what happens to me. I try to be honest about my feelings. I get upset easily when I'm reprimanded. Most people are better liked than I am. I always know what to say to people. I talk to people a lot about their deeper feelings. Thinks usually don't bother me. I can't be depended on. APPENDIX C SELF-EVALUATION QUESTIONNAIRE WEIGHTINGS FOR EACH ITEM AND LIE SCALE INDICATED 10. 11. 12. 13. APPENDIX C SELF-EVALUATION QUESTIONNAIRE WEIGHTINGS FOR EACH ITEM AND LIE SCALE INDICATED dreaming. else. VSD SD SA VSA I spend a lot of time day- 6 5 2 1 I'm pretty sure of myself. 1 2 5 6 I often wish I were someone 6 5 2 1 I'm easy to like. 1 2 5 6 I never worry about anything. 6 5 2 l I find it very hard to talk in front of a class. 6 5 2 1 I am a person who is comfortabl+ with emotions. 1 2 5 6 There are lots of things about myself I'd change if I could. 6 5 2 1 I'm a good friend to have when someone's in trouble. 1 2 5 6 I get upset easily at home. 6 5 2 l I always do the right thing. 6 5 2 1 I am proud of my ability to communicate with others. 1 2 5 6 I know what to do without 1 2 5 6 being told. 123 Lie Lie APPENDIX D EMOTIONAL WORD GUIDE APPENDIX D EMOTIONAL WORLD GUIDE Do Not Count as an Count as an Emotional Word W "feel” or "feeling“ "remember" (in any context) ”think" "sick" or "sickness" "consider" {Eififiiiida‘éfifiii ”decide" "couldn't” "plan” "shouldn't" “nice" "have to” “really" "had to" 25;:{2 (When £95 modified) "supposed to" "save" ”die" or "death" "experience" "beating" "opportunity" "won" "usual" "losing" "necessary" "doing well" ”possibly" "all by himself" "maybe" "alone" "familiar" "lonely" "shaking" "trYi“9" "boyfriend” or "girlfriend" "lost" “especially" [names of diseases] "friend" (when not modified) 126 APPENDIX E THREE SAMPLE STORIES WITH EMOTIONAL WORDS UNDERLINED APPENDIX E THREE SAMPLE STORIES WITH EMOTIONAL WORDS UNDERLINED Stopy 1 Picture Card Stimulus: Grandmother l ”Why,ywhy," she thought, "is this happening to me? 2 I never thought it would. I've always been so full of life 3 4 and so busy. Why is it happening to me?” As she sat there in her chair by the window watching 5 the world outside, she thought over her life and she felt 6 7 vegy sad that things had happened as they had. She EEEE very lonesome as she sat there and watched the children playing outside and yet nobody came to her door to play with her or to talk to her. She'd never thought that this would happen to her. Her life had been a veryfall one. She'd been pégy_as a youngster. She'd been active as an adult. She used to be on all of the committees and play in all the sports and go on all the picnics and do everything. And 12 now here she was, sitting here by her window, all alone. 127 128 13 Why was it? Could anybody explain why old people 14 15 were left alone when they got old? Her friends had died as she got old. Her children had moved away. They were a long ways from her and they didn't call very often anymore because she kept telling them how to raise their children and they didn}: like that. But it seemed like with all her years of experience and wisdom, that she ppght t67be able to have something to offer them. She didn't mean to inteifere, just to Eéip them. But they didn't see it that way. They saw it as an interferingoold lagy and they called less and less. 21 And now she was getting so it was hard to move. It 22 was hard to walk to the store and so she didn't go very 23 often. It was hard toygo out to the neighbor's for a cup 24 of coffee, so she didn't go very often. It was hard to make coffee or to make bread anymore or to make coffeecake, and so she didn't have neighbors in very often. So, here 25 2 . she sat, alone. She couldn't understand why it was that everyone had gone. 129 Story 2 Picture Card Stimulus: Cake Today was a very :pecial day. Mother had been telling John about the little girl that was going to be coming to live with them. Mother explained that the little girl's parents were not agle to take3care of her and that she had agreed to take the little girl in to stay with the family until the parents could take the little girl back to 4 live with them. So, for this very special occasion, John 5 had insisted that mother bake a cake and he asked, of 6 course, for his very favorite, banana cake with chocolate frosting. Mother had, indeed, baked the cake. 7 The little girl arrived. John sized her_gp and 8 thought, "Well, she's 0.x. for a girl. I probably can get 9 along_withh%r all right.” He could see that the little 1 . ll 12 girl was feeling very uncomfortable and verystrangg and 13 frightened. So he invited her into the kitchen and showed 14 her the surprise that his mother had baked. He asked his mother if he could cut her a piece and mother said, "Oh 15 certainly you can." So he went to the drawer and got out a knife and cut a piece of cake and offered it to the little 16 girl who veEy eagerly picked it up in her hands and began to eat it. John cut himself a piece and although no words were exchanged between the two of them, he knew that she 17 18 19 was feeling much better about being there and be innin to 20 II feel a part of the family and John was feeling good about 130 22 having her there, too. This would give him a chance to really be a . . . Story 3 Picture Card Stimulus: Birds I see this as a trip on a boat. The person has binoculars and is looking at the birds that are flocking around, coming, flying at the back of the boat. The mountains are off in the distance from the boat. I see 1 this as one of the lovliest experiences a person could have and this happens to be we're taking a boat trip across the Atlantic . . . and one of the nice . . . it was a change between vocations and there was noprissure from the last one and noprzssure from the next one. One of those ggigp, hgzpy times that come so ragely it seems in life and what a really rewaZding and enjoyable 2xperience it was to observe the limited . . . things from the environment as far as the 9 10 distractions is the word I want. The limited number of ’11 12 13 distractions. It was a Jery meditative, relaxing time; one of observing birds. It makes me, reminds me of the story of Jonathan Livingston Seagull and how it really did . . . I don't know: I think that an experience like a boat trip is something that if evegybogy iguld have one, it would 21:32 them see themselves and £122 them really. . . . It gives such a géim.feiging.. At least our experience on this particular boat at that particular time did in that the 19 seas were so calm and everything was sort of like an 131 20 interlude in life before the next step and how every little 2] thing that comes around is so noticed. . . . The flourescence 2 in the water and the fish jcmping up and the birds and the beigty and the waves, the different color of the water. It's a veryggeply emotionalzzxperience. It was at least for us. It was sort of like symbolic of :6projection from one stage of life to another or from one experience to another. APPENDIX E SUBJECT CONSENT FORM APPENDIX F SUBJECT CONSENT FORM I) 10' __ - I have agreed to participate in the present research study of my own free will. I understand that I am free to discontinue my participation in the project at any time. I understand that the data collected in this project will be treated with strict confidence, and any data collected based on my performance will be used completely anonymously. I understand that any aspects of the project that are not explained before it is carried out will be explained, upon request, at its completion. I realize that I am expected to refrain from discussing the experiment and my participation in it until May 19, 1973. Signed If you wish to receive an abstract of the results of the study, please fill in your summer address below. 132 APPENDIX G DEMOGRAPHIC DATA SHEET APPENDIX G DEMOGRAPHIC DATA SHEET Sex: Female Male Age: Under 20 41 - 45 21 - 25 26 - 30 31 - 35 36 - 40 51 55 56 + College: Education Human Ecology Other (Please specify) Specialization area or major: Degree program for which you are enrolled: Bachelor's Master's Doctoral Other (Please specify) 133 APPENDIX H CARTOON STIMULUS PICTURES 135 136 H. , m” H H i n .0-" . :mm \ | -‘_ 9‘... 138 139 140 141 142 as... Or eOeoo \ 0.90. 143 144 146 147 1- MM*——- --_—- ~..1..- __ i A - 148 APPENDIX I STORY RECALL CHECKLIST APPENDIX I STORY RECALL CHECKLIST STORY NUMBER Think back over the story you have just related. Try to remember all of the components that you included in that story, but, in par- ticular, think about the number of feelings that you talked about. Indicate below your best estimate of the number of times that you expressed an emotional word in the story you've just completed. 1 - 3 16-18 31-33 4 - 6 19-21 34-36 7 - 9 22-24 37-39 10-12 25-27 40-42 13-15 28-30 43 + 149 APPENDIX J STORY COMPARISON FORM APPENDIX J STORY COMPARISON FORM STORY NUMBER Were you able to incorporate more feelings or fewer feelings into the story you've just completed? When compared to the story immediately preceding it, was this story greater in the number of emotional words expressed, about the same, or smaller in the number of emotional words expressed? Greater About the same Smaller 150 APPENDIX K IMAGERY RATING FORM APPENDIX K IMAGERY RATING FORM Rate the imagery scene that you have just pictured by circling the appropriate numbers below. CLARITY OF THE IMAGE 1 2 3 4 5 6 7 Almost Very Poor Nut Clear Very Vivid no poor clear clear image or poor SPEED OF IMAGE ATTAINMENT 1 2 3 4 5 6 7 No Very Slowly Not Rapidly Very Immed- Image slowly rapidly rapidly iate 151 APPENDIX L COMPLETE LIST OF IMAGERY SUGGESTIONS APPENDIX L COMPLETE LIST OF IMAGERY SUGGESTIONS Positive Imaggg Imagine that you are sitting by the fireplace with your loved one. Music is playing softly on the stereo. Your loved one gives you a tender glance and you respond with a kiss. You think to yourself how wonderful it is to care for someone and to have someone care for you. Imagine that you have been unexpectedly invited to visit some really great friends at a summer home in the mountains. You are now arriving for a worry-free weekend. Imagine that you have been a bit sad and depressed recently, but that you've just received a phone call from an old friend whom you haven't seen in a long time. This unexpected surprise has turned the tone of your day com- pletely around! Imagine that you have just completed a difficult job. Your superior comes by and praises you highly for "a job well done.” He also makes it clear that such good work is going to be rewarded very soon. 152 153 Imagine that you have just been named a grand winner in the Michigan State Lottery. How might you spend all this money? Imagine that you can make yourself invisible when- ever you wish. Think about how you could use this power. Imagine a warm summer evening by a mountain lake. You are watching a spectacular orange-red sunset as it lights up the sky and reflects on the lake. Imagine that you are at a lively party. Somebody walks across the room to you, smiles in a friendly way, and says, ”I'm glad to meet you. I've heard so many good things about you. Do you have a moment to talk?” Imagine that you are given free access to any food or merchandise in a large supermarket for a period of fifteen minutes. You have one shopping cart and you must take everything that you want to a certain designated area within the specified time limit. Imagine that you are walking along a mountain pathway with your dog by your side. You notice attractive lakes, streams, flowers, and trees. You think to yourself, "It's great to be alive on a day like this!" Imagine that you are lying on the soft warm sand of a sunny beach. Feel the sun soak into your body. 154 Imagine that an admired prof has just commended you for an outstanding performance on a very difficult test. You had worked very hard preparing for this examination. Imagine that you are eating the biggest, most delicious ice cream sundae you've ever seen (and you're not getting full!). Imagine that you are floating, relaxed above the clouds. You are aware of the fluffy white clouds below you and the warm sun and blue sky above. Imagine that it is a bright, warm summer morning and you are on vacation. You can do anything you may wish for the entire day! Imagine that you've worked a long hard day and that your muscles have been very tired and sore. Now, at the end of the day, you are soothed and relaxed as you soak in a tub of hot water. Neutral Images Imagine a gray rock. Imagine an ice cube. Imagine a pencil sharpener. Imagine a dim movie theater with an empty screen. Imagine Imagine Imagine Imagine Imagine Imagine Imagine Imagine Imagine Imagine Imagine Imagine Imagine Imagine breathe without 155 a white bowl half filled with water. a single piece of chalk. a banana peal. a empty blackboard. a pile of brown sand. the dotted line in the middle of the road. an empty piece of white paper. a man walking. a white t-shirt. an empty ash tray. an electric fan. Aversive Images that you are walking by numerous dead fish as you walk around the edge of a polluted lake. a starving baby abandoned on a war-torn that you are in a room so filled with cigar smoke that your head aches, your eyes hurt, and you can't coughing. 156 Imagine that you have an examination in an hour on a three hundred page book which you have not read. Imagine that you are in the town dump on a swelteringly hot summer afternoon. Imagine that you are covered with a slimy, puss- filled rash that itches and is nearly unbearable. Imagine that you are in a pitch-black cave on your hands and knees. You put your hand into something that is cold, slimy, and bug-infested. The passageway is so small that you cannot turn around. Imagine that your dog has had diarrhea all over the living room rug and there is no one else around to clean it up. Imagine that you have become so obese that you can no longer fit into any of your favorite clothes. Imagine that you have awakened three times during the same night to answer a phone which is immediately hung up on the other end. Imagine that you have just been in a car accident and you look down to see that you are covered with blood. Imagine that you have stopped at an unfamiliar diner. You bite into a hamburger and find it is crawling with maggots. Imagine that you have meat and you are now vomiting Imagine that you have rejection letters from places position. And yet, you still Imagine that you are, you greatly dislike, and they you. recently eaten some rotten all over your kitchen. collected a file of fifty where you have applied for a have no job! by chance, alone with someone make a sexual advance toward APPENDIX M FINAL EVALUATION FORM APPENDIX M FINAL EVALUATION FORM What, in your opinion, was the probable purpose of this Study? Briefly outline the types of thoughts or ideas that occurred to you as you participated in the experiment. Did you attempt to maximize the number of feeling/emotional words in the stories that you related during the exper- iment? Yes No Did you try to keep track of the number of feeling/emotional words that were included in each story? Yes No Did you compare the performance in the number of feeling/emo- tional words included in one story with the number of these words in other stories? Yes No 158 “1111111111111“