ABSTRACT ACCURACY OF SOCIOMETRIC PERCEPTION AND ITS RELATION TO THE ACTUAL AND PERCEIVED DYADIC RELATIONSHIPS AND SOCIOMETRIC POPULARITY AMONG BOYS' GROUPS IN A CAMP by Ken Takeda The purpose of this investigation was to explore the relationships between the actual and perceived "like- dislike" between members in groups and the accuracy of the members' predictions of each other's feelings and socio- metric p0pularity. The subjects were forty-nine male campers, from three different cabin groups, and their ages ranged from eleven to thirteen. They all came from lower socioeconomic backgrounds and were referred to the camp by various social welfare agencies. The subjects were interviewed individually with the interview structured around three picture-sociometric-iike questions: One, the subject was asked to indicate his like, dislike and neutral feelings toward others; secondly, ~he was asked to guess others' feelings toward him; and third, he was asked to guess others' sociometric popularity. The accuracy of predictions was derived from the dis- crepancies between the actual sociometric variables (obtained from the first question) and the perceived or guessed sociometric variabled (obtained from the second and third questions).‘ ————'——_———1 Ken laKeda Fisher's exact probability test was run on the data from each individual. The median test was used on the data from the group. Both tests were used to determine whether a significant difference existed in the subject's prediction depending upon the presence or absence of a given sociometric variable. The significance of a series of Fisher's exact probability tests was tested with the use of the table provided by Sakoda, Cohen and Beall‘. The significance of overall combined results was tested with the use of the chi-square model reported by Jones and Fiskea. The equi- valent form reliability of Rho .72 was obtained. For validity Rho .69 was obtained between the sociometric rank order and two counselors' averaged popularity ranking of subjects. Both were significantly different from zero at the .Oi level of confidence. Some of the major findings of this research are as follows: 1. There is no significant difference in the accuracy of one's prediction of the other's feeling toward him whether it is accompanied by congruent feelings (one feels toward the other as the way he perceives the other feels toward him) or not. 2. The accuracy of one's prediction of the other's atti- tude toward him is moderate but it is significantly more accurate when he perceives himself as receiving a "positive" rather than "neutral-negative" choice from the other. one'a estim negat ;5.. When reciprocal feelings (one feels toward the other as the way the other feels toward him) exist between indi- viduals, one of them is significantly more correct in recog- ni zing the other's feelings toward him than when non- reciprocal feelings exist between them. rhu. The accuracy of one's prediction of the other's atti- ttade toward him is moderate but it is significantly more a1:curate when he perceives a ”positive" rather than "neutral- negative" choice from the other. 5. There is no significant difference in the accuracy of 0119.8 prediction of the other's papuiarlty whether there aJ'e congruent feelings between them or not. 6.. There is no significant difference in the accuracy of one's prediction of the other's papuiarlty whether one estimates himself as receiving a "positive" or "neutral- negative" feeling from the other. 7. There is no significant difference in the accuracy of One's prediction of the other's popularity whether there are reciprocal feelings between them or not. 3- There is no significant difference in the accuracy of 009's prediction of the other’s p0puiarity whether he is Chosen or rejected by the other. 9. There is no significant difference in one's prediction Of his own sociometric popularity regardless of whether he Perceives himself as enjoying a "high" or ”medium-low" sociometric status. l‘C'I Iuncua iC). The accuracy of the subject's prediction of the other's sc>ciomstric status is moderate but it is significantly more a<:curate when the other enjoys a "high" rather than "medium- iovv" sociometric status. I- Sakoda, James M., Cohen, Burton H., and Beail, Geoffrey. "Test of Significance for a Series of Statistical Tests," Psychological Bulletin, l954, 5i, i72-l75. 2. Jones, Lyle V. and Fiske, Donald w. "Models for Testing the Significance of Combined Results,” Psycholo- gigal Bgilgtin, I953, 50, 375-382. ACCURACY OF SOCIOMETRIC PERCEPTION AND ITS RELATION TO THE ACTUAL AND PERCEIVED DYADIC RELATIONSHIPS AND SOCIOMETRIC POPULARITY AMONG BOYS' GROUPS IN A CAMP by Ken Takeda A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY College of Education Guidance and Personnel Services l962 ‘\ (ya-I’ll i a), K 1"." ‘fi f fir. Dedicated to my parents Ken Takeda Candidate for the degree of Doctor of Philosophy F’inal Examination: May l4, l962, College of Education Thesis: Accuracy of Sociometric Perception and its Relation to the Actual and Perceived Dyadic Relationships and Sociometric Popularity among Boys' Groups in a Camp Outline of Studies: Major subject: Guidance and Personnel Services Minor subject: Social Work Biographical Items: Born: January l9, l932; Tokyo, Japan . Undergraduate Studies: Kwansei Gakuin University, l950-l954, B.A. Degree, March, l954. Graduate Studies: Kwansei Gakuin University, l954- l956, M.A. Degree March, l956. University of Toronto, i956-l95S, M.S.W. Degree, November, l958. The Merrill-Palmer Institute of Child Devel0pment and Family Relationship, l958-l959; Cornell University,_l959-l960; Michigan State University, l960-l962, Ph.D. Degree, June, l962. Experience: Graduate Assistant, Kwansei Gakuin University, l954-I960; Graduate Teaching Assis- tant, Cornell University, l959-i960; Graduate Assistant, Michigan State University, l960-l96l; Assistant Instructor, Michigan State University, i96l-l962. Professional Memberships: Canadian Association of Social Workers Japanese Association of Social Workers iii Eu of Co.- the of F li‘lt burst to Mr cm; SIUdy ACKNOWLEDGEMEN TS Sincere appreciation is hereby expressed to Dr. John E.. Jordan, for his kind guidance throughout the planning, eatecution and reporting of this research. Gratefulness is also expressed to Dr. Walter F. Johnson, Dr. David R. Krathwohl, Dr. John D. Krumboltz, and Dr. Myrtle R. Final, who supplied helpful suggestions. The writer is indebted to Dr. Kingsley Montgomery, clirector of Camp Oakland, for permitting the research to be held at the summer camp and making the necessary sched- Lnie adjustments, and to the camp counselors who assisted it: the arrangement of interviews and other phases of the irwvestigation. Thanks are also extended to Mr. John Paterson of the Bureau of Educational Research for assistance in selection Of’ statistical procedures and to Mrs. Georgia B. Reed of the computer Laboratory for assistance in programming. Special thanks go toMrs. Helga Kaplan for examining the writer's English. Gratitude is expressed to Dr. Irving Torgoff, Professor 01’ Psychology at Merrill-Palmer Institute, who inspired the wri ter in the selection of the research area and in the pursuit of the investigation. Appreciation is also expressed to Mr. Marvin Goodman for his constant encouragement and critical comments which enabled the writer to complete this study. Chapter II. III. IV. VI. VII. TABLE OF CONTENTS Page THE PROBLEM A. Statement of the Problem ...... .............l 8 Purpose of the Study.......................S C. Definitions. . .............. . ..... ...6 D. Limitations of the Study.. ......... ..... 7 REVIEW OF LITERATURE ........................ ..IO HYPOTHESES... ................. . ............... 29 RESEARCH PROCEDURES A. sat-ting Of the.StUdYOO0.0.0.0:00000000000031* 8. Research Population ..................... 34 C. Data Collection: Sociometric Interview...36 Preparatory Period. ...... ... ......... 36 Individual Interview ................ ..36 D. Sociometric Questionnaire. ...... .......37 E. Validity and Reliability Procedures ....... 38 Validity Procedure. ... .. ....... . ..... 38 Reliability Procedure. ............ 38 F. Treatment of the Data ................... ..39 Tabulation of Data on the Matrix Table ................................ 39 Combined Relationship Table .......... .39 Statistical Treatment ..... . ....... ... .40 RESULT-80.0.0.000.00.000.000...O000.00.00.0000049 DISCUSSION AND INTERPRETATION.................69 SUMMARY, CONCLUSIONS AND SUGGESTIONS FOR FURTHER RESEARCH A. Summary of Problem and Design... ........ ..94 The Problem........ .......... . ........ 94 Hypotheses .......... . ................ .95 The Subjects.............. ........ ....97 Collection of Data. ................... 97 Treatment of Data ..... .. ...... . ..... ..98 8. Summary of Findings....... ................ 99 C. Summary of Discussion...... ..... .. ....... l02 D. Reconsideration and Implication for Further Study .................. .. ....... lO7 V APPENDIXOOCOOOOOOOOOOOOOO.OOOOOO-OOOI ....... CO... ..... l'l BIBLIOGRAPHYOO0.0.0.0000.........OOOOOOOOO0.0.0.000...'5 vi NJ; LIST OF TABLES Table Page Analysis of Variance for the Subjects' Predictions of the Objects' Attitudes toward the Subjects.......... ..... ....... ......... 50 Analysis of Variance for the Subjects' Predictions of the Objects' Sociometric Status. ....... .... ....... The Number of Subjects Whose Predictions of Objects' Attitudes are Significantly Accurate and not Accurate in Relation to "Congruency" and "Non-Congruency?.... ..... . ....... 52 The Number of Subjects Whose Predictions of Objects' Attitudes are Significantly Accurate and not Accurate in Relation to the Subjects' Estimates of the Objects' Attitude toward them..... ......... .............. The Number of Subjects Whose Predictions of Objects' Attitudes are Significantly Accurate and not Accurate in Relation to "Mutuaiity" between them.......... ................ 56 The Number of Subjects Whose Predictions of Objects' Attitudes are Significantly Accurate and Not Accurate in Relation to the Objects' Attitudes toward the Subjects ........ 58 The Number of Subjects Whose Predictions of Objects' Sociometric Status are Significantly Accurate and not Accurate in Relation to "congruency" and "Non- Congruency" .................. The Number of Subjects Whose Predictions of Objects' Sociometric Status are Significantly Accurate and not Accurate in Relation to the Subjects' Estimates of the Objects' Attitudes toward them ............. 6| The Number of Subjects Whose Predictions of Objects' Sociometric Status are Significantly Accurate and not Accurate in Relation to "Mutuality" and "Non- Mutuality" between them ........................... 63 vii i4. l6. '9. The Number of Subjects Whose Predictions of Objects' Sociometric Status are Significantly Accurate and not Accurate in Relation to the Objects' Attitudes toward them ..... ............ ...... ................65 The Number of Subjects Whose Predictions of their Own Sociometric Status Are Accurate and not Accurate in Relation to their Self-Confidence ................................... 66 The Number of Subjects Whose Predictions of Objects' Sociometric Status are Significantly Accurate and not Accurate in Relation to the Objects' Sociometric Status ........................... ..... .... ....... 67 The Result of Fisher's Exact Probability Test for Individual Subjects for Hypotheses I to VIII and X (Group I) ...... . ..... . ...... . ..... 73 The_Result of Fisher's Exact Probability Test for Individual Subjects for Hypotheses I to VIII and X (Group II).. ........ ..............74 The Result of Fisher's Exact Probability Test for Individual Subjects for Hypotheses I to VIII and X (Group III) ...... ....... .......... 75 Relationships between Mutuality and 8's Prediction of 0's Attitude to 8 When S Perceived Positive Chohces from 0 (SS who showed si nificant accuracies in Hypothesis II?................... .......... .......77 Relationships between Mutuality and S's Prediction of 0's Attitude to 3 When S Perceived Positive Choices from 0 (33 who did not show significant accuracies in Hypothesis II)................................78 Relationships between Mutuality and S's Prediction of 0's Attitude toward 8 When 8 Perceived Neutral-Negative Chokzes from 0 (33 who showed significant accuracies in Hypothesis II)......................Bl Relationship between Mutuality and S's Prediction of 0's Attitudes toward S lW1en S perceived Neutral-Negative Choices from 0 (33 who did not show sig- xwificant accuracies in Hypothesis II).. .......... 82 viii 20. The Number of Subjects Whose Predictions of Objects Attitudes toward them are Significantly more Accurate and not Accurate in Relation to the Coexistence of TMutuality" and,"Congruency" ..... ..............85 ix j _ 0f kn uh '8 89c: f~ ear?! 4 on. 1. CHAPTER I THE PROBLEM A. Statement 21_the Problem ‘ Most of our lives are spent in what appears to be fairly well coordinated interactions with other people. We. choose the most apprOpriate way in which to act among numer- ous possible interpersonal situations. The relative smooth- ness of operation in our day-to-day living is a reflection of the fact that we are to a certain degree aware of what others do, feel, think and are ready to do. Asch (2zl39) says: To act in the social field requires a knowledge of social facts -- of persons and groups. To take our place with others we must perceive each other's exis- tence and reach a measure of comprehension of one another's needs, emotions and thoughts. The perception of others' existence and comprehension of others' needs is mainly automatic; thus we behave without knowing and thinking very much about the "principles" by which we operate. This is often in spite of levels of capacities and skills which we may have in appraising ¢>thers. We usually engage in the process without paying iflUCh attention to our role in it. Two factors underlie this ”automatic" process. First, we».simply gather various information through perception; secondly, we infer prOperties and potentialities of the perception which are not immediately evident. The process of perception has many aspects. In order to behave appropriately we may have to assess others' traits, intentions, feelings, attitudes, capacities, role- taking behaviors, and the situation in which the behavior occurs. We may need to know differences of roles or we may have to differentiate a temporary state from a more permanent one.- Tagiuri and Petrullo (53:xiii-xiv) distinguish three major elements in the perceptual process: There is the it ation in which the person to be judged is embedfiea. We do not need to "look" at the person who has lost a loved one to make a pretty good guess about how he feels. Then there is the pgrgon, apart from the situation. If we did look at h m or talk to him, we would probably be able to conclude -- at least within any one culture -- that he is sad, even without knowing about the circum- stances mentioned above. Usually, however, the cues from both of these sources point in the same direc- tion, thus increasing the likelihood that our judgment will be correct. But there is a third major source of variation in this system: the perceive; himself. He is selectively tuned to perce ve certain events in preference to certain others, and when the task is ambiguous enough, as complex events can often be, he will select and interpret the evidence accordingly. It'is from this third element that the researcher's hypothetical construct was developed. He questioned the ways an individual sees himself and others; how others actually perceive him; the actual and perceived relation- shit: between one individual (subject) and another (object); and how the accuracy of the subject's perception may be affected by the sociometric pOpuiarity of the object and the subject himself. fl!) "4*“. E ha is ill COT. rel The interpersonal relationship varies greatly accord- ing to the degree of social articulation between the sub- Ject and the object or the perceiver and the perceived. At one extreme we can think of a highly emotionally charged case; at the other, we can think of the subject "perceiving" a complete stranger. This level of emotional involvement or "distance" between the perceiver and the perceived is another important variable affecting his perception of others. In selecting relevant variables in human perception it is important to keep in mind that assessment of human environment begins and continues to occur in active inter- action with important others. Thus, the variables studied must have consequences for all interpersonal relationships. Sentiments of like and dislike between persons fit this criterion, but the interpersonal relationship involves many other dimensions as well. Why, then, does the researcher focus his interest on this particular fact of human rela- tions? It is because feelings of like and dislike are the coummm denominator of most interpersonal situations, a fact Irefiected in various areas of the social sciences. Tagiuri (53:3i6-3l7) describes: Newstetter and Felstein, nearly thirty years ago, suggested that social adjustment may be under- stood as the level of mutually satisfactory inter- action between individual and group and that this, in turn, is based primarily upon the group acceptance of the individual, on the one hand, and the indi- vidual's acceptance of the group, on the other. Homans, in his sociological theory of group behavior, selected actLyit , interaction, and sentiments of like and dis__ke for a systematic description 0?— human interactTa . Factor analytic studies of mutual ratings by members of small groups concur on the presence of three basic factors: influence and initiative, task competence, and like-dislike. While the first two are not applicable to every group, the last one is always present. From a methodological point of view, standard socio- metric procedures simultaneously provide two types of data about any member of a group: (a) information about his affective response to others, and (b) information about others' affective response to him. But traditional socio- metry does not include the member's perception of a relation- ship between two persons: his perception of others' feeling toward him or the feelings between two other members. One's attitude and behavior do not necessarily consist of objec- tive knowledge of the field but rather consist of specula- tion or inference from cues received from the object. Though sociometric data provides useful information, inter- personal behavior can be more fully understood if the subject's view is obtained. In dealing with "inference" or "speculation", the social scientist is concerned with how close the subject's "guess" is to "realityJ. Differences may exist between the subject's feelings and his expression of them but as such discrepancies cannot be satisfactorily resolved, the subject's description of his feelings will be used in this investigation. In present-day studies of "person perception" the fissue of accuracy is often the focus of interest. Unfor- tunately, however, "accuracy" has been a problem fraught with pitfalls, as Cronbach pointed out (l8). Consequently, the results of these studies have been inconsistent and difficult to interpret. The present researcher does not attempt to resolve all these problems, but rather hopes to add some light to the studies of "accuracy" in person perception. He investi- gates how accuracy of perception is related to the actual and perceived like-dislike relationships as well as the sociometric status of the perceiver and the person per- ceived. 8. Purpose 31 the Study The purpose of this study is to explore the relation- ship between some of the sociometric variables existing in groups and the accuracy of group members' predictions of them. More specifically the study concerns: l) the relationship between the perceived like-dislike of two group members and the accuracy of their pre- dictions of each other' s feelings for one another; 2) the relationship between one's estimate of another group member's feelings toward oneself and the accuracy of that prediction; 3) the relationship between the actual like-dislike of two group members and the accuracy of their pre- dictions of each other' a feelings for one another; 4) the relationship between one' s feelings for another group member and the accuracy of that member's pre- diction of one' s feelings' 5) the relationship between the perceived like-dislike of two members and the accuracy of their predictions of each other' 3 pOpuiarity; 6) the relationship between one' 3 estimate of another group member's feelings toward oneself and the accuracy of prediction of that group member's popularity; 7) the relationship between the actual like-dislike of two members and the accuracy of their predictions of each other' a popularity; 8) the relationship between one' s feelings for another group member and the accuracy of that group member's prediction of one' s popularity; 9) the relationship between one' 8 estimate of his own p0pularity and the accuracy of that prediction; ID) the relationship between actual popularity in the group and the accuracy of another' 8 prediction of that papularity. C. Definition; Subject: A person who makes a sociometric "choice" or who perceives his or other's socio- metric papularity and attitude. Obiegt: A person sociometrically "chosen" by the subject or one whose sociometric pOpularity and attitude the subject perceives. Attitude: A person's expressed or perceived socio- metric "choice" such as "positive choice", "rejection" and "neutrality". Congrugncy: The subject's tendency to predict an object's attitude for the subject as the subject feels for the object. For example, if A likes B, A thinks that 8 also likes A. Mutuality: Accuracy: Self—anfidence: D. The similarity of sociometric attitudes between the subject and the object. In other words,.a subject feels toward an object in exactly the same way as the object feels toward the subject. For example, if A likes B, 8 also likes A. The degree of discrepancy between a subject's prediction (guess) of his own or other's sociometric pepularity and their actual sociometric popularity, or the discrepancy between a subject's perception (guess) of an object's sociometric attitude and the object's actual sociometric attitude. For exam- ple, if A predicts B enjoys high socio- metric papularity while 8 actually has low or medium sociometric pOpularity, A's prediction is ggt_accurate. A person's prediction (guess) of his own sociometric popularity; in other words, a feeling of being chosen or rejected. Limitations g_, he Study The subjects of this study were boys ranging in age from edeven to thirteen years. They were from the oldest group in each of three two-week camp periods. They were chosen because it was felt that they were old enough to answer a rather complex sociometric-like questionnaire. However, the researcher does not have any scientific evi- dence to prove this point. The size of the sample used was forty-nine. Though statistics were run for the data from individual subjects, the lack of random sampling and the smallness of the sample size limits the generalization of findings. Furthermore, the homogeneity of campers' socioeconomic background may limit the possibility of generalizing the findings to different socioeconomic p0pulations. In this study the researcher was concerned mainly with the relations between the subject-object dyadic relationship and the subject's perceptual accuracy. There might be, however, many other possible related variables which are not considered in this study, such as age, family back- ground, intelligence, group experience in school and neigh- borhood, behavior, personality characteristics and the length 0f association, and counselors' influences. Subjects were to be unacquainted when they came to camp, This requirement was not completely satisfied, how- ever, because the iocal social welfare agencies sent more than one camper from the same community. Nine campers knew Path other prior to the camp period.. The camping session lasted for two weeks and it may be difficult to generalize the findings to permanent associations. The sociometric-like data was obtained through a single interview. Different results might have.occurred if the subject had been interviewed several times, though test- retest rank order correlation was as high as .82, signi- ficantly different from zero, at the .Ol level of signifi- cance in the pilot study. In the preceding paragraphs a general description of the rationale of the present study has been given. The (specific purpose of the study has been presented. The definitions of the terms used have been described. Finally, the limitations of the study have been discussed. In Chapter two, a review of the literature will empha- size the studies on empathy and sociometric perception. The tthd chapter presents the hypotheses to be tested in this.study. The fourth chapter contains the descriptions of the setting of the study, the subjects used, the method by which the sociometric data was collected, and the validity and reliability procedures. The steps in the treatment of the data will be explained. The fifth and sixth chapters contain an analysis of the data, an examination of the stated hypotheses in the light of this data, and a discussion of the results. A summary of the study, findings and conclusions will be presented in the final chapter. In addition to these points some suggestions, which might prompt further research, will be offered. E. . .Hfi ......ml'i 4 t; I" all” CHAPTER II REVIEW OF LITERATURE: Background g;,Thggry an Research 1_ t e Field Several converging lines of theory are involved in this study. First, the study originated from various theories of "self-concept" (l6, 33, 40, 4#, Si). Secondly, the study is stimulated by theories of interpersonal relationship (2, 32, 50). Thirdly, and more specifically, this research is motivated by empathy studies by Cottrell (l7) and Dymond (20, 2i) and by sociometric perceptual studies of Ausubel and his associates (4, 5, 6, 7, #6), Trent (65, 66), Boggarta (8, 9, lo), Campbell (l3), Taylor (6i, 62) and Tagiuri and his associates (l2, 37, 52, 53, 54, 55, 56, 57, 58, 59, 60). While the work of the first two groups of theorists are not directly related to the perceptual study of socioe metric dyadic relationships in a small grOUp, the last group «If researchers have more specifically contributed to the theory and method of this study. The researcher will re- view, therefore, only the relevant literature in empathy study'and person perception by the third group mentioned above. The "like-dislike" interpersonal relationship and its perception is a focus of social psychology, but is still relatively unexplored. Serious attention to sociometric IO perception began at the turn of the century with the develop- ment of the concept of empathy by Theodore Lipps (ii). The implication of theory in sociometric perceptual research, however, did not become possible until the devel0pment of the Freudian emphasis on insight and transference, the con- cept of role-taking (40), and particularly role playing (4i) and the concomitant impetus of the new "sociometric tests"(42). Moreno's concept of "tale” synthesized the development of the new sociometric test and is singularly important in this area of its conceptualization. It is interesting that "teie" has had at least three definitions. The first refers to the minimum transfer of meaning between two persons, namely more awareness. The reference, thus, is to the minimum condition for interpersonal relations, and the focus is on process and mutuality. The second defini- tion refers to the increased rate of interaction between members of a group. The third definition is more explicit andis best identified as "two-way empathy." This emphasis on mutuality characterizes Moreno's approach to the social through the psychological, or more accurately, the analysis of the psychological in the situational context. A scheme for the empirical observation of empathy was deveIOped by Cottrell (l7) and Dymond (20, 2|). The latter proposed a scale for the measurement of empathic ability. The instrument was made up of four parts, each containing the same six items. In the first part the individual was asked to rate himself. In the second part he was asked to l2 rate some other individual. In the third he was asked to rate the'other individual as he believed this other would rate himself. In the fourth he rated himself as he thought the other would rate him. This scheme used ratings rather than sociometric questions. The essential difference between the interest in sociometric perception and empathy as de- ' scribed by Dymond, is that those emphasizing empathy take as their datum the discrepancy between the "actual situa- tion" and the perception of it, while workers in social perception are interested in the inter-relationships of perception and actual situations. A unique study of "tele as two-way empathy” in actual situations was made by Toeman (63, 63). A double is an auxiliary ego who attempts to co-experience with the subject in "situ". Toeman prOposed a scheme for observation and tried to measure the accuracy of double empathy and tele ability. Most empathy studies have dealt with the accuracy of perception of ratings of personality characteristics rather than with accuracy of perception of sociometric relation- ships. However, much experience has been reported which is of interest to the student of sociometric perception. Probably the prime reason for inclusion of the empathy approach here is the twofold criticism which has deveIOped: on the one hand, it has been domonstrated that empathic ability corresponds closely to the "conventional" or normal patterns of response, and that a high empathy score may '3 occur because "...conventional people get good scores on empathy tests because most of their partners (or referents) in the test are also conventional” (38). Projection, appar— ently, accounts for the relationship (38). This immediately throws doubt on approaches that deal with the identification of (stereotype) response patterns as measures of empathy rather than conventionaiity (49).. There remains, of course,_ the question bf empathy's existence. The more recent writers cautiously agree that it exists (29, 3B). A second criticism is implicit in Gage's (29) suggestion for the use of "stana dard persons“ in the measure of empathy analogous to the use of the "auxiliary ego" in psychodramatic research (4i), and his questioning whether the accuracy of perception of strangers can be carried over in interpretation to non- strangers. This questions the comparability and meaning of ratings when certain factors of interaction are or are not held constant. This same problem has been raised by Moreno and others regarding sociometric choice scores and other types of indices. Recently, Ausubel and his associates (3, 4,.5, 6, #6) have concerned themselves with an area they call "soci- empathy" which focuses on the use of questions concerning sociometric perception and sociometric choice reports. They asked high school students to rate each other on a five point acceptance and rejection scale, predicting how every other student would rate them on the same scale, and predicting the sociometric status each classmate would I4 receive on the scale. The term ”sociempathy” was intro- duced to describe a particular variety of social perception ,- "an individual's awareness of his own and others' socio- metric status in a given group of which he is a member" (sun). I) 2) 3) Some of their findings were: Sociometric status of the perceiver was not sig- nificantly related to accuracy of perceiving own or others sociometric status. In the case of girls, however, a slight but significant degree of positive relationship prevailed between accuracy of perceiving own status and the socio- metric status of the perceiver. The ability to perceive own sociometric status and the ability to perceive others' sociometric szague were completely unrelated in these studies : 3 . The perceptual response set measures derived in these studies are sufficiently stable and have sufficient generality over individuals and over related judgmental tasks to warrant their use for purposes of individual prediction. These measures may be interpreted as generalized per- sonality trends within an individual, i.e., as indices of typical, self-consistent modes of perceiving the interpersonal and hierarchical aspects of social situations. a) Self-underestimators perceive themselves as neither very accepting nor as very accept- able persons. They tend to be better ad- justed than self-overestimators and to modify their aspirations for future per- formance more readily and realistically in line with objective experience. They are more accurate in perceiving their own sociometric status, but have less self- regard. b) Self-overestimators, on the other hand, perceive themselves as highly accepting individuals and as on the receiving end of interpersonal relationships. They set and maintain higher aspiration levels than are warranted by their actual performance ability. They are more strongly motivated by strong needs for success and for acceptance by '5 others; yet despite this fact they are less accurate in their perceptions of others. c) Whether perceptual judgments tended to be extreme or in the middle range of the dis- tribution was found to be related to moti- vational orientation. Extreme judgments appeared to be reflective of strong sub- jectively oriented needs, whereas non- extreme judgments were more typical of subjects who responded more in terms of group norms (46:226-227). More recently Ausubel (6) investigated to what extent an individual's sociometric status in the group affects both the accuracy with which he perceives his own and others' status, and the accuracy with which his own status and his sociometric acceptance of others can be perceived by other group members. Conversely, he studied to what extent sociempathic ability and relative transparency of sociometric attitudes may affect sociometric status in the group. He found that ability to perceive the sociometric ratings received from others varied directly with the sociometric status of the latter (i.e., the subjects per- ceived) and accuracy of perceiving the sociometric status 0f others also varied directly with the sociometric status 0f the latter (i.e., the subject perceived) (4:83). Trent (65, 66) investigated the interrelationships between anxiety and accuracy of perception of self and Others for sixty-three institutionalized delinquent boys. Accuracy of perception was measured by discrepancy scores r'elai'esenting the sum deviation of how others would rate l6 the individual on a five-point acceptance-rejection scale as compared to the individual's actual rating. Accuracy of perception of others' status was determined by a discrepancy score which was the sum of deviations of the individual's sociometric status ratings of others on a five-point scale, as compared to the other lndividual's actual rating. Anxiety was measured by the Children's Form of Manifest Anxiety Scale. It was found that there is no relationship between the accuracy of perception of self-status and the accuracy of perception of others' status. It may be considered that the ability to perceive one's own status and the ability to perceive the status of others are separate skills. They involve distinctly different perceptual processes. Or perhaps, real relationship between perception of one's own and others' status is contaminated by pervasive influ- ences such as the desire of subjects to react to others in a manner deemed apprOpriate within the particular living condition, or the interaction of individual estimates with reference to group estimates (65). The findings of Ausubel and his associates as well as Trent's are related to a large body of literature indicating that perception is largely influenced and distorted by the i166d8 of'the perceiver and by his emotional attitudes toward persons, objects, and events in his perceptual world. Evidence has been accumulated, for example, that better adjusted and more intelligent, insightful and self-accepting l7 individuals are more apt to rate themselves and others realistically, i.e., in closer accord with group rating (34, 47). Furthermore, persons rated high in a given trait are more accurate in judging themselves and others in rela- tion to this trait (l5, l9). Frenkel-Brunswik (26,27) has described a large number of mechanisms of self-deception (e.g., "distortion into the Opposite", "exaggeration”, "omission", "rationalization”) which Operate to reduce the accuracy of self-perception. She concluded that we do not always see ourselves as we are but instead perceive the environment in terms of our own needs. Self-perception and perception of the environment actually merge in the service of these needs. Thus, the perceptual distortions of ourselves and the environment fulfill an important fun- ction in our psychological household (27). After establishing that subjective factors played an important role in the evolution of interpersonal perceptions, the next relevant question was the influence which various aspects of these perceptions (i.e., "accuracy", "assumed similarity") exert on social behavior and its effectiveness. Gage (30) reviewed the inconsistent results in various studies of the relationship between (a) differential ability to predict the attitudes of others, and (b) sociometric status and leadership ability. Perhaps the crucial factor is that superior perceptual acuity enhances social effec- tiveness only when perceptual superiority is manifested in those specific areas impinging upon the structural or l8 functional prOperties of a particular group (l4). The positive aspect of this factor of relevance was strikingly illustrated in the studies of "assumed similarity" by Fiedler (23, 24, 25). i Borgatta (8, 9, l0, ll) studied extensively the inter- relationships within and among various classes of measures such as the observation of interaction, affective choices reported and expected by persons, leadership ratings of self and "buddies", leadership ratings given by superiors, intelligence measures, two projective measures, three items of identification, age, education and rank among l26 air force enlisted personnel divided into fourteen groups of nine persons each. Among the measures which are related to the choice- rejection interactions, a strong relationship was found between the tendency to choose and the tendency to expect to be chosen. No appreciable relationship was found between papularity and the tendency to choose. A significant relationship was found between pOpularity and the expec- tancy that one will be chosen. Borgatta concluded that at least two factors were Operating in choice behavior: (a) There is accurate perception of social situations: and (b) there is projection (in the non-analytic sense) or recipro- city between the expectation of being chosen and choosing (II). Campbell and Yarrow (i3) examined the ways in which processes of interpersonal perception and action are related '9 to children's successful and unsuccessful functioning in peer groups. They systematically observed and interviewed 26D pre-adoiescent children at the beginning and end of summer camp sessions to obtain detailed information on the children's perception of their peers, their actions, and reputable measures of social effectiveness. They found little evidence that children who differed in social effectiveness differed systematically in the con- tent of their views of others: yet they did differ signi- ficantly in interpretative quality of their perceptual reports. Appraisal of interactional data showed that acceptance by peers was associated with behavior patterns indicating greater freedom of action. . Campbell and Yarrow tentatively concluded that the simultaneous consideration of both the subject's behavior and the qualitative properties of his perception would lead to a more refined prediction of social valuation (l3:l8-i9). Taylor (6i) considered a three-dimensional frame of reference for emotional interactions in small groups. He asked the members of therapeutic groups and student classes to rank their companions and themselves with regard to their popularity. The data suggested that while group members had approx- imately the same knowledge of others' dominance status as of their own, they were generally less aware of their own papularity than that of others (6l:656). 20 In the second set of questions,.Taylor requested the subjects to rank their eUbjective preference for their partners in the groups. 'The data indicated that the group members understood and obeyed the different test instruc- tions, and were able to distinguish between the introspective assessment of their own interpersonal feelings toward group partners and their objective estimates of their public status (6l:658). The sociometric questionnaire was also applied to rneasure the subjects' love-hate feelings toward each other. 111s data indicated that group members were influenced in tateir feelings for group companions by their conscious or tueconscious awareness of the papularity which these com- panions publicly enjoyed in the group, and that this irrfluence was so marked that it almost obscured the fact 'that;genuineiy personal feelings, unaffected by partners' FHJblic status, did exist in many dyadic relationships (Si :659). Taylor devised rank matrix of guessed self-appeal (i .e., his estimates of the feelings or preferences he believes himself to have aroused in his partners) and measured the degree of interpersonal friendliness which "as attributed to individuals by the group members generally. He also devised a matrix of guessed self-appeal. The r'fisults of the two kinds of matrices were combined for At‘lzributed Interpersonal Friendliness. Interestingly, the relationship between Attributed Interpersonal 2| Friendliness and Publicly Exhibited Friendliness was not found to be as close as the relationship between Inter- personal and Public Popularity. This led the investigator to conclude that group members based their guesses chiefly on consideration of their dyadic relations and failed to take adequate account of any public phenomenon which -- as in the case of love-hate feelings -- might have introduced some semblance of agree- ment into the divergencies of their dyadic scores. What- ever general friendliness a partner exhibited to the group as a whole, seems to have impressed his companions less than the emotions he displayed -- or was felt to have displayed -- in his private dyadic dealings (6l:662). In the later report the same investigator concerned himself with a distinction between the recognition of emo- tions in others generally and the recognition of those emotions when directly intended for the subject. He also concerned himself with the difference between global and dyadic judgments of others. It was found that unpOpular persons tended to be quite unaware of their dyadic appeal. It was speculated that the unpOpularity of these peOple was caused by their misjudgment of the feelings of others toward them, and by the gauche and ill-adapted behavior which their misjudgment may have provoked. But it may also be that both unpOpularity and ignorance of dyadic self-appeal originated because the person concerned had been unable, for one reason or another, to establish suffi- 22 cientiy close and congenial relations with his group com- panions (62:67l). In their earlier studies, Tagiuri and his associates (39, 52, 53, 54, 55, 56) concerned themselves with investi- gating whether members of a group can perceive their feel- ings for each other more accurately than might be predicted by chance; how their perception of others' feelings toward them related to their own feelings for others; whether the mutuality of feeling in a group is higher than would be expected by chance; and how these phenomena of "accuracy", "congruency", and "mutuality" relate to each other. In more recent studies (37, 53, 57, 59) Tagiuri and his associates concerned themselves with the manner in which the preferences of the members of a group for each other are known to their companions, what the members' distinctive behaviors are, and what factors determine the degree to which the sociometric choices of a particular member would be known. Observations were made on some sixty well-acquainted groups, ranging in size from six to thirty-five members, the average age of members varying fnom eight to forty years. The groups, composed mostly of then, were of several kinds: naval crews, summer campers, seminar participants, semi-therapeutic groups. All the lled conditions are suitable for examining separate components of the process. The present researcher emphasized the former situation to satisfy his research design and purpose. Camping, which' prrn/ides twenty-four hours of infOrmai living for children, was considered an adequate (or a genuine) real-life situa- titan. The sociometric-like questionnaire was constructed \u 0\ so that the necessary components of the interaction process were revealed in sociometric matrix. C. Data-Collection: Sociometric Interview Once the researcher determined the variables and the situation, his approach to the collection of data was rele- tively simple. He depended on the conscious or verbalized preferences of an individual for others in the manner made familiar by sociometry, and then proceeded to ask the sub- ject about his perception or evaluation of others' feelings toward him and their sociometric status. l. Preparatogy Period: Before starting the sociometric interview with the individual subject, the researcher Spent about ten days with the subjects in the camp setting so that they felt at home with him. The informal singing, swimming, camp fire, cook-out and various athletic programs provided Opportunities for the researcher to establish rapport with them. 2. Individual Interview: Each sociometric interview was conducted individually with the subject and the interviewer * facing each other at a table. The questionnaire was handed to the subject. After the researcher explained the purpose of the interview, the subject was given the pictures of his cabin- mates and asked to sort them into three differently colored boxes according to the instructions. The red box was for *See Questionnaire in Appendix A. "positive choice" and "high sociometric status". The blue box was for "negative choice" and "low sociometric status". The white box was for "neutral choice" and ”medium socio- metric status". Each picture was given a number on the back to identify the name Of the camper The interviewer read the identical questionnaire consisting of three questions to each subject. After each question, the subject was asked to tell the interviewer what he was supposed to do so that the interviewer was certain the subject fully understood the nature of the question. After the subject completed each question, the inter- viewer took out the pictures from the three boxes and capied the number on the back of each picture in the apprOpriate place on the interviewer's questionnaire. The length of interviews ranged from ten to twenty- eight minutes with an average of fifteen minutes. D. Sociometric Questionnaire A picture sociometric-like questionnaire was prepared ‘UD obtain the information on five variables: the subject's sociometric attitude, the object's sociometric attitude, the subject's perception of the object's sociometric zattitude, the object's sociometric status and the subject's perception of his own sociometric status. The researcher first asked the subject to sort the picrtures of his cabin-mates into three differently colored 38 boxes, namely those who he likes, dislikes, and those toward whom he feels neutral. The next step in the procedure consisted of obtaining from each subject the names of group members who, in his Opinion, like, dislike, and feel neutral about him. Since the researcher asked initially, "Put the pictures of boys whom you like into the red box, those you don't like in the blue box and those you don't particularly care whether you like or dislike in the white box," he now asked the subject, "Put the pictures of boys you think like you in the red box, those you think don't like you in the blue box, and those you think don't care whether they like you or not in the white box." The third step consisted of ob- taining the subject's guess of the pOpularity ranking of the rest of the group members as high, medium and low. E. Validity and Reliability Procedures l. Validity Procedure: The researcher looked for an outside criterion by which to judge the validity of the subject's sociometric answers. For this purpose two coun- selors were asked to rank the campers in terms of pOpularity with the group. The two different ranks were averaged. The Spearman rank order coefficient of correlation was obtained as .69 significantly different from zero at .Ol level of significance in the pilot study. 2.. Reliability Procedure: The researcher used Equivalent Form Reliability in the pilot study. Set A and Set 8 of 39 the questionnaire were measuring the same variable. The Spearman rank order coefficient of correlation of .72 was significantly different from zero at .Ol level of signifi- can CE . F. Treatment 21_ he Data l. Tabulating Data on the Matrix Tables: There were three levels of questions in the questionnaire: the subject's attitude, the subject's perception of the object's attitude toward the subject, and the subject's perception of the sociometric status of the object. These were plotted separately in three N X N matrix tables. Each matrix table contained the data for each cabin group To simplify plotting, the figure "3" was arbitrarily given to "positive choice" and "high sociometric status". Figure "2" was given to "neutral choice" and "medium sociometric status". Fig- ure "l" was given to "negative choice" and "low sociometric status". The first matrix table indicated the subject's atti- tudes and his actual sociometric status. The second matrix table indicated the subject's perceptions of others' socio- metric status. The third matrix table indicated the subject's perceptions of others' sociometric status 2. Combined Relationship Igng: Three matrix tables were transformed into a Combined Relationship Table. The table contained the subject and the object in dyads in the left hand column. The columns indicated, from left to right: 40. l. The subject's attitude toward the object. 2. The object's attitude toward the subject. 3. The object's attitude toward the subject per- ceived by the subject. 4. Discrepancy between (2) and (3). 5. The object's sociometric status perceived by the subject. 6. The object's sociometric status. 7. Discrepancy between (5) and (6). Figures " ", and "2", and “l" were again used to indicate sociometric attitude and sociometric status. There were three combined relationship tables as the pop- ulation consisted of three cabin groups. In dealing with the actual sociometric status and the perceived sociometric status, the researcher arbitrarily used only "positive choices". If a subject received or was perceived as receiving "positive choices" from more than two-thirds of the group members, he was defined as enjoying a "high" sociometric status. If a subject received or was perceived as receiving "positive choices" from more than one-third and less than two-thirds of the group members, he was defined as having "medium" sociometric status. If a subject received or was perceived as receiving "positive choices" from less than one-third of the group members, he was defined as having "low" sociometric status. 3.. Statistical Treatment: To test the first eight and the tenth hypotheses, Fisher's exact probability test was Inga“ i .» In”. LE '57.»! 4: run on the data for each subject. The median test was run on the data in Hypothesis IX. There were nine kinds of relationships between the sub- ject and,the object in terms of the subject's attitude toward the object and the object's attitude toward the subject. They were: Positive-positive, positive-neutral, positive- negative, neutral-positive, neutral-neutral, neutral-negative, negative-positive, negative-neutral, negative-negative. These were dichotomized as "mutual" (positive-positive, neutral-neutral, and negative—negative) and "non-mutual" (positive-neutral, positive-negative, neutral-positive, neutral-negative, negative-positive, and negative-neutral). There was another set of nine combinations between the subject's attitude toward the object and the subject's estimate of the object's attitude toward the subject. These were also dichotomized as "congruent" and non- congruent". There were (N - l) relationships which each subject had established or perceived himself as having established with the rest of the group members. The re- searcher used these dichotomized categories in order to use Fisher's exact probability test. In the following analyses, a 2 X 2 table was used for ‘the individual subject where Fisher's exact probability test was used for individual data, the chance probability obtain- iru; at least n statistics significant at the .05 level from the 1&9 calculated statistics was tested (#5). When it was founud to be beyond the .05 confidence, the chi-square model 42 for testing the significance of combined results was used to test the overall significance of the final analysis (35). Analysis Ig£.Hypothesis‘L The (N - l) relationships, which each subject perceived as having established with the group members, were dichoto- mized in the rows as "congruent" and "non-congruent". The columns, which were also dichotomized as "accurate" (dis- crepancy "0") and "inaccurate" (discrepancy more than "l"), indicated the degree of accuracy of the subject's prediction of the object's attitude toward him. The degree of accuracy was obtained by finding a discrepancy between the subject's prediction of the object's attitude toward him and the object's actual attitude toward him. Since one subject related to (N - l) group members, there were (N - l) dis- crepancy scores. Fisher's exact probability test was applied to determine whether the accuracy of the subject's predic- tion of the object's attitude toward him was significantly different from one row to another. The analysis will reveal how many of the forty-nine subjects were significantly different in accurately pre- dicting object's attitudes toward them, depending upon their congruency or Eggyconqruency with the objects. Analysis f3; Hypothesis 2.1.- The (N - l) sociometric attitudes, which each subject perw:eived as having received from the group members, were tomized as "positive" and"neutral-negative" in the row The columns of the table, which were also dichotomized ”.214-“ ; M‘L J . , . 43 as "accurate" (discrepancy "0") and "inaccurate" (dis- crepancy more than "l"), indicated the degree of accuracy of the subject's prediction of the object's attitude toward him. The degree of accuracy was obtained by finding a dis- crepancy between the subject's prediction of the object's attitude toward him and the object's actual attitude toward him. Since one subject related to (N - I) group members, there were (N - l) discrepancy scores Fisher's exact probability test was applied to determine whether the accuracy of the subject's prediction of the object's atti- tude toward him was significantly different from one row to another. The analysis will reveal how many of the forty-nine subjects were significantly different in accurately pre- dicting object's attitudes toward them, depending upon perceiving themselves as receiving positive or neutral- negative choices from the objects. .Analysip for Hypothesis III The (N - l) relationships, which each subject establ- lished with the group members, were dichotomized in the rows as "mutual" and "non-mutual". The columns, which were also dichotomized as "accurate" (discrepancy "0") and "inaccurate" (discrepancy more than "l"), indicated the degree of accuracy of the subject's prediction of the object's attitude toward him. The degree of accuracy was obtained by finding the discrepancy between the subject's prediction of the object's attitude toward him and the 44 object's actual attitude toward him. Since each subject related to (N - l) group members, there were (N - l) dis- crepancy scores. Fisher's exact probability test was applied to determine whether the accuracy of the subject's prediction of the object's attitude toward him was signi- ficantly different from one row to another. The analysis will reveal how many of the forty-nine subjects were significantly different in accurately pre- dicting objects' attitude toward them, depending upon their mutuality and pppymutuality with the objects. Analysis £g£,HypothesL§'yy The (N - l) sociometric attitudes, which each subject received from the group members, were dichotomized in the rows as "positive" and "neutral-negative". The columns of the table, which were also dichotomized as "accurate" (dis- crepancy "0") and as "inaccurate" (discrepancy more than "l"), indicated the degree of accuracy of the subject's prediction of the object's attitude toward him and the ob- ject's actual attitude toward him. Since each subject related to (N - I) group members, there were (N - l) dis- crepancy scores. Fisher's exact probability test was applied to determine whether the accuracy of the subject's prediction of the object's attitude toward him was signi- ficantly different from one row to another. The analysis will reveal how many of the forty-nine subjects were significantly different in accurately pre- dicting objects' attitudes toward them, depending upon 45 receiving positive or ppptral-neqative choices from the objects. Analysis jg; Hypothesis 1 The (N - l) relationships, which each subject per- ceived as having established with his group members, were dichotomized in the rows as "congruent" and "non-congruent". The columns, which were also dichotomized as "accurate" (discrepancy "0") and "inaccurate" (discrepancy more than "I"), indicated the degree of accuracy of the subject's prediction of the object's sociometric status. The degree of accuracy was obtained by finding a discrepancy between the subject's prediction of the object's sociometric status and the object's actual sociometric status. Since there were (N - l) group members besides the subject, each sub- ject had (N - l) discrepancy scores. Fisher's exact pro- bability test was applied to determine whether the accuracy of the subject's prediction of the object's sociometric status was significantly different from one row to another. The analysis will reveal how many of the forty-nine subjects were significantly different in accurately pre- dicting objects' sociometric status, depending upon their '£2£££2£Q£y,or ggpycongruency with the objects. Mammy; The (N - l) sociometric attitudes, which each subject perceived as having received from group members, were dicho- tomized in the rows as "positive" and "neutral-negative". The columns of the table, which were also dichotomized as 46 "accurate" (discrepancy "0") and "inaccurate" (discrepancy more than "l"), indicate the degree of accuracy of the subject's prediction of the object's sociometric status. The degree of accuracy was obtained by finding the dis- crepancy between the subject's prediction of the object's sociometric status and the object's actual sociometric status. Since there were (N - I) group members besides the subject, each subject had (N -l) discrepancy scores. Fisher's exact probability test was applied to determine whether the accuracy of the subject's prediction of the object's sociometric status was significantly different from one row to another. The analysis will reveal how many of the forty-nine subjects were significantly different in accurately pre- dicting objects' sociometric status, depending upon per- ceiving themselves as receiving positive or neutral- negative choices from the objects. Analysis Ip£_Hypothesis gill The (N - l) sociometric attitudes, which each subject received from the group members, were dichotomized in the rows as "positive" and "neutral-negative". The columns of the table, which were also dichotomized as "accurate" (discrepancy "0") and "inaccurate" (discrepancy more than "l"), indicated the degree of accuracy of the subject's prediction of the object's sociometric status. The degree of accuracy was obtained by finding a discrepancy between the SUbject's prediction of the object's sociometric statue. 47 "“_5" .—- M a < a... a and the object's actual sociometric status. Since each student related to (N - I) group members, there were (N — l) discrepancy scores. Fisher's exact probability test was applied to determine whether the accuracy of the subject's prediction of the object's sociometric status was signifi- cantly different from one row to another. The analysis will reveal how many of the forty-nine subjects were significantly different in accurately predic- ting objects' sociometric status, depending upon receiving positive or neutral-negative choices from the objects. Analysis 19; Hypothesis ll_ 0 The forty-nine subjects were categorized in a 2 x 2 table. The rows were divided as "high" and "low" in terms of the subject's self-confidence. The subjects who expected "positive choices" from more than half the group members were considered as "high" and those who expected "positive choices" from half or less than half of the group members were considered as "low" in self-confidence. The columns, which were dichotomized into "accurate" (discrepancy "0") and "inaccurate" (discrepancy more than "l"), indicated the degree of accuracy of the subject's prediction of his own sociometric status. The degree of accuracy-was obtained by finding the discrepancy between the subject's perception of his own sociometric status and his actual sociometric status. , The median test was applied to determine whether the accuracy of the subject's prediction of his own sociometric gigs! .. .. plan unfit” 48 status is significantly different depending upon his high or low self-confidence. Analysis 12; Hypothesis,1 A 2 x 2 table was used for each subject as was done in the analyses of Hypotheses I to VIII. The objects' actual sociometric status was dichotomized as "high" and "medium-low" in the rows. The group members who received "positive" choices from more than two-thirds of the group members were considered as "high", between more than one- third and less than two-thirds as "medium", and less than one-third as "low". The columns, which were also dichotomized as "accurate" (discrepancy "0") and "inaccurate" (discrepancy more than "l"), indicated the degree of accuracy of the subject's prediction of the object's sociometric status. The degree of accuracy was obtained by finding the discrepancy between the subject's prediction of the object's sociometric status and the object's actual sociometric status. Since each subject perceives (N - l) objects, there were (N - l) dis- crepancy scores. Fisher's exact probability test was applied to determine whether the accuracy of the subject's prediction of the object's sociometric status was signifi- cantly different from one row to another. The analysis will reveal how many of the forty-nine subjects were significantly different in accurately pre- dicting objects' sociometric status, depending upon'pygp or medium-low status which the objects were enjoying. CHAPTER V RESULTS In this chapter the data will be presented showing how each of the specific hypotheses was statistically tested. For hypotheses I to VIII and X, Fisher's exact probability test was applied to the data from the individual subject. The median test was applied for hypothesis IX. The .05 level of significance was used for both tests as the criteria to reject the null hypothesis. Before the hypotheses were tested, however, analysis of variance was applied to test whether or not a significant difference exists in the accuracy of the subject's predic- tion of the object's attitude toward him and the sociometric status among three different cabin groups. Individual accuracy scores in predicting the object's attitude toward the subject and the object's sociometric status were obtained as the number of "right" guesses. Both accuracy scores were sub-grouped separately according to cabin groups. Bartlett's test was applied to test the homogeneity of sample variances. In the prediction of the object's attitude toward the subject, chi-square of l.533 was obtained. In the prediction of the object's sociometric status, chi-square of 2.026 was obtained. Since the criter- ion value of chi-square for the .05 level of significance 2+9 SO and two degrees of freedom is 5.99l, the null hypothesis is accepted and the assumption of homogeneity is considered justified in both accuracy scores Table l shows the results of analysis of variance for the subjects' accuracy scores in predicting the objects' attitudes toward them. The criterion value of F associated with the .05 level of significance for 2 and 46 degrees of freedom is 3.20. The obtained value of l.24 is less than 3.20. The null hypothesis was accepted. It was concluded, therefore, that there was no significant difference in the accuracy of the subjects' predictions of the objects' atti- tudes toward them among three cabin groups. Table l. Analysis of Variance for the Subjects' Predictions of the Objects' Attitudes toward the Subjects _:— rt. -_ _:‘ r Component of Variability SS df V F Fc Between Groups l7.04 2 8.52 l.24 3.20 Combined Within Groups 3l4.88 46 6.85 Total 33i.92 48 Table 2 shows the results of analysis of variance for the subjects' accuracy scores in predicting the objects' sociometric status. The criterion value of F associated with the .05 level of significance for 2 and 46 degrees of freedom is 3 20. The obtained value of l.63 is less than 3.20. The null hypothesis is accepted. It was concluded, «————. - fl...“__n—‘f 9| therefore, that the difference of accuracy of the subjects' predictions of the objects' sociometric status among three groups, as reflected by the means, was not significant. Table 2. Analysis of Variance for the Subjects' Predictions of the Objects' Sociometric Status mepgnent of Variability SS df V . V Fc Between Groups l5.05 2 7.53 i.65 3.20 Combined within Groups 2l2,0l 46 4.6l Total 227.06 l. Hypothesis,L Hypothesis: Accuracy of the subject's prediction of the object's attitude toward him is unrelated to congruency between them. Table 3 shows the number of subjects whose predictions of objects' attitudes toward them are significantly more , accurate when they have congruent rather than non-congruent relationships. It also shows the number of subjects whose predictions of the objects' attitudes are not significantly more accurate regardless of the "congruency" or "non- ‘congruency" of the relationship. The analysis shows that only two out of forty-nine subjects (4%) were significantly more accurate in their predictions of the objects' attitude toward them when they had "congruent" rather than "non-congruent" relationships 52 . r——.‘Mn r.--—.- 4-“- Table 3. The Number of Subjects whose Prediction of Objects' Attitudes are Significantly Accurate and not Accurate in Relation to "Congruency" and "Non- Congruency" Cabin Significant Not Total Significant Congruent Non-Congruent I l 0 l4 l5 II 0 0 l6 l6 III I 0 l7 ’ l8 Total 2 O 47 49 with the objects, while none of the subjects were signi- ficantly more accurate in their prediction of the objects' attitudes toward them when they had "non-congruent" rather than "congruent" relationships with the objects. Forty- seven out of forty-nine subjects (96%) did not show sig- nificantly better accuracy in predicting attitudes toward them either when they had "congruent" or "non-congruent" relationships with the objects. The researcher tested the chance probability of obtaining at least two out of forty- nine subjects who are significantly more accurate (the .05 'level of confidence) in predicting objects' attitudes toward them when they have "congruent" rather than "non- congruent" relationships with the objects.* It was found A— *See Chance Probability of Obtaining at Least n Statistics Significant at the .05 Level of N calculated Statistics in Sakoda, James M., Cohen, Burton H., and Beall, Geoffrey. "Test of Significance for a Series of Statistical Tests." Dnunhnlnninal DHIIA+§A Cl ’3 Inc). A l'71 .to be not beyond the .05 level of confidence. It would appear then that there is no difference in the accuracy of a subject's prediction of the object's attitudes toward him regardless of whether the subject has "congruent" or "non-contruent" relationships with the objects. Hypothesis I was supported. 2. Hypothesis._L Hypothesis: -Accuracy of the subject's prediction of the object's attitudes toward him is unrelated to his esti- mate of the object's attitude toward him. Table 4 shows the number of sUbjects whose predictions of objects' attitudes toward them are significantly more accurate when they perceive themselves as receiving "posi- tive" rather than "neutral-negative" choices from the objects and vice-versa. It also shows the number of sub- jects whose predictions of objects' attitudes toward them are not significantly more accurate regardless of whether the subjects perceive themselves as receiving "positive" or "neutral-negative" choices from the objects. The analysis shows that fourteen out of forty-nine (28%) were significantly more accurate in their predictions of objects' attitudes toward them when they perceived them- selves as receiving "positive" rather than "neutral-negative" choices. None of the subjects were significantly more accurate in their predictions of the objects' attitudes toward them when they perceived themselves as receiving 54 Table 4 The Number of Subjects Whose Predictions of Objects' Attitudes are Significantly Accurate and not Accurate in Relation to the Subjects' Estimates of the Objects' Attitudes toward Them. Cabin Significant Not Total Significant Positive Neutral- Negative I 2 0 I3 IS II 7 0 9 l6 III 5 0 l3 l8 Total l4 0 35 49 "neutral-negative" rather than "positive" choices Thirty- five out of forty-nine subjects (72%), however, did not show a significantly better accuracy in predicting the objects' attitude toward them whether they perceived them- selves as receiving "positive" or "neutral-negative" choices from the objects. The researcher tested the chance pro- bability of obtaining at least fourteen out of forty-nine subjects who are significantly more accurate (the .05 level of confidence) in predicting objects' attitudes toward them when they perceive themselves as receiving "positive" rather than "neutral-negative" choices from the objects. It was found to be beyond the .00l level of confidence. The chi-square transformation for testing the signi- ficance of the joint probability was l98.7|9 at 98 ‘4... ... ‘_‘__._-_‘ '. _‘ fl. 4 ‘1: degrees of freedom." This was also beyond the .00l level of confidence. The researcher may conclude that the accuracy of the subject's prediction of the object's attitude toward him is significantly more accurate when he perceives him- self as receiving a "positive rather than "neutral-negative" choice from the object. Hypothesis II was rejected. 3. Hypothesis 1;; Hypothesis: Accuracy of the subject's prediction of the object's attitude toward him is positively related to ' mutuality between them. Table 5 shows the number of subjects whose predictions of objects' attitudes toward them are significantly more accurate when the subjects have "mutual" rather than "non- mutual" relationships with the objects and vice-versa. It also shows the number of subjects whose predictions of the objects' attitude toward them are not significantly more accurate regardless of their "mutua or "non-mutua relationships. The analysis shows that twenty-four out of the forty- nine subjects (49%) were significantly more accurate in their prediction of the objects' attitudes toward them when they had "mutual" rather than "non-mutual" relation- ships. None of the subjects were significantly more * The composits chi-square is given by Jones and Fiske (35) as x2 = -2 ,., loge Pi . with 2 k degrees of freedom, where k is the number of independent probability values to be combined. 56 Table 5. The Number of Subjects Whose Predictions of Objects' Attitudes are Significantly Accurate and not Accurate in Relation to "Mutuality" Between Them. r Cabin Significant Not Total Groups - Significant Mutual Non-mutual I 8 C 7 '5 II 5 ' 0 ll l6 III ll 0 7 l8 Total 24 D 25 49 accurate in their predictions of the objects' attitudes toward them when they had "non-mutual" rather than "mutual" relationships. Twenty-five out of forty-nine subjects (5l%), however, did not show a significantly better accuracy in predicting objects' attitudes toward them whether they had "mutual" or "non-mutual" relationships with the objects. The researcher tested the chance probability of obtaining at least twenty-four out of forty-nine subjects who are significantly more accurate (the .05 level of confidence) in predicting objects' attitude toward them when they have "mutual" rather than "non-mutual" relationships with the objects. It was found to be beyond the .00l level of confidence. The chi-square transformation for testing the significance of a joint probability was 278.200 at 98 57 degrees of freedom. This was also beyond the .00l level of confidence. The researcher may conclude that the accuracy of a subject's prediction of the object's attitude toward him is significantly more accurate when he has a "mutual" rather than "non-mutual" relationship with the object. Hypothesis III was supported. 4. Hypothesis I__ . Hypothesis: The accuracy of the subject's prediction of the object's attitude toward him is positively related to the object's actual attitude toward him. Table 6 shows the number of subjects whose prediction of objects' attitudes toward them are significantly more accurate when they receive "positive" rather than "neutral- negative" choices from the objects and vice-versa. It also shows the number of subjects whose predictions of objects' attitudes toward them are not significantly more accurate regardless of whether the subjects receive "positive" or "neutral-negative" choices from the objects. The analysis shows that ten out of forty-nine subjects (20%) were significantly more accurate in their prediction of the objects' attitudes toward them when they received "positive" rather than "neutral-negative" choices from the objects. None of the subjects were significantly more accurate in their predictions of the objects' attitudes toward them when they received "neutral-negative" rather than "positive choices from the objects. Thirty-nine 53 Table 6. The Number of Subjects Whose Predictions of Objects' Attitudes are Significantly Accurate and not Accurate in Relation to the Objects' Attitudes toward the Subjects. Cabin Significant Not Total Significant Positive Neutral- Negative' I 5 0 l0 l5 II l 0 l5 l6 III 4 0 l4 l8 subjects (80%), however. did not show a significantly better accuracy in predicting the objects' attitudes toward them whether they received "positive" or "neutral-negative" choices from the objects. The researcher tested the chance probability of obtaining at least ten out of forty-nine subjects who are significantly more accurate (the .05 level of confidence) in predicting the objects' attitudes toward them when they received "positive" rather than "neutral-negative" choices from the objects. It was found to be beyond the .00l level of confidence. The chi-square transformation for testing the sitnificance of a joint probability was l56.20 at 98 degrees of freedom. This was also beyond the .00l level of confidence. The researcher ,—~.—.~ (m5 ...p-a- 3d.) .. . 59 may conclude that the aCCuracy of the subject's prediction of the object's attitude toward him is significantly more accurate when he receives "positive" than"neutral-negative" choice from the object. Hypothesis IV was supported. 5. Hypothesis M Hypothesis: Accuracy of the subject's prediction of the object's sociometric status is unrelated to congruency between them. Table 7 shows the number of subjects whose predictions of objects' sociometric status are significantly more accu- rate when they have "congruent" rather than "non-congruent" relationships with the objects, and vice versa. It also shows the number of subjects whose predictions of objects' sociometric status are not significantly more accurate regardless of whether they have "congruent" or "non- congruent" relationships with the objects. The analysis shows that only two out of forty-nine Subjects (4%) were significantly more accurate in their predictions of the objects' sociometric status when they had "congruent" rather than "non-congruent" relationships with the objects. None of the subjects were significantly more accurate in their predictions of the objects' socio- metric status when they had "non-congruent" rather than "congruent" relationships with the objects. Forty-seven OUt of forty-nine subjects (96%) did not show a signifi- cantly better accuracy in predicting the objects' socio- metric status either when they had "congruent" or 60 Table 7. The Number of Subjects Whose Predictions of Objects' Sociometric Status are Significantly Accurate and not Accurate in Relation to "Con- gruency" and "Non-Congruency" Cabin Significant Not Total Significant Congruent Non-congruent I l 0 l4 l5 II 0 0 l6 l6 III I 0 l7 l8 Total 2 0 47 49 "non-congruent" relationships with the subjects. The . researcher tested chance probability of obtaining at least two out of forty-nine subjects who are significantly more accurate (the .05 level of confidence) in predicting the objects' sociometric status when they have "congruent" rather than "non-congruent" relationships with the objects. It was found to be not beyond the .05 level of confidence. It would appear, then, that there is no significant dif- ference in accuracy of the subject's prediction of the object's sociometric status regardless of whether the subject has a "congruent" or "non-congruent" relationship with the object. Hypothesis V was supported. 6. Hypothesis._L Hypothesis: Accuracy of the subject's prediction of the object's sociometric status is unrelated to his estimate of the object's attitude toward him. Table 8 shows the number of subjects whose predictions of objects' sociometric status are significantly more accurate when they perceive themselves as receiving "posi- tive" rather than "neutral-negative" choices from the ob- jects and vice-versa. It also shows the numbers of subjects whose predictions of objects' sociometric status are not significantly more accurate whether the subjects perceive themselves as receiving "positive" or "neutral-negative" choices from the objects. Table 8. The Number of Subjects Whose Predictions of Objects' Sociometric Status are Significantly Accurate and not Accurate in Relation to the Subjects' Estimates of the Objects' Attitudes toward Them. Cabin Significant Not Total Significant Positive Neutral- Negative I o 0 l5 I5 II o 0 l6 I6 111 o 0 I8 I8 Total 0 o 49 49 l“.: 1‘.u‘_"w .4 _ u . 62 The analysis shows that no Subjects were significantly more accurate in their predictions of the objects' socio- metric status whether they perceived themselves as receiving "positive" or"neutral-negative" choices from the objects. It would appear that there is no significant difference in accuracy of the subjects' predictions of the objects' sociometric status regardless of whether the subjects per- ceive themselves as receiving "positive" or "neutral- negative" attitudes. Hypothesis VI was supported. 7. Hypothesis Ell Hypothesis: Accuracy of the subjects' prediction of the objects' sociometric status is unrelated to mutuality between them. Table 9 shows the number of subjects whose predictions of the objects' sociometric status are significantly more accurate when they have "mutual" rather than "non-mutual" relationships with the objects and vice versa. It also shows the number of subjects whose prediction of the ob- jects' sociometric status are not significantly more accurate regardless of whether the subjects have "mutual" l" relationships with the objects. or "non-mutua ' The analysis shows that none of the subjects were significantly more accurate in their predictions of the Objects' sociometric status when they had "mutual" rather than "non-mutual" relationships with the object. . W‘MC'J‘A ‘1... 63 fl‘t:“.a-‘ Table 9. The Numbers of Subjects Whose Predictions of Objects' Sociometric Status are Significantly Accurate and not Accurate in Relation to "Mutuality" and "non-Mutuality" between Them. Cabin Significant Not Total __ Significant Mutual I ' Non-Mutual I 0 l I4 l5 II 0 0 l6 l6 III 0 0 l8 l8 Total 0 l 48 49 One out of forty-nine subjects (2%) was significantly more accurate in his prediction of the objects' sociometric status when he has "non-mutual" rather than "mutual" rela- tionships with the objects. Forty-eight out of forty-nine subjects (98%) did not show significantly better accuracy in predicting the subject's sociometric status whether they had "mutual" or "non-mutual" relationships with the objects. The researcher tested chance probability of obtaining at lease one of forty-nine subjects who are significantly more accurate (the .05 level of confidence) in predicting the object's sociometric status when he has "non-mutual" rather than "mutual" relationships with the object. It was found to be not beyond the .05 level of confidence. It would appear that there is no significant difference 64 in accuracy of the subject's prediction of the object's sociometric status regardless of whether the object has a "mutual" or "non-mutua relationship with the object. Hypothesis VII was supported. 8. Hypothesis VIII Hypothesis: Accuracy of the subject's prediction of the object's sociometric status is unrelated to the quality or kind of the object's attitude toward him. Table l0 shows the number of subjects whose predic- tions of objects' sociometric status are significantly more accurate when they receive "positive" rather than "neutral- negative" choices from the objects and vice versa. It also shows the number of subjects whose predictions of objects' sociometric status are not significantly more accurate regardless of whether the subjects receive "positive" or "neutral-negative" choices from the objects. The analysis shows that none of the subjects were significantly more accurate in predicting the objects' sociometric status regardless of whether the subjects received "positive" or "neutral-negative" relationships from the objects. It would appear that there is no difference in accuracy of the subjects' predictions of the objects' sociometric status regardless of the objects' attitudes toward the subjects. Hyposhesis VIII was sup- ported. 0\ U1 Table ID. The Numbers of Subjects Whose Predictions of Objects' Sociometric Status are Significantly Accurate and not Accurate in Relation to the Objects' Attitudes toward Them. Cabin Significant Not Total Groups Significant Positive Neutral- Negative I 0 0 l5 l5 II 0 0 l6 l6 III 0 0 l8 I8 Total 0 O 49 49 9. Hypothesis I_r. Hypothesis: Accuracy of the subject's prediction of his own sociometric status is unrelated to his self- confidence. Table ll shows the number of subjects whose prediction of their own sociometric status are accurate and not accu- rate when they have "high" and "low" self-confidence. There were twelve subjects whose predictions of their own sociometric status were accurate when they had "high" self-confidence, while eleven subjects were not accurate. There were eight subjects whose predictions of their own sociometric status were accurate when they had "low" self- confidence, while eighteen subjects were not accurate. 66 .. H‘s. “' Table ll. The Number of Subjects Whose Predictions of ' Their own Sociometric Status are Accurate and not Accurate in Relation to their Self- Confidence. Subject's prediction of his own sociometric status accurate not accurate Total "High" self ‘i confidence l2 ll 1 23 "Low" self- § confidence 8 1 l8 i 26 20 29 49 Chi—square value of l.5l3 was obtained from these data. Since the criterion value of chi-square for the .05 level of significance at one degree of freedom is 3.84l, the null hypothesis was accepted. It would appear that there is no significant difference in the subject's prediction of his own sociometric status regardless of whether he per- ceives himself as enjoying "high" or "low" sociometric status. Hypothesis IX was supported. l0. Hypothesis X Hypothesis: Accuracy of the subject's prediction of the object's sociometric status is positively related to the object's sociometric status. Table l2 shows the number of subjects whose predictions of objects' sociometric status are significantly more accurate when the objects are enjoying "high" rather than "medium-low" sociometric status and vice versa. It also ( 7‘. «q ‘ 43 Q? T. - shows the number of subjects whose predictions of objects' sociometric status are not significantly more accurate regardless of whether the subjects are enjoying "high" or "medium-low" sociometric status. Table l2. The Number of Subjects Whose Predictions of Objects' Sociometric Status are Significantly Accurate and not Accurate in Relation to the Objects' Sociometric Status. Cabin Significant Not Total Groups Significant High Medium- Low I 3 I ll l5 II 8 0 8 l6 III 5 0 l3 l8 Total l6 l 32 49 The analysis shows that sixteen out of forty-nine subjects (33%) were significantly more accurate in their predictions of the objects' sociometric status when the objects were enjoying "high" rather than "medium-low" sociometric status. Only one of the subjects (2%) was significantly more accurate in his predictions of the ob- jects' sociometric status when the objects were enjoying "medium-low" rather than "high" sociometric status. Thirty-two out of forty-nine subjects (65%), however, did not show significantly higher accuracy in predicting the objects' sociometric status either when the objects 'were enjoying "high" or"medium-low" sociometric status. The researcher tested the chance probability of obtaining at least sixteen out of forty-nine subjects who are sig- nificantly more accurate (the :05 level of confidence) in predicting the object's sociometric status when the object' is enjoying "high" rather than "medium-low" sociometric status. It was found to be beyond the .00l level of confidence. The chi-square transformation for testing the - significance of a joint probability was 274.30 at 98 degrees of freedom. This was also beyond the .00! level of confi- dence. The researcher may conclude that the accuracy of (the subject's prediction of the object's sociometric status is significantly more'accurate when the object enjoys a "high" rather than "medium-low" sociometric status. Hypo- thesis X was supported. CHAPTER VI DISCUSSION AND INTERPRETATION Hypothesis pp the relationship between congruency and th accuracy pi the subject's prediction pj,the object's att tude:' Hypothesis I predicted that there is no significant difference in accuracy of the subject's prediction of the object's attitude toward him whether he has a congruent or non-congruent relationship with the object. The hypo- thesis was supported. 3' Hypothesis pp the relationship between congruency and t e accuracy 2: the subject's prediction pi the object's sociometric status: Hypothesis V predicted that there is no significant difference in accuracy of the subject's prediction of the object's sociometric status whether he has a congruent relationship with the object or not. The hypothesis was supported. It is interesting to question whether congruency is a tendency for pepple to feel chosen by those whom they choose, or contrariwise, to choose those they feel have selected them. Undoubtedly, there must be individual differences in the order of this cycle. In terms of subjective experience, a person's awareness of his affec- 69 e i- A‘u’ nnmrm' - i 7O tive response toward another seems to come sooner or with greater certainty than his awareness of another's feelings for him. The observations made above are probably reflected in the fact that even when group members have congruent relationships with objects, the accuracy of their predic- tions of objects' feelings do not exceed those made when they have non-congruent relationships. The congruent set does not Operate only in connection with one's own feelings and perceptions but it also occurs in judging feelings between others as well as feelings and perceptions held by others about a third person. There is also no doubt that a general projection upon others of one's own tendency toward congruency is oper- ating here (53). This kind of strong subjectivity is probably what prevented group members from accurately perceiving others' sociometric attitudes and status. Hypothesis pp the relationsh p between the object's atti- tude perceived 21 the subject and the accuracy piIthe_ subject's prediction pj,the object's attitude: Hypothesis II predicted that there is no significant difference in accuracy of the subject's prediction of the <3bject's attitude toward him whether the subject perceives. himself as receiving a positive or neutral-negative choice from the object. Hypothesis II was rejected. Fourteen (NJt of forty-nine subjects did show significantly better laccuracy in predicting others' attitudes toward them when 7| they perceived positive rather than neutral-negative choices from the objects. The prediction was made on the basis of a finding by Tagiuri (55) that reciprocal relationships are more trans- parent. Tagiuri (53) also reported that if mutuality of feelings happens to be absent the subject's perception may be inaccurate. The researcher naively postulated Hypothesis II on the assumption that the subject's estimate of the object's attitude toward him does not involve mutuality between them. The researcher, therefore, hypothesized that the subject's estimate of the object's attitude toward him is unrelated to the accuracy of the subject's prediction of that attitude. The above hypothsis was naive from two points of view. First, the instrument itself might not have been precise enough to obtain the necessary data. Secondly, neither mutuality nor non-mutuality was proved to be unrelated to the subject's estimate of the object's attitude toward the subject. A question still remains, however, as to why fourteen out of forty-nine subjects did show significantly better accuracy in predicting others' attitudes toward them when they perceived positive rather than neutral-negative choices from the objects. To explore the above question, the researcher analyzed inme data on three different levels, First, it was found that nine out of fourteen subjects, who were significantly more correct in predicting the objects' feelings when they perceived positive rather than neutral-negative feelings from the objects (Hypothesis II), were also significantly more correct in predicting objects' feelings when they had mutual feelings with them. (Hypothesis III). Tables l3, l4 and l5 show the results of analyses I to VIII and x which indicate the subjects who predicted the objects' attitudes and sociometric status significantly better in one condition than the other. The figure indicates the level of confidence in Fisher's exact probability test while no figure indicates that the result did not reach the 5% level of confidence. The above comparative analysis between Hypothesis II and III led the researcher to the conclusion that mutuality is not solely accountable for the fact that fourteen out of forty-nine subjects did predict the objects' attitudes toward them significantly more correctly when they felt chosen rather than rejected or given neutral choices. It was felt that dyadic relationships should be analyzed in the second level. The above fourteen subjects' dyadic relationships with objects j;pm_pppm_;ppy perceived positive choices were analyzed. Fisher's exact probability test was run to test whether there was any difference in the accuracy of subjects' predictions of objects' attitudes toward them when they had mutual feelings as contrasted with when they 73' Table l3. The Result of Fisher's Exact Probability Test for Individual Subjects for Hypothesis I to VIII and X (Group I) .- Hypoth. I II III IV V VI VII VIII x §ppjects l .025 2 .025 .05 3 -.025 4. 5 6 .0l 7 -.05 8 .0l .025 9 .025 .005 ID .025 .025 ll .0l .0l .05 I2 .05 .0l l3 .05 .025 .025 .005 I4 - .05 “ma-.33! “MI-h n’x‘. (. «L_‘ 74 Table I4. The Result of Fisher's Exact Probability Test for Individual Subjects for Hypothesis I to VIII and X (Group II) Hypoth. I II III IV V VI VII VIII X Subjects l .005 2 ' .05 3 .025 .005 .05 4 5 .05 6 .005 .0l .005 7 025 8 .025 9 .05 .0! .05 l0 .0l .025 ll I2 .005 l3 I4 .025 l5 .005 .025 .005 75 Table l5. The Result of Fisher's Exact Probability Test for Individual Subjects for Hypotheses I to VIII and X (Group III) Hypoth. II III IV VI VII VIII X Subjects l .025 .005 2 .05 3 4 .05 5 6 .05 7 .05 8 .005 9 .005 .005 ID .005 .05 .005 II .025 .005 '2 .005 .0l .005 '3 .0l l4 .005 .Ol .05 I5 .0l .025 l6 l7 .05 l8 .Ol .05 .Ol 76 did not. It was found that only four out of fourteen sub- jects were significantly more correct in predicting objects' feelings toward them when they had mutual relationships rather than when they did not. The same analysis was made for the remainder of the subjects. Thirty-five subjects' dyadic relationships with objects iggg 3223,;ppy perceived positive choices were analyzed. It was found that ten out of thirty-five sub- jects (who did not show significantly more correct predic- tions in objects' feelings toward them when they perceived positive choices from the objects instead of neutral- negative choices in Hypothesis II) were significantly more correct in predicting the objects' attitudes toward them when they had mutual rather than non-mutual relationships. It would appear that there is little difference between the two groups ip_;pp preportions pi subjects 332 gpgg si - nificantly pp;p_correct in predicting objects' attitudes toward them. The frequencies in each of the 2 x 2 tables for Fisher's exact probability test in both groups were accumulated separately. Four cells indicates: (2) the subjects who predicted objects' attitudes significantly more correctly when they had mutual relationships with objects, (b) the subjects who did not predict objects' attitudes toward them significantly more correctly when they had mutual relationships with objects, (c) the subjects who predicted objects' attitudes toward them significantly more correctly when they did not have mutual 77 i relationships with the objects, and (d) the subjects who did not predict objects' attitudes toward them significantly more correctly when they did not have mutual relationships with them. lp_both groups only dyadic relationshipp lg ppi£p_lpg supjects perceived positive choices jgpfl objects p352 analyzed. In Table l6 there were ill dyadic relationships in the group in which subjects showed significantly more accurate predictions of ojbects' attitudes toward them when they perceived positive rather than non-positive attitudes from objects. Cell "a" indicates seventy-eight which was 70% of lll dyadic relationships. Cell "b" indicates only 2, which was 2% of ill dyadic relationships. Cell "c" indi- cates l8, which was l6% of III dyadic relationships. Cell "d" indicates l3, which was l2% of ill dyadic relationships. Table l6. Relationships between Mutuality and S's Predic- . tion of 0's Attitudes to S when S perceived Positive Choices from 0 (83 who showed signi- ficant accuracies in Hypothesis II) accurate inaccurate Mutual a 78 b 2 80 (70%) (2%) (72%) Not c l8 d l3 3| Mutual (l6%) (l2%) (28%) , 96 - l5 Ill (86%) (I476) “00%) 78 In Table I7 there were 293 dyadic relationships in the group in which the subjects did pp}_show significantly more accurate predictions when they perceived positive rather then non-positive attitudes. Cell "a " indicates l52 which was 52% of 293 dyadic relationships. Cell "b" indicates only 2 which was .6% of 293 dyadic relationships. Cell "c" indicated 44 which was l5% of 293 dyadic rela- tionships. Cell "d" indicates 95 which was 32.4% of 293 dyadic relationships. Table I7. Relationships between Mutuality and 8's Prediction of 0's Attitudes to S when S Perceived Positive Choices from 0 (83 who did not show significant accuracies in Hypothesis II) accurate inaccurate Mutual a l52 b 2 l54 (52%) . (.6%) (52.6%) Not c 44 d 95 l39 (I5%) (32.4%) (47.4%) l96 97 293 (67%) ‘ (33%) (IOO%) It appears that cell "a" of the first group is consid- erably larger in preportion than the second group, while cell "b" in the second group is considerably larger in proportion than that of the first group. In total, however, there was much similarity between the two groups with reference to each cell's prOportion. It was remarkable that in both groups cells "D" had extremely 79 small preportions (2% and .6%). This finding led the researcher to speculate that when subjects perceive positive attitudes from objects accompanied by mutuality between them, it leads the subjects to high accuracy in their prediction of objects' attitudes toward them. But is the subject's prediction more accurate when he perceives (a) a positive attitude or (b) a neutral-negative attitude from the object of a mutual relationship? Does the mutuality still play an important role in the subject's prediction of the object's attitude toward him even when he perceives neutral-negative choice from the object? Fourteen subjects' (those who showed significantly p more accurate predictions of objects' attitudes toward them when they perceived positive rather than neutral- negative choices) relationships with objects, I;pm_gppm, Lgpy,perceived neutral-negative choices, were analyzed. Fisher's exact probability test was run to test whether there was any difference in accuracy of subjects' predic- tions of objects' attitudes toward them when they had mutual feelings with objects as contrasted with when they did not. It was found that none of the fourteen subjects were significantly more correct in predicting objects' feelings when they had mutual relationships with the ob- jects as compared with when they did not. The same analysis was made for the remainder of the subjects (those who did pp;_show significantly more accurate predictions of the objects' attitudes toward them when they 80 perceived positive rather than neutral-negative choices from objects). Thirty-five subjects' dyadic relationships with objects m EDP—".1 3.9.21 perceived neutral-negative choices were analyzed. It was found that only two out of thirty-five subjects were significantly more correct in predicting objects' attitudes toward them when they had mutual relationships with the objects as compared with when they did not. There appears to be little difference between the two groups in the accuracy of the subjects' prediction of objects' attitudes toward them when the dyadic relation- ships in which the subjects perceived neutral-negative choices from the objects were singled out. Apparently there is little difference in the accuracy of subjects' predic- tions of objects' attitude toward them when they perceived neutral-negative choices regardless of whether mutual or non-mutual relationships existed. The frequencies in each of the 2 x 2 tables for Fisher's exact probability test in both groups were accumulated separately. Four cells indicate: (a) subjects who predicted objects' attitudes toward them significantly more correctly when they had mutual relationships, (b) subjects who did not predict objects' attitudes toward them significantly more correctly when they had mutual relationships, (c) subjects who predicted objects' atti- tudes toward them significantly more correctly when they did n0t have mutual relationships, and (d) subjects who 8i did not predict the objects' attitudes toward them signi- ficantly more correctly when they did not have mutual relationships. In both groups only dyadic relationships in which subjects perceived neutral-negative choices from objects were analyzed. In Table l8, there were l05 dyadic relationships in the group in which subjects showed significantly more accurate predictions of objects' attitudes toward them when they perceived positive attitudes as contrasted with when they did not. Cell "a" indicates l3, which was l2% of l05 dyadic relationships. Cell "b" indicates l7, which was l6% of l05 dyadic relationships. Cell "c" indicates 7, which was 7% of I05 dyadic relationships. Cell "d" indicates 68, which was 65% of l05 dyadic relationships., Table l8. Relationships between Mutuality and S's Prediction of 0's Attitudes toward 8 when S Perceived Neutral-Negative Choices from 0 (85 who showed significant accuracies in Hypothesis II) accurate inaccurate Total Mutual a l3 b l7 30 (I2%) (I6%) (28%) Not c 7 d 68 75 Mutual (7%) (65%) (72%) 20 85 l05 (I9%) (85%) (IOO%) In Table I9 there were 247 dyadic relationships in the group in which the subjects did not show significantly more accurate predictions when they perceived positive attitudes d‘i'Tr 82 from objects as compared with when they did not. Cell a indicates 55, which was 22% of 247 dyadic relationships. Cell "b" indicates 49 which was 20% of 247 dyadic rela- tionships. Cell "c" indicates 4i, which was l7% of 247 dyadic relationships. Cell "d" indicates l02, which was 4l% of 247 dyadic relationships. Table I9. Relationships between Mutuality and S's Predic- tion of 0's Attitudes toward 3 when S Perceived Neutral-Negative Choices from 0 (33 who did not show significant accuricies in Hypothesis II) accurate inaccurate Total Mutual a, 55 b 49 IO4 (22%) (20%) (42%) Not c 4| d lO2 I43 Mutual (l7%) (4l%) (58%) 96 ISI 247 (29%) (6I%) (IOO%) It appears that cells "a", "b" and c of the second group, (Table I9) are considerably larger in prOportion than are those of the first group, (Table I8) while cell "d" in the first group is larger in pr0portion than that (of the second group. In total, however, the proportion in each cell showed a great deal of similarity between the two groups. (It is interesting that cell "d" is larger in its proportion than are those of cell "c".) This leads the researcher to speculate that when the subject perceives l1eutraI-negative attitudes from the object with whom he lnas a non-mutual relationship, his prediction of the object's 83 attitude is often misleading. Furthermore, when the subject perceives neutral-negative choige from the object, the presence of mutuality of feelings between them does not lead the subject to greater accuracy of predicting the object's attitude toward him as it did when he perceived positive choice. The above finding may not completely explain the fact that fourteen pp;,pi,forty-nine subjects were significantly more accurate in predicting the objects' attitudes toward them when they perceived positive rather than neutral- negative choices from the objects. It appears, however, that when the subject perceives himself as receiving posi- tive choice from the object, it leads him to greater accuracy in predicting the objects' attitude toward him only when mdtuality of feelings exists between them. In the combined relationship tables, the researcher examined "mutual" relationships between subjects and objects which were accompanied by the subjects' estimation of posi- tive choice from objects. It was discovered that the above described dyadic relationships were dominantly accompanied by "positive" mutuality and "positive" congruency. It is self-evident that the subjects' predictions of the objects' attitudes towrd them are I00 per cent correct if the sub- jects perceive positive choices from the objects and also if positive mutuality and positive congruency coexist between them. 84 The researcher, therefore, examined to what extent the subjects' prediction of objects' attitudes toward them are significantly more accurate when the subjects have both mutual and congruent relationships with the objects as contrasted with when they do not. Fisher's exact proba- bility test was applied to the data from individual sub- jects. Table 20 shows the number of subjects whose predictions of objects' attitudes toward them are significantly more accurate when they have both "mutual" and "congruent" relationships as compared with when they do not and vice versa. It also shows the number of subjects whose pre- dictions of objects' attitudes toward them are signifi- cantly more accurate regardless of whether they have both "mutual" and "congruent" relationships with the objects or not. The analysis shows that thirty-nine out of forty-nine 9 subjects (80%) were significantly more accurate in their predictions of objects' attitudes toward them when they had txath "mutual" and "congruent" relationships with the objects as compared with when they did not. None of the subjects were significantly more accurate in their predictions of theicabjects' attitudes toward them when they had either "nuituality" or "congruency", or neither, as compared with when. they had both. Ten out of forty-nine subjects (20%) did not show a sigywificantly better accuracy in predicting object's 85 Table 20. The Number of Subjects Whose Predictions of Objects' Attitudes toward them are Significantly more Accurate and not Accurate in Relation to the Coexistence of "mutuality" and "Congruency". Cabin Significant Not Total Groups p, Significant Mutuality & Mutuality & Congruency Congruency Coexisted Not Coexisted I _, I2_ 0 3 l5 II II 0 5 l6 III l6 0 2 l8 Total 29 0 IO 49 attitudes toward them regardless of whether they had both "mutual" and "congruent" relationships with the objects or not. The researcher tested the chance probability of obtaining at least thirty-nine out of forty-nine sub- jects who are significantly more accurate (the .05 level of confidence) in predicting object's attitudes toward them when "mutuality" and "congruency" coexisted between subjects and objects. It was found to be beyond the .00l level of confidence. The chi-square transformation for testing the significance of a joint probability was 459.l90 at 98 degrees of freedom. This was also beyond the .00l level of confidence. It would appear then that one is 86 able to predict the feelings of another toward himself with a greater degree of accuracy when both "mutual" and "con- gruent" relationships exist. When Fisher's exact probability test was applied to the individual to test whether he is significantly more accurate in his prediction of the object's attitude toward him when positive mutuality and positive congruency exist rather than when they do not, only twenty out of forty- nine subjects (4l%) were significantly more accurate when the two relationships coexisted than when they did not. The chance probability of obtaining twenty out of forty-nine subjects who are significantly more accurate at the .05 level of confidence was beyond the .00l level of confidence. The chi-square transformation for testing the significance of a joint probability was 339.90 at 98 degrees of freedom. This was also beyond the 00l level of confi- dence. While the coexistence of positive mutuality and posi- tive congruency enables the subject to predict the object's feelings toward him perfectly, the accuracy of the subject's prediction in the above condition does not exceed that occurring when mutuality and congruency (positive, neutral, and negative) coexist. This is due to the fact that in the second row, where positive mutuality and positive congruency do not coexist, it still includes the coexistence of neutral mutuality and neutral congruency, and negative mutuality and negative 87 congruency which also increases the accuracy of the sub- ject's prediction. Hypothesip,pp.the relationship between the object's attitude perceived py the subject and the accuracy pi the subject's prediction pj_the object's sociometric status: Hypothesis VI predicted that there is no significant difference in the accuracy of the subject's prediction of the object's sociometric status whether the subject per- ceives himself as receiving a positive choice, rejection or neutral choice from the object. The hypothesis was supported. The prediction was made on the basis of findings by Ausubel (3, 4, 5, 6), Schiff (46), and Trent (65, 66) that the abilitypto perceive one's own sociometric status and the ability to perceive others' sociometric status are completely unrelated. Adams (I) also found a very low correlation between abilities to rate self and others. Singer (ll) obtained a negative relationship between per- ception of one's own status and perception of preferred as ppposed to non-preferred persons. Dymond (2|) on the other hand, found a significant positive relationship between empathic ability and insight (the ability of an individual to rate himself as others rate him). Although there are some inconsistencies in various research results, it would appear that the type of sensiti- Vity that enables individual A to predict accurately how .‘u‘i’ U) 0) individual 8 feels toward him (A) is unrelated to the type of sensitivity that enables him to predict accurately how the entire group accepts 8. These two unrelated sensitivities, one of which is responsible for accuracy of predicting the object's feeling toward the subject and the other which is responsible for accuracy of predicting the other's sociometric status, seem to be accountable for the fact that there is no significant difference in the accuracy of the subject's perception of the object's attitude toward him whether he perceives posi- tive or neutralgnegative choice from the object. Hypothesis pp the relationship between mutuality and t e accuracy _I. e subject's piediction p: the object's attitude: Hypothesis III predicted that there is a significantly higher accuracy of the subject's prediction of the object's attitude toward him when there is mutuality between them as contrasted with when there is not. The hypothesis was supported. The effect of mutuality on the accuracy of prediction of the object's attitude toward the subject has already been examined and discussed. It was found that if congruency and inutuality of feeling are present between the subject and the <3bject, then accuracy will occur. When congruency is singled out alone, however, the subject showed equal accu- racy in predicting the object's feelings toward him whether his prediction was congruent with his feeling toward the object or not. On the other hand, accuracy is considerably higher under conditions where the individuals have mutual feelings for each other. But why does mutuality alone in- crease accuracy while congruency alone does not? Why does the coexistence of mutuality and congruence increase accu- racy to near perfection? These may be the questions which should be answered in future studies. Hypothesis pp,the relationship between mutuality and the accuracy 9: the subject's prediction 9: the object's socio- metric status: Hypothesis VII predicted that there is no significant difference in accuracy of the subject's prediction of the object's sociometric status when mutuality does or does not exist between them. The hypothesis was supported. The effect of mutuality on the accuracy of the sub- ject's prediction of the object's attitude toward him has already been examined and discussed. (It was reported that the type of sensitivity which can predict another's feel- ings toward oneself is unrelated to the sensitivity which predicts group acceptance of an individual member.) The accuracy of the subject's prediction of the object's feeling toward him largely depends on the coexistence of mutuality and congruency.. When mutuality exists alone, the subject's perception of the object's attitude toward him still comes considerably closer to the object's :f-al_- ..F; “a!“ In -“I 90 actual attitude. Since the subject's perception of the object's attitude toward him is unrelated to the accuracy of his prediction of the object's sociometric status, it is anticipated that there is little relationship between mutuality and the accuracy of the subject's prediction of the object's sociometric status. Hypothepis pp_lpp,relationship_between ppp,object's p531: lppp_toward gpp,subject ppp,;pp,accuracypj,;pp,subject's prediction 21,;pg object's attitude: Hypothesis IV predicted that accuracy of the subject's prediction of the object's attitude toward him will be significantly more accurate when the subject receives positive rather than neutral or negative choice from the object. The hypothesis was supported. This hypothesis was based on the common sense notion that pepple are more receptive to a positive rather than a negative feeling from others. The present researcher has already reported that accuracy of the subject's prediction of the object's attitude toward him is almost perfect when the "positiveness" of attitude is added to the coexistence of mutuality and congruency. This indicates that the "posi- tiveness" of the object's attitude plays an important role in the subject's prediction of the object's attitude. PeOple do not always see the environment as it is but instead perceive it in terms of their own needs. Doubt- lessly they become more receptive to others' positive rather than negative feelings toward them. 9i Hypothesis pp_the relationship between the object'p attitudp toward the subject and the accuracy pj,the pubject's re- diction pi the object's sociometric status: Hypothesis VIII predicted that there is no significant difference in accuracy of the subject's prediction of the object's sociometric status whether he receives a positive, neutral or negative choice from the object. The hypothesis was supported. While the "positiveness" of the object's attitude toward the subject played an important role in his pre- diction of the object's attitude, it did not play any significant role in his prediction of the object's socio- metric status. In summary, it would appear then that accuracy of the subject's prediction of the object's sociometric status is unrelated to congruency, how the subject perceives the object's attitude toward him, mutuality and the object's actual attitude toward the subject. It may be concluded that the ability to predict another's sociometric status is unrelated to the like-dislike relationship between the person predicting and the person to be predicted. Hypothesis pp.;pp,relationspip between Lpp,subject's ppljf confidence ppp.;pg,accurapyppj.;pp,pppject's prediction 2: pl§_gpp_spciometric status: Hypothesis IX predicted that there is no significant difference in accuracy of the subject's prediction of his 92 own sociometric status whether he considers himself as enjoying a high, middle, or low sociometric status. The hypothesis was supported. A facile generalization based on a doctrine of the defense of self-esteem might say that pepple who are, insecure in themselves and in their interpersonal relation- ships with others might distort the reality and exaggerate their own popularity by recognizing only cues of acceptance from others while cues of rejection are perceptually denied as an act of ego defense. 0n the other hand some peOple may be overly sensitive to neutral cues because of polite- ness Paranoid-like persons may be overly sensitive to cues of rejection. It may be considered that the self-awareness of one's own sociometric status is a function of the person chosen as well as the person choosing. This part of the question should be answered, however, in future studies. Hypothesis pp _t_l_i_e_ relationship m _t_l_i_e_ p_b_jp£_t_'_s_ pp_c_i_p- metric status ppp,;pp subject's pgediction pi LEE object's sociometric status: ' Hypothesis X predicted that accuracy of the subject's perception of the object's sociometric status will be sig- nificantly higher when the_object enjoys a high status rather than a low or middle status. The hypothesis was not proven but there was a tendency toward the prediction. Ausuble (4) reported that the subject's ability to predict the sociometric rating received from others varies ‘93 directly with the sociometric status of the person per- ceived. The common sense notion of sociometric perception also says that it is easier to predict the sociometric status of members who enjoy very high or very low pOpularity. The result in this study seems to indicate that the object's actual sociometric status is an important factor in the subject's prediction of the object's sociometric status. It is also considered, however, that other factors such as the object's self-confidence, his behavior in the group and his actual and perceived relationship may play a role just as important as the object's sociometric status. This point should be explored further in future studies. CHAPTER VII SUMMARY, CONCLUSIONS, AND SUGGESTIONS FOR FURTHER RESEARCH This last chapter consists of three sections. In the first, a restatement of the problem is made. Its method- ology and hypotheses are summarized. A series of findings and conclusions, resulting from an analysis of the data, constitute the second section. In the last section some implications and suggestions for further research are made. Summary of Problem and Design _p3_Problem In present-day studies of person-perception the issue of accuracy is often the focus of interest. The results of the studies have been inconsistent, however, and difficult to interpret. Consequently it is considered to be a rela- tively unexplored field in social psychology. This study was especially designed to explore: l) the relationship between the perceived like- dislike of two group members and the accuracy of their predictions of each other's feelings for one another; ID the relationship between one's estimate of another group member's feelings toward oneself and the accuracy of that prediction; 3) the relationship between the actual like-dislike of two group members and the accuracy of their prediction of each other's feelings for one another; 4) the relationship between one's feelings for 94 another group member and the accuracy of that member's prediction of one's feelings; 5) the relationship between the perceived like- dislike of two members and the accuracy of their predictions of each other's pOpularity; 6) the relationship between one's estimate of another group member's feelings toward oneself and the accuracy of the prediction of that group member's pOpularity; 7) the relationship between the actual like-dislike of two members and the accuracy of their predic- tion of each other's pooularity; 8) the relationship between one's feelings for another group member and the accuracy of that group member's prediction of one's pOpularity; 9) the relationship between one's estimate of his own pOpularity and the accuracy of that predic- tion; IO) the relationship between actual pOpularity in the group and the accuracy of another's pre- diction of that pOpularity. Hypotheses A hypothesis was formulated for each of the above described relationships. These hypotheses can be pre- sented as ten different combinations between sociometric variables and the accuracy of the subject's prediction of them. 96 . q. _— —-u- I“_‘I—n l . Hypothesis Sociometric Variable (SV) Accuracy of the S's Prediction (AP) Existence of Relation be- tween SV and AP I Congruency S's prediction of None 0's attitude toward 3 II S's estimate S's prediction of None of 0's atti- O's attitude tude toward 8 toward S III Mutuality S's prediction of Positive O's attitude toward 8 IV O's actual S's prediction of Positive attitude to- 0'8 attitude ward 8 toward 8 V Congruency S's prediction None of 0's socio- metric status VI S's estimate S's prediction None of 0's atti- of 0's socio- tude toward metric status 8 VII Mutuality S's prediction of None O's sociometric _status VIII O's actual S's prediction of None attitude 0's sOciometric toward 8 status IX S's perceived S's prediction of None sociometric own sociometric status status X 0'8 actual S's prediction of Positive sociometric status O's sociometric status 97 The Subjects The forty-nine campers who served as the research subjects were boys ranging in age from eleven to thirteen, who were attending-Camp Oakland, Oxford, Michigan. All the campers came to this camp through various social wei- fare agencies in Oakland County. Ail of the subjects were from lower socioeconomic backgrounds. Because of camp policy, the camp took new campers except a few special cases so that most of the campers were unknown to each other until they arrived at the camp. The prOportion of Negro and Caucasian campers was nearly even. These forty- nine subjects were the oldest group of campers. They lived in three different cabins during different camping sessions. The number of campers in each cabin was fifteen, sixteen and eighteen respectively. Collection 31‘, gati- Before starting the sociometric-like interview with the subjects, the researcher spent almost ten days exposing himself to the subjects in the camp site so that the sub- jects felt at home with him. Each subject was then interviewed individually, answer- ing to the sociometric-like questions. The subject was given the pictures of his cabin-mates and then was asked to sort them according to the instructions. The picture-sociometric-like questionnaire consisted (Df three questions. First, the subject was asked to 98 indicate his feelings toward others. Next, he was asked to guess others' feelings toward him. Third, he was asked to guess others' sociometric pOpularity. In other words, the subject,was asked to indicate "like", "dislike" and "neu- tral feeling" toward others and guess how others feel toward him. The subject was also asked to guess his own and others' sociometric popularity at three different levels such as high, medium and low. The subjects were interviewed once during the two-week camping session. Treatment 2123}; The sociometric data was transformed into matrices. Figures "I", "2", and "3" were given to "negative," "neutral", and "positive" choices or "low", "medium" and "high" popularities. The matrices were transformed into the combined relationship table which contained var- ious sociometric variables for an individual and his rela- tionship with each group member. From this table the accuracy of the subject's prediction of each object's socio- rnetric attitude and status was easily obtained by subtract- ing the predicted value from the actual value. To test all hypotheses, except hypothesis IX, Fisher's eexact probability test was run on the data from each sub- ‘jectq The median test was used on the over-all data for hypothesis IX. Both tests were used to determine whether a significant diihference existed in the subject's prediction depending {JpOTI the presence or absence of two different levels of a 99 given sociometric variable. When Fisher's exact probability test was used for individual data, the chance probability of obtaining at least n statistics significant at the .05 level from the forty-nine calculated statistics was tested. When it was found to be beyond the .05 confidence. the chi-square model for testing the significance of combined result was used to test the overall significance of the final analysis. Summary of Findings l. Hypothesis I was supported. It was concluded that there is no significant difference in the accuracy of one's prediction of another's feeling toward him whether it is accompanied by congruent feeling or not. That is, if-A feels chosen by 8, A is no more accurate in recognizing B's feeling when A also chooses 8 than when A does not. 2. Hypothesis II was rejected. It was concluded that the accuracy of the subject's prediction of the object's atti- tude toward him is moderate but significantly more accurate when he perceives himself as receiving a "positive" rather than a "neutral-negative" choice from the object. That is, if A feels chosen by 8, he is more accurate in reCOgnizing £3's feeling toward him than when he feels rejected by B. It should be noted, however, that only 28% of the 2m1bjects were significantly more accurate in their guess 01’ others' feelings toward them when they felt chosen rather than rejected. lOO 3. Hypothesis III was supported It was concluded that when a mutual feeling exists between individuals, one of them is significantly more correct in recognizing another's feelings toward him than when nonemutual feeling exists between them. That is, if A is chosen by B, A is more correct in recognizing B's feeling toward him when he also chooses 8 than when he rejects B. It should be noted, however, that 49% of the subjects were significantly more correct in their guess of others' feelings toward them when they had similar feelings for each other. 4. Hypothesis IV was supported It was concluded that the accuracy of the subject's prediction of the object's atti- tude toward him is moderate but significantly more accurate when he receives a "positive" rather than "neutral-negative" choice from the object. *That is, if A is chosen by 8, he is more correct in recognizing B's feeling toward him than iwhen he is rejected by B. I t should be noted, however, that only 20% of the sub- _jects were significantly more correct in predicting others' fkeelings toward them when they were chosen rather than rejected. 5. Hypothesis V was supported. It was concluded that truereiis no significant difference in accuracy of one's pyrtediction of another's popularity whether there is a (3011gruent feeling between them or not. That is, if A feels chosen by 8, he is no more correct in recognizing B's pOpularity when he also chooses B than when he rejects B. 6. Hypothesis VI was supported. It was concluded that there is no significant difference in accuracy of one's prediction of another's pOpularity whether one estimates himself as receiving a positive or negative feeling from another. That is, if A feels chosen by 8, he is no more correct in recognizing B's popularity than when he feels rejected by B. 7. Hypothesis VII was supported. It was concluded that even when a mutual feeling exists between individuals, one is no more correct in his prediction of another's pOpular- ity than when no mutual feeling exists between them. That is, if A is chosen by B, he is no more correct in recog- nizing B's pOpularity when he also chooses B than when he does not. 8. Hypothesis VIII was supported. It was concluded that there is no significant difference in the accuracy of one's prediction of another's pOpularity whether he is chosen or rwejected by the other. That is, if A is chosen by B, he is no more correct in recognizing B's popularity than when lwe is rejected by B. 9. Hypothesis IX was supported. It was concluded, there- ‘forwe, that there is no significant difference in the l02 subject's prediction of his own sociometric popularity regardless of whether he perceives himself as enjoying "high" or "low" sociometric status. That is, if one thinks" that he has a high pOpularity in his group, he is no more correct in recognizing his pOpularity than when he thinks he has a low pOpularity. IO. Hypothesis X was supported. It was concluded that the accuracy of the subject's prediction of the object's socio- metric status is moderate but significantly more accurate when the object enjoys a "high" rather than "medium-low" sociometric status- That is, if B is pOpular, A is more correct in recognizing B's popularity than when 8 is unpOpular. . It should be noted, however, that only 33% of the sub- jects were significantly more correct in their guess of others' pOpularity when the others were pOpular rather than unpOpular. Summary of Discussion Congruency seems strongly associated with "subjectivity" and this probably leads the subject to inaccurate predic- tions. When the subject perceives a positive attitude from the object, it seems to lead him to a moderately more accurate prediction than when he perceives a neutral- negative attitude. When the subject perceives a positive attitude from the object which is accompanied with mutuality between them, it seems to increase further the accuracy of 3’39": "I: . u - uca ill... v 5 -_ an IQ} P‘s - -.-—-4.- the subject's prediction of the object's attitude toward him. When the subject perceives a neutral-negative atti- tude from the object with whom he has a non-mutual relation- ship, his prediction of the object's attitude toward him often seems to be misleading. Furthermore, when the sub- ject perceives a neutral-negative choice from the object, the presence of mutuality of feelings between them does not seem to lead the subject to greater accuracy in pre- dicting the object's attitude toward him as it did when the subject perceived a posltive choice. When the subject perceives himself as receiving a positive choice from the object, it leads him to greater accuracy in predicting the object's attitide toward him only when mutuality of feelings exists between them. When a dyadic relationship accompanies both a "posi- tive” mutuality and a "positive" congruency, it is self- evident that the subject's prediction of the object's atti- tude toward him is lOO per cent correct. In other words, if the subject perceives himself as receiving a positive choice from the object and he also has a positive mutuality with the object, his prediction of the.object's attitude toward him is always correct. The researcher, therefore, examined to what extent the subject's prediction of the object's attitude toward him is significantly more accurate when the subject has both the mutual and congruent relationships with the object as |O4 contrasted with when he does not. The analysis shows that thirty-nine out of forty-nine subjects (80%) were signi- ficantly more accurate in their predictions of the object's attitude toward them when they had both "mutual" and "con- gruent" relationships with the object as compared with when they did not. On the other hand, none of the subjects were significantly more accurate in their predictions of the object's attitude toward them when they had either "mutual- ity" and "congruency", or when they had neither as compared to when they had both. Ten out of forty-nine subjects (20%) did not show a significantly better accuracy in pre- dicting the object's attitude toward them regardless of whether they had both "mutual" and "congruent" relation- ships with the object or not. Ausubel 3, 4, 5, 6), Schiff (46), and Trent (65, 66) reported that the type of sensitivity that enables indi- vidual A to predict accurately how individual 8 feels toward him is unrelated to the type of sensitivity that enables him to predict accurately how the entire group accepts B. These two unrelated sensitivities, one of which is respon- sible for the accuracy of predicting the object's feeling toward the subject and the other which is responsible for the accuracy of predicting another's sociometric status, seem to account for the fact that there is no significant difference in the accuracy of the subject's prediction of'the object's sociometric status regardless of whether l1e perceives himself as receiving a positive or neutral- A'--*‘ '4". 'I—nyum- t . 4 - '- iOS negative choice from the object. It was found that if congruency and mutuality of feeling exist between the subject and the object, then accuracy of prediction occurs. When congruency is singled out alone, however, the subject shows no difference in his accuracy in predicting an object's attitude toward him regardless of whether or not he perceives the object's feelings toward him to replicate his own feelings for the object. On the other hand, accuracy is greater under condi- tions of mutual feelings between individuals. But why does the coexistence of mutuality and congruency increase accu- racy to nearly perfect? These are questions which should be examined in future studies. The accuracy of the subject's prediction of the object's feeling toward him largely depends on the coexistence of mutuality and congruency, especially the former. In other words, if mutuality exists, the subject's perception of the object's attitude toward him becomes closer to the object's actual attitude. However, since the kind of atti- tude the subject perceives from the object is unrelated to the accuracy of his prediction of the object's socio- rnetric status, it is anticipated that there is little I'elationship between mutuality and accuracy of the sub- ghect's prediction of the object's sociometric status. Common sense tells us that most peeple are morehecepr~ ti‘ve to positive than negative feelings from others. It was also found in this study that the accuracy of the "3 .I 7-2 . lO6 Tr-‘r-Tr: Tu .: ~1- r subject's prediction of the object's attitude toward him is perfect when the positiveness of attitude is accompanied by the coexistence of mutuality and congruency. This seems to indicate that "positiveness" of the object's attitude plays an important role in the subject's prediction of the object's attitude. Pepple probably do not always see the environment as it is but instead perceive it in terms of their own needs. Many of them probably become more recen- tive to another's positive rather than negative feeling toward them. While the "positiveness" of the object's feeling or attitude toward the subject plays an important role in the subject's prediction of the object's attitude, it does not play any significant role in his prediction of the object's sociometric status. It would appear then that accuracy of the subject's prediction of the object's socio- metric status is unrelated to congruency, the kind of atti- tude the subject perceives from the object, mutuality and the object's actual attitude toward the subject. It may be concluded, therefore, that the ability to predict another's sociometric status is unrelated to the like- disllke relationship between the person predicting and the person to be predicted. The self-awareness of one's own sociometric pOpularity is unrelated to the accuracy of one's prediction of his sociometric popularity. One's subjectivity seems to play a large part in his prediction. It may be considered that lO7 self-awareness of one's own sociometric pOpularity is a function of the person chosen as well as the person choosing. The results in this study seem to indicate that the object's actual sociometric status is an important factor in the subject's prediction of that status. It is also considered, however, that other factors such as the object's self-confidence, his behavior in the group and his actual relationship with and perceived relationship of the group members may play a role just as important as the object's sociometric status. This point should be explored further in future studies. Reconsideration and Implications for‘Further Study The methodology used in this exploration of the relation- ship between feelings and perception of like-dislike and sociometric status has been limited. However, it proved complex enough to permit some quantitative measures from one specific point of view. But this study, as well as other similar studies, is just a starting point. It is clear that some detailed qualitative studies are necessary to uncover the variety of processes which the present researcher has treated. For example, nearly one half of the subjects were significantly more accurate in predicting the objects' feelings toward them when they had "mutual" rather than "non-mutuaJ' relationships with the objects. The other half of the subjects, however, did not show a significantly better accuracy either when they had "mutual" I: ._.-w _. '5‘ \. .‘u-T‘. - or "non-mutual" relationships with the objects What accounts for such a difference? The present study does not answer this question. It may be partially solved by careful clinical interviews with the subjects on the nature of their choices and guesses and the cues used by them in predicting others' feelings toward them. The present investigation has not limited the number of choices given and guessed by the subjects. It can be speculated that the mutual relationships of high priority choices would be stronger and therefore they would provide more salient cues as to a member's preferences than would the less intense relationships. The same logic may be applied to the mutually rejecting relationships though our culture definitely limits the expression of rejection and it may not provide the cues as clearly as would the positive relationships. This cultural modesty or politeness might have led the subjects to prefer the neutral over the nega- tive choice. The researcher might have gained a clearer picture by limiting the subjects' responses to only the t0p three or four positive and negative choices. While the common approach of sociometric studies is the description of structure, the present researcher intended to use a similar method to generalize the results for persons into indices of "tendency". Therefore, it was desirable to control certain situational factors for the indices to have the intended meaning. It has been noted that in the use of sociometric tests, the range and length lO9 of acquaintance are not necessarily equal for all persons in a group. The researcher could not completely eliminate acquaintance prior to the camp. It would be desirable to control this condition of "prior" acquaintance in future studies. The groups in the present research have been homogeneous with regard to age, sex and socioeconomic backgrounds. The researcher does not know what would happen to his findings if similar investigations were applied to different groups. The data in this study were collected inlone admini- stration of the sociometric-like measures. It would be interesting to know if any tendencies of and relations between sociometric variables obtained are stable and general over repeated interviews in the same group and at different times in different types of groups. It has been noticed that certain subjects showed con- sistent accuracy in their sociometric perception regardless of their relations with the objects. while some showed consistently inaccurate perceptions. It could be spec- ulated that there are some other important related factors which may affect the subjects' accuracy of sociometric perception such as family background, intelligence, group experience, behavior and personality characteristics. In the present study the researcher was simply con- cerned with the relations between various sociometric variables or relationships and the accuracy of perception. A question arises here as to what are the "independent .9 ..a'L. "4 5. _‘ 1 - ‘Y‘ ' i llO The researcher stated three of ten as directional hypotheses. Previous research led to the expectation that the null hypothesis should be accepted for the other seven hypotheses. These were tested with statistical models used to sense differences. To the extent that the sample size, and sensitivity of the models could be judged adequate to sense a difference that might be considered practically significant, one would argue that the test of these seven hypotheses in the null form is adequate. However, since this is a judgment difficult to make when the sample size is small and the metric without real meaning, the researcher recommends that further research in this area use correla- tional statistics so that the magnitude of the relation- ship may be Judged and the inferential logic leads more clearly to specific conclusions. APPENDIX SOCIOMETRIC SCHEDULE SOCIOMETRIC SCHEDULE I would like to know how well you can tell how you feel about the boys in your cabin and how the boys in your cabin feel about you. When we do things together with peOple, it is nice to know how we fee1 about them and how they feel about us, because then we can know better how to act with them. It is quite natural that we like some peOpie more than others. Some friendships take longer to deveIOp than others. No one beside myself will ever see the answers, so don't be afraid to tell me what you really think. There are three boxes here and you put the pictures of your cabinmates in these boxes. (Put the boxes and pictures in front of the subject.) It doesn't make any difference how many pictures you put in each box. |l2 ’m.__._---_ II. III. ll} All right? Here is the first question. Put the pic- tures of the boys: a) Whom you like in the red box. b). Whom you don't like in the blue box. c) Whom you don't care whether you like or not in the white box Before you start, would you tell me what you are supposed to do? It is natural that we can't please everybody. In other words, some peOple like you more than others. Put the pictures of the boys: a) those you think would like you in the red box. b) those you think don't like you in the blue box. c) those you think don't care whether they like you or not in the white box. Before you start, would you tell me what you are supposed to do? This time, I want to see how well you can tell how popular each boy is, that is, how many boys in the cabin like him. Put the pictures of the boys: a) whom you think that most of the boys in your cabin like in the red box (more than lO boys). b) il4 whom you think that very few boys in your cabin would like in the blue box (less than 5 boys). whom you think that about half of the boys in your cabin would like (between 6 and 9 boys). .w, ~ 1.: p‘. .3 warm" I . \fl l0. BIBLIOGRAPHY Adams, H. F., "The Good Judge of Personality," Journal 91_Abnormal and Social Psychology, I927, 22, l72-l82. Asch s. 5., Social Ps cholo Englewood Cliffs, N.J., Prentice-Hall, l95%. Ausubel, D. P. "Reciprocity of Acceptance amont Ado- lescents, A sociometric Study." Sociometry, I953, l6, 339-349. Ausubel, D. P. "Sociempathy as a Function of Socio- metric Status in an Adolescent Group," Human Relations, I955, 8, l75-l84. Ausubel, D. P., Schiff, H. H., and Tasser, E. 8., "A Preliminary Study of Devel0pmental Trends in Sociempathy: Accuracy of Perception of Own and Others' Sociometric Status," Child ngelogment, l952, 23, lll-l28. Ausubel, D. P., and Schiff, H. M. "Some Intrapersonal and Interpersonal Determinants of Individual Differences in Sociempathic Ability among Ado- lescents," Journal 21.Social Psychology, i955, 4' a 39-56- Ausubel, D. P., and Schpoont, S H.,"Prediction of Group Opinion as a Function of Extremeness of Predictor Attitudes," Journal g£_Social Psycho- logy, I957, 46, l9-29. "“"" Borgatta, E. F. "Analysis of Social Interaction and Sociometric Perception," Sociometry, I954, l7, l7-32. ' Borgatta, E. "The Stability of Interpersonal Judgments in Independent Situations," Journal gI_Abnormal and Social Psychology, l960, 50, l53-l94. Borgatta, E. "Rankings and Self-Assessments: Some Behavioral Characteristics Replication Studies," Journal gj_Social Psychology,_l960, 52, 279-307. Borgatta, E. "Analysis of Social Interaction and Socio- metric Perception," In J. L. Moreno et. al. (ed.), The Sociometr Reader, Glencoe, Ill: The Free Press, 9 D, 272-297. IIS l2. l3. l4. '9. 20. 2|. 22. 23. 24. l|6 Brunner, J. and Tagiuri, R. "The Perception of Pepple," In G. Lindzey (ed.), Handbook gj_Social Psychology, Cambridge, Mass: Addison-Wesley, '95:, 3 - 5 . Campbell, J. 0., and Yarrow, M. "Perceptual and Be- havioral Correlates of Social Effectiveness," Sociometgy, l96l, 24, l-20. Chowdhry, K. and Newcomb, T. M. "The Relative Abilities of Leaders and Non-Leaders to Estimate Opinions of Their Own Groups," Journal 2I|Abnormal Egg Social Psychology, l952, 47, 5l-57. H. L. "An School and Cogan, L., Conklin, A., and Hollingworth Experimental Study of Self-Analysis, Society, l9l5, 2, l7l-l79. h Combs, A. W. and Snygg, D. Individual Behavior, New York: Harper and Brothers, i959. Cottrell, L. 8., Jr. "The Analysis of Situational Field in Social Psychology,‘ American Socio- logical Review, l942, 7, 370-3 2. Cronbach, L. J. "Processes Affecting Scores on 'Understanding of Others' and 'Assumed Similarity',‘ -Psychological Bulletin, I955, 52, l77-i93. Dudycha,~G. J. "Self-Estimates and Dependability," Journal g£_Social Psychol0gy, l940, l2, 39-53. Dymond, Rosalind F, "A Scale for the Measurement of Empathic Ability," Journal gI_Consulting Psycho- 12314 I949, '3, '27-'33. Dymond, Rosalind F. "Personality and Empathy," -Journal gj_Consulting Psychology, I950, l4, 343-350. Eng, E. W. "An Approach to the Prediction of Socio- metric Choice," Sociometry, l954, l7, 329-340. Fiedler, F. E., Leader Attitudes and Grou Effective- Iliin ness, Urbana, Ill: University of ois Press, I953. Fiedler, F.E., Hartman, W., and Radin, S. The Rela- E tionshigs g: Intergersonal Percegtion-;_ ffect- veness ig_Basketball Team, Urbana, ll: Bureau of Research and Service, University of Illinois, Mimeographed, l952. 25. 26. 27. 28. 29. ll? Fiedler, F. E., and Senior, Kate. "An Exploratory Study of Unconscious Feeling Reaction in Fifteen Patient-Therapist Pairs," Journal of Abnormal gflg_30cial Psychology, l952,”Z7;'ZAB:A55. Frenkel-Brunswlk, Else. "Mechanisms of Self-Perception," Journal 2:,Socjal Psychology, I939, lO, 409-420. Frenkel-Brunswik, Else. "Personality Theory and Per- ception," in R. R. Blake and G. V Ramsey (eds.) Perception: An Aggroach Lg Personalit , New York, the RonaTd Press, l95l, 353-Ei9. Gage, N. L. "Explorations in the Understanding of Others " Purdue University Study 91 Higher Echation, '95i9 79’ -9 o . Gage, N. L. "Accuracy of Social Perception and Effect- iveness in Interpersonal Relationships," Journal 2i_Personality, l953, 22, l28-l4l. Gage, N. L., and Suci, G. "Social Perception and Teacher-Pupil Relationships," Journal 31 Educa- tional Psychology, l95l, 42, l44-l53. Goodnow, R. W., and Tagiuri, R. "Religious Ethno- centrism and Its Recognition among Adolescent Boys," Journal 9: Abnormal and Social Psychology, l952, 4?; 3'6’320- Heider, F. The Psychology 9: Intergersonal Relations, New York:, John Wiley and Sons, Inc , l958. Jersild, A. T. In Search 31 Self, New York: Columbia University Press, l952. Jackson, T. A. "Errors of Self-Judgment," Journal 9: Applied Psychology, l929, i3, 372-377. Jones, L. V., and Fiske, D. W. "Models for Testing the Significance of Combined Results," Psychological Bulletin, l953, 50, 375-382. . Katz, L., Tagiuri, R., and Wilson, T.R. "A Note on Estimating the Statistical Significance of , 0 Mutuality," Journal g1 General Psychology, l958, 58. 97-i03. Kogan, N. and Tagiuri, R. "On Visibility of Choice and Awareness of Being Chosen," Psychological ReDOFtS, l958, 4, 83-86. k)! {D 39. 40. 4|. 42. 44. 45. 47. 48. 29. 50. 5|. ll8 Lingdgren, H. C., and Robinson, Jacqueline, "An Evalua- tion of Dymond's Test of Insight and Empathy," Journal gj,Consultino Psychology, I953. l7, l72-l76. Luce,_R. D., Macy,_J. Jr., and Tagiuri, R. "A Statis- tical Model for Relational Analysis," Psycho- metrika, I955, 20, 3l9-327. Mead, G. H. Mind, Self, and Societ , Chicago: Univer- sity of Chicago Press, l93 . Morenoé J. L. Psychodrama, New York: Beacon House, Inc. I9 5. ' Moreno, J. L. Who Shall Survive? New York: Beacon House, Inc., l953, (Rev. Ed.) Polansky, N., Lippit, R. and Redl, F. "An Investigation of Behavioral Contagion in Groups," Human Rela- Ideas, '950. 3. 3|9-348. Rogers C. R CIient~Centered Theragy Boston: Houghton Mifflin co.,'l95i. ' Sakoda, J. M., Cohen, B. H., and Beal, G. "Test of Significance for a Series of Statistical Tests," Psychological Bulletin, l954, 5|, l72-l75. Schiff, H. M., "Judgmental ReSponse Sets in the Per- ception of Sociometric Status," Sociometry, I954, l7..207-227. Sheerer, Elizabeth T. "An Analysis of the Relationship Between Acceptance of and Respect for Self and Acceptance of and Respect for Others in Ten Coun- seling Cases," -Journal 2:.Consulting Psychology, |949. l3. l59-l7 . Siegel, S. Nonparametric Statistics for he Behavioral * Sciences, New York: McGraw Hill, Inc., I953? Speroff, B. J., "A Measure of Mass Empathy," in 133 Status 2: Emgathy gg g Hygothetical Construct Lg Psychology Today, A symposium of the American Psychological Association, Mimeograph, Psychology Department of Louisiana State University, I953. Sullivan, H. S. Interpersonal Theorylgg Psychiatry, New York: Norton, I953. Symond, P. M. The Ego and the Self ___ , New York: Apple- ton, Century, Crofts, Inc., l95l. 52. 54. 55. 56. 57. 58. 60. 6|. 62. ||9 Tagiuri, R. "Relational Analysis: An Extension of Sociometric Method with Emphasis upon Social Perception," Sopiometry, l952, l5, 9l-lO4. Tagiuri, R. "Social Preference and its Perception," In R. Tagiuri and L. Petrullo (eds.) Person Perce tion and Interpersonal Behavior, Stanford, Calif: Stanford University Press, l958, 3l6-336. Tagiuri, R., Blake, R.R., and Brunner, J. 3. "Some Determinants of the Perception of Positive and Negative Feelings in Others," Journal of Abnormal and Social Psychology, I953. 48, 585-5927 Tagiuri, R., Brunner, J., and Blake, R.R., "On Relation Between Feelings and the Perception of Feelings among Members of Small Groups," In E.E. Maccaby, T. E. Newcomb, and E.L. Hartley (eds.), Readings 21 Social P cholo , New York: Henry Holt and CO., l958, llO- I . Tagiuri, R., Brunner, J., and Kogan, N. "Estimating the Chance Expectancies of Dyadic Relationships within a Group," Psycholggical Bulletin, I955, 52, l22-l3l. Tagiuri, R , and Kogan, N. "The Visibility of Inter- personal Preferences,” Human Relations, I957, IO: 385‘390. . Tagiuri, R. and Kogan,-N. "Personal Preference and the Attribution of Influence in Small Groups," Journal 2: Personality, I960, 28, 257-265. Tagiuri, R., Kogan, N., and Brunner, J. S. "The Transparency of Interpersonal Choice," Socio- m£££XJ '955’ '89 368’379- Tagiuri, R., and Petrullo, L. Person Perce tion and Interpersonal Behavior, Stanford, Calif: Stanford Un versity Press, I958. Taylor, F. K., "The Three Dimantional Basis of Emotional Instructions in Small Groups, Part I and Part II." in J. L. Moreno et. al. (eds.) 122 Sociometry Read r, Glencoe, III: The Free Press, l9 0, 3504557. ' Taylor, F. K. "Awareness of One's Social Appeal," in J. L. Moreno et. al. (eds.) The Sociometr Reader, Glencoe, Ill: The Free Press, l9 0, - 7i. 65. 64. 65. 66. I20 Toeman, Z. "Clinical Psychodrama: Auxiliary Ego Double and Mirror Techniques," Sociometry, l946, 9, l78-l83. Toeman, Z. "The Double Situation in PSYCHOdrama," W, l95l ,,_I , “436-446. Trent, R. "The Relationships of Anxiety to Popularity and Rejection among Institutionalized Delinquent Boys," Child Development, I957, 28, 379-384. Trent, R. 0., "Anxiety and Accuracy of Perception of Sociometric Status among Institutionalized Delin- quent Boys," Journal p: Genetic Psychology, I959 94, 85-9l. RUGM USE ONLY