I...~. , ";z ‘ . . av. .~,|.‘. TY, 'PERSQNAL ‘ mvssr :QFZITHE' i r a IfG N A 3 ‘:i L ‘ ; a mum m, w a «i .m J 5x mm: magma a c All :EcERIAm This is to certify that the thesis entitled An hperimontal Investigation of Certain ABpectS of the Personality Theory of Karen Horney presented by James ‘w'ard t‘iilkins has been accepted towards fulfillment of the requirements for Ph.D. Psychology degree in Major professor Date flow ’4 /75:6 0-169 I- I”? V .4 R Y 4. ‘ “.‘|’~r Y. C ’ “O A!» ..K s: s. c . IV H ll. ‘1 $th 1' I ' f U nix/cats: 1y ~(‘I- m 5% 3.. 21‘ R W73!» AN EXPERIHENTAL INVESTIGATION OF CERTAIN ASPECTS OF THE PERSONALITY THEORY OF KAREN HORNEY BY James w; Wilkins, Jr. A THESIS Submitted to the School for Advanced Graduate Studies of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of 1 DOCTOR OF PHILOSOPHY Department of Psychology 1956 glagv-fl ACKNOWLEDGMENTS For their kind and thoughtful assistance in the preparation of this thesis the writer wishes to express his appreciation to Dr. G. M. Gilbert, chairman of the committee, to Dr. Alfred E. Dietze, and especially to Dr. Charles Hanley. VITA James W. Wilkins, Jr. candidate for the degree of Doctor of Philosophy rinal examination, June 11, 1956, 9:30 A.M., Office of the Chairman, Department of Psychology Dissertation: An Experimental Investigation of Certain ‘ Aspects of the Personality Theory of Karen Horney Outline of Studies Major Subject: Clinical Psychology Minor Subject: Personality and Social Psychology Biographical Items Born, August 1h, l92h, Warren, Ohio Undergraduate Studies, Biarritz American University, France, 19h6; Kent State University, 19h6-48 Graduate Studies, Kent State University, 19h9-50; Michigan State University, 1952-56 Experience: Staff Psychologist, Dayton State Hospital, Ohio, 1950-51; VA Trainee, 1953; Prison Psychologist, 195h-56 Member of Psi Chi, The Michigan Psychological Association, Midwestern Psychological Association; Associ- ate of the American Psychological Association f1, Abstract James H. Wilkins Jr. 1 This study undertook to assess through the application of systematic empirical procedures some of the concepts put forward by Horney on a clinical-intuitive basis. .A brief summary of Horney's theory was presented, which emphasized the three principal modes, or trends, of adjust- nwmt that she has described. These trends were regarded as generalized behavior patterns that find expression through an individual's moving toward, against, or away from.others, in reference to his interpersonal responses. Certain.relationships between these trends had been in- dicated by Horney. These were presented as indicating the general questions to which the investigation was addressed. Brief reviews of various methods for personality study were included. Consideration of the lack of specific relevance in available psychological tests resulted in a decision that such.tests would have to be devised for the study. Two scaling methods for the study of Barney's trends were described. One, a forced-choice set of scales, was dealt with.in detail. Preliminary investigations with this scale aided in developing its reliability to an acceptable range. Seventy-five married couples were tested with the devised scales. The four highest scoring persons of each.sex, on the three trends, provided a set of three groups for the experi- mental study of the validity of scale results in.terms of James W. Wilkins Jr. 2 Horney's three dynamic trends. The experimental subjects were submitted to a verbal participation experiment intended to evoke responses relevant to the theory and consistent with behavior patterns characteristic of the subjects. Three psychology graduate students were given training in order to rate responses of subjects tape-recorded in the verbal parti- cipation experiment. The adequacy of their judgments was assessed, which revealed a relatively high level of inter- rater agreement, except among the away-from sets of judgments. Results indicated that normal people, randomly selected, were found to be approximately normally distributed with respect to their scores on the scales depicting Horney's three vari— ables, except that scores were generally lower and less reliable with respect to detachment (moving away) than with the comp pliant (toward), and the aggressive (against) scales. High agreement with scale statements representing a particular trend was found to be predictive of behavior in the verbal participation situations, except detachment, which appeared to be predictive of hostile expression among females, or not predictive. Correlational study of the various scoring categories of the scales revealed that preference for compliant state- ments was negatively related with selection of aggressive statements. No reliable relation of either of these was found to hold with detachment. Differences between the sexes favored females over males with respect to compliance and the reverse James W. Wilkins Jr. 3 relation regarding aggressiveness. Intermate correlations revealed that mates tend to attribute similar qualities to one another, or to show no significant resemblance, rather than to be opposite (complementary) as expected in terms of the theory under test. I It was concluded that evidence in support of the use- Iulness of at least two of Horney's trends had been obtained, and that detachment, as she conceived it,appeared to be more applicable to men than to women. In general this trend pro- vided findings of less reliability and significance than the other two. TABLE OF CONTENTS PAGE I O INTRODUC TI on O O I O O 0 O O O O O O O O O I O l A. Horney's Theory of Neurotic Conflict . . . 2 B. Statement of the Problem . . . . . . . . . 7 II. THEORETICAL AND HISTORICAL BACKGROUND . . . . . . 10 A. Personality Dimensions . . . . e . . . . . 10 B. Personality Measurements . . . . . . . . . 15 C. Related Stu‘iies e e e e e e e e e e e e e 16 III. METHODOIIOGY . O O C O O . O O I O O O O O O O 20 A. The Construction of the Trend Scales . . . 20 Pilot Study of the Scales 23 Preliminary Study of the Validity of the Scales 27 Study of the Revised Dynamic Trend Scales 29 B. The Selection of the Basic Sample . . . . 30 Subjects 3O Selection Procedure 32 C. The Social Participation Experiment . . . 3h Procedure 3h 1 Apparatus 35 Experimental Instructions 37 Subjects . 38 Scoring 39 Derivation of Standard Participation Scores NZ IV C RESULTS 0 O O O O O O O O O O O O C O O O O I “A A. Results Obtained with the Dynamic Trend Scales oeeeee eeeeeeeoee ALL]. B. Results of the Social Parti ipation Experiment eeeeeeeeeeoeeee 50 TABLE OF CONTENTS (Cont.) V. DISCUSSION AND CONCLUSIONS . . . . . . . . . . A. Data Obtained with the Dynamic Trend Scales B. Implications of the Social Participation ‘ Experiment 0 e o o e e e e g C. Mate Differences and Similarities . . . . . z D. Other Theoretical Implications . . . . . . VI 0 SUWJARY O O O I O O O O O I O O O 0 O 0 O O I WCES . C O O O O O O C I O O O O C O O O O APPEMIX 1 O O O O O O O O C O O O O I O O O Q 0 mmlx 2 I O O O O O I O O O O O O O O O O O O APPEND Ix 3 O O O O O O O O I O O O O I O O 0 0 0 APPENDIX )+ O C C O O O 0 O O O I O O O O O O O . APPEINDIX 5 O O O O O O I O O O O 0 O O O O C O 0 APPme 6 O O O O O I O D O. Q I O I O O O C O O flPENDIX 7 O 0 O O O O C O O C O 0 O O O O O O . Page LIST OF TABLES mam ’ PAGE 1. The Results of the First (odd—even) Reliability Study of the Scales . . . . . . . . . . . . . 26 2. Test-Retest Correlations flor the Twelve Scores of the Dynamic Trend Scales (N=21) . . . . . . . 31 3. Intercorrelations among Judgments on Three variab le 8 O O O O O I O I I O O O I O O O O O Ll-3 h. Over all Correlation of Marriage Partners on Each of the Twelve Categories of the Dynamic Trend Scales O I O I O O O O I 0 O O O O O O O O “-5 5. Mean Scores and Standard Deviations of Husbands and Wives Separately on Each of the Twelve Scoring Categories of the Dynamic Trend Scales . . . . . h? 6. Intercorrelations of Self—rating Scale Scores with Other Categories of Scores for Husbands and Wives,Separately eeeeeeeeeeeoeee “.9 7. (a) Husbands Self-evaluations Correlated with Evaluations of them by their Wives; (b) Wives Self-evaluations Correlated with Evaluation of them by their Husbands . . . . . . 51 8. F Tests of the Sources of Variance . . . . . . . . S3 9. Average Standard Scores for Experimental Groups from Participation (P) and Dynamic Trend Scales (S)5h LIST OF FIGURES FIGURE PAGE 1. TYPOIOgical SOtheS o o o e o e o o e e e e o e 12 2. Arrangement of forced-choice scores . . . . . . . 2h 3. The experimental arrangement for social partiCipation O O O I O O O O O O O O O O O O 32 A. Form for raters' recording judgments as to category of S' participation responses . . . . ho - v u I w n n a . n n a a v a a o _ a a n n a q s u u n 1 . n a a a e u a . q u I o a o u e I. INTRODUCTION A survey of the literature reveals no empirical studies dealing with the personality theory of Karen Horney. It appears to the writer that references to her works (6, 7, 32, 51) are sufficiently numerous to justify such investiga- tions and that thereby significant contributions to current personality theory might be achieved. It was with this in mind that he undertook the present study. Best known, perhaps, of Horney's viewpoints is her classification of personalities according to trends of move— ment toward, against, and away from people. These trends, or modes of adjustment, are also referred to as compliance, aggression, and detachment. It is to the objective assessment of this aspect of her theory that this study is specifically addressed.* The present chapter deals with the following two topics in respective order: 1) Horney's theory of neurotic conflict, and 2) a statement of the research problems. In the following section a brief overview of Horney's theory is presented. For more detailed coverage the reader is referred to Newcomb (3h) or Munroe (32). *A study (31) reported during the preparation of this thesis does in fact specifically involve the operational use of these Horneyan concepts, but does not attempt their assess— ment, per se. A. Horney's Theory of Neurotic Conflict In brief,~Horney's personality theory is in many respects similar to Freudian psychology in that she espoused psychic determinism, unconscious motivation, resistance and the mecha- nisms of ego-defense (especially reaction formation, projection, repression). Horney's thinking was congenial to the importance of childhood experience in establishing the "basic" pattern for adult behavior. Her greatest disagreement with orthodox psychoanalysis is seen in her rejection of concepts that were grounded in biological and physico-chemical speculation. She was skeptical of the developmental sequence, oral-anal-genital, and antagonistic to the concepts of penis-envy and masculine protest. It is apparent that some of her arguments may reflect personal bias whereas others appear to have a more objective rationale. She denied that experiences are fundamentally sexual and accused Freud of a mechanistic, Darwinian approach in his use of polar opposites "a 19th century proclivity." (21) Horney proposed that the fundamental motive of humans is a quest for safety and security in a "potentially hostile world." Subsequently, the "basic conflict" in neurosis is an internal conflict between fundamentally incompatible "security systems." This latter formulation underwent some modification and her last statement (26) which said in effect that the begig inner conflict was between the "true healthy self" and the spurious neurotic idealized image a precipitate of the "security systems". There is considerable vagueness in her description of the "true self". Throughout the present study the emphasis is on her earlier formulations of incompatible neurotic "trends." Horney's trends appear to be acquired and generalized behavior patterns aimed at the alleviation of insecurity, or ”basic anxiety". The following quotations give general defini- tion to these three modes of adjustment. (26) "Moving_toward others." Group I, the compliant type manifests all the traits that go with "moving toward“ people...He shows a marked need for affection and approval and an especial need for a "partner" -- that is, a friend, lover, husband or wife "who is to ful- fill all expectations of life and take all responsi- bility for good and evil, his successful manipulation becoming the predominant task...Because of the indis- criminate nature of his needs, the....type will over- rate his congeiality....this type needs to be liked, wanted, desired, loved; to feel accepted, welcomed, approved of, appreciated....to be helped, protected, taken care of, guided....He tries to live up to the expectations of others"....etc. (23) This trend has also been called "The Appeal to Love" and will be referred to by various appropriate terms. Wflgging against others." The aggressive type takes it for granted that everyone is hostile, and refuses to admit that they are not. To him life is a struggle of all against all and the devil take the hindmost. His attitude is sometimes quite apparent but more of- ten it is covered over with a veneer of suave polite- ness, fairmindedness, and good fellowship....he needs to excel, to achieve success, prestige, or recognition in any form....he has a strong need to exploit others ....he regards all feelin s, his own as well as others, as "sloppy sentimentality ....love plays a negligible r019 e n (23) This trend has also been called "The Appeal to Mastery" (26), and is equally called the aggressive, dominant or as- cendant adjustment. "Moving away from" others. The third face of the basic conflict is the need for detachment....these people have an "onlooker" attitude toward themselves and toward life in general....there is an inner need to put emotional distance between themselves and others....a determination not to get emotionally in- volved with others in any way, whether in love, fight, cooperation, or competition...there is a striking need for self-sufficiency...he tends to shroud himself in secrecy....1ong term obligations are avoided if pos- sible....he will conform outwardly to avoid friction, but in hiw own mind he stubbornly rejects all conven- tional rules and standards....detachment and superior- ity are linked..." 2 ) This trend has also been called "The Appeal to Freedom" (26) which is often called aloofness, detachment or neurotic independence. These abridged descriptions are included to help the reader obtain the gist of their content without the need to consult the original sources (21, 22, 23, 2h, 25, 26). It should be noted that each "trend" embodies not only negative attributes but also desirable qualities that the possessor sees as virtues. The hallmark of neurosis is the compulsive utilization of these trends. The healthy person is assumed to flexibly oscillate between being properly loving and kind, assertive and efficient, independent and creative. Among her most declarative statements (23) are some that have the approximate status of hypotheses similar to those derived for the present study. These remarks are quoted in view of their relevance. Neurosis, it must be said, is always a matter of degree....I invariably mean, a person to the extent that he is "neurotic". su- Neurotic conflicts are concerned with the same general problems as perplex the normal person. ...any...examp1e of neurotic conflict would show an....incompatibility of conflicting drives and... their...compulsive nature. Awareness of conflict is a matter of degree in differentiating the neurotic from the normal. Regarding moving toward, against, and away from people....in a predominantly leaning and complying type we can observe aggressive propensities and some need for detachment. A predominantly hostile person has a compliant strain and needs detachment too. And a detached personality is not without hostility or a desire for affection. The predominant attitude, however, is the one that most strongly determines actual conduct. It represents these ways and means of coping with others in which the particular person feels most at home...that the potency of the submerged ten- dencies may be very great is evidenced by the many instances in which the attitude accorded predomi— nance is reversed.... From the point of View of the normal person there is no reason why these attitudes should be mutually exclusive. One should be capable of giving in to others, of fighting, and of keeping to oneself. If one predominates it merely means an overdevelopment along one line. It is not accidental that a conflict that starts with our relations to others in time affects the whole personality. Human relationships...mold the qualities we develop, the goals we set....the values we believe in. All these in turn react upon our relations with others and so are inextricably bound 0 I , The fact that under the conditions of our civilization this need for love obsession is more frequent and more apparent among women than men has given rise to the notion that it is a specifically feminine longing. Actually, it has nothing to do with femininity or masculinity but is a neurotic phenomenon in that it is an irrational compulsive drive. The aggressive type...is often drawn toward the compliant type - just as the latter is drawn toward him. ninnnnnnnn_lllin_ The detached person's goals are negative, he wants not to be involved, not to need any body, not to allow others to intrude on or influence him....only a limited number of tendencies in- trinsic to detachment as such can be formulated. Sometimes a person goes through periods of alternating between one extreme and another comp pliance to aggressiveness to detachment . Horney's thinking appears to be psychologically oriented. Her break from the structural and heriditary biases of ortho- dox medical psychoanalysis is incisive. Her ideas are thus largely in keeping with the general trend of American psychol- ogy. She emphasized the importance of social factors and the cultural milieu as determiners of personality organization and adjustment. This is compatible with the psychologist's tendency to View present behavior as an answer to immediate stimulus variables in terms often largely conditioned by prior experience or modification of the organism through learning, but not as a mere repetition of infantile response or solely dependent upon little understood constitutionally qualifying factors. As has been indicated, the particular aspect of Horney's ideas to be dealt with in the present paper is her notion that human behavior follows three major trends. For purposes of the study these trends or modes are considered to mean that all people may be seen as fitting one or another, or some combination of these patterns of behavior. In essence, this thesis attempts to fairly represent these aspects of Horney's theory in objective terms in order to determine whether they if 7 can be considered valid descriptions of normal members of this society. A further purpose is to investigate the merit of some of Horney's assertions about the interaction of her trends as they are manifested in mate selection or marriage. B. Statement of the Problem Horney's writings make lively reading and contain many descriptive passages with considerable "face validity." Her descriptions of inter- and intra-personal response are unfortunately not accompanied by suggested methods for tests of her hypotheses. Instead she carefully qualified the impli- cations of her views and resolved in artful and appealing literary style some seeming contradictions. However, the scientific validation of such views re- quires more rigorous procedures. A first step in this direc- tion might be the construction of appropriate measures of the dynamic trends that Horney described. The capacity of such measures to separate individuals according to the lines des- cribed may be accepted as evidence in support of the existence of these trends and their distinctive imbeddedness in different Personality structures. Our first attempt therefore is along this line, viz. to construct a test of dynamic trends toward Others, against others, and away from others. Two such at- teumts are herein described. A second approach might be the testing out in social or quasi-social situations of whether people do in fact behave in accordance with these trends. Accordiniy, an experiment was attempted in which the behavior of individuals could be assessed according to the directions toward, against, and away from others. The use of subjects who indicated charac- teristic use of a trend on the basis of the previously men- tioned tests provided not only a means for the experimental manipulation of variables, but also a means of exploring the validity of the tests. A third check on Horney's theory could be incorporated into the research design. This involves Horney's theory of the relationships of her dynamic trends to sex and mate selec- tion. Specifically her theory holds that individuals tending to be generally compliant are likely to select as mates per- sons who are more dominant or aggressive, and vice versa. She also asserts that ”in our culture" the compliant adjust- ment appears to be more associated with women than with men, and that the reverse holds for the aggressive mode. The use of married couples as subjects provides a means for checking out the correctness of this view. In summary, then, the following hypotheses were tested: 1. that a test could be constructed that would dis- tribute individuals according to Horney's dynamic trends of movement toward, against, and away from others, 2. that individuals who test high on a given trend will behave in a characteristic manner in quasi- social situations designed to elicit toward, against, and away from sorts of behavior, that individuals select mates in accordance with Horney's theory as it relates to mate selection and sex differences. 10 II. THEORETICAL AND HISTORICAL BACKGROUND Horney expressed the desirability of establishing a typology on the basis of the trends she described (23). Since in some respects at least her modes of adjustment might be so understood a brief review of the literature in this area was made. For more exhaustive treatments of the subject the reader is referred to any of several sources available, such as Stagner, Cattell, and others (1h, 19, 20, ’49). In view of the fact that concepts such as types, dimen- sions, factors, traits, etc. vary in their generality and specificity in the writings of different authorities onper- sonality theory the following section may only serve to point out the lack of agreement encountered in connection with them. In the second section of this chapter problems in connection With personality measurement are considered since they bare directly on the objectives of the current investigation. The third section reviews several studies reported in the litera- ture that are more or less closely related to the present in- vestigation either in content or method. A. Personality Dimensions Since typologies are probably the hoariest means of general- ‘ 121:“; about the attributes of man they are considered first in i r'— 11 _. sequence. What are types, and why have them? Stagner (uh) presents a summary of the substance of, plus some pros and cons regarding the issue of typologies: Type descriptions are...the technique of picking some outstanding feature of the individual and using that as a label for the totality. There are at least three different conceptions of psychological types....a) the mutually exclusive; b) the contrasting types /of a bi-polar continuum/;* and c) a multimodal continuum. (see Fig. 1 Among the notions favorable to type theories are: a)....it emphasizes the extent to which the parts of personalit (specific emotions, prejudices, traits and attitudes are determined by the whole. b) /Pre- diction of behavior may be more accurate if the judge allocates/ the personality to some appropriate type classification. c)....The greatest single value of type descriptions is /their serving as/ as anchoring points for reference frames regarding personalities. d) Types....call attention to certain processes in relatively pure form, uncontaminated by accidental and confusing factors. 0n the negative side, the following is cited from the same source: a) The very multiplicity of type theories.... belies their general applicability. b) Were there any universal system of typing which would fit a substantial number of cases or throw special light upon personality organization, it would be espoused b a greater number of psychologists than at present. c Each type theory represents the special interest of its inventor. The weaknesses of these criticisms respectively seem to be: a) a multiplicity of combustion theories did not prevent Priestley from developing one that was adequate to explain the data; b) general acceptance of any idea at any point in “Brackets mine a I : u ' I I . I . I b I l » I I a 3 t ‘_ c Fig. l. Typological schemes 13 time is no certain measure of its validity; and c) any psycho- logical, or other theory almost always represents the special interest of its inventor. Although some success has been achieved by Sheldon in attempting to demonstrate relationships between body build and temperamental traits (l, 9, 27, 42) these results are not striking nor do they lend themselves readily to interpretation. Psychologists generally are wary of this approach for at least two reasons. First, there is usually an implied mind-body dualism. And second, the irreversible determinism is incom— patible with the wide-spread atmosphere of pragmatic optimism in psychology. Not so alien to modern psychology, however, is the quest for some means of systematizing behavioral phenomena. In fact it may be said that in the early phases of the development of any science an adequate and consistent taxonomy would be desirable. In attempting to fulfill this supposed need Cattell (10), Eysenck and others have made empirical inVestigations aimed at uncovering consistent patterns of behavior. These researchers have used the well known methods of factorial analysis which they consider to be scientific and objective, it would seem, principally because their methods involve highly complex mathematical processes. Their general objectives seem to be the quantitative description of various traits, attitudes, temperaments, types, or dimensions of personality. Fe 14 _ Although the outcome of such studies may or may not have psychological meaning the method of factorial analysis appears to produce nothing not originally included in the measurement devices. It is possible however that certain general sources of variance may be isolated by these techniques. Regarding regularities in human behavior Cattell has suggested the fol- lowing: In typology there are two kinds of tasks: a) discovering what types exist in nature -- since a type defined by fiat is of little practical use; and b) defining to what extent a given individual belongs to the established types....Types are.... established by clinical forms of intensive obser- vation aimed at noticing repeating patterns....(lO) However, while the previously mentioned investigators point with pride to the establishment of "scientifically" respectable descriptions of the behavioral attributes of people their methods have come under serious attack. Along this general line of criticism McNemar has declared himself as fOIIOWs: Summarizing briefly, the factors in factoring behavior have to do with nabbing a small sample, ignoring other crucial sampling matters, treating the rotational problem irrationally, using tests of known unreliability, violating the requirement of experimentally independent measurements, pre- destinating the outcome, tossing in too much or not enough, choosing and ignoring tests when naming factors, struggling to make sense out of the re- sults, and varying all over the map in the use of hypotheses. (29) This brief mention of different authorities indicates that to date no definitely agreed upon method has been devised nnnnnnnn 15 for the satisfactory description of generalized traits or personality types, or "trends" as in the present study. It can be asserted however that a reliable and valid sorting of individual cases according to empirically estab- lished syndromes of behavior would be useful. B. Personality Measurements As is well known there are many available psychological tests. Appraisal of the usefulness of many of them however is disappointing. Whether it be a particular paper-pencil personality test, projective technique, or observational pro- cedure, the relative amount of variance that can be accounted for by its application is no cause for complacency with measurement methods in personality and clinical psychology. This generalization is based upon commonplace reports in the periodical literature dealing with the reliability and es- pecially the validity of these methods. Most contemporary in- struments cannot lead to the specific prediction of behavior based upon information obtained through their application. Generally they appear more adequate to a posteriori descrip- tion than to the task of prediction. One important difficulty with personality tests is that the "variables" singled out for attention vary from test to test. Were the clinician to take seriously the independent 16 existence of all the "factors", "trends," "variables", and "diagnostic categories", purportedly measured by such tests as the Guilford-Martin Test, the Guilford-Zimmerman Test, the Cattell 16PF Test, the MMPI, the Rorschach (depending on the system) and TAT, etc. he would find the task of psychodiag- nosis more perplexing indeed than it frequently appears. In view of the fact that the variables represented in the more commonly used instruments are not specifically rele- vant it appeared, as was mentioned, that a major task in "testing" Horney's theory would be to devise appropriate measures of the trends she described, that would be suffi- ciently reliable to justify their use. C. Related Studies Of greater relevance to the present study than either typologies or factorial techniques is the variety of person- ality and trait assessment procedures that have been employed. The classical study in this area is that of Allport and Vernon (5). Using types described by Spranger these investi- gators studied the validity of their scales by administering them.to selected groups (e.g. theological students and business- economics students, science and pro-med students). They found that the groups achieved relatively high scores on the value that corresponded to the particular type in question. 17 Rosenzweig and co-workers have worked out a triadic scheme of behavioral classification that is closely related to that of Horney. Efforts to demonstrate the validity of this conceptual scheme have not been highly successful. With a series of researches, G. W. and F. H. Allport have reported sex differences in response to their Ascendance-Submission Reaction Study. Validity problems were handled in a manner similar to that employed in the study of values. Prominent in Murray's personality motivational system (33) are concepts that closely correspond to those dealt with in the present study. For example, n succorance and n abase- ment compare with Horney's "moving toward"; n dominance, n aggression and n achievement with "moving against“; n autonomy with "moving away" from people. The validating procedure of studies in this framework vary in relevance to the present study. McClelland and co-workers have concentrated effort on the motives of achievement, affiliation and succorance by analyzing thematic material elicited from Subjects (Se) under controlled conditions while simultaneously varying independent variables such as 8' field of major study. Stephenson (MS) has applied "Q” technique to the study of Jung's "types". Three specific studies appear to be most closely related to the present study. The first is a series of investigations by Freedman, Leary and co-workers (33) who have also applied empirical methods within the guide lines of "neo- (or para-) 18 Freudian" psychology. The second is an "empirical elaboration of the theory of complimentary needs in mate selection" by flinch, Ktanses and Ktanses (AB). Third, is a recently re- ported study by Corsini (12) that deals with intramarital variables. The two former sets of investigations were speci- fically concerned with conceptualizations derived from the writings of H. S. Sullivan. The Winch gt 2l° studies will be dealt with further since their use of the theory of "complementary needs" in mate selection is especially pertinent to the present project. In essence, these researchers have martialed evidence in support of the notion that the motive clusters of marriage mates tend to be opposite. By this they mean that dominant persons tend to marry submissive ones, and sonon. It was felt that a study such as the present one would provide a check on these findings. While this manuscript was in preparation an additional study relevant to the above issue appeared. In this latter investigation Corsini (12) addressed himself to the problem of "Understanding and Similarity in Marriage". His data were obtained from a group of 20 married couples. Since data es- pecially relevant to this issue were obtained during the present investigation the topic will also be referred to again in a later section of this thesis (page 57 ). In contrast to most of the preceding studies (except 36) the present investigation is addressed to the assessment of l9 dynamic modes of adjustment in interpersonal relations. This objective is sought by means of a double barreled approach that includes two independent measures of the variables under observation, one of which seeks to measure the social, inter- personal correlates of questionnaire measures. 20 III. METHODOLOGY The material of this chapter is divided into three major groupings according to their chronological sequence in the investigation. These three steps are, respectively,, 1) the construction of scales to measure Horney's dynamic trends, 2) the selection of the Basic Sample, and 3) the development and application of a social participation experiment. A. The Construction of the Trend Scales An essential requirement is to fairly represent Horney's ideas in our measuring instrument. It is felt that this requisite was at least partially fulfilled by the following method: Her works were scanned for statements descriptive of her three trends of adjustment. Most of these statements were found in "Our Inner Conflicts - A Constructive Theory of Neurosis" (23), and "Neurosis and Human Growth" (26), the latter her final book. These volumes appeared to reveal most clearly her conceptions of three dynamic trends. The assortment of statements lifted from context was edited as little as possible, but as much as seemed necessary in order that they could be organized into parallel sets, representing each of the three trends, or some aspect of them. 21 This approach is similar to that suggested by Stephenson* as a method for testing personality theories by means of stort (AS). Two general types ofscaling procedure were considered. First, was the thermometer method of obtaining ratings. Second was a forced choice method similar to those employed by R. B. Cattell, Strong, and others, Both of these methods are described by Guilford (18). The aim in either case was to derive a measure of an individual's relative commitment to any one or combination of Herney's three trends. For the thermometer method two paragraphs were written describing each of the trends - six paragraphs in all. In each case the first paragraph emphasized the "virtues" of the trend, and the second its "vices". These sets of paragraphs were placed at the top of each of three successive pages so that each trend was represented separately on a single page. At the bottom of each page thermometer type scales were ar- ranged so that there werc two on each side separated by a series of phrases denoting degrees of similarity, from "extremely unlike" at the bottom, to "extremely similar" at the top of the scales. The order of the pages was randomized to offset possible position effects. The two scales on the left side were provided so that S could rate his similarity to each *The appiication of his entire method was considered, but it was decided that adequate results could be obtained through more direct means thus avoiding the internal complexities of the ”Q," method. 22 paragraph separately, and likewise that of the other person S ‘was requested to evaluate (see Appendix 1, page 77 ). The paragraphs were held to nearly equal lengths and apparent cogency. S accomplished the rating merely by marking one short line across each separate scale, four in all on each successive page. Scoring was accomplished by placing over the scales a matched transluscent sheet divided into intervals which were assigned numerical values. The forced choice Trend Scalesalso originated from the assortment of descriptive remarks taken from Horney's works. In this case they were assembled into groups of three each. Each set of statements in every case included one representing each of the dynamic trends. The sets were arranged so that statements representing each trend occurred nearly evenly in everypossible position and order. The initial version of the scales contained twenty-five such sets or items. In their construction an attempt was made to make all statements within any given item roughly parallel with respect to intrinsic desirability, intensity of expression, etc. The forced choice scales were administered by instructing S to deal with each item in the series separately, and to de- note agreement with one, disagreement with another, and to leave one statement unmarked in each item. Agreement was shown by marking an M, for most like, while marking L was expressive of being least like a given statement. Two columns 23 of blank spaces were provided, one on the left and one on the right side of each page. These were so arranged that one ap- peared on each side of any given statement as locations for the M or L ratings. The blank spaces on the left were to accommodate self ratings and those on the right the subject's evaluation of some other person (see Appendix 2, page 81 )o A typical example of a forced choice item.follows: Self Other I sometimes feel lost, alone, and helpless. I always try to match danger with courage and strength. I generally take a pretty detached view of life's risks. The scoring of the scales was accomplished by means of a key that denoted the trend from which each of the item statements originated. The M scores were obtained by simply counting the number of Ms that occurred in each category for both the self and other ratings. A similar procedure was followed to obtain L scores. These scores were then noted on the cover of each questionnaire in the order represented in Figure 2. Pilot Study of the Scales E administered the first forms of the two scales to an introductory psychology class (N=hl). Several difficulties immediately arose in connection with the thermometer scale. Self Other M L M L Toward Against Away Total Responses 25 25 ‘ 25 25 Fig. 2. Arrangement of forced-choice scores. 25 It appeared that ambiguity in the instructions had contributed to these, for in several cases subjects had drawn a single line across both scales on either side of the pages. Also noted was an obvious tendency in nearly all subjects to mark high similarity to all positive paragraphs, regardless of the trend involved and congruently to denote a more or less uniform lack of similarity to the negative paragraphs. These findings resulted in the decision to abandon further use of the thermometer scale. The results obtained with the forced choice scales led, on the other hand, to more hopeful expectations. Scores ob- tained by this method were analyzed using Pearson product- moment correlations.* Since scheduling problems precluded a second administration of the scales, its reliability was first studied by means of odd-even correlations. In Table l is presented the reliability coefficients based on the 3 between odd- and even-numbered items. The symbol system in the left hand column of this table follows the scor- ing scheme hitherto described and illustrated in Figure 2. One further comment regarding this figure will call attention to the "power" of the forced choice method in regard to the purposes at hand. Since S‘is free to mark an M and an L be- side two of the statements in each item set, he can be as con- sistent as he pleases in the selection of statements representing “For the total group, agreement with statements connoting compliance correlated -.76 with agreements with statements connoting aggression. TABLE 1 THE RESULTS OF THE FIRST (ODD-EVEN) RELIABILITY STUDY OF THE SCALES 26 Subscore r sMTo .58 SMAg .hh sMAw -.Q7 sLTo .h? sLAg .70 sLAw .196 0MTo °7u6 OMAg .667 OMAw .386 OLTo .55 OLAg .hhl OLAw .05 1 Standard error of zero g (N=ul): N-l = 0158 f .7 any of the three Horneyan trends. It is not possible for S, however, to get a high score on more than one trend, due to the fact that the more he distributes his choices among the different trends the fewer can he consign to any given category. If scores on each of the three trends were statistically independent of one another, Ss would be found who would be high on more than one trend. Whether or not to include them and how to include them in the experimental groups would be a difficult problem; one that is eliminated with the use of the forced-choice procedure. There were two outcomes of the pilot study with the forced choice scale that are worthy of comment in view of their re- lation to subsequent findings. First, ratings of another by a given S tended to be more consistent than his (or her) self-ratings. Second, there was an apparent lack of consis- tency in ratings involving the statements representing the trend of detachment. This latter outcome may have been due to an absence of detached persons in the sample, inadequacy of the statements, unreality of the variable presumably be- ing measured, some other source of error, or a combination of those mentioned. This problem will be considered further in a later section (page 65 ). Preliminary Study of the Validity of the Scales Since E had assembled the item-sets almost entirely on the basis of subjective judgments, it was considered appropriate to check on his assignment of statements to the three categories. 28 Accordingly, the individual statements were clipped from. one of the questionnaires, then were shuffled and put into a disordered bundle of 75 strips each containing one state- ment. A brief description of Horney's theory was then given to two graduate students of the psychology department at MSU. These students were asked to sort the statements into three groups of 25 according to the trend they felt each represented. The statements, thus sorted, were found to be in better than 90 percent agreement with E's designation. Though it was not felt on the basis of this outcome that the scale could be considered valid, some evidence of face validity was present. In view of the relatively thorough analysis to which the first form of the Trend Scales had been subjected, it had be- come apparent that extensive revisions were required. To better specify the items in greatest need of modifications, one further analytic procedure was used. The frequency and types of selections were tabulated for each statement in the scales. This simple method revealed that some statements were rated M by nearly all 85, whereas some other statements re- ceived almost no M or L marks at all. The statements were subsequently revised in varying degrees in accordance with this finding. The notions of social desirability and unde— sirability were kept in mind during this revision, although empirical study of the effects of these variables was not un- dertaken. The position order of the items was altered so that items among which both types of ratings had been most evenly 29 distributed were placed at the first part of the scale, on the assumption that their implications were less obvious to most" Ss. Against the possibility that it should again prove necesé sary to employ the odd-even method for the study of internal consistency of the revised scale, one further item was added, bringing the total to twenty-six. Study of the Revised Dynamic Trend Scales Since E was aware of the weaknesses of the first (odd- even) reliability study of the scales, arrangements were made for a group of volunteer Ss (N821) from.another introductory psychology class to participate in a test-retest study. Since these subjects, though about equally divided according to sex, were not all married, they were directed to follow the in- structions as were given (see Appendix 3) to subjects for the earlier reliability study. On the second occasion they ‘were requested to be sure to rate the'gggg "other" person. It is, of course, obvious that results obtained by the test-retest and split-half methods have their own special merits and shortcomings and cannot be held as comparable. Thus, the reliability coefficients presented in Tablsz are of a different nature from.those presented in Table 1. With- out furthor regard to the latter, however, Table 2 indicates that the Trend Scales, though not productive of impressively high correlations between scores of subjects on two occasions, one‘week apart, were nonetheless sufficiently consistent in 30 response evokation to permit its use. It could be further argued that analysis of interrelations between various cate- gories of response to the Trend Scales may thereby be justi- fiede B. The Selection of the Basic Sample From.Horney's frequent remarks to the effect that neurotic phenomena are continuous with patterns of behavior seen in ”normals," it was held to be reasonable that married student families would reveal this continuity. The revised Dynamic Trend Scales provided the measuring instrument, and it was accordinglyutilized for the collection of data from the Bash: Sample. Subjects The selection of Ss for the study proper was accomplished as follows: the revised scales were administered to a group of 75 married couples, aged 20 to 28 years, white, with no more than one child, who had been married from 8 months to 5 years. These couples were found among the two married hous- ing areas of the MSU campus. Seventy-five couples, or an N of 150, were arbitrarily selected as the size for the sample, for this number (in a nor- mal distribution) provides u persons of each sex, on each trend, who fall one or more sigma units gbgge the mean. These were intended to be the experimental Ss. TABLE 2 31 TEST-RETEST CORRELATIONS FOR THE 12 SCORES OF THE DYNAMIC TREND SCALES (N = 21) .eeu .793 .S9LL .700 .588 .389 .aue .906 .3514. .816 .719 9373 Estimate of probable G? = .013 32 Selection Procedure With the size and limiting attributes for 83' admissa- bility to the sampling distribution decided upon, and the second form of the Dynamic Trend Scales prepared, the popu- lation data were collected as follows: E proceeded from door-to-door (in the manner of the ambitious peddler) in a deliberately patternless fashion through each of the married student housing areas of MSU. The sample was in this manner about equally divided between old "barracks apartment" area, and the new "University Village" brick apartments. E hoped that no partially relevant selective factor could have en- dangered the singleness of these two subsamples. When E was admitted to an apartment and had determined through questioning that a given couple was eligible for in- clusion in the sample, he remained adamant, where necessary, in insisting that they participate in the "research project". This procedure netted a total of about 3 refusals. Very little reluctance was encountered, and many persons insisted that they were "happy to participate" for various reasons. This cooperative spirit in itself gave E some cause for alarm in view of his hypotheses regarding the attributes of the ag- gressive type of person. When a couple had agreed to participate E proceeded to explain in general terms that each S was admitting himself (or herself) to a "chance situation", in which roughly "one out of every six persons filling out the questionnaire (scales) 33 would be called for an additional experiment, to be conducted at the Psychology Building," within a few weeks." The state- ment that "chance" would determine their selection was a rationalization with a germ.of truth. 88 completed the scales as couples, in the presence of E, who discouraged any discussion of the items. E read the instructions aloud to each couple and then answered any relevant questions. Each scale blank was numbered in a se- quence from.l through 150, and each successive couple tested was given the next two progressively higher numbered blanks in the series. Husbands were given the odd-numbered, wives the even-numbered blank in each case. This procedure was emp ployed to avoid confusion in later separation and analysis of the completed scales. Each couple was promised by E that he would send them.a letter (Appendix 5) that would sump marize in general terms the nature and purpose of the research and suggest references should they have further interest in the project and its outcome. ‘With the twenty-six items on the revised scales, there was an exact total of 26 acceptances and 26 rejections by each 3, and a similar set by S's spouse. There were conse- quently two sets of responses presumably descriptive of each S (one subjective and one external). All items were scored by using the key that denoted the trend represented by each state- ment. Scores were tabulated under two groupings (self and other), according to two valences ("M", most like, and "L", 3h least like), on three variables (toward, against, and away). In other words, there were two complete evaluation groupings obtained from each S, composed of two subsets of three con- tingent scores. C. The Social Participation Experiment The central part of this project was the experimental investigation of correspondence between scale marking behavior and socially interactive behavior under standard conditions. It is assumed that if a person reveals a relatively strong tendency to approve of (mark M) items representing a given trend, say "toward", then the prediction would follow that he should be compliant, cooperative, etc., when confronted with other people. This would follow from Horney's ideas. The same could be said of agreement with items representing her other trends, that is, the "against" person should be socially dominant, aggressive, or hostile, the "away" person would appear aloof, stolid, bored, or withdrawn (passive). Procedure * A verbal participation experiment (a kind of auditory TAT) was designed to test reactions of life-like social situations. Twelve commonplace, socially frustrating situations *The design for this phase of the study was an elabora- tion and modification of a method suggested by John Teahan, a contemporary advanced graduate student in psychology at MSU. 35 were conceived, for example, "The Lackadaisical Nurse" (see Appendix A, page 101). These situations were transcribed onto electronic tape. E's voice described each of the twelve situations in neutral, rather expressionless tones. After each description a different voice made four successive and separate stimulus statements relevant to the situation, description immediately preceding. These four statements were written in a manner that attempted to anticipate a wide range of responses from Ss and to be equally appropriate to egch sex. An attempt was made to vary the "situations" in terms relevant to the theory under test, 1.6., the stimulus voices individually were members of a series of four groups, each including a "moving toward", "moving against", and "moving away" type (see Appendix M). These were internally randomized for position effects and sex (half were of each sex). Apparatus The apparatus consisted of two Revere magnetic tape re- cording devices (single speed, 3 and 3/h in/sec.). These devices were placed on a table, one (#1) facing S, the other (#2) facing E (see Figure 3). The former "played" the stimulus tape, the latter recorded both the stimulus- statements and subsequent response, but not the situation- descriptions. The recorders were manually operated (using a built-in mechanical stopping device) in such a way that while 36 Fig. 3. The experimental arrangement for Social Participation. 37 recorder #1 played the descriptive material, recorder #2 was inactive. With the onset of the first stimulus-statement recorders #1 and #2 were both in action. When the voice (from recorder #1) had completed its stimulus-statement to S, recorder #1 was stopped and E allowed S up to 60 seconds to respond, however very few latencies were that long. After the first response, recorder #1 was again activated to emit the second stimulus-statement and was again stopped, and so on until all four stimulus-statements for that particular situation had been emitted, and subsequent responses had been elicited; then, recorder #2 was stopped for the duration of the next description in the series, etc. Experimental Instructions Prior to the performance of the experiment E told each S which might be expected, in general terms, roughly as follows: "You have been selected, on the basis of chance, for participation in this, the second phase of the research project. I am going to ask you to be seated here, in front of this table. You will notice that there is a microphone in front of you, as well as two tape recoding machines. From the machine on your right you will hear descriptions of comp monplace social situations. After each situation is described a voice (now on the tape) will speak to you as though you were both present in the situation. Your "task" is merely to answer, or respond to the voice just as you would if you were in such 38 a situation. I know this seems artificial, but if you close your eyes and use your imagination I think you will find it quite easy to do. The voice in each case, will speak to you four times, and you are expected to respond each time. If you find that in some of these situations you would not re- spond, just say so. Try to pay as little attention to me as possible as I shall be busy manipulating the machines and following the material. I "Do you have any questions?" (Usually S did not.) "Then we shall proceed. First there will be two practice situ- ations to help you get the 'feel' of what you will be doing." Subsequent to these general instructions the stimulus tape was played through to the first "voice-remark" of the first practice situation. The recording machine was activated after the description and prior to the onset of the "voice" stimulus. Thus the recording tapes contained the relevant "situational stimuli" plus the audible aspects of 8' response, or participation. Subjects In order to test the validating.predictions, the four highest scoring persons of each sex on each trend were selected from the Basic Group. This made eight "high" persons on each trend, or twenty-four in all. The next highest persons in each case were also separated from the Basic Sample for possible use as alternate experimental Ss. Although data were collected 39 from all such alternate 83 it became necessary to make only one substitution of an alternate for an experimental S. The latter occurred in the Ag group of females. Scoring It has been noted (page 3h) that all experimental Ss were confronted with 12 recorded situations, each of which was intended to elicit h responses, or a total of AB in all. The raters in turn were instructed to evaluate (or force) each response in terms of the three trends (To, Ag, Aw). The evaluation of the social participation responses was carried out by E and two other advanced graduate students in, clinical psychology at MSU. Responses were scored in terms of the three Horneyan categories, plus a fourth category for unclassifiable responses (see Figure h). Each rater, or judge, was provided with a long (18 in. x no in.) sheet that included the population identification number and thus the sex of each experimental S, in a column down the left side of the sheet. Proceeding to the right across the scoring sheet by each S was a row of 12 grid sets, one for each of the‘ situations. Each grid set was comprised of h columns, one for each trend plus a fourth (no category), unidentified, column. The four rows of each grid were provided for the successive scoring of each response (total of h) in each situation. Utilizing the method just described the raters met on three occasions to listen to and judge the tape-recorded #0 .memnodmes soapsmHOthwd .m ho hsowopwo on ma mpdoEwndn museuooen.maopsa mom Shem .: .wfim .opo QGOHB GGOQB UGOHB UCOoHH. “yahvfie NH all a m m a gossasflm gospefiam mafia «spam 33m 3036 ’ v H 41' participation responses of all experimental Ss in consecutive order, according to the three Horneyan trends. The recorded stimulus and response material filled the upper and lower recording surfaces of four and one-half 1200 foot reels of Scotch magnetic tape. The total listening time was approxi- mately ten hours. Training of the judges was informal, and consisted of describing the three "trends", and reading brief relevant descriptive passages from "Our Inner Conflicts" (23). Prac- tice in making independent judgments was carried out by scoring the responses of S's originally selected to serve as control Ss but whose responses were no longer intended for inclusion in the analysis of the experimental data. The raters were thorough in their consignment of responses to the prescribed categories. This is indicated in the aver- age total of classified responses per subject, which was uh out of the possible h8, or 92 percent. Reliability of Participation Scores In order to assess the consensus among the raters inter- correlations were computed. The result of this analysis is reported in Table 7. Although the agreement among judges is generally high (all coefficients significant beyond the 1% level) there was generally higher agreement noted between raters g and b than between 2 and either of the other two. This inter-judgmental disparity is most apparant among judg- ments of responses on the third variable (Aw). file 1.2 Derivation of Standard Participation Scores Since there were three raters (scorers) there were, in consequence, three sets of raw participation scores on each of the three trends for all Ss. These raw scores were ob- tained by summing the check marks in each separate response category assigned by individual raters. This procedure re- sulted in three sets of raw scores for all Se in each response category. Standard participation scores were obtained by averaging for each S the scores obtained from the three judges. This averaging procedure increases the reliability of such scores. Since the judges differed in the frequency with which they assigned responses to the three categories, it was necessary to convert each judge's distribution of ratings on a given trend into standard score form using the familiar equation: ~ z = X - X 6' The three standard scores each S received on a given trend were then averaged to yield his Standard Participation Score. The final scores were obtained by application of the formula (2 + 3)10 to these averaged scores. #3 TABLE 3 INTER CORRELATIONS -AMONG JUDGMENTS ON THREE VARIABLES (To, Toward, Ag, Against, Aw, Away), BY THREE RATERS (Wilkins, a; Teahan, b; Martin, c) Variable Raters r TO a b .90“ a c .88 b C 079 Ag a b .85 a c .80 b C .79 AW a b 073 a c 053 b C 057 ufiAn g of .h70 is significant at the p .01 level. IV. RESULTS The results of the application of the methodology described in the preceding chapter arex'eported in two major sections of the present chapter. These are, respec- tively, the results obtained with the Dynamic Trend Scales, and the results of the Social Participation Experiment. A. Results Obtained with the Dynamic Trend Scales Statistical analyses of data obtained through the use of the Dynamic Trend Scales are presented in Tables h, 5, 6 and 7. As was mentioned previously there are twelve cate- gories of scores for the scale performance; six expressing self-directed judgments and six evaluative of the marriage partner. These in turn are subdivided into two equal groups, one expressing a most preference (M), and the other a least (L) preference for any given two statements out of the three each per twenty-six items. The presentation of the scale data follows the same system of notation presented earlier. Thus, M denoted ac- ceptance, L rejection, of item-statements. s denotes assess- ment in terms of self, and 0 denotes evaluation of the mar- riage partner. To, Ag, and Aw refer to the Horneyan trends f 15 TABLE h OVER ALL CORRELATION OF MARRIAGE PARTNERS ON EACH OF THE TWELVE SCORING CATEGORIES OF THE DYNAMIC TREND SCALES Scoring Category E sMTo .OZh sMAg .263* SMAw -.057 sLTo .033 sLAg oIEh sLAw .OOO 0MTo -.l2h OMAg 0033 OMAw -.l99 oLTo o115 oLAg .201 OLAw .205 *Significant at 5% level (.228). 46 being represented, e.g., toward, against and away from people, respectively. Since these data were obtained from marriage partners in the same room.at the same time the independence of their scale marking behavior was investigated. It was felt that these intramarital data could be regarded as independent if they failed to correlate significantly. That this turned out to be the case is revealed in Table h, which contains only one 3 (significant at the 5% level) which might possibly suggest "influence". This coefficient, , will be referred to SMAg later in connection with related findings. The means and standard deviations for the distributions of both husbands and wives separately are presented for comp parison in Table 5. N for each distribution is 75. There are several striking outcomes to be observed in Table 5. Some of these are as follows: The mean values for each of the six scoring categories of the g variety for a given sex is nearly equivalent to the values appearing in the g grouping of the other sex. Notable differences between the sexes are seen in regard to their relative preferences for statements describing trends To and Ag. Wives attained a mean value four full units high- er than their husbands on trend To, while the men exceeded the women by over three points on both s and o distributions for trend Ag. These differences are statistically significant well beyond the .001 level of confidence. Trend Aw showed LL? TABLE 5 MEAN SCORES AND STANDARD DEVIATIONS OF HUSBANDS AND WIVES SEPARATELY ON EACH OF THE TWELVE SCORING CATEGORIES OF THE DYNAMIC TREND SCALES Husbands Wives p Means S.D. Means S.D. BMTO 10.8 3 0914-1 1).).09 3 057(4- 0001 en‘s 8.5 3.807 5.2 2.569 .001 SHAW 607 ZQSBLI. 509 20MB .05 BLTO 7.0 3.181 5.2 2.607 .001 SLAB 8 08 3 0363 10 09 2.85“. 0001 810‘,“ 10.1. 20757 909 20587 nos. ONTO 114.02 14.0585 1002 14.018“. 0001 oMAg 5.8 3.890 8.6 n.1h2 .001 OMAR 6.0 2.353 7.2 2.7hl .01 OITO 507 3.159 703 3.395 .01 OLAg 10.u 3.615 9.0 3.321 .02 oL 10.0 2.536 9.7 2.977 n.s. h8 similar differentiation in terms of sex grouping but less significantly, and without significance in L judgments. Another aspect of Table 5 is the comparative variability. Husbands showed greater variability than wives with g judg- nmnts while wives showed the greater variability within the set of‘g judgments. These differences are not great, it is their uniformity that is noteworthy. There is only one ex- ception (oMTo) among the two sets of six categories. In Table 6 can be seen the Pearson product-moment coef- ficients of correlation between the self-ratings of Mggt like on the cmmpliant trend (sMTO), and all other categories of scores, for husbands and wives separately. In both sections of Table 6 further coefficients are as follows: the remaining two self-evaluative scores of the M type were intercorrelated with scores resulting from ratings of the marriage partner as Meat like each of the trends. The significant inverse relationship between assent to compliant vs. aggressive statements is probably most accurately revealed in the coefficient indicating the correlation between BMTo and sLAg. This is due to the contingency between scores within each set of three. It has been previously indicated that as choices accumulate under a given trend within a par- ticular set, M or L, the remaining scores in the set are cor- respondingly lowered. Thus, the best indicator of the likely negative relationship is apt to be positive coefficient .59h, com. eonwoamfiswfim Ho>oH Ratt mNN. ooseoawfidem Ho>oa MM# 9 seem. efi... ewe... u .H ease he mos.-**omm. ammm.- emo.u u a maze omo.n seem. mea.n ems. seesm. seams.u m:o.a oea.s ems. semee.- seems.u u e ease Qua: pepeaeasoo osoom mo>a3 was. aasm.u sea. a seam m moo. *pmm. #omm.| aswm.n u a «Sm a ease Spa: popdfleamoo eaoom mnnwpmsm 0N0. 00H. #w:m.u Nod. **:om. **:®m.: OHN.| Ono.l boa. **omm.l *tmnw.l 3d w< 09 no game mwdmm OBQm 3420 MAEO 0820 adzmo mszm .L eeeehnem meaaomepso wsfisoom wqmaamammm .mm>H3 eza mezammem mom ammoom so mmHmoemaeo mmmeo meH3.m3a_.s43< 92 34 .3233. as £538 20 .EE $05 328340 mamas ammHs use 2H mmmoom mq_up_Mea_ns_ _ _ _ _l2_._5 _ _ _53__ __ _ __l7_ _ Group Aw M q 23 21 SM r l? 26 5h 2 2 7 i 20 23 11:7 (alternate) 25 . 23 h? Means 22 2h 50 F u 0 AA 59 v 8 37 55 w 11 no #7 x 16 33 M7 (alternate) 16 he 39 ‘Means 10 39 #9 Group Means 16 31.5 h9.5 . 197 APPENDIX 7 Participation Scores of Experimental Subjects ‘High Scale Scoring Participation Response Type Subject Groups by Trend To Ag Aw Group To - M a 38 27 . 23 b 37 16 ' 39 c 3h 29 2h d 3h 20 3h (alternate)* M3 2h 23 Means 38 23 29 F e 32 31 2h f 30 29 33 g h2 21 28 h 52 15 an (alternate)* 33 27 32 Means 37 25 28 Group Means 38.5 23.5 28.5 Group Ag M 1 12 50 28 J 16 51 21 k 33 32 28 1 9 39 3h (alternate) 22 33 37 Means 20 Al 30 F m. 21 #3 23 n 2 1 2 o O 2 3 s»? (alternate 32 29 31 Means 31 29 27 __ _'lGropp M°€ES._ ‘_ _' _' _’25.5_' _' 35 g. _"-28:5'_ Group Aw 8 M 1 l 3 i? 3% 3: t i g; E (alternate) 29 51 Means 27 26 39 F : g2 a; i w x 25 33 (alternate 36 32 21 Means 27 36 26 Group Means 27 31 32.5 *These alternates were not originally intended for inclusion in the analysis except through the loss of an experimental subject. #The only alternate included in the variance analysis was a female replacement in group Ag. In this way the symmetry of the design was maintained. 1'11 “7 1M I'll: ' l I! H H 3129310 W! . 2.3.2..