THE EFFECTS OF CREATIVITY, INTELLIGENCE AND COMMUNICATOR EXPERTISE ON CONFORMITY ’ THESIS FOR THE DEGREE OF PH. D. MICHIGAN STATE UNIVERSITY YVON I’EPIN . n. i'fi‘nl I'v‘ ' ll mmuynzlwul 11“le I! mu IIIIIIHJIILII ', LIBRARY Michigan State University This is to certify that the thesis entitled The Effects of Creativity, Intelligence and Communicator Expertise on Conformity presented by Yvon Pepin has been accepted towards fulfillment of the requirements for /A. FD degreein gY(/‘/d.j/ : 5M4 Major professor Date fly /5: /(/7/ ,mmishw 0..., O 3"; : I wbk-fi ABSTRACT THE EFFECTS OF INTELLIGENCE, CREATIVITY, AND COMMUNICATOR EXPERTISE ON CONFORMITY By Yvon Pepin On the basis of a theoretical framework developed by the au- thor, this research investigated the interdependence of intelligence, creativity and communicator expertise as they affect conformity behav- ior. It was hypothesized that intelligence, as measured by available tests, is an index of intellectual conformity and therefore should be positively related to reliance on information provided by others, espe— cially when they are presented as experts. Creativity, supposedly an index of independent thinking, was expected to show the reverse pat- tern. Similarly it was hypothesized that the effects of differential levels of expertise of a communicator would very much depend upon the intelligence and creativity of the person receiving the communication, with those of lowbintelligence and high-creativity being relatively in- sensitive and those of high-intelligence and low creativity being high- 1y sensitive to such variations. On the basis of a median split on both indexes, four groups differing in their level of creativity and intelligence were formed. Yvon Pépin The subjects in each of those groups were then assigned in equal num- ber to either a low or high confederate expertise condition. Confor- mity was assessed by a perceptual judgment task developed by Crane (1970), in which the subject and a confederate give their estimates of the number of dots flashed onto a screen. The results failed to support the hypotheses concerning the main effects and simple interactions of the three variables. However the significant (p (.02) three way interaction revealed a high degree of interdependence among the independent variables. When that inter- action was investigated by way of a simple - effect analysis, it was found that differences in intelligence significantly affect conformi- ty only when the subject is exposed to a high expertise communicator; and even then its effects may be reversed depending on the subject's concurrent level of creativity with a high level of creativity lead— ing to a negative relationship and a low level of creativity leading to a positive relationship between I.Q. and conformity. Creativity significantly lowered conformity only when both intelligence and com- municator expertise were low. High expertise significantly increased conformity and low expertise similarly reduced it only when the sub— jects displayed discrepant levels of intelligence and creativity, with the low intelligence — high creativity group behaving contrarily to predictions. An interpretation of these findings was presented in the discussion section. Yvon Pépin A If g . '1'; [2' I [7/ ApprovedJK/A/v—e /:I~é a,“ 4x. Jeanne E. Gullahorn \g Committee Chairman Date December 15, 1211 THE EFFECTS OF INTELLIGENCE, CREATIVITY, AND COMMUNICATOR EXPERTISE ON CONFORMITY By Yvon Pépin A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1971 4' ACKNOWLEDGMENTS When one turns back to his graduate career, he is always impressed by the number and the quality of the people who partici- pated in it. The person I am most indebted to is certainly Dr. Jean— ne Gullahorn. Since the moment when I made the happy decision to elect her as the Chairman of my graduate committe, she has cons- tantly and without any reservation provided help and understand- ing at every step until the moment of my graduation. It is very seldom that one finds from a professor such readiness and willingb ness to be available in spite of the many tasks her status of In- terest Group Chairman imposes on her. I especially appreciated the fact that she was able to help me to develop and refine my own ideas rather than forcing me into a more conventional theoretical position. I am also especially indebted to Professor William Crane who was very influential, both by his writing and his comments, in planning the experimental phase of this study. I also appreciate his providing me with the facilities of his own experimental mate- rial for the purpose of this research. Dr. Lawrence Messé's comments, especially at the prepa- ratory stage, were essential to the successful completion of my 11 research. Although I first failed to recognize its beneficial ef- fects, I deeply appreciate his curbing of those usually unrealis- tic tendencies and projects of mine. Finally, long informative discussions with Drs Ellen Strom- men and John Paul McKinney permitted me to develop the theoretical framework which triggered this research. I acknowledge the wonderful collaboration I received from the teachers, the counselors and the secretarial staff of Haslett High School and especially the principal, Mr. Donald Ualdron. My colleagues Ralph E. Cooper and Robert K. Goodman gene- rously sacrificed some of their own busy time in helping me prepare the computer analysis of the data. Lars Olson, Craig Kobylka and Mark Calder also deserve com- mendation for the competence they manifested in their role as confe- derates in this experiment. Finally, the most important person in my graduate career is my wife, Michele, who, according to circumstances, acted as a coworker, a counselor, a source of ideas and discussions and a secretary, in ad- dition to being a companion beyond compare. Her support through this whole enterprise was far greater than what could realistically be ex- pected. If formal requirements did not exist, she would be as much entitled as I am to that Ph.D. diploma. iii TABLE OF CONTENTS Page LIST OF TABLES .............................................. vi LIST OF FIGURES ............................................. vii CHAPTER I. INTL‘ODUCTION COOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOCC 1 General Cognitive Framework .................. 2 Creativity, Intelligence and Conformity ...... 6 Expertise and conformity eeeeeeeeeeeeoeeeeeeee 10 Differential Effects of Expertise Depending on Creativity and Intelligence eeeeeeoeeeeeeeoeeeeooeeoeeeoo 12 II. MEWOD OOOOOOOOOOOOOOO00.00.0000...0.00.00.00.000. 15 subjeCtS eeeeeooeeeeeeeeeeooeoeeeeeooooeeeeeee 15 Materials eeeeeeeeeeeeeeeeeeeeeeeeeeeoeeoeeeeo 15 Procedure .................................... 2O DeSign eeeeeeeeeeeeeeoeeeeeeeeeeeeeeeeeeeeeeee 23 III. MULTS .OOOOCCOOCOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO. 24 Creativity and Intelligence .................. 24 Compliance and Conformity .................... 26 Acquaition and TranSfer eeeeeeeeeeeeeeoeeeeeo 37 Control SUbjeCtS Differences oeeeeeeeeeeeeeeee 37 31mm 0f Results 0.00.0000000.000.000.000... 39 Iv. DISCUSSION .OOOOOOOOOOOOOOOOOOOO0....0.0.0.0000... 41 creatiVity and Intelligence oeoeeeeeeeeeeeeeee 4'1 corlformity .00....COCOOCOOOOICOCOOOCOOOOCOO... 42 Intelligence and Conformity .................. 43 iv O s w o . a O o I Q C L O I a U I O O o O O a O O u I F C D O u C -) J 9 v i - er 0 Q t ~ ' I I' O i I o O I e o . O o a U 0 O 9 I I ‘ ‘ e 0 ~ 0 t c o _ - s v ' o n . C b I K O I . Iv ‘ I l O u C I ‘ ‘ l I a J I I O a d c d Creativity and Conformity .................... Expertise and Conformity ..................... Learning Conformity Dispositions ............. Implications and Suggestions ................. BIBLIOGIiAPHY O...0.0.0.0....0.000.000.0000.000000000000000COC APPENDIX A. B. C. D. Correlations Between the I.Q. Tests Employed and RegTGSSion Equations eeeeeeeeeeeeoeeeeeeee Letters, Test Booklet and Correction Keys 00000000000000000000000...e00000000000000 Reliabilities of the Creativity Tests mployed OOOOOOOOOOOOOOOQQO0000000000.0000000. Intelligence and Creativity Means of the Different Groups Employed .................... Detailed Account of the Experimental seSSion eeeeeeeeeeeeeeeeeeeeeeeeeeeeoooooooeeo Page 45 46 50 51 54 59 6O 82 83 a r g a r r o I O O I D I D O C C O I ' 0 - u s a 1 s u n t I . I a x o . . e - a O I v - v a a O o 0 Table 1. 2. 3. 4. 5. 7. 8. LIST OF TABLES Creativity Tests Used in Three Studies .............. The Relationship Between Creativity and Intelligence Test Scores ................... Correlations Among Creativity, Intel- ligence, and Teachers'Assessments .............. Summary of Analysis of Variance ..................... Group Mean Conformity Scores According to Conditions ........................ Significant Simple Interactions and Simple-Simple Main Effects for AchiSition .OOCOOOOCCOOOOOOOOOOOOOOOOOOOOOOOO.. Significant Simple Interactions and Simple-Simple Main Effects for Transfer eeeeeeeeeeeeeoooeoeoooooooeoooooooooeee Mean Estimates of the Confederates, the Control Group, the Total Ex- perimental Group and the High and Low Expertise Groups in the.Acqui- sition and Transfer Phases ..................... Page 17 21 25 27 32 33 36 | Q . - O O < o 4 r I O I O .. o e o e . ~ v r o o no. LIST OF FIGURES Figures 1. Interaction of Expertise and Trials for the Acquisition Phase OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOCOIOOOO. 2. Comparison of Trial Blocks Effects for the Experimental and Control Group .COCOO0.0.0000000000000000000000...0.0... Page 30 O... CHAPTER I INTRODUCTION The concept of conformity is central in current social-psyh chological theory. In fact, rather than arguing that individuals in a given society come to behave in a similar manner and manifest simi- lar views of the world because they face the same kind of reality, current theories focus on how a particular social group imposes its ways of behaving, values and attitudes on individual members. Actual- ly, if we examine Asch type experiments, it looks as if they are right at least half of the time since, on the average. SQ“ of the subjects conform. In general, social-psychological studies of conformity have manipulated such situational determinant as the ambiguity of the sti- mulus (Asch 1956; Patel and Gordon 1960; Kelley and Lamb, 1957), group size (Asch 1951; Gerard and Uilhelmy, 1968), size of the majority (Asch 1956; Tuddenham 1961), group acceptance (Kelley and Shapiro 1954: Har- vey and Consalvi 1960) and group expertise, status or prestige (Cohn Yes and Brown, 1961). In interpreting these studies, some infer that the results should apply to most persons, if not everybody. Such a gratuitous assumption overlooks the fact that in Asch's initial expe- riment, only 32% of the subjects conformed. Situational manipulations have increased that percentage to 80% (Allen, 1965) but never totally. By manipuling such personal variables as occupation however, Crutch- field‘s experiments (1962) have increased conformity responses to a perfect percentage (100% for army officers) and decreased it to prac- tically zero (for scientists). While research in the area of personality deals only indi- rectly with conformity, a number of studies (Crutchfield 1962; Rak- each 1960; Barron 1963; Adorno et a1. 1950; Uitkin et al. 1954 etc.) indicate that individual differences in conformity are sufficiently large to challenge an approach to social phenomena based primarily on the concept of conformity. Although social-psychologists espouse the view that social psychological phenomena cannot be explained unless one approaches them from the point of view of an interaction between the person and social reality, still most of the studies focus on one or the other aspect separately, thus neglecting their transactional character. The present study constitutes an attempt to investigate the transactional relationship between personal and situational varia- bles as they affect conformity behavior. In reviewing pertinent literature, I shall first discuss ge- neralities regarding cognition, followed by a consideration of how certain cognitive variables may affect conformity and may interact with situational variables in such an effect. General nggitive Framewor .1 The interaction of external and internal factors is central in the study of cognition, especially 1Amore detailed theoretical treatise appears in Pépin (1971). cognitive development. Piaget (1952, 1964), for example, shows how mature cognitive structures are attained by a constant and complex interplay between internal (schemes) structures and external (reali- ty) features. He describes how more effective cognition develops through a progressive adaptation or equilibration of internal sche- mas and external structures as a subject encounters difficulties and adaptively realigns his activity. The Optimum equilibrium is attai- ned when there is an effective counter balance between the two func— tions of assimilation of the given to internal schemas and accommo- .d§tigg,of the schemes to pertinent reality features. By over-empha- sizing the role activity of the subject in cognitive development howe- ver, Piaget tends to consider the process as an exclusively inductive, spontaneous and maturational conceptual system formation where other persons have little to do except interfere. Piaget's theory would be more adequate if it took into ac- count the cultural milieu in which the human organism develops. The child is born in a society which has already constructed and imposed a conceptual system of categories over physical, as well as social, reality. Communication, one of the bases of any society, is possi- ble only if every individual is made to learn these categories and their logical interrelationships. Most of the time, that conceptual system is so advanced, extensive and complex that an individual could not, in a life-time, discover even a small part of it if he were to personally experience every fact and induce concepts out of that experience. So he has to be instructed and made to accept certain conceptualizations without actually being able to verify their va- lidity. Vigotsky (1962) presents an alternative point of view. He distinguishes between fggggl and spgntaneous concepts; the first being those actually used in the society, the second those developed out of personal experience. He argues that instruction accelerates cogniti- ve development because it structures cognitions by setting models. Instruction then provides formal concepts which first are empty struc- tures that experience progressively fills in. From Vigotsky's view- point therefore, spontaneous concepts benefit from the high degree of structuring of the formal concepts and, at the same time, formal con- cepts acquire utility through actual experience. If we apply such a framework to the theory of socialization, an interesting hypothesis emerges. We may hypothesize that sociali- zation agents differ in the degree to which they emphasize formal or spontaneous conception when dealing with the socializee. One aspect of such a question is to what extent are they willing to let sponta- neous conceptualization interfere with formal logic in the hope that the socializee will solve the problem by himself. Another aspect is to what extent will they let the socializee question the validity of formal concepts. To the extent that formal concepts are emphasized in socia- lization, the locus of validity evaluation.will shift from an inter- nal to a cultural perspective. Because of their lack of cognitive experience, children are apt to make errors and to be corrected fre- quently with the result that they may come to rely on others to 6V8! luate their concepts. Another consequence is that they will focus their learning attempts at understanding cultural knowledge and ga- thering information rather than at adapting to their personal rea- lity. Their level of self-esteem will depend on the extent to which others acknowledge their ability at the job. Schactel thus descri- bes the long term consequences of this same phenomenon: In addition to the closure of the world which results from the transmission of a familial and / or cultural viewpoint, parental curbing of the child's exploratory drive can also be a factor that interferes with the world openness of the child and often leads to a more or less powerful strongh- tening of the tendency to avoid the unknown and remain em- bedded in the familiar (1959, p. 187). Hess and Shipman (1965) feund that a type of control based on formal laws (in that particular case ”status oriented control“) is correla- ted with a cognitive orientation that is geared toward social adjust- ment rather than objective conceptualization. If spontaneous conceptualization is emphasized in sociali- zation however, the locus of validity estimation stays internal ra- ther than external. As Piaget describes it, a feeling of competen- ce and control thus arises from the fact that a spontaneous concept proves, for the moment at least, functional and adaptive. Intellec- tual development actually functions in the same way homeostatic pro- cesses do. Being less coached by socialization agents, the child will focus his learning attempts to those aspects of the environment which cause him adaptation (in the Piagetian sense) difficulties, thus centering on the discovery of what is not understood rather than on the understanding of what is discovered. His selfbesteem will de- pend an the degree to which he is actually effective and competent in doing so. Pbr purposes of exposition, I have described extremes on a differential vs. cultural control continuum. In the real world, however, we may expect to encounter different combinations of the two dispositions. Creativigy, Intelligence and Conformigz. Having conside- red some consequences of formally vs. spontaneously oriented socia- lization, let me now relate these concepts to the research on intel- ligence and creativity. In Guilford's theory of convergent and di- vergent thinking (1950), as well as in Getzels and Jackson's book on Creativity and Intelligence (1962) and in Barron's (1963, 1967) Tor- rance's (1963), Crutchfield's (1962) and Hanson and Kogan's (1965) works, it is assumed that the creative individual is more independent, less conformist, less socially oriented and more original than the less creative person. By contrast, the highly intelligent but less creati- ve person is pictured as a more socially dependent, conformist and less original individual. Such generalizations are plausible, since 1.0. tests consist of "a sample of the cognitive function tasks that are‘gr gggl.and egpected in any large cultural group or subgroup” (English and English, 1958; italics not in original). In general, findings from studies regarding creativity and intelligence concur with the previous- ly outlined consequences of what we may call ”differing socialization styles". For example, the generalizations regarding the locus of e- valuation of concepts constitutes the core of Rogers' (1959) theory of creativity. Weisberg and Springer (1967), in a study of environ- mental factors leading to creativity, underline many family patterns which lead the child to rely on himself and to be independent of others in dealing with reality. Getzels and Jackson (1962) also pre— sent evidence indicating that socialization patterns differ markedly in the families of highly intelligent adolescents as ccnstrasted with highly creative youth. With reference to personal functioning, Cattel (1963) COD! cludes that creative researchers are more schizothyme (critical, ex- act, precise, skeptical), have a higher level of dominance (asserti- veness, independence, unconventionality) and are more self-sufficient than individuals in the average papulation. Crutchfield, after a stu- dy of 45 research scientists in the area of missile develOpment, sum- marizes: Taken together, these findings from current studies of crea- tive persons and the other findings on personality attribu- tes of conformists offer consistent evidence for a signifi- cant empirical relationship between conformity tendencies in the person and lack of manifest creativity (1963, p. 135). Golan (1963) also concludes that creative people are more inp dependent, self-accepting, assertive, impulsive and complex than non— creative people. Wallach and Kogan (1969) also note some effects of creativity on unconventional behavior in the classroom with young chil- dren. They also mention a factor that has been observed in other studies like Cattell's -— that is, the apparent introversion of the creative person which, interestingly enough, does not hinder them from being popular (Torrance, 1963). Getzels and Jackson (1962) find the highly creative adolescent to be much more independent in setting his values, less conventional in choosing a career and mo— re critical in evaluating cultural standards; thus creative youth appear less ready to take formal cultural conceptualization for gran, ted. A reverse pattern is found for highly intelligent, but not so creative, students. In a more direct test of the effects of crea- tivity on conformity, Barron (1967) reports a difference significant at the .001 level between originals and non-originals (as measured by Guilford's tests) on a modification of Asch's procedure and one at the .05 level on an Independence of Judgment scale. However, except for Wallach and Kogan and Getzels and Jack- son, the findings on the effects of intelligence on conformity are not so clear and the study of the effects of intelligence and the persona- lity characteristics of the highly intelligent person have been biased by the value emphasis long associated with a high I.Q. Crandall et a1. (1958) found no significant relation between peer compliance and I.Q. in either preschool or elementary school children. Berenda (1950) also found no significant relationship between yielding and 1.0. In another study by Iscoe, Williams and Harvey (1964), eliminating I.Q. from among the independent variables used in that study did not significantly chan- go the efficiency of predicting conformity behavior. In a previous ex- periment (1963) they had computed correlations between I.Q. and confer- mity separately for different age and sex groups. Ten of the twelve correlations reported were negative, but none were significantly dif— ferent from zero. So, generally speaking, the relation between I.Q. and conformity is not very strong. There is some indication in the attitude change literature that it is more difficult and requires better prepared manipulations to persuade more intelligent persons (Garment, Miles and Cervin, 1965; Hovland Janis and Kelley, 1953). unfortunately, these studies offer no data regarding how the more intelligent subjects score on creativi- ty measures; and since the two variables appear independent, we lack laboratory evidence regarding the effects of creativity on persuade- bility.2 In order to relate creativity, intelligence and conformi— ty, we thus prepose the following hypotheses regarding the effects of the first two factors on the third. Hypothesis 1: Level of creativity should decrease level of conformi- ty in a specific task (because the creative person is a usual non-conformer). Hypothesis 2: Level of intelligence should increase level of confor— mity in a specific task (because the highly intelligent person is a usual conformer). 2 There has been a controversy regarding the relationship of intelligence and creativity. Some studies show creativity to cor— relate positively with intelligence whatever the criterion (Golan 1963; Mc Nemar, 1964). Even the Getzels and Jackson study is not convincing in regard to the independence of these measures; in fact, Wallach and Kogan are the only ones whose data support such an assumption. Thornp dike (1963) argues that the available evidence does not permit the ge- neralization that there exists a pervasive dimension of individual dif- ferences called ”creativity" that is quite distinct from intelligence. Perhaps the most cogent, resolution of this controversy is provided by Maer and Stein (1955), who found creativity and intelligence to covary significantly up to a superior range of intelligence. Above that level, they found no correlation between the two factors. 10 These hypotheses lead to the conclusion that the highest le- vel of conformity in a specific task should occur when the individual combines a high level of intelligence with a low level of creativity; and conversely conformity should be minimal when the individual com- bines a high level of creativity with a low level of intelligence. With these hypotheses as a foundation, we shall now consi- der the interaction of personal and situational variables. Expertise and Conformity. One of the best documented situap tional variables affecting conformity is the expertise of the persons with whom the subject interacts. Whether expertise presumably emana- tes from the general status of the person or persons to whom the indi- vidual complies or whether it involves specific expertise in a given task, this variable has consistently been found closely related to conformity. From the status point of view, we may say that the status hierarchy of a society is that society's‘gggmal conception of exper- tise; that is, it gives to the individual who needs it a quick and ex— peditious way of assessing the expertise of another individual without actually testing that expertise. Status thus constitutes an accepted and even promoted, although oversimplified, way to evaluate the exper- tise of a communication source. Asch's data (1956) indicate that the more a group is perceived as prestigious, the more an individual will conform to its norms. High prestige children are particularly suc— cessful as sources of behavior contagion among their classmates and playmates (Polansky, Lippitt and Redl, 1950; Lippitt, Polansky and.Rosen, 11 1952; Cohn, Yes and Brown, 1961). Patel and Gordon (1960), while des- cribing a decrease in conformity from the 10th to the 12th grade, re- mark that this decline is much less evident when high prestige norms are involved than when subjects are exposed to low prestige norms. Hanipulations of credibility in the area of attitude change are often made along a status criterion, and generally such a manipulation pro- ves successful (e.g. Zagona and Barter 1966; Kelman and Hovland, 1953; and many others). In general, except for a few more discriminative studies that we will consider later, the high rank or prestige of a person in a social group gives him power to induce conformity in those of lower status. Expertise may also be considered in specific context when it is related to a given task. Especially in novel and problem situa- tions, the individual who does not bring from prior experience any im- mediate solution is likely to be more sensitive to social sources of information and imitate the behavior of others (Ferguson, 1970). There is much evidence that an individual who is perceived as being competent or expert at a given task will be more readily imitated (and hence will induce more conformity) than an individual who does not demonstrate such expertise (Hausner 1954; Crane 1970). Since this aspect of expertise will be manipulated in the cur- rent investigation, the following hypothesis is prOposed: Hypothesis 3: A person who is perceived as competent in a relevant task will be more readily imitated and will induce more confor- mity than a person who is not perceived as competent in that relevant task. 12 Differential Effects of Expertise Depending on Creativity and Intelligence. Actually, the process underlying both status - produced and task-specific expertise appears to be the same. Camp- bell (1963) describes it as a decreased reliance on internal sources of information paralleled by an increase in reliance on interperso- nal sources of information, so that responses learned by interperso- nal mediation like linguistic mediation or model imitation are more likely to be conforming than responses learned by personal experien- ce, such as trial and error. In a specific situation, any event weakening perceptually dependent or strenghtening socially dependent modes of response acquisition will increase conformity. Thus what expertise manipulations really manipulate is the quality of inter- personal sources of information. However, the picture of the highly creative individual that emerges from the preceding discussion is one of a person who has learned throughout the circumstances of his socialization to rely up pon himself and his personal experience in responding to different si- tuations. It is likely that such an individual will tend to accord more importance to ”perceptually dependent” modes relative to "so- cially dependent" modes than the average individual, regardless of how much the socially - dependent modes are stressed in a particular situa- tion. 0n the other hand, it appears that in order to have a high I.Q., much importance must be given by the subject to social as Opposed to more personally dependent types of information. ,Actually, the highly intelligent person has effectively internalized the cultural model of 13 reality and thus has learned to rely on socially - dependent modes of learning. So it seems likely that variations in the quality of the in- terpersonal sources of information should not affect the highly crea- tive person as much as the highly intelligent person, inasmuch as the creative individual attends more to the task rather than the inter- personal aspect of the situation. For example, certain personal cha- racteristics that are theoretically as well as empirically related to creativity have been found to neutralize the effects of situational variables. wilscn (1960) found that subjects labeled as ”socially ac- commodative” changed their attitude more when influenced by liked peers than by disliked peers. Subjects who were labeled as "concerned with information accuracy”, however, did not respond differentially accor- ding to attractiveness of the source. In an experiment by Jazubczak and Walters (1959) a group of dependent subjects proved to be more res— ponsive than less dependent subjects to social influence emanating from peers, suggesting that independent subjects vary their behavior less in accordance with the status of the source of influence. Similarly Mc Da- vid (1959) differentiated one group who were particularly attentive to the information contained in socially communicated message from another group particularly oriented to the source of the message. He found the source-oriented group to be more susceptible to peer influence, more affected by large discrepancies between their judgments and the judgb ments of others, and more likely to agree totally (rather than compro- mise) with the false norms presented when they yielded to influence. l4 Vidulich and Kaiman (1961) even report that low dogmatic subjects con- form more readily with the judgments of the low rather than the high status source. The characteristics of the socially dependent subjects in these studies are far from being congruent with the self-sufficiency and the need for accuracy characteristic of the creative person. Data regarding the relationship between intelligence and res- ponsiveness to variations in the quality of social influence are scarce and inconclusive. However, because of the assumed sensitivity of the high I.Q. person to interpersonal sources of information, we would ex- pect task specific expertise to produce more conformity in highly intel- ligent persons than in less intelligent persons. Thus the following hy- potheses are proposed. Hypothesis 4: varying levels of expertise of the communicator will have more effect on the conformity responses of less creative than of the more creative persons. Hypothesis 5: varying levels of expertise of the communicator will have more effect on the conformity responses of more intelli- gent than less intelligent persons. These hypotheses lead to the conclusion that varying levels of expertise will have the least effect on highly creative and low I.Q. per- sons and the most effect on low creative and high 1.0. persons. These hypotheses were tested in an experiment combining the testing methodology of personality theorists with the more situational laboratory approach of social psychologists. CHAPTER II METHOD The present study employed a factorial quasi-randomized blocks design relating intelligence, creativity, and communicator expertise to subjects' conformity responses in a laboratory situa- tion. Subjects. The subjects in this study were sixty male junior and senior students enrolled at Haslett High School, Michigan. All subjects volunteered to participate in the experiment. Materials. The materials used were of three types: measures of intelligence, measures of creativity, and measures of conformity. 1.2, scores were based on results from the Iowa Test of Basic Skills which were obtained from the students' records. Such results were not available for eight of the students involved; therefore for these subjects a comparable score was computed from a regression equation relating either the Large-Thorndike (verbal scale) or the DAT, which most of the students had in their files concurrently with the test employed. (See Appendix A for 1.0. test data) 15 16 The choice of creativity measures was based on several con- siderationss. Getzels and Jackson (1962) used five tests mainly de- rived from the works of Guilford and Cattell. Wallach and Kogan (1965) adapted Guilford's and Torrance's tests for individual testing, arguing that the limitations inherent in group testing did not allow for real creativity to show up. Barron (1963) also used the same type of ins- truments in experiments relevant to this one. Table 1 lists the ins- truments employed in those three studies. The tests that correlate the least with intelligence in the Getzels and Jackson study are the Uses for Things Tests (r=.18), the Fables Test (r=.13) and Make up Problems (r=.24). However, their crea- tivity tests appear to correlate no more between themselves than with intelligence. Wallach and Kogan, on the other hand, claim their tests form a reliable whole which is independent of intelligence both as a whole and as parts. In the current study, tests measuring many of the same facets of creativity assessed in these previous investigations were incorpo- rated in a battery for group administration. The Uses for Things Test 3There is still a controversy over the criteria and the measurement of creativity as expressed by Klausmeier and Harris (1962). Several attempts to develop such criteria have been made, not the least being by Jackson and Messick (1965). This study however, used measures that have generally been used to assess creativity. 17 Table 1 Creativity Tests Used in Three Studies Getzels Hallach 8‘ & Barron Jackson Kogan Uses for Things x .X X Word.Associations x ,X (a) Consequences X Plot Titles X Hidden Shapes X Similarities X Interpretation of ‘x X (b) Visual Patterns Fables x Make—up Problems X Word Rearrangement X Anagrams X (8) concepts." (b) Wallach and Kogan use line forms and other visual patterns. Barron uses TAT and Rorschach figures. variation of the test called: "Instances of class 18 (Getzels and Jackson, 1962) was selected to measure semantic sponta- neous flexibility, or the subject's capacity to shift frames of refe- rence and to use the environment in an original manner. A word asso- ciation test (Thigg;Categories Test by ETS) was chosen to measure idea— tional fluency, or the facility to call up ideas wherein quantity and not quality of ideas is emphasized. The third test chosen for this study is the Hidden Rifles Test (ms), designed to measure flexibili- ty of closure, critical exactness and, to a certain extent, field in- dependence. As a fourth test, an adaptation of Guilford's Conseguen— ces Test was employed to measure originality in thinking, or the abi- lity to produce remotely associated, clever or uncommon responses. Fi- nally, on the basis of Pelletier's data on self-concept (1970), indica- ting that selfeconcept may be the result either of an inductive concept formation process or of the internalization and application to oneself of social labels and categories, an adaptation of the Who Am I Test was included, on the expectation that a creative person would produce non-stereotypic responses. Scoring guidelines for this unique adapta- tion of the test appear in Appendix B, along with the other creativity tests and their scoring guidelines. A summated index of creativity was calculated by adding the subject's standardized scores on each of those five tests. Conformity, the dependent measure in this study, was as- sessed in the same way as in Crano's experiment (1970), with basical- ly the same setting and instruments. l9 Crano's experiment uses a perceptual judgment task in which the subject estimates the number of dots flashed onto a screen. In the ”acquisition” or ”compliance” phase of the experiment, the sub- ject always precedes or follows an experimental accomplice in announp cing his judgment of the number of dots. (While Crano used 5 blocks of 10 slides each in this phase, the current study shortened this to 4 blocks of 10 slides). In the ”transfer” or ”conformity” phase the subject and the confederate write their judgments without communicsp ting, thus eliminating direct social influence. Demand effects are minimized by a guarantee of anonymity. Each block of slides is pre- arranged so that the mean number of dots on each slide is 40 and the variances do not differ significantly from block to block. According to a preliminary study, a stimulus interval presenp tation of 5 seconds maximises accuracy while minimizing judgmental con- fidence in this situation; furthermore respondents consistently tend to underestimate the number of dots projected on each slide (1970). Since a different age grouping of students is involved in the current study, Crano's last finding was checked with a control group of 12 subjects from the same pool, who judged the stimuli without social influence. As in Crano's study, this control group also demonstrated a tendenp cy to underestimate; therefore his procedure was adOpted and all the confederate estimates were programmed to be 30% overestimations. Thus, higher mean estimate scores tend to reflect greater influence by social sources of information and less reliance on internal sour- ces. Crano also reports a significant (p«<.001) effect of res- ponse order in the acquisition phase. In order to maximize obser- ved conformity effects in this experiment, the subject always fol- lowed the confederate in announcing his estimates. Proceduge. Upon approval of the project by the adminis- tration of Haslett High School, a letter was mailed to 136 prospec- tive subjects and to their parents requesting their agreement to participate (see Appendix B). I.Q. scores were then obtained from the students'files. The creativity scores were obtained in a group session of 60 minutes. Only 56 subjects actually took those tests. Apparently there was some reluctance in the subject population ba- sed on previous bad experiences with psychology experiments; in ad- dition, many were absent from classes on the morning of testing. A median split was applied to the I.Q. and creativity distributions, and subjects were assigned to one of four possible groups depending on their scores. The distributions so obtained is illustrated in Table 2. 21 Table 2 The Relationship Between Creativity and Intelligence Test Scores Creativity Above Below Median Median T Above Median 16 12 28 Intelligence Below Median 12 16 28 T 28 28 56 In each of the groups which did not display a discripancy between their level of creativity and their level of intelligence, four subjects were randomly rejected and assigned to the control group, along with four other students who had not taken the creativity tests; thus the number of control subjects as well as of subjects in each cell was equal (12). The subjects in each of the four groups were then assigned to one of two possible expertise treatments with one of three possible con- federates in each condition. These subgroups were prearranged so that the means and variances in creativity and intelligence were not signifi- cantly different in the high and low expertise subgroups (see Appendix D for scores). 22 As in Crano's study (1970), the expertise manipulation in- volved influencing the subject's perception of the confederate's high or low level of competence in the task, based presumably on his prior experience in a similar experiment. A detailed account of the experi- mental procedures appears in Appendix B. These procedures followed Crano's very closely, with the following modification. Due to the fact that Haslett High School isva relatively small school, the students tend to know one another, at least by sight. It did not seem ethically and technically advisable to use peers as confederates; therefore young look- ing freshmen acted as confederates. The manipulation was planned so that each confederate would appear to the subject as a recent transfer student from a neighboring school. It appears that this procedure was successful on the basis of post-experimental interviews. After the experimental session, the subjects were partially de- briefed (especially regarding the identity of the confederate) and they were asked to c00perate with the experimenter by not discussing what had happened in the experimental room with their peers. The subjects res- ponded favorably to this request, and several even spontaneously formula- ted good reasons why they should keep the secret. 0n the basis of post- experimental interviews it appears that no diffusion occurred. The sub- jects were totally debriefed by a pamphlet explaining the purpose, me- thod and results of the experiment a few weeks following the experiment. A control group (including the 8 subjects rejected from the experimental groups and 4 volunteer students from the same grades) jud— ged the stimuli without social influence, so that the amount of 23 conformity induced by the conditions could be assessed. That control group was the first to be tested; therefore the risk of diffusion was initialy minimized by letting the first impact of the experiment on the students be of no significance for the results. ‘Qggigg, With respect to the compliance phase of this expe- riment, a 2XZXZX3X4 (repeated measures on the fifth factor) design was used. With respect to the conformity trials, the design was 2X2X2X3X5 (repeated measures on the fifth factor). The first four factors invol- ved two levels of intelligence, two levels of creativity, two levels of expertise and the three confederates. The last factors were the 4 Ac- quisition Blocks and the 5 Transfer Blocks. CHAPTER III RESUDTS Before discussing the analyses pertinent to the hypotheses regarding the impact of creativity, intelligence, and exposure to ex- pert communication on conformity behavior, let us consider some of the data regarding the tests employed in categorizing subjects. Creativity and Intelliggnce. As can be observed in the in- tercorrelation matrix provided in Table 3, the relationship between creativity and intelligence found in this study is somewhat at va- riance with the data reported by Getzels and Jackson (1962). An over- all correlation of .46 between the summated creativity index and the intelligence test scores appears to contradict the assumption of in- dependence between the two variables; furthermore the correlations of each particular creativity test with intelligence is higher than ex- pected. However, except for the Hidden Figures Test, the creativity tests appear to correlate better between themselves than with intel- ligence - expecially the Uses for Things, Consequences and Who Am I tests. This is reflected in the reliability estimate (alpha 2 .79) obtained for the overall index of creativity. .Also interesting are the teachers' evaluation of the creativity and intelligence of their students, which correlate much more with intelligence test scores (r=.63) than with the creativity index (r=.42). The teachers also 24 25 Table 3 Correlations Among Creativity, Intelligence, and Teachers' Assessments l 2 3 4 5 6 7 8 9 Category ' -1 18 45 30 38 62 10 01 19 Figures -2 46 38 40 65 43 44 52 Uses -3 68 65 87 37 43 54 Consequences -4 55 78 31 49 45 Who Am I -5 80 33 36 35 Creativity TOTAL -6 42 46 56 Creativity TEA -7 39 63 (teacher assessment) Intelligence -8 58 Intelligence TEA (teacher assessment) -9 26 tend to see a higher correlation between creativity and intelligence (r=.63) than actually exists. ‘Qggpttancefignd Conformity. Subject's estimates regarding the number of dots in the stimulus arrays during the acquisition and transfer trials, the two dependent variables in this study, were ana- lyzed by means of a 2 (Intelligence) by 2 (Creativity) by 2 (Communi- cator expertise) by 3 (Confederate) by 44 (Blocks of Trials) analysis of variance. Since none of the interactions of the Confederate fac- tor with other factors approached significance, they were pooled in- to the error variance. The results appear in Table 4. We had hypothesised significant effects of intelligence, creativity and communicator expertise as well as significant first order interactions in cases where expertise was involved. A glance at Table 4 generally shows lack of support for the hypotheses as they were presented, suggesting that they are not precise enough. Except for the expertise effect which was marginally signi- ficant (p‘<.07) in the transfer phase, none of the main effects nor simple interactions approach the significance level if we consider on- ly the important factors in this experiment (Intelligence, Creativity, Expertise). However the absence of an expertise effect in the acqui- sition phase could be explained by its interaction with the Blocks of Trials factor which was marginally significant (p (.07). That 4 5 in the case of transfer trials. Summary of Analysis of Variance 27 Table 4 Acquisition Transfer Source df ms F df ms r Intelligence (1) 1 17,214.19 1.74 1 234.04 < 1 Creativity (C) 1 3,152.51 < l 1 9,487.84 216 Expertise (E) 1 4,351.02 < l 1 15,990.34 3.63 (a) Confederate (s) 2 10,130.10 1.03 2 1,138.44 < 1 I x C 1 1,271.02 < 1 1 53.20 <1 1 x E 1 2,745.19 < 1 1 168.34