SELF CONCEPT CHANGES AS A FUNCTION OF PARTICIPATION IN SENSITIVITY TRAINING AS MEASURED BY THE TENNESSEE SELF CONCEPT SCALE Thesis for the Degree of Ed. D. MICHIGAN STATE UNIVERSITY ROBERT CHARLES BROOK 1968 TH LSIS This is to certify that the thesis entitled Self Concept Ghanges as a Function of Participation in Sensitivity Training as Measured by the Tennessee Self Concept Scale ' ‘ presented by Robert Charles Brook has been accepted towards fulfillment of the requirements for Ed D degree injigheLEfiucation Major essor Any—we? ABSTRACT SELF CONCEPT CHANGES AS A FUNCTION OF PARTICIPTATION IN SENSITIVITY TRAINING AS MEASURED BY THE TENNESSEE SELF CONCEPT SCALE by Robert Charles Brook The specific problem investigated in the study was to determine the extent to which sensitivity training may influence the self-concept of the participants of the 1967 Fifth Provincial Training Laboratory in the EpiSCOpal Church as measured by the Tennessee Self-Concept Scale. The experimental group (N=70) and the control group (N=32) were comprised of laymen and Episcopal Priests. The population for the experimental group was drawn from mid- western states and the control members were from central New York. Initial equivalence between the groups was established demographically. However, the groups were not equivalent on pre-test data. The experimental population experienced seven consecutive days of sensitivity training. Two components of an individual‘s self—concept were evaluated in this study, namely, self-esteem and defensive- ness. Self-esteem, that is, an individual's enhancing or depreciative self evaluation was measured by the Counseling Form of the Tennessee Self—Concept Scale. Defensiveness was measured by an individual's ability to accept mildly derogatory statements about himself. An index of the Robert Charles Brook individual's capacity for self-criticism was attained from the Tennessee Self-Concept Scale. The initial research design was a modified Soloman A group design attained by randomly assigning members from both populations to the respective group (Campbell and Stanley, 1963). The unique- ness of this design would have permitted 2 x 2 statistical analysis of variance of post-test scores. Because pre-test equivalency between the experimental and control popula- tions was not established, alternative statistical proce- dures were employed in analysing the data. Specfically, analysis of covariance, the pre-test being the covariant (Kerlinger, 1966). This statistical method makes the necessary modifications in sampling error. The first hypothesis stating that self-esteem will increase as a func- tion of training was not accepted. The second hypothesis that training will reduce the level of defensivensss was likewise not accepted. The conclusion of this study is that participants in this particular sensitivity training laboratory did not significantly alter their perceptions of themselves as a result of treatment as measured by the Tennessee Self-Concept Scale. SELF CONCEPT CHANGES AS A FUNCTION OF PARTICIPATION IN SENSITIVITY TRAINING AS MEASURED BY THE TENNESSEE SELF CONCEPT SCALE By Robert Charles Brook A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF EDUCATION Department of Administration and Higher Education 1968 657752 ACKNOWLEDGMENTS I would like to express my sincere appreciation to the following people for their assistance and support during my studies. The Right Reverend Richard S. Emrich, Bishop of Michigan, for granting me his approval and sanction to continue advanced education. Dr. Don Hamachek, Dr. Max Raines, Dr. Gerhard Linz, and the committee Chairman, Dr. Vandel Johnson for their assistance and support at every stage of my studies. The dissertation would have been considerably delayed if it were not for the secretarial assistance which Vera Kean and the inmate clerks of the State Prison of Southern Michigan provided; to them I-am greatly indebted. The dissertation would have been an impossibility if it were not for the encouragement and support I received from my parents. My special gratitude is expressed to my wife, Suzanne, for her patience, assistance, and constant sup— port during all my studies. ii TABLE OF CONTENTS Page ACKNOWLEDGMENTS . . . . . . . . . . . . . ii LIST OF TABLES . . . . . . . . . . . . . . iv Chapter I. THEORY AND PURPOSE . . . . . . . . . . 1 Statement of the Problem‘. . . . . . . . 1 Theoretical Definitions -. . . . . . . . l Rationale . . . . . . . . . . . . . 10 Theoretical Hypotheses . . . . . . . . 10 II. REVIEW OF LITERATURE. . . . . . . . . . 11 Introduction . . . . . . . . . . . ll summary of Previous Studies . . . . . . . 11 Specific Studies . . . . . . . . . . 1A Conclusions . .’ . . . . . . . . . 23 III . METHODOLOGY 0 O O O O O O O O O O . 214 Operational Definitions . . . . . . . . 2A Groups . . . . . . . . . . . . . . 29 Research Design . .. . . . . . . . . . 30 IV. RESULTS OF THE STUDY . . . . . . . . . 36 Demographic Data . . . . . . . . . . 36 Pre-TeSting Data 0 o o o o o o o o o 37 V. DISCUSSION . . . . . . . . . . . . “5 Implications for Future Research . . . . . 50 Summary . . . . . . . . . . . . . 51 BIBLIOGRAPHY . . . . . . . . . . . . . . 54 iii TABLE 10. ll. 12. 13. LIST OF TABLES Demographic data . . Pre-test scores for dependent variables . . Comparison of pre- post-test difference by groups . . . . Test for homogeneity of regression for self—criticism . . Analysis of covariance self-esteem . . . Analysis of covariance identity . . . . Analysis of covariance self satisfaction . Analysis of covariance behavior . . . . Analysis of covariance physical self . . Analysis of covariance moral self . . Analysis of covariance personal self . . of covariance self . . . Analysis family of covariance self . . Analysis social for dependent for dependent for dependent for dependent for dependent O O O ‘ O O for dependent for dependent for dependent for dependent iv variable variable variable 0 O 0 variable variable variable variable variable variable Page 37 38 A0 Al Al Al A2 A2 A2 A3 A3 A3 AA CHAPTER I THEORY AND PURPOSE Statement of the Problem The specific problem investigated in this study was to determine the extent to which sensitivity training may influence the self-concept of the participants of the 1967 Fifth Provincial Training Laboratory in the Episcopal Church as measured by the Tennessee Self-Concept Scale. Theoretical Definitions Self—Concept Self, as a psychological construct, is an effort to understand human behavior, and self-concept has been and continues to be variously defined. Hall and Lindsey (1957), say the term self . . . has come to have two distinct meanings. On the one hand it is defined as the person's attitudes and feelings about himself, and on the other hand it is regarded as a group of psychological processes which govern behavior and adjustment. They say that both of these constructs are helpful; however, they believe much of the confusion in the literature today stems from a mixing, overlapping, and interchanging of these definitions. Regardless of this confusion, they contend that no current 1 self theorist conceives of self as an "inner manikin" or "man within the breast." In discussing the origin of self, Hall and Lindsey say that the various self theories observe or are governed by the principles of causality and also that self emerges as the consequence of the organism's interaction with its environment (pp. A67—A69). Snygg and Combs (1950) speak about the phenomenal self as differentiated out of the individual's environment. The phenomenal self, then, includes all those parts of the phenomenal field which the individual experiences as part or characteristic of himself. It is his self perception as derived from his environment. When speaking about self- concept and behavior, they say, . . there is no reason why the psychologist, in looking for the immediate cause of behavior, needs to go back to find an historical cause outside of the present perceptual field. Certainly, the events of an individual's life effect his behavior. But it is important for us to recognize that it is the perceptions of these events and not the physical events themselves which are the immediate causes of behavior (p. 80). Carl Rogers (1951) defines self-concept as ". . . an organized configuration of perceptions of the self which are admissible to awareness" (p. 501). The self, as the nuclear concept in Rogers' theory of personality, has numerous properties, some of which are these: (a) it develOps out of the organism's interaction with the environment, (b) it may introject the value of other people and perceive them in a distorted fashion, (c) the self strives for consistency, (d) the organism behaves in ways that are consistent with the self, (e) experiences that are not consistent with the self structure are perceived as threats, and (f) the self may change as a result of maturation and learning. Robert White (1956) in his discussion of self—concept synthesizes the theories of Allport, Murphy, Goldstein, Sherif, and Cantril. White (1956) says, "The concept of the self helps us to bear in mind the basic fact of the unity of the organism" (p. 156). He goes on to say that without the concept of self "we would have no point of anchorage for the personal patterns or tendencies that is characteristic of each individual" (p. 156). He further defines the self by saying, "Its nucleus appears to be what is experiences as "I" and "me", as distinguished from every— thing else that is "not me" (p. 157). For White the self emerges from social interaction and the self continues to be influenced by a person's experiences. The self is sub— ject to continuous change. The pattern is formed and reformed many times in the course of life (p. 158). Each self theorist might offer an alternative defini- tion to self, but two principle characteristics seem to emerge. The first is the recognition of the importance of social interaction in the emergence and development of an individual's self-concept and his self-evaluation or level of self—esteem. Secondly, these self theorists believe an individual's self-concept is open to change. For example, Rogers (l96l, 1967), Gividen (1959) and Ashcraft and Fitts (196A) have demonstrated that therapy can positively influ- ence the way in which a person sees himself, that is, change his self—concept and influence his self—evaluation. Fitts (1965) says, "Psychotherapy or other positive experi- ences would be eXpected to result in enhancement of the self-concept, while stress or failure would be expected to result in lower self-esteem" (p. 28). Specific evidence to substantiate this position will be discussed in the review of literature. In discussing self-esteem Ruth Wylie (1961) says, "The most commonly studied class or aspects of the phenome— nal self includes such attitudes as self-satisfaction, self— acceptance, self-esteem, self-favorability, congruence between self and ideal self, and discrepency between self and ideal self" (p. A0). Wylie points out that these terms are not synonymous and says that for some authors, "self- acceptance means respecting one's self including one's admitted.faulmswhile self-esteem or congruence between self and ideal self means being proud of one's self or evaluating one's attributes highly" (p. A0). Self-esteem is an aspect of one's self-concept, more specifically, self-esteem is the "general evaluative attitude toward self" (Wylie, p. AO). Ford and Urban (1965) also indicate that self-esteem is one's evaluation of himself and they define self—esteem as learned positive regard. They also indicate that an indi- vidual's self-esteem or self—regard is influenced by new experiences (pp. A08-Al3). In Mullahy's discussion of Sullivan, self-esteem is defined as an individual's self— appraisal. He says, "self—esteem may be said to be made up of, or at least circumscribed by, reflective appraisal" (p. 297). Robert White (1956) discusses self-esteem in the context of personal competence and says "that self—esteem is tremendously effected by the income of esteem that one receives from others" (p. 161). He goes on to say that the "experience of competence, based on the effectiveness of one's own activity in dealing with the environment, is a vital root of self—esteem" (p. 161). Defensiveness as used in this study has reference to the individual's inability to accurately assess himself or his environment. Defensiveness is variously defined, but it always includes elements of self or environmental dis- tortion (White, 1956; Hall and Lindsey, 1957). Carl Rogers (1961) defines defensiveness as being the organism's reSponse to experiences which are perceived or anticipated as threatening, as in- congruence with the individual's existing picture of himself, or of himself in relationship to the world. These threatening experiences are tem- porarily rendered harmless by being distorted in awareness or being denied to awareness (p. 187). Rogers (1961) perceives one of the objectives of therapy to be the reduction of defensiveness. He says, "a large part of the processes of therapy is the continuing discovery by the client that he is experiencing feelings and atti— tudes which heretofore he has not been able to be aware of, which he has not been able to 'own' as being a part of himself" (p. 187). Defensiveness then, is the individual's inappropriate protection of self which interferes with and prevents the person from being open to accurately perceiving his experiences (Rogers, 1961). Sensitivity Training A succinct explanation of sensitivity training is pro- vided by Leland Bradfor (Driver, 1958). He says, " The T-group training provides fluid, unstructured groups in which every person becomes deeply involved, in which problems of leadership, decision making, interpersonal problems, and hidden agenda are seen in slow motion. In such a group, the individual becomes more diagnostically sensitive to what happens in groups; becomes aware of the ways in which groups grow; gains in self—awareness. The T-group pro- vides opportunity for trainees to behave; to secure feedback on their behavior; to experiment with new ideas of leadership and membership . . . to get-~on the "feeling" as well as "intellectual" level--a real awareness of the problems of group organization, functioning and growth (p. 3A5). The Reading Book: Twenty-first Annual Summer Labora- tories in Human Relations Training (1967) states the purpose of sensitivity training is to "help each individual realize his own potential for growth more fully and to increase his ability to work effectively with others in a variety of situations" (p. 2). The authors continue by saying the following five factors are important broad objectives of sensitivity training: (1) self—insight; (2) better understanding of other persons and awareness of one's impact on them; (3) better understanding of group processes and increased skill in achieving group effectiveness; (A) increased recognition of the characteristics of larger social systems; and (5) greater awareness of the dynamics of change. More specifically, in regard to self changes, training groups create a climate encouraging of the follow- ing: (1) Own feelings and motivations; (2) Correctly perceiving effects of behavior on others; (3) Correctly understanding effect of others behavior on self; (A) Hearing others and accepting helpful criticism, and; (5) ApprOpriately interacting with others (p. 2). These learnings are not easy to attain, nor does the usual social environment provide the conditions for such learning. Whereas a training group is a unique social community where the following conditions need to be met in various ways if participants are to reach personal goals of improvement and change in insights, understandings, sensitivities, and skills: (1) Presentation of Self: Until the individual has an opportunity to reveal the way he sees things and does things, he has little basis for improve- ment and change. (2) Feedback: Individuals do not learn from their experience. They learn from bringing out the essential patterns of purposes, motives, and behavior in a situation where they can receive back clear and accurate information about the relevancy and effectiveness of their behavior. They need a feedback system which continuously operates so that they can change and correct what is inappropriate. (3) Atmosphere: An atmosphere of trust and non defensiveness is necessary for people both to be willing to expose their behavior and purposes and to accept feedback. (A) Cognitive Map: Knowledge from research, theory, andwexperience is needed and important to enable the individual both to understand his experiences and to generalize from them. But generally, information is most effective when it follows experience and feedback. (5) Experimentation: Unless there is opportunity to try out new patterns of thought and behavior, they never become a part of the individual. With- out experimental efforts relevant change is difficult to make. (6) Practice: Equally important is the need to practice new approaches so that the individual gains security in being different. (7) Application: Unless learning and change can be applied to back home situations, they are not likely to be effective or lasting. Attention needs to be given to helping individuals plan application. (8) Relearning How to Learn: Because much of our academic experience has led us to believe that we learn out of listening to authorities, there is frequently need to learn how to learn from presentation—feedback-experimentation (p. 11). Matthew Miles (1959) says, "Any training aims at change in the person." He further develops the concept of change by saying, ". . . a training group's objective is change in its members' way of doing things, their proce- dures, their practices . . ." (p. 35). Miles also says that changes in the person are attained by providing for and practicing the following themes: (1) A focused con— cern with skills, that is, the tools a person needs to bring his actions into line with his intentions; (2) Whole person learning. This implies that training focuses upon several behaviors; namely, thinking, feeling, choosing, and acting. (3) Guided practice. Training also implies practice-~repeated performance of particular skills with explicit, immediate information on the results of a try, and (A) psychological safety. The atmosphere of the training group is characterized by being protective and to the extent possible, free from threat (pp. 28-33). To further define sensitivity training groups it will be helpful to differentiate sensitivity training groups from psychotherapy groups. To a real extent the differentiation appears to be one of degree. The unifying theme which characterizes the groups is that both groups are concerned with changes in their members behavior (Miles, 1959). However, for the sake of differentiation, it can be said that members of therapy groups are patients. They are disfunctioning. Something in the person has gone wrong; the person is troubled or suffering and needs care, whereas training groups members are functioning, that is, they are not clinically diagnosed as patients. They may, however, have dissatisfaction about their own social or personal skills. Matthew Miles (1959) cites what we can consider to be the principal difference between therapy and training groups. He says, . . . in a training group, . . . there is more emphasis on his "outer workings"--the way he relates to people. A training group is usually less concerned with the inner reasons for why someone does something; and more concerned with how he does it, what the impact is on others, and how he can improve what he does to become more skillful (p. 30). By skillful he means more aware of himself, his feelings, his behavior, and his effect upon other people. lO Rationale The specific concern, then, was to facilitate or extend our understanding of how sensitivity training effects the self-concept of participants. The rationale can be stated in the following way: one's self-concept, that is, his organized perceptions about himself, are influenced by positive or new learning experiences (Rogers, 1951; Fitts, 196A; Sanford, 1966). Sensitivity training is viewed as a positive learning experience, organized to facilitate new self learnings (Miles, 1959). Thus, we would expect that sensitivity training will effect the participant's self-concept. The most direct question is, then, can we measure this effect upon self-concept as a consequence of participation in a sensitivity training laboratory? Theoretical Hypothesis The principal hypothesis to be investigated in this study is that an individual's self-concept will be positively influenced as a function of his participation in sensitivity training. It is also hypothesized that as a function of sensitivity training a participant's level of defensiveness will decrease. These hypotheses are presented in operational terms in Chapter III. CHAPTER II REVIEW OF LITERATURE Introduction The review of literature is organized to accomplish two objectives. The first objective is to present an over- view of psychotherapy outcome studies. The objective will be accomplished by summarizing previous reviews of the literature to indicate the trends, difficulties, and con- ditions of outcome research. This procedure is followed because in general this study can be considered an outcome study. Outcome studies are concerned with the consequence of treatment and do not take specific account of the pro- cesses of treatment (Strupp, 1966; Myers, 1966). The second objective is to review current and specific research studies which indicate self—concept variables change as a function of sensitivity training or psychotherapy. Summary of Previous Studies Ruth Wylie (1961) in an early review of the literature regarding self-concept, encountered an area which was pro— fuse, disorganized, and prohibitive of meaningful synthesis. She summarized her impressions of the literature by saying, We have noted that the empirical researches on constructs concerning the self cannot be classified according to theoretically relevant categories 11 12 because the theories are vague, incomplete, and overlapping; and, because no one theory has received extensive, empirical exploration . ... there is a good deal of ambiguity in the results, Considerable apparent contradiction among the findings of various studies, and a tendency for different methods to produce different results. In short, the total accumulation of substantiative findings is disappointing, especially in prOpor- tion to the great amount of effort which obviously has been expended (p. 317). The common errors in the studies reviewed by Wylie were the following: (1) vague methodological procedures which prevented replication of the study reported; (2) insuffi- cient numbers of different control groups and the absence of randomization or matching among the groups; (3) the absence of objective and independent judgements of patient improvement; and (A) the researchers' tendency to over- generalize the conclusions and implications of their study. Schmidt and Pepinsky (1965) reviewed the counselling research literature in 1963 concluding that it was not a "bumper crop" and that apparently the researchers were more concerned with expediency than the ends of scientific knowl- edge. They go on to say, "We regret to report that many of the studies reviewed have suffered from lack of adequate forethought and from careless workmanship,such as their findings are left Open to serious question" (p. A25). Robert Carkhuff (1966) in his article "Counselling Research, Theory and Practice—-1965" characterizes the counselling research literature by saying 13 At the present time, there is a lack of any compre- hensive, integrative approach to research and theorizing. The systematic, inductive building and deductive testing processes, in continual pro- cess of qualification and modification, are absent. There is no differential weighing and relating of the multitude of counselor, client and contextual variables to the multitude of indices of construc- tive client change or gain (p. A76). Hans Strupp (1966) in his review of outcome research agrees with Kiesler and Wylie that psychotherapy research is disorganized and that results tend to be either non— significant or contradictory. He says Experimental comparisons of therapy and no therapy patients consistently reveal two major findings: (1) Some therapy patients improve and some worsen, and (2) The no therapy or control patients receive "therapeutic" assistance from clergymen, physicians and friends (p. 1A2). Carl Rogers' (1966) impressions of psychotherapy research takes cognizance of the confusion and contradictory result of outcome studies but he is more optimistic in his outlook. Psychotherapy at the present time is in a state of chaos, it is not, however, a meaningless chaos, but an ocean of confusion, teeming with life, spawning vital new ideals, approaches, procedures, and theory at an incredibly rapid rate. Hence the present is a period in which the most diverse methods are used and in which the most divergent explanations are given for a single event (p. 11). Recognizing the confusion and the conflicting results of research is important. The importance lies in identify- ing errors and methodological difficulties so as to prevent the same failings in the current research (Frank,.l967; Krumboltz, 1965; Stock, 196A; Mezzano, 1966). However, it 1A is also important to recognize the favorable finding in psychotherapy outcome research. Even though Wylie (1961) specified the errors and failings of psychotherapy research as it relates to self-concept, she also concludes that "On the whole we have found that there are enough positive trends to be tantalizing" (p. 317). Specific Studies Psychotherapy An early study by Rogers (1961) which is character— istic of his research indicates positive results as a con- sequence of psychotherapy. Using the Q sort technique as a pre- post-test measure of self— ideal-self congruence, Rogers hypothesized that clients receiving psychotherapy would increase in self—esteem. He employed two control groups, the first group was a non-therapy group matched for age, sex and social economic status, and the second control was attained by using the therapy group as an own- control group. The results were as follows: the non— therapy control group on the test-retest measure did not change. The own-control group had a mean correlation of -.01 sixty days prior to therapy and at the beginning of therapy the groups' mean correlation was unchanged. The mean correlation at the completion of therapy was .3A. He concluded from this evidence that "One of the changes associated with client centered therapy is that self- perception is altered in a direction which makes the self 15 more highly valued" (p. 258). The study included a twelve month follow-up to investigate whether these changes in self-esteem were transient or continued over time. The follow-up mean correlation for the therapy group was .31 which indicates that regression to a pretherapy condition did not occur and that the changes were permanent. In‘a more recent study, Rogers (1967) evaluated whether the same positive effects which occurred in neurotic clients as a consequence of therapy were also true for a more disturbed patient population. This was an ex- tensive study designed to measure the relationship between therapist and process variables and therapy results. However, only the specific portion of the study which relates to the outcome of therapy will be discussed. The subjects of this study were hospitalized chronic and acute schiZOphrenic patients who were matched with identical hospitalized patients for control purposes. The type of therapy employed was non-directive and in some cases extended over several years. One of the procedures employed to measure self changes as a function of therapy was to have independent raters score the patients' Thematic Apperception Test responses not knowing which tests were pre— or post-therapy or which tests were from the control or experimental groups. One of the marked changes shown by the therapy group was that their emotional distance from the experience they 16 described in the TAT picture showed significant alterations from pre- to post—test. The therapy patients reduced their need to deny or emotionally distance themselves from their experience. The control group in the hospital, on the other hand, showed some tendency to become more defensive, more distant from the experience they were describing. It is also of interest that the therapy group showed signifi- cant improvement in the appropriateness of their emotional expression whereas the control group showed a trend toward more extreme expression of emotion. The therapy group also gave evidence of improvement in the capacity to handle interpersonal relationships in satisfying ways. Rogers concluded that the therapy group was less vulnerable, psychologically, and more capable of facing themselves and their environment than were the control group. Caroline Ashcraft and William Fitts (196A) using the Tennessee Self-Concept Scale undertook the following study to indicate self-concept changes occur in psychotherapy. More specifically, they found that an individual tends to value himself more highly as a function of psychotherapy as measured by the Tennessee Self-Concept Scale. The study was conducted at the Nashville Mental Health Center with out-patients. The experimental group consisted of 30 patients who had been in psychotherapy for 3 months or more and the mean therapy time was once a week for 6 to 8 months. The control group was composed of 2A patients who 17 had been waiting for psychotherapy 3 months or more with a mean waiting period of 6 2/3 months. The types of therapy which the experimental group experienced varied from environmentally oriented case work, intensive individual and/or group therapy. All subjects were measured on a test-retest basis with the Tennessee Self-Concept Scale. The hypothesis was that the experimental group would Show predictable and significant changes but that the control group would show no change. Of the 21 variables reported, the experimental group changed in the predicted direction. The change was significant at the .05 level for 17 of the 20 variables while the control group showed little signifi- cant change. Of the same 21 variables the control group changed only on two variables; a decrease of psychotic tendency and in perceptual conflict. Ashcraft and Fitts concluded that the experimental "subjects who received psychotherapy reported self-concepts that were more posi- tive in all areas, more consistent, with less evidence of deviation and pathology" (p. 118). The Rogers (1961, 1967) and Ashcraft and Fitts (196A) studies indicate that as a consequence of psychotherapy subjects tend to value themselves more highly, that is, positively increased their self-esteem. l8 Sensitivity Training The following studies indicate that similar self- concept or personality changes occur as a function of sensitivity training. Richard Burke and Warren Bennis (1961) report a study which does not employ a control group, and consequently the generalizations which we may draw from this study are limited. However, the Burke study is referred to because their results are typical of several sensitivity training studies. Burke and Bennis studied Human Relation Training groups during a summer session at the National Training Laboratory, in Bethel, Maine. The experimental population numbered 8A subjects with 13 to 15 members in each of six training laboratories. The experimental subjects repre- sented a heterogeneous background and a wide variety of occupations from several geographic areas within the United States.. The subjects were of both sexes with an age range of 25 to 60 years, and a medium age in the upper thirties. The laboratory experience lasted for three weeks during which time delegates participated in an intensive program of instruction and training, including daily two hour meetings of the T-group as well as skill exercise group sessions on theory, individual and group consultation and counseling, and special clinics organized around particular topics of interest. To measure self-satisfaction, Osgood's 19 Group Semantic Differential test was administered to the subject twice during the laboratory experience. The pre- test was administered during the middle of the first week and the post-test was administered during the latter part of the third week. Burke and Bennis conclude that members of training groups during the course of their training experience seem to become more satisfied with their per- ception of self, move their actual self percept in the direction of their ideal, become more congruent in their perceptions of others and come to see others more as these individuals see themselves. A more recent study by Sherwood (1965) found similar results as the Burke and Bennis study. However, they were more specific as to when the person moves in the direction of accepting the perceptual evaluations of other group members. Using 68 members of the 1961 summer session of the National Training Laboratories at Bethel, Maine, he also found that there was a tendency for the groups as a whole to change their individual self perceptions in the direction of the groups' perceptions. This was especially true when the person perceived the group as a strong referent group and the group evaluated him highly. However, when the person was not evaluated highly he tended to dis- regard their standards and continued to use his own. Irwin Rubin's (1967) research indicated that individ— uals participating insensitivity training increase their 20 self acceptance and likewise show a decreasing level of ethnic prejudice. He tested the following hypotheses: (l) the higher an individual's level of self acceptance, the lower will herds level of prejudice; (2) as a result of participation in sensitivity training an individual's level of self acceptance will increase and his level of prejudice will decrease; and (3) changes in self acceptance will be associated with changes in prejudice. Self acceptance, as it was used in this research project, involved the subject's willingness to confront ego-alien as well as ego—syntonic aspects of the self and to accept rather than to deny their existence. He employed the Dorris, Levenson, and Hanfmann Sentence Completion Test to measure the level of an indi- vidual's self acceptance. Prejudice was defined as the extent an individual was willing to accept others in terms of a common humanity and the Harding and Schuman test of prejudice was employed. The subjects were 50 participants in the Osgood Hill summer program in sensitivity training. The sample population was composed of 30 females and 20 males with an age range from 23 to 59 and a mean age of 33. The experimental group was randomly split into two groups of unequal size, the smaller group serving as their own control group andtested by a mailed questionnaire two weeks prior to their sensitivity training. The entire experimental population was tested upon their arrival and the morning next to the last day. Using T-tests to 21 evaluate the difference between the means, Rubin reports the following results. Participant's self acceptance increased from 55% to 67% which was statistically signifi- cant at the .01 level. Prejudice decreased from A6.2 to A2, also significant at the .01 level. Rubin continued the analysis by asking the question, "Did the experimental group change more than the control group?" To ascertain this difference a Mann—Whiteney U-test was performed on the difference between the difference within the experi— mental and control groups. The result of the test was significant atthe .0A level, and Rubin concluded, "In other words, not only do the experimentals change while the controls do not, but also, the experimentals change significantly more than the controls as a function of sen- sitivity training" (p. 237). Harrison (1966) studied concept changes as a function of participation in a sensitivity training laboratory at Bethel, Maine. He states that The concepts which the individual has available for structuring his interpersonal relationships are held to determine the kinds of responses which are possible for him since he can respond only along perceptual or behavioral dimensions for which he possesses concepts. Events, feelings, and behavior which are not so conceptualized do not have meaning for the person (p. 518). He found that participants changed their perceptions of others away from concrete-descriptive statements to de- scriptions oriented to interpersonal processes. He also found a direct and positive relationship with the 22 hypothesized change and the participant's level of involve— ment in the training experience. Another important finding of the study was that the hypothesized changes reached significance three months after the completion of the training experience. He concludes There is a significant change in concept usage following the training experience. This change appears to be progressive. It is slight when measured shortly after the experience, reaching significant proportions only when measured three or more months after the end of the laboratory experience. This pattern of change would be expected if the participants went home with a readiness to perceive others along more expressive and inferential dimensions, and if they required some time and inter— action with the others in order to obtain new infor- mation which was needed to activate the concepts (p. 520). Delaney (1966) found that subjects who experienced sensitivity training increased in sensitivity and displayed a greater awareness of non-verbal emotional expressions of other people. His subjects were graduate students whom he randomly assigned to two different groups. The experimental subjects were assigned to a sensitivity training group and the control subjects were assigned to a group which experienced lectures on non—verbal communication. Pre- post-test measures were attained from the subject's responses to filmed interviews where they were asked to respond to the video-taped subject's emotional expressions. He concluded that those subjects who experienced sensitivity training were more able to prOperly identify emotional expressions than were the control subjects. 23 Conclusion Reviewers of research literature have indicated that outcome studies reveal conflicting results (Schmidt and Pepinsky, 1965). However, the contradictory results are profitable in encouraging new research,and researchers need not be discouraged by this condition (Rogers, 1966). The studies reviewed indicate that neurotics (Rogers, 1961) and more disturbed patients (Rogers, 1967) change their self perceptions in positive directions as a function of non-directive psychotherapy. Ashcraft and Fitts (196A) found that subjects alter their self-concept in positive directions as a consequence of group psychotherapy. Other studies indicate that sub- jects alter their self perceptions to comply with group norms (Burke and Bennis, 1961; Sherwood, 1965) or increase their self-acceptance (Rubin, 1967), or display greater awareness of emotional expressions in others (Delaney, 1966) as a consequence of their participation in a sensi- tivity training laboratory. CHAPTER III METHODOLOGY Operational Definitions Self-Concept For the purpose of this study self-concept was oper- .ationally defined as the person's self description as measured by the Tennessee Self-Concept Scale developed by William Fitts (Kerlinger, 1966, p. 3A). In discussing the nature and purpose of the scale Fitts (1965) says Over recent years a wide variety of instruments has been employed to measure the self-concept. Nevertheless, a need has continued for a scale which is simple for the subject, widely applicable, well standardized, and multi-dimensional in its description of the self-concept. The Tennessee Self-Concept Scale . . . was developed to meet this need (p. 1). The contents of the scale is described by Fitts (1965) as consisting of . . 100 self-descriptive statements which the subject uses to portray his own picture of him- self. The Scale is self—administering for either individuals or groups and can be used with subjects age 12 or higher and having at least a sixth grade reading level. It is also applicable to the whole range of psychological adjustment from healthy, well adjusted people to psychotic patients (p. 1). In discussing the development of the scale Fitts (1965) says, 2A Sub 25 . . . the first step was to compile a large pool of self-descriptive items. The original pool of items was derived from a number of other self- concept measures including those developed by Balester (1956), Engel (1956), and Taylor (1953). Items were derived also from written self— descriptions of patients and non-patients. After considerable study, a phenomenological system was developed for classifying items on the basis of what they themselves were saying. This evolved into the two-dimensional, 3 x 5 scheme employed on the Score Sheet of both forms. . .,. After the items were edited, seven clinical psychologists were employed as judges to classify the items . . . The final 90 items utilized in the Scale are those where there was perfect agreement by the judges (p. l). -Scales and Their ReliabIIitIes are : The sub-scales which were evaluated in this study 1. The self-criticism score which is composed of ten items that are mildly derogatory. Low self-criticism score is an indication of defensiveness, where a high score generally indicates a normal, healthy openness and capacity for self-criticism (R .75). 2. Self-esteem scale, which is made up of the following sub—scales with a test-retest reliability of .92. a. The Identity Scale, which indicates "this is what I am" (R .91). b. The Self-satisfaction Scale, which is an indication of self-acceptance, that is, how 26 does the person feel about the self he describes (R .88). The Behavior Score is an index of how the person describes how he acts, namely, the person describes himself by saying, "this is what I do" (R .88). The Physical Self, here the person is describing his feelings about his body, his health, his physical appearance, skills and sexuality (R .87). The Moral-ethical Self. This scale provides us with an indication of the person's percep- tion of his moral worth, relationship to God, feelings of being a good or bad person (R .80). The Personal Self. This scale indicates how the person evaluates his personal worth or his feelings of adequacy as a person (R 085). The Family Self. Here we see something of the person's feelings of adequacy, worth, and value as a family member (R. 89). The Social Self. Here we see how the person perceives himself in relation to others. 27 Hypothesis The following hypothesis of self-concept changes as a function of sensitivity training were investigated in this research: I. Self-esteem will increase. a. identity will increase b. satisfaction will increase 0. behavior will increase d. physical self will increase e. moral self will increase f. personal self will increase g. family self will increase h. social self will increase 11. Self-criticism will increase. Validity The validation procedures employed for the instrument (TSCS) are of four kinds. Regarding content validity, Fitts (1965) says, The purpose here has been to ensure that the classification system used for the Row Scores and Column Scores is dependable . . . an item was retained in the Scale only if there was unanimous agreement by the judges that it was classified correctly. Thus, we may assume that the categories used in the Scale are logically meaningful and publicly communicable (p. 17). The instrument was also validated by its ability to differentiate between groups. 28 Statistical analyses have been performed in which a large group (369) of psychiatric patients have been compared with the 626 non- patients of the norm group. These demonstrate highly significant (mostly at the .001 level) differences between patients and non-patients. . . . The author has also collected data from the other extreme of the psychological health continuum-—from people characterized as high in personality integration. The basic hypothesis which was established here was that this group would differ from the norm group in a direction opposite from that of the patient group (Fitts, 1965, p. 17). "Another way to assess the validity is to determine the correspondence between scores on the Scale and other measures for which correlations should be predicted" (Fitts, 1965, p. 2A). Using the Pearson product moment correlation coefficient and the ETA coefficient, the Tennessee Self-Concept Scale and the Minnesota Multiphasic Personality Inventory are significantly correlated "in ways one would expect from the nature of the scores" (Fitts, 1965, p. 2A). In a study done by Sundly (1962) the Edwards Personal Preference Schedule and the T808 "indi- cated rather clear nonelinear relationships between scores on the two tests" (Fitts, 1965, p. 2A). This, Fitts contends, was to be expected. SensitivitygTraining Sensitivity training was Operationally defined as the activity engaged in by members of groups who were assemble at Gull Lake, November 5 through 11, 1967, for the purpose of sensitivity training. The program was 29 under the auSpices of the Fifth Province of the Episcopal Church. Participation Participation was operationally defined as member— ship in one of the groups at Gull Lake. There was no effort to differentiate quality of group membership. Groups The control and experimental groups were intact. Each member voluntarily signed up to participate in sensitivity training in his respective geographical region. By intact, it is meant that the Training Com- mittees of each Province recruited and assigned the members for their respective Training Laboratory. The training for the Experimental Group was under the auspices of the Fifth Province of the Episcopal Church and the training for the Control Group was under the direction of the Second Province of the Episcopal Church. In each case, the groups were comprised of both laymen and Priests. To assist in establishing equivalence between the groups each person filled out a personal information question- naire. The questionnaire provided information regarding age, sex, education, churchmanship, and whether the person is lay or ordained. Equivalence was also established by the use of a pre-test which both the Experimental and Control groups were administered. If questions remain 30 regarding equivalence of the groups, analysis of covariance will further equate the groups. The Control Group was selected primarily so as to have a group which signed up to participate intraining but be at a waiting stage. This is expected to control for desire or expectations to change if changes occur in the Experimental Group. The effect of test-retest was controlled by randomly selecting from the Control and Experimental Groups one— half of the pOpulation which received the pre-test. The pre-test was administered to the entire populations in such a way so as to see the effects of taking the test twice. The pre-test was mailed in both cases and a high rate (75-100%) was returned (90%). Only one test was discarded from the Experimental population and that was because it was improperly marked. Research Design Description The design employed in this research project was a modified Solomom four—group design. The design is modified because it lacks true randomization. However, by using this design the effects of testing and treatment interaction are determinable. The design also included a replication of treatment. The principle effect of employ- ing the Solomon four-group design was to increase general— izability (Campbell and Stanley, 1963, p. 25). 31 The following is a graphic representation of the design. Pre-Test Treatment Post-Test l. O X 0 Experimental 2. X 0 3. O 0 Control u 0 l. Represents one-half of the experimental popula- tion which was randomly selected to receive test, treatment, and retest. 2. Represents one-half of the experimental pOpula- tion which receives treatment and post-test. 3. Represents one-half of the control population which was randomly selected to receive test- retest. A. Represents that portion of the control popula- tion which receives only the post-test. The four-group design was achieved by randomly dividing both the Experimental and Control Groups in half for the purposes of pre—testing. It was by this method that we are able to control for testing and treatment interaction effects. 32 The following table represents the time sequence for the research project. Pre-Testl Treatment Post-Test Oct. 23 Nov. 5-12 Nov. 11 Nov. 12 Experimental 0 X 0 Control 0 O 1. One-half of both the Experimental and Control populations received and took the pre—test by mail. 2. The same experimenter administered the post- test for both populations in a group setting. Internal Validity Stanley and Campbell suggest the following two shortcomings regarding internal validity of the non- equivalent control group design: selection—-maturation interaction and regression. These weaknesses are also applicable to the modified Solomon four-group design used in this study. The illustration provided by Campbell and Stanley (1963) for the selection--maturation interaction weakness is where a therapy group is contrasted to a "normal" group. They say, ". . . a gain specific to the Experimental Group might well be interpreted as a Spontaneous remission process specific to such an extreme group, a gain that would '33 have occurred even without treatment" (p. A8). This weak- ness is not applicable to the current study. The selection process for membership in either group was not differenti- able. The principle variable which needed to be controlled was intention to participate in training. This has been accomplished. Both the Experimental and Control Groups were comprised of members who voluntarily enrolled for training. The Control Group, however, did not have train- ing between the pre- and post—test, whereas the Experi— mental Group did. The second potential difficulty of this design is regression. However, extreme scores were not a factor in the selection of the groups and consequently regression toward the mean did not Operate or differentiate the two groups. An additional internal validity issue for this design is "history." In the prescribed design, the Control and Experimental Groups run concurrently. However, to achieve equivalence between groups, it was necessary to collect the Control Group data approximately a week prior to the Experimental Group. The groups did not overlap and communication between group members did not occur. The operation of "history" as an invalidity phenomena was not expected to influence the groups and there was no indication that it did. 3A In meeting or protecting for these shortcomings, Campbell and Stanley (1963) say that the study may approach true experimentation which was the objective in indicating pre-experimental sampling equivalence. External Validity Campbell and Stanley (1963) in connection with exter- nal validity discuss a principle shortcoming of the non- equivalent control group; namely, the interaction of testing and treatment, a difficulty which can be overcome by the elimination of the pre-test. This has been done, at least in part, by the utilization of the four—group design for this project. However, the complete elimination of the pre- test for the research would have seriously limited our ability to measure sampling equivalence and also the original groups were not randomly selected. Consequently, the implications or generalizations from the study will have to be restricted to the populations under investiga- tion or equivalent populations. However, the selection committees for the provinces contend they employed similar standards for selection. If; this is the case, we can generalize our results to other voluntary sensitivity training groups within the Episcopal Provincial System. This is especially true since it was established that no significant testing-treatment inter- action was operating. 35 Statistical Analysis The statistical analysis employed in the study was a 2 x 2 analysis of variance. Graphically, the analysis resembled the following: 01 X 02 O3 . 0A X 05 06 No Treatment Treatment Pretested 0“ 02 No Pretest O6 05 From the column means, one estimates the main effect of treatment; from row means, the main effect of pre- testing; and from the cell means, the interaction of testing with treatment. If the main and interactive effects or pre-testing are negligible, an analysis of covariance of 0A versus 02, pre-test scores being the covariate, will be utilized. The following two tables were constructed to provide additional information: (1) a table of means and standard deviation for each group, and (2) a table to summarize the biographical data for the groups. CHAPTER IV RESULTS OF THE STUDY Demographic Data The following demographic information was obtained from a personal questionnaire which each participant com— pleted during the group administration of the post—test. The information is presented to indicate demographic equivalence between the experimental and control popula— tions. The experimental population numbered seventy participants drawn from the immediate tri-state area including Michigan, Ohio, Indiana and Illinois. The control population was composed of participants from central New York state and numbered thirty-two. The mean ages for the control and experimental populations was 38 and 37 respectively. In both groups the female population was older than the male members. For the experimental population the average age of males was 33; the average age of the females was 39. The control pOpulation dis— played the same trend; males 36 and females A1. Table 1 summarizes additional demographic information obtained from the personal questionnaires. 36 37 TABLE l.-—Demographic data.* Experimental (N=70) Control (N=32) Marital Status: Married 7 86 78 Single, divorced 1A 22 Education: College graduate 77 72 H.S. graduate and some college 23 28 Ordination: Priest 37 53 Layman or laywoman 63 _ A7 *numerical values are percentages Pre-Testing Data Further indication of equivalence between the experi- mental and control populations was attempted by administer- ing a pre-test. The Tennessee Self-Concept Test was employed for this purpose. Table 2 summarizes the pre- test means, standard deviations and t test for both popu- lations indicating the groups are statistically different for several dependent variables. lkn:only were the groups essentially different but an additional unexpected result was that on several dependent variables, both populations declined, made slight in- creases, or remained constant between pre- and post-tests. The non-significant pre- post-test decline was evident for the following dependent variables: self-esteem, TABLE 2.--Pre-test scores for dependent variables. 38 Experimental Control Mean Sd Mean Sd t Self Criticism 38.71 6.07 39.92 5.51 0.60 Self Esteem 353.17 26.35 33A.08 26.56 2.15* Identity l2A.96 9.63 125.38 6.56 0.1A Self Satisfaction 112.35 12.12 100.8A 1A.99 2.62* Behavior 115.85 10.05 107.8A 9.10 2.AA* Physical Self 72.92 7.A5 67.30 7.66 2.1A* Moral-Ethical Self 73.57 6.63 67.30 9.A8 2.AA* Personal Self 66.53 6.26 60.31 8.27 2.60* Family Self 71.10 7.33 69.62 7.A6 0.60 Social Self 69.03 7.72 69.5A 8.50 0.20 * Significance level .05 self-identity, behavior, physical self, family self, social self, and moral-ethical self. For the following dependent variables the experimental population non- significantly declined while the control pOpulation made slight increases: self-satisfaction and personal self. For the dependent variable self-criticism the experi— mental population made slight increases and the control population non-significantly declined. The changes for the dependent variables are not significant. The 39 hypothesis self-criticism and self—esteem would increase as a function of sensitivity were nor accepted. Table 3 summarizes pre- post-test changes for the populations investigated. A statistical test was made to adjust for group dif- ferences as indicated by the pre-test. Specifically analysis of covariance, the pre-test being the covariant. Downie and Heath (1959) say, In effect, analysis of covariance adjusts the means for the effect of the uncontrolled variable and makes the necessary modifications in sampling error. The corrected sampling error is then used to test for the significance of differences among adjusted means" (p. 186). The first test in covariant analysis is a test for homo- geneity of regression which was not established for the dependent variable self-criticism. This means analysis of covariance is not appropriate for this particular dependent variable. The data is reported in Table A. Regression was parallel for the remaining dependent vari— ables; however, no significance was attained. The variable physical self being the only exception. Results are reported in Tables 5-1A. Ao No.0: mo.m oo.me om.e em.mt mm.ou mo.oH ea.me Ne.e mo.me caem Heeeom 03.0 mm.m mH.He ee.a me.me me.ou Ho.m oe.me mm.e oa.ae Lflem gasses me.o Hm.m mm.ae me.m Hm.oe mo.e Sm.m ee.eo em.e mm.ee caem Hmeompea. 03.0- mm.m He.me me.m em.ee He.ou .m.e me.me Sm.e em.me caem Heeeeemnfiesoz ma.on ma.m em.e© ee.e om.ae mm.H mw.m we.;e ne.e mm.me cflem Heeemaea ma.on e:.oa mm.eefi 9H.m em.eoa me.a- ee.HH NS.HHH mo.oa mm.mHH sea>eeem me.o ao.ma og.moa ma.efl ew.eea mm.o om.ma am.mfifl ma.ma mm.mHH cofieeeemaemm Laem em.ou mm.m em.mma em.e wm.mma me.o OH.HH ea.amfl me.m em.emfi seeeceefl mo.ou Hm.em mm.mmm mm.em mo.emm HH.on ee.am mm.mmm mm.em ea.mmm Eeeemm caem em.on mm.e eo.am Hm.m ma.eg mm.a :e.m mm.oe eo.e He.mm Emeeaeato caem .em see: .em see: em see: .em see: p emteuemoa emeeueea e emeeuemoa eweeueea Hopscoo Heeeeefieeexm .mdsonw an mocmLEMMHU pmmplpmoq Iona no somHLMQEOOII.m mqm