A STUDY OF WHITES' ATTITUDES. COMMIFMENT AND OVERT BEHAVIOR TOWARD MEMEERS OF A MINORITY GROUP Thesis for {'50 Degree of DE. D. MICHIGAN STATE UNIVERSETY James Max Fendrich 1965 LIBRARY Michigan State University WWWWW WW WW WWW WWWWWW 3 1293 00875 7605 This is to certify that the thesis entitled A Study of Whites' Attitudes, Commitment and Overt Behavior Toward Iiembers of a Minority Group presented bg James Mix Fend rich has been accepted towards fulfillment of the requirements for Doctor of Philosophy degree in Sociology . WMM Major professor Santo F. Camilleri Date August 133 196; 0-169 SDECIaI Permission NO renewal 3519. 2— 5 m THESIS WWWWWWWWWWWWWWWWWW1~ 31293 00875 7605 LIBRARY Michigan State University This is to certify that the thesis entitled A Study of Whites' Attitudes, Commitment and Overt Behavior Toward Members of a Minority Group presented by James Max Fendrich has been accepted towards fulfillment of the requirements for Doctor of PhilOSOth degree in SOCiOlOW-EY /’ . Major professor Santo F. Camilleri Date August 13, 1963; 0-169 n—’-'-”-'mmmp.roi~g.in, ,, r—.-— .v--~ -0 'Spectal Permission N0 renewal 3119.273 W99 W beti PTO of acc 10 C15 ABSTRACT A STUD! OF WHITES' ATTITUDES, COMMITMENT AND OVERT BEHAVIOR ‘ TOWARD MEMBERS OF.A MINORITY GROUP by James Max Fendrich The objective of this thesis was to study the relationship between attitude, commitment and overt behavior. One general proposition that had frequentLy been used to explain patterns of overt behavior was: behavior is a function of both individually acquired characteristics and the social context in which the behav- ior is expressed. A specific derivation of this preposition was culled from the literature. In its final form.the proposition stated: Behavior = Commitment (antecedent characteristics); Attitude (definition of the situation and antecedent characteristics). It was used as a guideline to examine white students' attitudes, commitment and overt behavior toward Negro students. The three antecedent characteristics were authoritarianism, past intergroup contact with Negroes and perceived support from significant others for engaging in interaction with Negroes. Two factors that influenced the definition of the situation were postulated. They were: the normative expectations involved in the role that subjects play as research respondents and the subcultural normative eXpec racial attitude ticam alter t consisted of a] continent and commitment befi affect the exp defined as int Variables and 11WPOtheses we: and the relat: used as the s White 3 Pilot (N = t° measure ti to be mliab “Mimic,r13 h Small group mm used a: valid rem.- The a"Weeds“ contingentl definition James Max Fendrich normative expectations for college students to express liberal racial attitudes. Experimental conditions were designed to systems- tically alter the definition of the situation. The conditions consisted of altering the order in which students would express commitment and attitude. It was hypothesized that the expression of commitment before the expression of attitude would significantly affect the expression of attitude. Commitment and attitude were defined as intervening determinants between the four independent variables and overt behavior. Eight propositions and twenty-six hypotheses were developed to explain the resultant behavior and the relationships between determinants. Kendall's Taub was used as the statistic to test relationships. White students at Michigan State University were used in a pilot (N = 46) and major study (N = 263). Scales were designed to measure the five determinants of overt behavior. They were found to be reliable and valid. In the pilot study the experimental conditions had the predicted effect on the expression of attitude. Small group discussions designed to improve interracial relations were used as the measure of overt behavior. It was found to be a valid resultant of the determinants. The general preposition was found to be true. The three antecedent and the two intervening variables were found to be contingently necessary conditions for overt behavior. The definition of the situation was not found to be an important deteminant of attf IWexpression of c: attitude being nor cmment and ove the definition of tion for the arms the limbo of uncor suggested that the controlled Bxperi: Single deteminan James Max Fendrich determinant of attitude in the major study. EXperimental Condition I (expression of commitment before attitude) did not result in attitude being more consistent with the antecedent variables, commitment and overt behavior. Prepositions IV and V stated that the definition of the situation was an important contributory condi- tion for the expression of attitude. They were not relegated to the limbo of unconfirmed theoretical statements. The pilot study suggested that the propositions might be true in a carefully controlled experimental environment. In the major study the best single determinant of overt behavior was commitment. in: A STUIX CF WHITES' ATTITUDES. COI‘DETMENT AND OVER!‘ BEHAVIOR TOIIARD MEMBERS OF A MINORITY GROUP By James Max Fendrich A THESIS Subnitted to Michigan State Univeristy in partial fulfillment of the requirements for the degree of IDCTOR OF HIILOSOPHI Department of Sociology 1965 Harv peo author would li helped in many her assistance “0’31 support t Witnesses, deg rent and critic The mem‘ D30 Santo F. W ACKNOWLEDGEMENTS many people contributed to the completion of the thesis. The author would like to express his appreciation to his wife. Judith helped in many ways. Within the community of friends and associates her assistance in the mechanics of preparation is well known. Her moral support was more significant. Her wareness of strengths and weaknesses, desires and frustrations coupled with words of encourage- ment and criticism made the thesis become a reality. The members of the guidance committee deserve recognition. Dre. Santo F. Camdlleri, Eugene Jacobson, James McKee and William Faunce offered helpful suggestions during the course of development. Dr. Camilleri's aid Should be singled out. As chairman of the committee he was willing to spend many hours going over details of the thesis and offered valuable criticism. He also stepped beyond his professional role. His empathy and wit simplified the task. Dr. A. O. Heller was also helpful in the initial stages of development. The assistance of several graduate students who served as sounding-boards for several ideas incorporated into the thesis is appreciated. Chuck Tucker deserves special credit for helping to gather the data and serving as a lending library. Frank Holland also assisted with the data collection. Lee Sloan read portions of the draft copy. pointing out errors and suggesting improvements. Dick Sturgis, Larry Sneden and Jon Reiger allowed me to use their classes ii to gather data. L1 Burke‘ exceeded his ra‘ care in prepari I would (3379113 Chapter eIaecially hel; racial bebavim to active CODC' to gather data. Al Burkett's assistance in obtaining output from the 3600 exceeded his rate of reward. Joe Curtin and Rose Curtin exercised care in preparing the input. I would like to express aw gratitude to the members of the Campus Chapter of the NAACP. Melvin Moore and A1 Williams were expecially helpful. Although we did not significantly change inter- racial behavior on campus, a few more people passed from inactive to active concern for the American Dilemma. iii LIST W TABLES LIST CF FIGUE LIST Q“ APPENI Chapter I. mac 111, DP TABLE OF CONTENTS IIIST OF TAM-IE O O O O O O O O O O O O O O O O O O O O 0 O O vini LIST “I. FIGURE 0 O O 0 O O O O O O O O O O O O O O O O O O n LIST m APPENmCfi O O O O O O O O O O O O O O O O O O O O O Xii Chapter I. INTRowCTION O O O O O O O O O O O O O O O O O O O l The PrOblem O O O O O 0 O O O C O O O C O O O O O O l PurposesoftheStudy............... 3 The Significance of the Problem . . . . . . . . . . 5 LimitationsoftheStudy............. 10 O O O O O O O O O O 12 AnOverview.......... II. THEORETICALORIENTATION.............. 11+ PropositionalTheory............... 11+ Review of the Literature on the General Pr0position 16 The General Proposition Selected as a Guideline . . 22 III. DEVELOPMENT OF SPECIFIC PROPOSITIONS AND HIPOI‘HESE 25 Introduction . . . . . . Antecedent Variables . . Authoritarianism . . . . Past Intergroup Contact . Perceived Support from Significant Others The Definition of the Situation . . . . . The First Position: Attitude and Situation The Second Position: Attitude vs. Situation Summary of the. Theoretical Significance of the SocialSituation................ 58 Factors that Affect the Expression of Attitudes . . 6O Attitudes Toward Members of a Minority Group . . 66 Commitment to Become Involved in Possible Activities with Members ofa Minority Group . . . . . . . . 71+ Overt Behavior Directed Toward Members of a MinorityGraup................. 81 Theoretical Hypotheses Derived from the Propositions 81 SW1? 0 O O O O O O O O O O O O O O O O O O O I O 87 O O O O coco C... 0000.. ooo‘oooo ,5 O O O O 0 O O O O O O O O O C 0 iv Chapter IV. RESEA] Intro Exper Sampl Rando Int; C0D: Chapter IV. Table of Contents cont. RESEARCH DESIGN AND PROCEIIIRES . . . . Introduction ......... Experimental Design . . . . . . sample 0 O O O O O O O O O O O O Randomization Procedures . . O O O O O O O 0 Operational Definitions, Reliability and Validity Of the scales 0 O O O O O O O O O O O O O O O The califonfi-a F-scale O O O O O O O O O O O O O The Past Intergroup Contact Scale . . . . . . . . Scale The Perceived Support of Significant Others Definition of the Situation . . . . The Attitude Scale . . . . . . . . The Commitment Scale . . . . . . . The Overt BahBViOr Scale 9 o o o 0 Reliability of Research Procedures Conclusions . . . . . . . . . . . . ANAHSISOFTHERESULTS........ Introduction . . . . . . . . . . . . . Analysis of Pr0positions and Hypotheses Proposition I . Proposition II Preposition III Preposition IV Proposition V . Proposition VI Preposition VII Preposition VIII Interpretation of the Results 811mm 0 O O C O O O O O O O O O O O O 0011011131011 0 o c o o o o o o o o o o o O O O 0 O o o o c c c o o o c o o o o o o c o o o o c o o o o o o o o c c o c c o o o o c o o O O O O O SUMMAIU AND CONCLUSIONS . . . . . . . . Overview ............ MajorFindings . . . . . . . . . Limitations . . . . . . . . . Inplications for Future Research 0 O O O O O O O O O O O O O O O O 0 9O 9O 9O 92 98 100 107 111 115 117 120 1 22 1% 128 131 131 132 132 132 136 1&9 151» 157 161+ 164 177 182 181+ 181+ 187 192 191+ APPENDIX A . APPENDIX B . APPENDIX C . EIBIIOGBAHIY Table of Contents cont. Page APPENDIX A O O O O O 0 O O O O O O O O O O O O O O O O O O O 199 Awmm B O O O O O O 0 O O O O O O O 0 O O O O O O O O O O 8“9 APPEnmX C O O O O O O O O O O O O O O O O O O O 0 O O O O O 222 HMQGMM O O O O O O O O O O O O O O O O O 0 O O O O O 0 22a Table 10. II. 12. 13. Results Modif Results Past Results Perce RBSults Atti' Results Comm Results Over The R81 Atti The Re] Coym The Re: Beh- The Re: Atti The Re; 00m; The Re: The Re: 31,} (2) 10. 11. 12. 13. LIST OF TABLES Results of the Item-Total Score Analysis of the Mdified F‘scale O O O O O O O O O O O O O O O O 0 Results of the Item-Total Score Analysis for the Past Intergroup Contact Scale . . . . . . . . . . Results of the Item-Total Score Analysis for the Perceived Support from Significant Others Scale . Results of the Item-Total Score.Analysis for the Atti tude scale 0 O O O O O O O O O O O O O O 0 O 0 Results of the Item-Total Score Analysis for the commtment scale 0 O O O O O O O O O O O O O O I 0 Results of the Item-Total Score.Analysis for the Oven Betlavior scale 0 O O O O O O O O O O O O O O The Relationship Between Authoritarianism.and Attitude for Treatments (1), (2) and (3) . . . The Relationship Between Authoritarianism and Commitment for Treatments (1), (2) and (1+) . . . . The Relationship Between Authoritarianism and Overt Behavior for all Treatments . . . . . . . . . . . The Relationship Between Past Intergroup Contact and Attitude for Treatments (1), (2) and (3) . . . . . The Relationship Between Past Intergroup Contact and Commitment for Treatments (1), (2) and (4) . . . . The Relationship Between Past Intergroup Contact and Overt Behavior for all Treatments . . . . . . . . The Relationship Between Perceived Support from Significant Others and Attitude for Treatment (1). (2) and (3) O 0 0 C O O O C O O O O O O O O O O O Vii Page 105 110 114 119 122 125 133 134 135 137 138 139 140 This lb. 1L 1?. EL 19. The Rela‘ Signifi (1). ( The Rela Signif Treats The Rel; Inter; The ReL Perce Treat The Rel Peres Treat The Re] Auth< Fem. Comm: The Re‘ Auth Atti The Re Ant} Per: 9] The R: Undt The 1». Con- The {1 C011 List of Tables cont. Table Page 1h. The Relationship Between Perceived Support from Significant Others and Commitment for Treatments (1), (2) and (u) 0 O 0 O O O O O O O O O O O O O 0 11+]- 15. The Relationship Between Perceived Support from Significant Others and Overt Behavior for’all Tm tment Groups 0 O O O O O O O O O O O O 0 O O O luz 16. The Relationship Between Authoritarianism and Past Intergroup Contact for all Treatment Groups . . . . 144 17. The RelationShip Between Authoritarianism and Perceived Support from Significant Others for all Treatment Groups . . . . . . . . . . . . . . . . . 145 18. The Relationship Between Past Intergroup Contact and Perceived Support from Significant Others for all Tmatment Groups 0 O O O O O O O 0 O O O O O O O O 1’46 19. The RelationShip Between the Combined.Affects of Authoritarianism, Past Intergroup Contact and Perceived Support for Significant Others on Cthe nt 0 O O O O O O O O O O O O O O O O 0 O O 1% 20. The Relationship Between the Combined.Affects of Authoritarianism, Past Intergroup Contact and Perceived Support for Significant Others on Attitude s O O O O O O O O 0 O O O O O O O O O O 0 O 11+8 21. The Relationship Between the Combined.Affects of Authoritarianism, Past Intergroup Contact and Perceived Support of Significant Others on we rt &haVi o r O O O O O O O O O O O O O O O O O 0 lug 22. The Relationship Between.Attitude and Authoritarianism Under'Two Experimental Conditions . . . . . . . . . 151 23. The RelationShip Between Attitude and Past Intergroup Contact Under Two Experimental Conditions . . . . . 152 2G. The Relationship Between.Attitude and Perceived Support from Significant Others Under Two Experimental condj-ti 0118 O O O O O O O 0 O O O O O O O O O O O O l 5 3 viii ‘ o Table 25. 27. 32. 33. List of Tables cont. Table Page 25. The Relationship Between Attitude and Commitment Under Two Ehrperimental Conditions . . . . . . . . . 155 26. The Relationship Between Attitude and Overt Behavior Under Twa Experimental Conditions . . . . . . . . . 156 27. A Comparison of Findings from the Pilot Study and Major Study for the Relationship Between Attitude and Commitment and Attitude and Overt Behavior . . 157 2B. The Relationship Between Attitude and Commitment for Treatments (1) and (2) . . . . . . . . . . . . 159 29. The Relationship Between Attitude and Overt Behavior for Treatments (1), (2) and (3) . . . . . . . . . . 160 30. The Relationships Between Authoritarianism and Commitment and Authoritarianism and Attitude for Students in Treatments (1) and (2) . . . . . . . . 161 31. Relationships Between Past Intergroup Contact and Commitment and Past Intergroup Contact and Attitude forTreatments (1)and (2) . . . . . . . . . . . . 162 32. The Relationships Between Perceived Support from Significant Others and Commitment and Perceived Support from Significant Others and Attitudes for Treatments(l)and(2) .............. 163 33. The Relationships Between Overt Behavior and Commit- ment and Overt Behavior and Attitude for Treatments (1) am (2) O O O O O O O O O O O O C O O O O O O O 165 34. The Relationship Between Authoritarianism and Overt %havinor . O O O O O C O C C C C O . O O C . 180 35. The Relationship Between Past Intergroup Contact and wart mmuor 0 O O O O O O O O O O O O O O O O O 180 36. The Relationship Between Perceived Support from Significant Others and Overt Behavior . . . . . . . 181 Table 37. The Relatic Behavior List of Tables cont. Table Page 37. The Relationship Between Attitude and Overt Behavior . 181 38. The Relationship Between Commitment and Overt Rhavior O O O O O O O O O O O O O O O O O O O O O O 18 2 Figure l. Newcomb Betwe 2. A Modif of th Behav LIST OF FIGURES Figure Page 1. Newcomb's Diagram Illustrating the Relationship Between Attitude and Overt Behavior . . . . . . . 18 2. A Modified Presentation of Newcomb's Description of the Relationship Between Attitude and Overt khaflor O O O O O O O O O O O O O O O O O O O O 23 Append Append Append LIST OF APPENDICES Apmndi-x A O O O O 0 O O O O O O O O O O %ge 199 AppendixB............... Pag6219 prm C O O O O O O O O 0 O O O O O 0 figs 222 xii The Problem An old a: Psychology is t] There have been These defining. W m that Chas 0f 0 These Clispositi a result of ex; that guide, bia Presented a. lie that experience CHAPTER.I INTRODUCTION The Problem An old and persistent problem in the field of Social Psychology is the relationship between attitude and overt behavior. There have been literally hundreds of attempts to define attitude. These definitions have frequently omitted the gggg§.praximum or the differentia spgcifica of the concept. Attitude is derived from that class of concepts called acquired behavioral dispositions. These dispositions refer to the fact that behavior is modified.as a result of experience. Individuals retain residues of experience that guide, bias or otherwise influence later behavior. Campbell presented a list of 80-odd terms that called attention to the fact that experience has modified the behavioral tendencies of the organism, e.g., attitude, belief, habit, interest, motive, set, expectation, value , etc .1 The distinctiveness--dif§gggg§yg sgcifica-u-of the concept is that an attitude'is a behavioral orientation toward a class of objects based on cognitions having varying degrees of affect loading. 1Donald T. Campbell, ”Social Attitudes and Other Acquired Behavioral Dispositions," Psychology} A Study of a Science, ed. Sigmund Koch (New York: McGraw-Hill, 1963). pp. 97-101. The cone! precedes social of behavior a p conclusion. If becomes signifi direction for t and eXpression “Pom the situat that guide and Provided in the have been defix hum“ PWsical ObJ'ects that i: The con inVleement in the attitude, of directneSs indirectly eXp pattems may t 2 The concept orientation is used to refer to that which precedes social acts. Before engaging in a recognizable pattern of behavior’a person prepares to carry out the act to its perceived conclusion. If, in the course of the act, the pattern of behavior becomes significantly altered, new orientations develop to provide direction for the patterned activities. The particular formulation and eXpression of the acquired behavioral diSposition is dependent upon the situational context. The residues of past experiences that guide and affect behavior are influenced by the stimuli provided in the context of eXpression. The objects of attitude have been defined to include people, institutions, ideals or non- 2 In this study only social objects, i.e., human physical things. objects that interact with the subject will be considered. The consequent forms of overt behavior'are defined as involvement in or commitment to interaction with the object of the attitude. The involvement or commitment varies in degrees of directness and Specificity. The behavior is either directly or indirectly eXpressed toward the object of the attitude. Interaction patterns may be directly oriented toward the object of the attitude or indirectly toward other people who are also concerned with the zThis general definition of the object of attitudes is derived from.Allan L. Edwards in Technigues of Attitude Scale ggnstruction (New York: Appleton-Century-Crofts, 1957), p. 2. Edwards agrees with Thurstone's emphasis on the wide range of possible attitudinal objects. object of the atti the United States policy in Southeas that time several attitude toward th who were concerned of activities to Party others were 5 “1° Shared similar-l “n“ 6°Cpress the attitude. The act:' Wcificity. For ¢ orientations towan date Any attractim a M” specific pa attractive girls w‘ sea °f the St in.“ ovzgltf Himmel oriented haV'io 3 object of the attitude.3 For example, in numerous universities in the United States there was a growing concern over governmental policy in Southeast.Asia during the first six months of 1965. .At that time several university faculty members had an unfavorable attitude toward the governmental policy. Of those faculty members who were concerned, a few committed themselves to various patterns of activities to express publicly and officially their discontent. Many others were satisfied to express their discontent to people who shared similar'attitudes toward the governmental policy and thus did not express their discontent directly toward the object of the attitude. The action pattern may also vary in the degree of its specificity. For example, two college students may have favorable orientations toward dating attractive girls. One may be content to date any attractive girl who is agreeable while the other may have a more specific pattemn of activity in seeking to date only those attractive girls who are Jewish. Eggposes of the Study The purpose of this study is to formalize a “middle-range" theory that is related to a general theoretical approach to social 3Ulf Himmelstrand, a Swedish sociologist, emphasized the fact that overt behavior resulting from attitudes need not directly be oriented toward.the object of the attitude in 'Verbal.Attitudes and Behavior: {A Paradigm for the Study of Message Transmission and T;gnsfonmation,' Public Qpinion Quarterly, XXIV (Summer 1960), 2 -231 O psychology.“ concerned with consequent for. cal approach i is the product This propositi however, can; theoretical p theory it, was relati on Ship: 4 . psychology.” The content area of the middle-range theory is concerned with the relationship between social attitudes and consequent forms of social behavior. The more general theoreti- cal approach is centered on the following proposition: behavior is the product of both individual characteristics and environment. This proposition is frequently associated with Field Theory, however, variations have been incorporated into different theoretical perSpectives. In order to formalize the middle-range theory it was necessary to gather a body of knowledge about the relationships between social attitudes, antecedent and consequent “The phrase middle-range theory is borrowed from Robert K. Horton in his book Social Theo and Social Structure (Glencoe, Illinois: The Free Press, 1 3 , pp. 5 . herton defines middle- range theories as '. . . theories intermediate to the minor work- ing hypotheses evolved in abundance during the day-by-day routines of research, and the all-inclusive speculations comprising a master conceptual scheme from which it is hoped to derive a very large number of empirically observed uniformities of social behavior." Hans L. Zetterberg prefers the term partial to middle-range in On Theo and Verification in Sociolo (Totowa, New Jersey: The Bedminister'Press, 19335, pp. 2—5. Partial theories are non- contradictory parts of theoretical sociology; The distinctiveness of these theories is their usefulness to both general theory and empirical research. As a body of interrelated propositions they can be related to more abstract conceptualizations and to Specific hypotheses. A formalized theory can provide explanation as well as description of empirical regularities. Zetterberg distinguishes this particular theoretical perspective from the sociological classics, e.g., Theories of Society; sociological criticism, e.g., Modern Sociological Theogyfin Continuity and Change; and sociological taxonomies, e.g., Toward A General Theory of.Action. forms of social positions were from the propo: cation study w: The Sigifi can The val and overt beha Of an article to NEWCOnb's e plus or minus DeFleu of studies on they found a action Corre] problem is 8i a major PeSe; SEVeI-E 5 forms of social behavior. From this knowledge a series of pro- positions were constructed. Specific hypotheses were developed from the propositions. After the theory was developed, a verifi- cation study was designed to test the hypotheses of the theory. The Significance of the Problem The value of studying the relationship between attitudes and overt behavior can be illustrated by contrasting the findings of an article by Newcomb and one by DeFleur and westie. According to Newcomb's estimate in 1956, there were 9,426 articles and books, plus or minus 2,712 on the topic of attitudes.5 DeFleur and westie agreed with Newcomb on the steady stream of studies on the verbal dimension of attitudinal behavior; however, they found a striking paucity of investigations concerned with overt- action correlates of such verbal behavior.6 Thus, the study of this problem is significant because it focuses on an area where there is a major research gap in the knowledge of attitudes. Several additional reasons can be cited to demonstrate the significance of this problem. The research is timely and related 5Theodore H. Nchomb, “The Prediction of Interpersonal Attraction,“ The American Psychologist, 1:: (November 1956) 575. During the nine-year period Since his estimate, there has been an accelerated pace of research and publication. It would be safe to estimate that the number of books and articles is now over 10,000. 6Melvin L. DeFleur and Frank R. westie, 'Verbal.Attitudes and Overt Acts: An Ehtperiment on the Salience of Attitudes,” American Sociological Review, XXIII (December 1958) 667. to an important in which Negroes becoming fully d been called the The importance 0 attitudes toward types of victori the attitudes an Practical proble attitudes toward Studies 3‘ faVOI‘ability of article by Hyman dating back to 1 attitudes,7 6 to an important practical problem. A large scale social movement in which Negroes are demanding their equal rights as citizens is becoming fully developed in the United States. This movement has been called the most important domestic issue in the United States. The importance of the issue requires precise information on whites' attitudes toward Negroes. The direction of the movement and the types of victories that will be achieved are in part dependent on the attitudes and actions of whites. There is, moreover. the practical problem of evaluating information gathered on whites' attitudes toward Negroes. Studies Spanning forty'years have reported the increasing favorability of whites' attitudes toward Negroes. In a recent article by Hyman and Sheatsley the results of public opinion polls dating back to 1942 reveal a remarkable Shift to more favorable attitudes.7 Bogardus has studies the phenomena of social distance over'a span of thirty years. In 1956 he obtained data from widely scattered groups of college students and adults. Although this was not a nationwide sample, it compared with the groups studied in 1926 and 1946. The various ethnic groups employed in the social distance scale maintained their relative positions; however, on a seven-step scale, the greatest social distance fell from 3.91 7Herbert H. Hyman and Paul B. Sheatsley, QAttitudes Toward Desegregation,’ Scientific American, Vol. 211 (July l96h) 23. (1926) to 2.83 (1' to 2.09 (1956).8 louis Harry American college a findings in W= in their attitudes Almc app the app: 85 i socz' mem‘: of : EVidence w‘ shifts t'0 more fa 7 (1926) to 2.83 (1956). The average score fell from 2.14 (1926) to 2.09 (1956).8 Louis Harris and Associates polled a representative sample of American college and university students and reported some interesting findings in Newsweek. They found the students extremely favorable in their attitudes toward Negroes:9 Almost unanimously'(98 percent) they say they approve of Negro and white students eating in the same cafeteria. Ninety—three percent approve of their living in the same dormitory, 85 percent of their belonging to the same social club, 47 percent of their dating a member of the other race, and 36 percent approve of intermarriage. Evidence which has been presented would indicate significant shifts to more favorable attitudes. However, a pertinent question can be raised, ”What is the significance of the results of these studieST' The interpretation of the above findings is dependent in part on which questions were asked. In a nationwide sample, Brink and.Harris found wide variations in the responses of whites. Agreement with commonly expressed stereotypes about Negroes varied from 68 percent (Negroes laugh a lot) to 31 percent (Negroes care less for the family). The vast majority of whites also thought Negroes should be guaranteed their civil rights; however, white 8Emory Bogardus, “Racial Distance Changes in the United States During the Past Thirty'Iears,' Sociology and Social Research, Vol. 42 (NovemberADecember 1958) 127-135. 9"Campus 65,“ Newsweek Ma azine (March 22, 1965), p. 48. The exact criteria for selec ing the sample were not given. suooort for cth (Federal vote-enf bill). 0'1 questi Negroes there was. they would objectl 90 percent said t a Negro. The ma: the Negro revolut Sit-ins and lie.d Bettlehei: Wild favorable asked. A questio intelligent as 7? Percent betWee asking if the sub Oily Varied 3 Per The diffs: 8 support for civil rights legislation varied from 57 Percent (Federal vote-enforcement law) to 66 percent (Public-accomodations bill). On questions concerning whites' feelings about contact with Negroes there was a wide range of opinion. Only 17 percent said they would object to working next to a Negro on the job, whereas, 90 percent said they would object to their teen-age daughter dating a Negro. The majority of whites were also opposed to the speed of the Negro revolution and the tactics emplqyed, e.g., lunch-counter sit-ins and lie-downs in front of trucks on construction sites.10 Bettleheim and Janowitz also found variations in the trends toward favorable attitudes, depending upon which questions were asked. A question asking if the subject thought Negroes were as intelligent as whites showed a favorable change from 42 percent to 77 percent between the years 1942 to 1956; however, a question asking if the subjects thought Negroes were being treated fairly only varied 3 percent during the same period.11 The different degrees of favorability appear to be dependent upon: (1) the degree of personal intimacy of contact and (2) the degree to which the attitude statements imply action on the part of the individual and/or others to bring about change. Those state- ments that imply little or no intimacy and no planned action to bring fl. lok‘illiam Brink and Iouis Harris, The Ne ro Revolution in America (New York: Simon and Schuster, 19345, pp. 138-153. llBruno Bettelheim and Morris Janowitz, Social 0 e and Prejudice (New York: The Free Press of Glencoe, 1934-5, pp. ll-13. about change 1 istic of the 1 not provide d able attitude strongly fave ment of attit behavior will attitudes. ‘- thi: range theory This theoret 3P9°Hlative social Psych Using 2612128: °f 1983 inc“ 9 about change received the most favorable responses and are character- istic of the greatest shifts in attitude. Although this study will not provide direct evidence to evaluate the trend toward more favor- able attitudes, it will provide evidence to evaluate the meaning of strongly favorable attitudes. The critical evaluation of the measure- ment of attitudes, the experimental design, and the measure of overt behavior will provide information that can be used to evaluate attitudes. A third reason for doing this study is to develop a middle- range theory to verify a general proposition in social psychology. This theoretical approach is more precise than the comprehensive, speculative systems that have been prominent in the development of social psychology; The theory was put to an exacting empirical test. Using Zetterberg's book as a model also helped to escape the pitfalls of 1983 inclusive approaches. The theory avoids the simplistic idea of analyzing the isomorphic relationship between attitude and.overt behavior. Specific hypotheses are interwoven with.more general propositions. A concern with the basic concepts and the relationship between concepts is another reason for executing this study. Literally, thousands of studies have used the concept attitude. Most studies have only been concerned.with the antecedents of attitudes. A few studies, however, have been directly concerned with.attitude as an intervening variable and the consequences of attitudes. The knowledge produced from a careful operationalization of the basic concepts and the propositions may on the major conc The final sethodological . most massive inc: of social scienc. ”9330““ to have on attitudes to“ This stud cant Variable of to the expreSSlc attitude and mm 10 concepts and the relationship of these concepts to hypotheses and propositions may aid in developing a cumulative body of information on the major concepts. The final justification for undertaking this study is methodological. Samuel A. Stouffer was responsible for one of the most massive increments of attitude study material in the history of social science. In his work on The American Soldier, he is reported to have said, "I would trade a‘ half-dozen Army-wide surveys on attitudes toward officers for one good controlled experiment);2 This study uses an experimental design to test the signifi- cant variable of the definition of the situation as it is related to the expression of an attitude and the relationship between attitude and overt behavior. Each of the other independent and control variables are measured in the four treatment groups in the research design. The emphasis on the situational variables as they are related to the expression of an attitude provides guide- lines for evaluating the usual methodology employed in a testing situation. Limitations of the Study The content of the attitude is limited to the orientations of whites toward Negroes. Although the dominant white group has orientations toward numerous minority groups, Negroes were selected as the object. Negroes as objects of whites' attitudes 12'John Madge, The Tools of the Social Science (Garden City, New York: Doubleday and Company, Inc., 15355. p. 75. represent a unin States. A combi1 istics; regional States; size of strongly held wh of willingness o the dominant grc cC’ilpleted, exec; attitudes of m to their attituc The Specf ll represent a unique case of interracial relations in the United States. A combination of factors such as phenotypical character- istics; regional isolation, during the early history of the United States; size of the minority group, their non-European origins; and strongly held whites' negative sentiments have influenced the degree of willingness of whites to allow Negroes to become assimilated into the dominant group. To date the assimilation process has not been completed, except possibly at the cultural level.13 Thus, the attitudes of whites toward Negroes cannot necessarily be generalized to their attitudes toward other minority groups. The specificity of both the subject and object of the attitude is limited to undergraduates in a northern liberal college environ- l4 ment. The extent of association between two or more variables studied may not be characteristic of more general, heterogeneous W 13m1ton H. Gordon, Assimilation in American Life (New York: Oxford University Press, 19345, p. 73. This recent book is a penetra- ting analysis of the problems of assimilation. Gordon conceives of American society as a series of subsocieties and subcultures. He analytically distinguishes between seven types of assimilation: cultural; structural; marital; identificational; attitude; behavior; and civic. The only type of assimilation that Negroes have exper- ienced is the cultural, and even this type of assimilation is limited to variation by class. In contrast the Jews have been substantially assimilated at the cultural level, mostly assimilated at the civic level, and partially assimilated at the marital and behavior levels. luThe reason the limitation was placed on the subjects of attitudes one a lack of financial resources. Students readily available at the Michigan State University campus were used as subjects. The reason for restricting the object was to provide an exact social object with which the subjects previously had either direct or indirect eXperience. The specific object provided a stimulus to which the subjects could meaningfully respond. samples. The e. tions can only enpirical evide in Overview The body theoretical or'l and hypotheses 2 Procedures and 1 conclusions. c] guideline for d rosearch, proce the major pmpc literature on 1 Proposition is pmpositions a. the deLIe -mn 12 samples. The extension of the findings to broader based pepula- tions can only be attempted by logical reasoning rather than by empirical evidence. An Overview The body of this study is divided into three parts: (1) the theoretical orientation, review of the literature and propositions and.hypotheses relevant to the theory; (2) the research design, procedures and the analysis of results; and (3) the summary and conclusions. Chapter II, a theoretical overview, presents the guideline for developing theoretical propositions for empirical research, procedures for developing theoretical propositions and the major*proposition of the middle-range theory. .A review of the literature on the variables that are interrelated in the major proposition is presented in Chapter III. The separate theoretical propositions and specific theoretical hypotheses that are part of the middle-range theory are also presented. The design of the experiment and research procedures which include: (1) the description and rationale for the various treat- ment groups, (2) a description of the pilot study that led to the major research, (3) an evaluation of the sampling techniques and the randomization procedures, (4),the operationalization of major variables in the theoretical propositions, and.(5) a demonstration of the reliability and validity of the scales designed to measure the variables is presented in Chapter IV. Chapter V presents the analysis of the results and the relationships be The ma jo.‘~ and conclusions . 13 analysis of the results, i.e., Specific methods of analyzing the results and the tables and tests of statistically significant relationships between variables. The major findings are evaluated in Chapter VI, the summary and conclusions. Implications for future research are considered. Msitionafi The cent the statement 0 Objective of so will be Phrasec‘ Although mathen Dover, it is dj that meet Striz and at the Same defined SOCial characteristic: Proposi- Variates must - least “’0 diff. CHAPTER,II THEORETICAL ORIENTATION Propgsitional Theory The central characteristic of the theoretical orientation is the statement of the theory in propositional form with the ultimate objective of seeking explanation. The propositions of the theory will be phrased in common language instead of mathematical language. Although mathematical models have precision, elegance and inferential power, it is difficult to construct propositions in social psychology that meet stringent rules for manipulating mathematical expressions and at the same time retain the complexity and diversity of crudely defined social phenomena. There are, however, several fomal characteristics of propositions stated in common language. Prepositions consist of related variates (variables). Variates must vary, i.e., they must be able to be observed in at least two different states. When the direction in which the variates influence each other is known or'assumed, the variates are subdivided into determinants and resultants. A determinant is a cause or independent variable and the resultant is the effect or dependent variable. Propositions must contain at least two variates but many propositions contain more than two. They may have multiple determinants and/or multiple resultants. It is not uncommon to have multivariate prepositions in the social sciences. l4 The complexity ( rigid experimen‘ propositions . '1 constructing mu: There an between detemi taxonomy of the l.A.. ther no; ZoAr; or: 3.11:4 ac in? of col 5. A: 3*) cos 15 The complexity of the phenomena and the difficulty of establishing rigid experimental controls limit the utility of simple bivariate propositions. They are used, however, as intermediary steps in constructing multivariate propositions in a theoretical framework.1 There are several possible attributes of the relationship between determinants and resultants. Zetterberg provides a taxonomy of the variety of causal relationships:2 1. A relation may be reversible (if x, then y; and if y, then x) or irreversible (if y, then y; but if y, then no conclusion about x). 2. .A relation may be deterministic (if x, then always y) or stochastic (if x, then probably y). 3. A relation may be sequential (if.x, then later y) or a coextensive (if x, then also y). 4. .A relation may be sufficient (if x, then y regardless of anything else) or contingent (if’x, then y, but only if z). 5. A relation may be necessa (if’x, and only if x, then y) or substitutable if x, then y, but if z, then also y). Propositions can possess different combinations of the above five attributes. The most frequent types of proposition in the social sciences have causal linkages that consist of irreversible, stochastic, sequential, contingent and substitutable relations. 1Zetterberg, op. cit., pp. 63-65. 2Ibid.. pp. 69-74. It should be noted that Zetterberg's description of causal relationships is not consistent with other discussions of causality. What Zetterberg defines as a necessary condition (if x, and only if x, then y) is defined as a necessary and sufficient condition. When Zetterberg defined the substitutable condition (if x, then y, but if 2, then also y), he was defining a sufficient condition. For further clarification see Claire Sellitz, Marie Jahoda, Morton Deutsch and Stuart W3 Cook, in Research Methods in Social Relations (New Ibrk: Henry Holt and Company, Inc., 1960), p. 81. Review of the Li In attem; ship between whi Segro students, reviewed to dis guideline in se propositions . that would be i Variations of 2 authors . lewii 16 Review of the Literature on the General Proposition In attempting to develop a theory to explain the relation- ship between white students' attitudes and overt behavior toward Negro students, theoretical literature in social psychology was‘ reviewed to discover a general proposition which could serve as a guideline in selecting variates from which to construct specific propositions. There was a specific attempt to discover a guideline that would be independent of particular time and Space sequences. Variations of a general proposition have been suggested by many authors. Lewin's proposition is the most widely known. He stated:3 In principle it is everywhere accepted that behavior (B) is a function of the person (P) and the environment (E), B = F (P,E), and that P and E in this formula are interdepen- dent variables. Deutsch believed that Lewin's emphasis on the relation of the individual to the situation led the scientific investigator to a more explicit realization that understanding behavior required not only a knowledge of the person, e.g., his past experiences, present attitudes and capabilities but also a knowledge of his immediate situation. In testing situations interpretation of the responses of a subject required a knowledge of the significance of the total response to the testing situation.“ Other social psychologists 3Kurt Lewin, 'Formalization and Progress in Psychology," Field Theo in Social Science, ed. Dorwin Cartwright (New York: Harper and Row, 1951), p. 25. “Norton Deutsch, "Field Theory in Social Psychology,‘I Handbook of Social Psychology, ed. Gardner Lindsey (Reading, Massachusetts: Addisonewesley, Inc., 1954), p. 185. have stated simi When desq one's self -conc tion of social a neural elements central nervous general ordered was involved.5 Newcom‘o and overt betxav a Simple and pe Product not on: The Other was . multiple. New 17 have stated similar positions. When describing the development of behavior directed toward one's self—concept. Mead stated it was dependent upon the organiza- tion of social attitudes. The organization was the result of both neural elements and their interconnections in the individual's central nervous system (past experience stored in memory) and the general ordered pattern of social behavior in which the individual was involved.5 Newcomb attempted to explain the inconsistency between attitude and overt behavior.6 He presented two reasons for there not being a simple and perfect correspondence. One was that behavior is a product not only of attitudes but of immediate situations as well. The other was that attitudes relevant to a situation are often multiple. Newcomb stated:7 There is nothing surprising in this fact [much of the variation in behavior is a result of variation in the immediate situa- tion of the individual] of course, and we would be sadly maladjusted if our behavior had some total momentum of its own that ignored the ongoing realities of the developing situations in which the behavior was occurring. 0n the other hand, it is clear that human behavior is not responsive only to the direct stimuli of the immediate situation. 5George Herbert Mead, Mind, Self and Society, ed. Charles W. Morris (Chicago, Illinois: The University of Chicago Press, 1959), p. 238. 6Theodore M. Newcomb, Ralph R. Turner and Philip E. Converse, Social P cholo (New York: Holt, Rinehart & Winston, Inc., 1965), pp. 6639. 7Ibid., p. 67. 59 provided a df attitude and 0‘“ Fig. l-- Between Attitud Subject's 'eX sum of St Althoug‘. Hewoomb‘s. Thc analyzing patt required: (1) the antecedent situations) of the couectior indiVidual's I Concurrent Co} in “hid“: the ‘ lal Sitnat \8 9%. York: ‘14qu harps: O °f PSr J' kl 18 He provided a diagram to help describe the relationship between attitude and overt behavior.8 Fig. l-—Newcomb's Diagram Illustrating the Relationship Between Attitudes and Overt Behavior. Subject'siiexperience .>> Subject's current- (sum of past situations)l 'attitudes ‘Subject's 4>avbehavior Although not as specific, the Sherifs' position is similar to Newcomb's. The Sherifs were concerned with the most useful way of analyzing patterns of behavior. Two essential sets of data are required: (1) data on the behaviors being studied and (2) data on the antecedent and concurrent conditions (correlated stimulus situations) of the behaviors. The antecedent condition refers to the collection of internalized characteristics that reflect the individual's personality conditioned within a social melieu. The concurrent conditions refer to the characteristics of the situation in which the behaviors are manifested.9 Yinger stated that the social psychological research is only rarely designed to handle simultaneously the variables that derive from inner tendencies and those that stem from the influences of the social situation.10 The proper subject matter for social psychology 8Ibid., p. 68. ’ 9Muzafer Sherif and Carolyn w. Sherif, Reference Group; (New York: Harper and Row, Inc., 1961+), p. 78 10J. Milton Yinger, ”Research Implications of a Field View' of Personality,'.American Journal of Sociology, LXIV (1963) 581-58“. is the explanat setting. He de simple mathemat factors located located in the bebavior will 1 The imp< filmy esta‘oli Heller. He de attitude in ov facil-‘“££k’tec] by Situation. P e as a Single c] of attitude it hinders the e: form the prop. AC : ( “he 19 is the explanation of behavior of the individual in the social setting. He described the social psychological point of view in simple mathematical terms. Behavior is the result of "predisposing" factors located in the individual and ”precipitating” factors located in the situation. If either set of factors equal zero, behavior will not result.11 The importance of the above theoretical orientations became firmly established after discussing two working papers with A. O. Haller. He deve10ped the proposition that the expression of an attitude in overt behavior is constrained, impeded or differentially facilitated hy characteristics of the personality and of the social situation. Personality and social situational variables are treated as a single class of independent variables that affect the expression of attitude in overt behavior. This class of variables aids or hinders the expression of attitude in overt behavior. In symbolic form the proposition stateszlz Ac = (Att; Fl;.Att; F2 . . .‘Att; En) where Ac = the overt action variable Att - the attitude toward the object F each facilitational variable including both personality and social structural variables In the area of race relations two excellent books have stated the need for a multivariate perspective. Simpson and D'Ibid" p. 585. lgA. O. Haller, ”A Point of View on Social Psychology," Paper presented to Alpha Kappa Delta of Michigan State University, October 25, 1963, p. 14. See also ”Some Principles of Attitudes and Behavior,” a working paper, 1960, pp. 1-8. finger's text references on in the study : prejudice has sociology and of research er. and discrimina aspects, they factors. They at any particu‘ sandman inform: analysis varia] °f the Possible Q18 of 1 the area of me They evaluated 20 Yinger’s text, Racial and Cultural Minorities, is replete with references on the importance of taking a multivariate perSpective in the study of race relations. They stated that the study of prejudice has been a fruitful approach to many problems of general sociology and social psychology.13 Although there has been a deluge of research and commentary on the personality aspects of prejudice and discrimination and the consequent neglect of social and cultural aspects, they recognized the legitimacy of studying personality factors. They were, however, opposed to one factor eXplanations at any particular level of analysis. A full understanding requires maximum information on all levels. For purposes of scientific analysis variables are separated. The scientist needs to be aware of the possible interaction among the variables.14 One of the best critical reviews of theory'and researCh in the area of race relations is the work of Suchman, Dean and Williams. They evaluated the relationship of attitude and overt behavior toward members of minority groups and stated a number of conclusions: 1. From the 1920's into the iste'lguo's there was a large number of studies that attempted to describe individuals' stereotypes of social and cultural groupings, their feelings of social distance, their eXpressions of liking and disliking. The great majority of these studies were too static and sheerly descriptive and they tended to treat “prejudice" as an unanalyzed aggregate of individual attitudes divorced from the functioning of real person- 13George Eaton Simpson and J. Milton linger, Racial and Cultural Minorities (New York: Harper & Row, Inc., 1965), p. 49. lulbid. , pp. 70—80. 2. Bgfigmhdoamamg ('9' FT. mom :1 momH-crcroon' 5'11 gmfl-g’nd SH'MrbOng 21 alities and from enduring social relations and tn? structural properties of groups and communities. 2. There is a need for research specifying the functional role of prejudice i2 the total dynamics of personality- in-social-setting.1 3. There is a pressing and urgent need to expand a vigorous and imaginative program of further research on the situa- tional variables as they are related to expressions of attitude and overt behavior. This need consists of developing techniques of field investigation and of controlled eXperimentation that will permit rigorous testing of the rich store of questions that can now be formulated in overcoming major obstacles--the inacces- sibility to observe many important real-life situations, the difficulty of relating survey data to those situa- tions we are able to observe, the ethical and operational hazards of situations contrived for research purposes, and the clumsiness of available techniques of recording the kind of data needed.17 4. One of the major difficulties in research has been the inadequacy of coming to grips with the problem of posing questions which clearly focus upon the antecedent-- consequent relation of unambiguous variables. The state of scientific development in this area is still typified by descriptive explorations . . . . The laws, generaliza- tions, or credible hypotheses of greatest immediate social interest and applicability are often not those which state a predictable relationship between two variables with all other variables constant. Rather they are those which are cast in what might be called a ”complex-adequate" form. The complex-adequate hypothesis typically deals with several interrelated variables, all of which are necessary conditions for the consequent condition to be predicted or controlled, but none of which alone is a sufficient condition. This approach attempts an optimal decision between the rigor of the simple, analytic, abstracg hypothesis and the complexity of concrete events.l 15EdwardA. Suchman, John P. Dean and Robin M. Williams, Jr.. Dese re ation: Some Pro ositions and Research Su estions (New Yerk: Anti-Defamation League of B'nai B'rith , pp. 90-91. 16 . .lhiQOO P0 9#- l7Ibdd.. p. 10a. lerid., p. 110. from t. irport' suggee depend: and a c and pas of the charact individ sociolo t 9 Th Gen general System no, ar Currest “era: flavio: tangent ° 1 ‘Odows . 22 Although each of the above orientations can be distinguished from the other's, there are similarities. They all recognize the importance of a multivariate approach. They implicitly or explicitly suggest that attitude might be studied as both an independent and dependent variable. It is an independent variable affecting behavior and a dependent variable affbcted by characteristics of the situation and past experiences. Although a few authors stress the individuality of the person's past experiences, they indicate that individual characteristics develop within a number of social contexts. Many. individual characteristics can be considered either psychological or sociological variables. The General Proposition Selected as a Guideline In this study a modification of Newcombfs preposition was selected as a general guideline. His approach contained both the general elements of the other prOpositions and a classification system of the types of determinants that influence behavior. His approach was modified because he neglected to emphasize the fact that current attitudes are affected by the current situation. The subject's current attitude is not independent of his definition of the current situation, i.e., stimuli present in the current situa- tion are associated with reconstructed past experiences to form the current attitude. Subjects do have acquired behavioral dispositions (general attitude), however, the particular orientation toward overt behavior (specific attitude) is partially dependent upon stimuli present in the situation. Newcomb's diagram should be modified as follows: 23 Fig. 2--A Medified Presentation of Newcomb's Description of the Relationship Between Attitude and Overt Behavior. Subject's relevant st e erience _A . . ' [Subject's currentgl Subject's . '8 attitudes --> behavior Subject's definition of the current _situation Symbolically the modification of Newcomb's diagram can be represented as: B =.Att (definition of the situation and antecedent variables). where B = observable behavior directed toward a class of objects. Att = attitude toward the class of objects. Definition of situation = the subjects' perception and interpretation of current stimuli organized into meaningful units. Antecedent variables = relevant variables derived from past experiences. This class of variables can include traditional psychological, social psychological and sociological variables. The definition of the situation is not equivalent to attitude toward the object. An attitude consists of an orientation to respond. A definition of the situation is the active organization of relevant stimuli in the social context that aids in defining the appropriate reaponse that the subject is required to make in the particular social context. The definition of the situation and antecedent variables are independent variables that affect the expression of verbal attitudes and overt behavior. Verbal attitude is an intervening variable that affects overt behavior. The at the researche explanation c information o predict overt require diffe into proposit selecting var is dependent 1 empirical faci fellowing Che; will be examir 2b The above proposition was a useful guideline that helped the researcher select classes of variables that provided an explanation of overt behavior. It did not, however, provide information on the specific independent variables that would predict overt manifestations of behavior. Different content areas require different sets of independent variables which are develOped into propositions to eXplain and to predict overt behavior. In selecting variables in any particular content area, the researcher is dependent upon his own intuitions, previously accummulated empirical facts and theoretical conceptualizations. In the following chapter the various components of specific propositions will be examined. ‘ DEVELCI Introduction In th:‘ explain three of a minorit; into three a: Situation . II 1. Au 2. Pa CHAPTER.III DEVELOPMENT OF SPECIFIC PROPOSITIONS AND HYPOTHESES Introduction In this study four independent variables were selected to explain three different manifestations of behavior toward members of a minority group. The independent variables were subdivided into three antecedent variables and a characteristic of the current situation. They were: 1. Authoritarianism, 2. Past intergroup contact with members of a minority group, 3. Perceived support from significant others for engaging in overt behavior with members of a minority group, and 4. Definition of the situation. The three different manifestations of behavior toward members of a minority group were: 1. Attitudes expressed toward members of a minority group. 2. Commitment to become involved in possible activities with members of a minority group, and 3. Overt behavior*directed toward members of a minority group. Both attitude and commitment were intervening variables. They were influenced by the four independent variables and influenced decisions to interact with members of the minority group. Overt behavior was the dependent variable. It was a pattern of action oriented directly toward minority group members. 25 intecedent Var; Authoritafianis The vari the California measured and c2 of articles ha1 utility in the. Authori The first reas California F4 toward minor‘r correlations . averaged abou of available relattionship developed in Silbs equent p Antecedent Variables Authoritarianism The variable authoritarianism (and its operational definition-- the California F-scale) has become one of the most frequently measured and criticized variables in the social sciences. Hundreds of articles have been written to attack, defend or qualify its utility in theory and research.1 Authoritarianism was selected as a variable for two reasons. The first reason was the persistence with which responses to the California F-scale has been found to be related to negative attitudes toward minority groups. In the initial development of the F-scale correlations of Forms 40 and 45 with the ethnocentrism scale averaged about .75.2 In 1954 Christie concluded that on the basis of available evidence, ”the general point of view regarding the relationship between personality characteristics and ethnic prejudice developed in The Authoritarian Personality has been substantiated by subsequent research. '3 lRespective examples of an attack, defense and qualification are: Herbert H. Hyman and Paul B. Sheatsley, ”The Authoritarian Personalityaqa Methodological Critique,’l in Continuities in Social Research: Studies in the Soc 9 and Method of iThe Authoritarian Personalit T eds. Richard Christie—a nd Marie J' ahoda (New York: The Free Press of Glencoe, l95fl), pp. 50-122; Nevitt Sanford, ”The Approach of the Authoritarian Personality” in P3 cholo of Personalit , ed. J. L. McCary (New York: Grove Press, Inc., 19EZS, pp. 253-319: and George Eaton Simpson and J. Milton Zinger, op. cit.. PP- 62-79. respectively. ' .M 2T. W: Adorno et al., The.Authoritarian Personality:(N6W'YOrk= Harper 8: Brothers, 19505, pp. 262-263,. 3Richard Christie, "Authoritarianism Re-examned,” Continuities in Social Research: Studies in the Scope and Method of "The Authori- tarian Personality”, eds. Richard Christie and Marie Jahoda (New York: The Free Press of Glencoe, 1954), p. 166. 27 Sanford reported that a number of studies have found substantial correlations between the E— and F-scales. He stated that 25 to 30 percent of the total variance of the overt expression of prejudice could be safely ascribed to authoritarianism.“ Martin and Westie found that scores on an independent scale measuring the degrees of racial tolerance were inversely correlated with scores on the F-scale.5 Pettigrew found that the F-scale correlated with prejudiced attitudes for different ethnic groups in South Africa and in the Northem and Southern regions of the United States .6 Photiadus and Bigger studied the relationship between nine variables and Borgardus' social distance scale. The nine variables were: religious orthodoxy; extrinsic belief; church participation; formal education; anomia, status concern; conservatism; authoritarianism; withdrawal and anti-social tendencies. The relationship between each of the six personality variables and ethnic distance was found to be positive and significant. When the nine variables were controlled, however, authoritarianism was the only personality variable that continued to be significantly related to ethnic distance.7 One of the more care- fully designed tests of the F-scale was the work of Chapman and “Sanford, op. cit., pp. 295-297. 5James C. Martin and Frank R. Westie, I'The Tolerant Personality,” American Sociological Review, 21+ (1959) 524-525. 6Thomas F. Pettigrew, "Personality and Sociocultural Factors in Intergroup Attitudes: A Cross-National Comparison," Journal of Conflict Resolution, II (1958) 31-32 and 37.40. 7John D. Photiadus and Jeanne Biggar, " Religiosity, Education End—6Ethnic Distance," American Joumal of Sociology, LXIII (1962) 70 72. l: Campbell. Aftr reSponse set , total score cc content and th the acquiesce: The see its relative : festations to:~ evidence demon Pmiadiced at; the F-Scale wi group was Comp Strength of th Theoret interpretatior ”Sponge patte i.en the Uni: contexts . Al‘ centrated 0n . 28 Campbell. After controlling for the effect of the acquiescence reaponse set, they found that individual item analysis as well as total score correlations indicated a relationship between the F-scale content and the ethnocentrism over and above that accounted for by the acquiescence set.8 The second reason for selecting the F—scale was to determine its relative importance in explaining and predicting overt mani- festations toward minority group members. The preponderance of evidence demonstrating the relationship between the F—scale and prejudiced attitudes cannot be dismissed. When the association of the F-scale with the overt manifestations toward members of a minority group was compared to the other independent variables, the relative strength of the F-scale as an independent variable was evaluated. Theoretically, particular’emphasis was given to a qualified interpretation of authoritarianism as measured by the F-scale. The response pattern to the F-scale is assumed to be social in origin, i.e., the unique pattern of a personality characteristic that develops is largely determined by the variety of experiences in different social contexts. Although the authors of The Authoritarian Personality con- centrated on the individual rather than on the group determinants of attitudes, they recognized the importance of social factors. The formative influences on the personality are external as well as internal. Personal attitudes become standardized because they 8Loren J. Chapman and Donald T. Campbell, ”The Effect of Acquiescence Response-Set Upon Relationships Among the F-scale, Ekhnocentrism and Intelligence," Sociometry, 22 (1959) 160. themselves are ccmformity.9 Stewart authoritariani roles that are high levels of limited numbex aged, the fur; antic religim isolates and 4 environment}: they are poor understand or tend to Cling the authors 0 Score high on SPoups and 01; qualities to 29 themselves are the product of social experiences that enjoin conformity.9 Stewart and Hault have suggested that the measured level of authoritarianism is negatively correlated with the number of social 10 They noted that many studies have found roles that are mastered. high levels of authoritarianism among those who have only played a limited number of social roles, e.g., the less well-educated, the aged, the rural, members of disadvantaged minorities, the more dog- matic religious groups, the lower socio—economic strata, social isolates and those who have been raised in an authoritarian family environment.11 When the above types of people meet new situations, they are poorly trained in taking the role of the other and cannot understand or'sympathize with members of different groups. Thus, they 12 Steiner and Johnson agreeing with tend to cling to the familiar. the authors of The Authoritarian Personality; characterized those who score high on the F-scale as people who tend to assign people to in- groups and outgroups attributing good qualities to the former and bad qualities to the latter. Subjects view people in a dichotomous fashion and they resist evidence which tends to contradict their simple conception of society.13 Newcomb recognized two important character- 9Adorno et al., op. cit., p. 7u7. loDon Stewart and Thomas Hault, 9A Social-Psychological Theory of the Authoritarian Personality,” American Journal of Sociology, 65 (1959). 274. ' 11Ibid., p. 277. lzIbid . , pp. 2714-279. 13Ivan EL Steiner and Homer H. Johnson, QAuthoritarianism and Conformity,” Sociometpy, 26 (1963) 21. istics in ' tinctions to perceiv particular threatened keep the < Th: to positi' importanc orientati Who score environ“ eIW'lronme ch53 of of the s of °bjec manifest Perceive °f the : perceim Thus 1 m. The be re \ Holt an 3o istics in the development of prejudice. One was the sharp dis- tinctions between ingroups and outgroups and the other was learning to perceive members of certain groups as sources of threat. When particularly important ingroup relations were perceived to be threatened, behaviors were elicited that manifested the attempt to keep the outsiders at a distance or to injure the outsiders.lu The two major characteristics that are thought to be related to positive reSponses to the F-scale are the rigid adherence and importance assigned to highly valued ingroups and the threat orientation to those outgroup groups in the environment. Those peOple who score high on the F-scale want to live in a stable, ingroup environment and do not want to change those qualities of the social environment (self and social order) that are defined as good. A class of objects that is associated with changing the good qualities of the social environment are perceived as threatening. The class of objects that is perceived as a threat is the object of unfavorable manifestations of overt behavior. Members of minority groups can be perceived as a class of objects involved in changing those qualities of the social environment that are defined as good. They can be perceived as demanding or seeking entrance into important ingroups. Thus, members of minority groups can be perceived as threatening. Therefore, the first proposition is: l“Theodore M. Neweomb, Social Psychology (New Ybrk: Henry Holt and Company, Inc., 1950), p. 579. Past Integ r01; The sec comitment and minority group reveals the v; a hiStOl'y as g the interacti Specific situ J. other‘s prese intergroup c attitudes an 31 Those people who have a high level of authoritarianism as measured by the F- scale will eXpress less favorable attitude, commitment and overt action toward members of minority groups. Past Intergroup_Contact The second independent variable that affects attitude, commitment and overt behavior is past eXperience with members of a minority group. Theory and research in the area of race relations reveals the variable intergroup contact has almost as wide and varied a history as authoritarianism. Intergroup contact can be defined as the interaction between both majority and minority group members in specific situations which require those involved to react to the other's presence and to communicate thoughts and feelings either directly or indirectly. The major question is, ”How effective is intergroup contact with members of a minority group in changing attitudes and behavior toward that group?“ Answers provided for the question have an interesting history. Blumer's comments on how social scientists are affected by the larger society are relevant. He pointed out that the study of race relations has Sprung from and is sustained by a melioristic interest in the improvement of the relations between racial groups. Social scientists have stressed the importance of putting the democratic ethic into practice.” Early research on the effects of intergroup contact l5Herbert Blumer, ”Research on Race Relations in the United agates of.America,' International Social Science Journal, 10 (1958) 5. tended to refl stated that in attitudes and continued to b hypothesis res Diring ijudice whit units of diffe the greater th plat°°n (50 me other cases, N regiment, (3,0C in sitIii-Itions attitudes t ow: favorable at t: t° Neglees as BI‘Ophy seamen “Ward had 8hj-Dped c Hegl‘oes Prior 32 tended to reflect this source of bias. In essence, the hypothesis stated that intergroup contact resulted in producing more favorable attitudes and behavior toward members of a minority group. As research continued to be published, qualifications were added to the original hypothesis resulting in the issue becoming more complicated. During Wbrld war II, Stouffer §§_gl, examined the amount of prejudice white infantrymen expressed toward Negroes in military units of different sizes. It was assumed that the smaller the unit the greater the contact. The greatest contact occurred when a Negro platoon (50 men) was assigned to a white company (200 men). In other cases, Negro units comprised parts of larger units, e.g., regiment (3,000 men) or a division (13,000 men). They found that in situations involving greater contact the whites had more favorable attitudes toward Negroes. EVidence indicated, however, that these favorable attitudes in more integrated units were somewhat confined to Negroes as fellow soldiers in combat.16 Brophy found major differences in the attitude of white seamen toward Negroes depending upon the number of times the seamen had shipped out with Negroes. Seamen who had shipped out with Negroes prior to the hostilities in world war II expressed less prejudice than merchant seamen whose service was limited to wartime. Among the seamen who had nevershipped out with Negroes, 33 percent were rated as unprejudiced. This percentage increased to #6 for léSamuel A. Stouffer et al., "Negro Infantry Platoons in White Companies,” Readings in Social ngphology, eds. Eleanor E. Maccoby, Theodore M. Newcomb and Eugene L. Hartley (New York: Holt, mnehart and Winston, Inc., 1958), pp. 598-601. 33 those who had shipped out once with Negroes, 62 for those who had shipped out twice, 80 for those who had shipped out three times and 82 for those who had shipped out five or more times.17 MacKenzie found that contact between workers having the same or nearly the same economic and social status improved friendly relations between whites and Negroes.l8 Watson's study of M5 residents of New York drew the same conclusion.19 Additional studies in work situations have also made quali- fications of the contact hypothesis. Harding and Hogrefe studied white workers in two department stores. They divided the white workers into three groups: (1) equal status contact group: (2) unequal status contact group (Negroes all had lower status); and (3) a group who had never worked with Negroes. Those workers who had equal status contact had more favorable attitudes toward Negroes on the job, but not with reSpect to public transportation, eating facilities, 20 The research of-Palmore, Minard and housing or friendship choices. Lohman and Reitzes confirmed Harding and Hogrefe's finding that more favorable attitudes may develop toward Negroes on the job but the l7Ira M. Brophy, 'The Luxury of Anti-Negro Prejudice," Public Opinion Qggrterly, 9 (1945) 1:624:63. 18Barbara K. MacKenzie, “The Importance of Contact in Deter- mining Attitudes Toward NegroeS." WM 13191121231. 43 (19%) 107-438. 19Jeanne Natson, "Some Social and Psychological Situations Related to Change in.Attitude,' Human Relations, 3 (1950) 39. 2°John Harding and Russell Hogrefe, "Attitudes of White Department Store Employees Tomrd Negro Co-workers,“ Journal of Social Issues, 8 (1952), pp. 19-22. favorable attil Tue stun effects of con. in integrated ‘I be friendly to In segregated 1 follow-up Stud Freeway cu “melted ho percentages < “eié‘nborly a “‘5 separate prOdettts we; Percentages Hare Complj 31+ favorable attitudes are not transferred to off the job experiences.m' Two studies of interracial housing have shown the positive effects of contact experiences. Deutsch and Collins found that in integrated housing about two—thirds of the housewives wanted to be friendly toward Negroes and one-third wanted to avoid contact. 22 In a In segregated housing the same respective ratio was 1:11. felloweup study Wilner §p_§1, provided supporting evidence for the previously cited research. When comparing the integrated and segregated housing patterns in the two studies, they found that the percentages of white housewives sharing at least one kind of neighborly activity with Negroes were remarkably similar. In the two separate studies the percentages in the integrated housing projects were 54 and 50. In the segregated housing projects the percentages were 3 and 5.23 Wilmer gp_§;, suggested that there were complicating factors that affected the outcome of intergroup contact. Initial attitude, the influence of social pressure, official and public policy, the outcome of the contact eXperience v—_ ZlErdman B. Palmore, ”The Introduction of Negroes into White Departments,“ Human Organization, 14 (1955) 27-28. R. D. Minard, 'Race Relations in the Pocahontas Coal Field, ” Journal of Social Issues, 8 (1952) 29-h4. Joseph P. Lohman and Dietrich C. Reitzes, lDeliberately Organized Groups and Racial Behavior,” American Sociological Review, 19 (1951+) 2&2. 22Morton Deutsch and Mary E. Collins, Interracial Housi . (Minneapolis, Minnesota: University of Minnesota Press, 1951), p. 79. 23Daniel M. Wilner et ___a__l., Human Relations in Interracial Housing (Minneapolis, Minnesota: University of Minnesota Press, 1955 9 p0 114'30 and size of th contact?“ A eerie youthful pop environment. tith Indian an States did not thcse students' SGHSitive area: country, they te In her a knowing I“egress contacts with 1 Studies qualifications controlled eXpe The two groups were bmught ix PBCippoCal pre gmups. The b< Sinations 35 and size of the minority group can all affect the outcome of contact.24 A series of studies using the contact hypothesis have measured youthful populations either in or outside the formal educational environment. Lambert and Bressler found that intergroup contact with Indian and other students from Asia who were in the United States did not automatically lead to favorable attitudes. whenever these students were interacting in situations discussing certain sensitive areas that involved implications of low status for their country, they tended to reapond negatively.25 In her sample of university students, MacKenzie found that knowing Negroes of professional status and having a variety of contacts with Negroes resulted in more favorable attitudes.26 Studies of younger age groupings have resulted in several qualifications of the contact hypothesis. Sherif in a carefully controlled experiment studied intergroup relations at a boys' camp. The two groups of boys lived separately at an isolated camp site. They were brought into competitive and frustrating interaction. Strong reciprocal prejudices and stereotypes developed between the two groups. The boys were presented with seven unstructured contact situations. These situations did little to reduce stereotyping, reduce conflict or and group distinctions. When the two groups 2“Ibid. , p. 6. 2531chard D. Lambert and Marvin Bressler, "The Sensitive- Area Complex: ‘A Contribution to the Theory of Guided Culture Contact,” American Journal of Sociology, 60 (1955) 584. 26M'acK’enzie, op. cit.. p. 435. were forced ir (raise enough flowing again the prejudice and outgroup c found similar Hogrefe, Eva an integrated ‘i‘fer from a measuring Soc test did Show play center . segregation attended a f Significant: significant,“ had more ag They felt t 36 were forced into situations of working for superordinate goals (raise enough money to bring a movie to camp or get the water supply flowing again after it had been disrupted), the group lines blurred, the prejudice was reduced and the differential rating of intergroup and outgroup disappeared.27 In the camp setting YarrOW'§§_§;, found similar results in NegroAWhite interracial relations.28 Hogrefe, EVans and Chein found that white children who had attended an integrated play center once a week for several.months did not differ from a control group in attitudes toward Negroes on a scale measuring social distance. However, the results of a projective test did show a larger proportion of the students who attended the play center either were favorable or strongly opposed to racial segregation in play situations.29 Mussen studied 106 white boys who attended a four-week interracial camp. He found that 27 boys became significantly more prejudiced against Negroes and 28 boys became significantly less prejudiced. Those boys whose prejudice increased had more aggressive feelings and greater need to defy authority. They felt themselves victims of aggression, felt that others were not kind and helpful and were more dissatisfied with the summer camp. 27M'uzafer Sherif, Inter rou Conflict and Coo eration: The Robbers Cave Experiment, ZNorman, Oklahoma: University of Oklahoma Book Exchange, 1961), p. 210. , 28Marian Yarrow et al., ”Acquisition of New Norms: A Study of Racial Desegregation in Interpersonal Dynamism in.a Desegregation Process,” Journal of Social Issues, ed. Marian Radke Yarrow, 1“ (1958) 29Russell Hogrefe et al., ”The Effects on Intergroup Attitudes of Participation in an Interracial Play Center,I reported in Thg| Handbook of Social Psychology, ed. Gardner Lindzey (Reading, Massachusetts: Addisonawesley, Inc., l95fi), p. 1053. Tnose whose pr feeling of pun toward those i and were satis In the desegregation Junior and set: in Qik Ridge, Emmy attit ‘i-‘Lr‘t’:ction,3l interesting d1 (those most di were aSked, I} been affected I"espondemm re Azong Norther;- 37 Those whose prejudice decreased had few aggressive needs and less feeling of punishment and retaliation. They had favorable attitudes toward those in their social environment, were accepted by others and were satisfied with the eamp.3° In the broader social context the effects of school desegregation are not clear. Campbell studied the attitudes of junior and senior high school students before and after desegregation in Oak Ridge, Tennessee. He used four scales to measure anti- minority attitudes and found a significant shift to a negative direction.31 In a national sample, Hyman and Sheatsley found interesting differences among those in the 21 to 24 age group (those most directly affected by school integration). Respondents were asked, "In what ways have you or any members of your family been affected by integration?” More than four-fifths of the respondents reported no effects. Not a single Southerner of this age group Spontaneously reported any kind of favorable effect. Among Northerners in the same age group, 5 percent volunteered an answer describing the personal effects of integration in favorable terms.32 Coles studied white adolescents who were in integrated schools and found significant changes toward more favorable attitudes. He found also that simple stereotypes began to break 30Paul H. Mussen, "Some Personality and Social Factors Related to Changes in Children's Attitudes Toward Negroes,“ Journal of Abnormal and Social Psychology, 45 (1950) h28-4#l. 31Ernest o. Campbell, ”On Desegregation and Matters Sociological,” Phyion (Summer 1961) 141. 32Hyman and Sheatsley, op. cit., p. 23. don. even anon cents.33 One of hypothesis was by Brink and H attitudes amor- i‘iesroes. This were thl‘ougho his cause. T3 about Negroes With Negroes Thus, Eleanor are with membe rs of Snob Cont; negative att reduce deve] 38 down even among many of the more segregationist-minded adoles- cents.33 One of the largest studies concerned with the contact hypothesis was a nationwide study of more than 1200 white people by Brink and Harris. They found consistently more favorable attitudes among whites who had some previous social contact with Negroes. This group comprised 25 percent of the total sample and were throughout the survey the most sympathetic to the Negro and his cause. They were less likely to agree with white stereotypes about Negroes and less likely to object to having future contacts with Negroes over'a variety of possible contact situations.34 Thus, the evidence indicates that negative attitudes and behavior are sometimes explained as a result of the Lack of contact with members of a minority group and sometimes explained as a result of such contact. Unpleasant contacts can increase the strength of negative attitudes and behavior;and certain kinds of contact can reduce develOped negative attitudes and behavior. A series of outlines have been developed to account for the different findings. These outlines proposed variables that should be taken into account in the analysis of intergroup contact between members of different groups. Allport's outline covered: (1) quan- titative aspects of contact; (2) status aSpects of contact; 33Robert 00193, The Desegregation of Southern Schools: A Ps chiatric Stud (New York: (Anti-Defamation League of B'nai B'rith, l9 3 . p. 10. 31" Brink and Harris, op. cit., pp. 140-148. (3) role a contact; ( and (6) th Dean and in related tc the effect intergrou; that the i interactic t‘ne {31‘8qu finger dex "mowledge l. 2. 39 (3) role aSpects of contact; (4) social atmosphere surrounding contact; (5) personality of the individual experiencing the contact; and (6) the areas of contact.35 In a detailed evaluation Suchman, Dean and Williams outlined a large number of variables that are related to intergroup interaction: the nature, the determinants, the effects and the patterns of communication associated with intergroup interaction.36 They retained, however, the proposition that the integration of whites and Negroes in situations of social interaction is accompanied and followed by an overall reduction in the prejudices of white people toward Negroes.37 Simpson and finger develOped four related propositions to sum up the present knowledge on the effects of contact on prejudice:38 l. Incidental, involuntary, tension-laden contact is likely to increase prejudice; 2. Pleasant, equal—status contact that makes it unnecessary for the individuals to cross barrier of class, occupa- tional and educational differences as well as differences in symbolic group membership represented by such symbols as “race” is likely to reduce prejudice; 3. Stereotype-breaking contacts that show minority group members (as individuals) in roles not usually associated with them reduce prejudice; u. Contacts that bring people of minority and majority groups together in functionally important activities reduce prejudice. 3scvordonAllport, The Nature of Prejudice (New York: Addison- Wesley, Inc., 1951+), pp. 262-263. 36Suchman et al.,.gp‘_git,. pp. 47-51. 37Ibid., p. 90. 38Simpson and'Yinger, op. cit., p. 510. In this qualifications suggests that lead to more f Inpothesis car. Romans in his behavior has 5 the contact ‘nyh ms Pmpositi 40 In this study the contact hypothesis will be accepted with qualifications for three reasons. First, the weight of evidence suggests that in most cases favorable intergroup interaction does lead to more favorable attitudes and behavior. Second, the contact hypothesis can be related to a more general theoretical perspective. Homans in his attempt to develop a parsimonious theory of social behavior has stated a two variate proposition that is related to the contact hypothesis. The proposition stated:39 If the frequency of interaction between two or more persons increases, the degree of their liking of one another will increase and vice—versa. This proposition has been reformulated by Malewski. He stated:40 If the costs of avoiding interaction are low and if there are available alternative sources of reward, the more frequent the interaction, the greater the mutual liking. The third reason is related to the second. In this study a series of propositions are developed to test a multivariate theory of the relationship between attitude and overt behavior. If the contact hypothesis was the only proposition to be tested many of its quali- fications could be thoroughly examined. ‘When the major objective is to interrelate a series of propositions, it becomes necessary to remain at a general level of analysis. 39George C. Homans, The Human Group.(New”Yorkx Harcourt, Brace and World, Inc., 1950), p. 112. uoThis recent derivation of Homans' proposition was reported in Hans Zetterberg, op. cit., p. 66. Taking intergroup co: depressions of were considerr;l contact. Con‘ of structural forced to sit Although contzl voluntary, y a wide range tions are dif seeking Compa as an indepen second propos Support from the large EI‘Oups. Of an inched 41 Taking into consideration the theory and research on intergroup contact, certain qualifications were made. Favorable expressions of overt behavior toward members of a minority group were considered to be influenced by the extent of varied intergroup contact. Contact could be frequent but only occur within one set of structural conditions. For example, a white student might be forced to sit next to a Negro student because of a seating chart. Although contact would be frequent, it would be imposed rather than voluntary. Varied intergroup contact refers to experiences across a wide range of possible situations. The structuring of the situa- tions are differentiated, e.g., from room assignment to voluntarily seeking companionship. Varied intergroup contact was considered as an independent variable affecting overt behavior. Thus, the second proposition is: If the costs of avoiding interaction are low'and if there are available alternative sources of reward, the more varied the intergroup contact, the greater the degree of favorable attitude, commitment and overt behavior toward members of a minority group. PerceivedSuppgrt From Significant Others The third independent variable selected was perceived support from significant others. The variable was selected from the large body of social psychological literature on reference groups. The general proposition that the attitude and behavior of an individual toward a class of objects develop in response to the attitud has been demon In the z the influence ( tmard members union members ' and Who also a Sherif stated men following main emphasis #2 to the attitudes and/or behavior of positive reference group members has been demonstrated frequently.ul In the area of race relations a number of authors have reported the influence of reference groups on whites' attitudes and actions toward members of a minority group. In explaining the phenomenon of union members who practiced non-segregation in their union activities and who also actively participated in the Detroit race riots of 19h3, Sherif stated that this inconsistency was the result of the union men following the dictates of conflicting reference groups. His main emphasis was reference groups. He statedzu2 Our idea of what we are, what other groups are, what is desirable for a person to be, what is a desirable position to occupy, who are desirable persons to associate with, who are the persons who should be put at a distance, are derived in their major out- lines from our reference groups. Deutsch in his study of interracial housing patterns discovered patterns of inconsistency in freely expressed attitudes and two different patterns of behavior. Deutsch stated that the most u;A few good examples of general research are: Theodore M. Newcomb, FAttitude DevelOpment as a Function of Reference Groups,” Readings in Social ngchology, eds. Eleanor E. Maccoby, Theodore M. Newcomb and Eugene L. Hartley (3rd ed.; New York: Holt, Rinehart and Winston, Inc., 1958), pp. 265-275; W} W} Charters, Jr; and Theodore M. Newcomb, ”Some Attitudinal Effects of Experimentally Increased Salience of a Membership Group," Ibid., pp. 276-280; Theodore M. Newcomb, The Ac uaintance Process (New York: Holt, Rinehart and Winston, Inc., 19 l ; Muzafer Sherif and Carolyn W} Sherif, op. cit., pp. 260—273; and Frank Miyamoto and Sanford M. Dornbusch, ”A Test of the Interactionist Hypotheses of Self-Conception,” American Journal of Sociology, LXI (1956) 399-403- uZIsidor Chein et al., (eds.), "Consistency and Inconsistency in Estergroup Relations,a Journal of Social Issues, Vol. V, No. 3 19 37. striking and u out of the in‘. groups which Pearli a model for hi college women toward Negroe groups unfavo, favorable to In the and Welt acts Scaled reSpons this EPOUp the quartiles, Tl covert action have onets pi level of agre would allow t ProfeSSional racial integr Obtained for 43 striking and most frequent instances of inconsistent behavior arise out of the individual's conformity to the standards of differing groups which function in relative isolation from one another.)+3 Pearlin used Newcomb‘s classic Bennington College study as a model for his study of white attitudes toward Negroes. For college women the process of changing to more favorable attitudes toward Negroes involved both disattachment from previous reference groups unfavorable to Negroes and attachment to new reference groups favorable to Negroes. In the study of the relationship between verbal attitudes and overt acts DeFleur and Westie obtained favorable and unfavorable scaled responses toward Negroes from 250 college students. From this group they matched 23 subjects from the highest and lowest quartiles. These matched subjects were asked to take part in an ”overt action opportunity”. It consisted of a signed agreement to have one's picture taken with a Negro of the Opposite sex. The level of agreement was obtained by the extent to which the subject would allow to have his picture publicized-—from a small group of professional sociologists to a nation—wide campaign advocating racial integration. The mean of the signed level of agreement was obtained for the total 46 subjects. Certain discrepancies became apparent. Nine of the students who expressed favorable attitudes toward Negroes fell below the mean and five of the students who exPressed unfavorable attitudes toward Negroes fell above the mean. “BIbid. , pp. 45-146. ”#Leonard I. Pearlin, "Shifting Group Attachments and Attitudes Toward Negroes," Social Forces, 33 (l95h) £7-50. The variable between eXpre decision to p was influence reference gro Linn r. _—_——— He did not fi Willingness t and level of of the questi the discrepan “9P8 attempti Student, 1.6, members. How norms "as 1110 meting DOm detemined by imprinted o te l I | | PGPSonal invol I over the Weak- ha The variable reference group was used to explain the discrepancy between expressed verbal attitudes and the action opportunity. The decision to participate in the various levels of action opportunity was influenced strongly by the perceived agreement of the peer reference group.45 Linn modeled his study after the DeFleur and westie research. He did not find a significant relationship between the degree of willingness to have a picture taken (measured by an attitude scale) and level of signed agreement (overt behavior after the administration of the questionnaire). The major explanation Linn provided to explain the discrepant behavior was the cultural milieu of the students. They were attempting to play the social role of a "liberal“ college student, i.e., expressing liberal attitudes toward minority group members. However, playing the role that conformed to the college norms was inconsistent with the more prevailing norm of avoiding personal involvement with Negroes. When confronted with two con- flicting normative expectations the role-playing behavior was determined by the stronger norm. The more stable, comfortable, imprinted, tested and experienced role became Operative and dominant over the weaker role. Linn agreed with the DeFleur’and Westie finding that peer groups are an important variable in explaining inconsistent behavior. However, he stated that the conceptual framework should include more than peer groups. The stronger A hjflelvin L. DeFleur and Frank R. Nestie, "Verbal Attitudes and Overt.Acts: ‘An Experiment on the Salience of Attitudes," American Sociological Review; 23 (1958) 667-673. nonnative family, c a picture comrdtmen pation ir. various I Ir used inst referean applied <' msearch and negl‘ 0f reSea have use the effe Who is p attitude to the a Signifie moan study or L “5 normative expectations and associated role are influenced by the family, community, school and friendship groups. Agreeing to have a picture taken with a Negro of the opposite sex involved a high commitment to personal social involvement with Negroes. Partici- pation in this level of social involvement was not supported by various reference group norms. In this study perceived support from significant others was used instead of reference groups. There has been confusion between reference groups and reference individuals. The concept has not only applied to groups but to individuals as well. As the theory and research developed, it tended to focus on the term reference groups and neglect the concept of reference individuals.“7 Many examples of research in and outside the area of minority group relations have used the term reference group when they were actually measuring the effects of reference individuals. It is defined as a person who is positively evaluated by an individual. The individual's attitude and behavior toward a class of objects develops in re3ponse to the attitudes and overt behavior of positively evaluated significant others toward the class of objects. Thus, the third preposition is: The greater the extent to which significant others are perceived to have favorable uéLawrence S. Linn, "Verbal Attitudes and Overt Behavior: A Study of Racial Discrimination," Social Forces, #5 (1965) 358-364. “7Merton, op, cit., pp. 284 and 302-303. The Defini The variable. variables independe] in the te. de‘-'eloped hemeen t, affects t and overt As PSYcholog Thomas ar “when, finds bit: with the #6 attitudes and actions toward members of minority groups, the more favorable the attitude, commitment and overt behavior toward members of a minority group. The Definition of the Situation The definition of the situation is the fourth independent variable. Its relationship to the independent and dependent variables is different from the relationships of the other three independent variables. First, it refers to a concurrent variable in the testing situation, rather than an antecedent variable developed in past experiences. Second, it affects the relationships between the three antecedent variables and attitude. Third, it affects the relationships of attitude and commitment and attitude and overt behavior. As a theoretical concept of sociologically oriented social psychology, the definition of the situation was first defined by Thomas and Znaniecki. They used it to refer to the individual's subjective attempt to orient himself to the context in which he finds himself, ascertain his interest, and then proceed to cope with the circumstances.”8 The concept was develOped to explain 48William I. Thomas and Florian Znaniecki, The Polish Peasant in Euro e and America, Vol. II (New York: Alfred A. Knopf, Inc., 1927;. pp. 13n6-18E9 and also William I. Thomas, Primate Behavior (New York: McGraw—Hill Book Company, 1937), p. 8. This material was reported by Tamotsu Shibutani in Societ and Personalit (Englewood Cliffs, N.J.: Prentice—Hall, Inc., 1961), pp. :l-hz. + how general a o ~ould be carri Althou interpretatioi Person' 3 Subj conditions in the Social 8.1, to be a CI‘uci positions hav the attitude of situationa Variables aic‘ behaviw- T} the sullatio, 1&7 how general attitudes (tendencies to respond in overt behavior) could be carried into specific patterns of overt behavior.”9 And the definition of the situation is a necessary preliminary to any act of the will for in given conditions and with a given set of attitudes an infinite plur- ality of actions is possible, and one definite action can appear only if these conditions are selected, interpreted and combined in a determined way and if a certain systematization of these attitudes is reached so that one of them become predominant and subordinates the others. Although the concept refers to the subject's subjective interpretation of his environment, it stresses the fact that a person's subjective perceptions are strongly influenced by conditions in the social situation.50 The important effects of the social situation on manifestations of behavior has continued to be a crucial issue in the social sciences. Essentially two positions have been taken. The first can be characterized as the attitude and situation_approach. It stresses that knowledge of situational variables along with attitude and personality variables aids in the understanding and the predicting of overt behavior. The second can be characterized as the attitude vs. the situation approach. It stresses that situational variables are the primary causes of behavior in the testing situation and overt ugThomas and Znaniecki, op. cit., Vol. I, p. 68. The quote was reported by John Madge in The Origins of Scientific Sociology (New York: The Free Press of Glencoe, 1962), p. 71. 50Yinger, Op. cit., p. 583, and Leonard S. Cottrell, Jr., 'The.Analysis of Situational Fields in Social Psychology,? American Sociological Review, 9 (19“3) 377. behavior, and not important in Chapter II Suchnan, Dean and situation contributions psycholOgy t, interracial the attitude Structure Spe‘ attitudes in Allpor and situation stated :53 #8 behavior, and thus, knowledge of attitude is either irrelevant or not important in understanding and predicting behavior. The First Position: Attitude and Situation Several authors who preferred the first position were presented in Chapter II. Deutsch, Newcomb, the Sherifs, Yinger, Heller and Suchman, Dean and Williams recognized the legitimacy of the attitude and situation approach.51 Additional authors have also made important contributions. Pettigrew criticized the traditional thinking in psychology that led to the neglect of situational variables in interracial behavior and the restricted interpretation and use of the attitude concept. Pettigrew stated that situations not only structure specific racial behavior, but they also change Specific attitudes in the process.52 Allport desired a reciprocity between individual theories and situational theories of causation of racial prejudice. He stated:53 . . . it grows more and more apparent that it is the varying situational context that set off varying action tendencies. Unless we admit the situation in our total analysis leor a review of the authors' positions, see pp. 16-21. 52Thomas F. Pettigrew, "Social Psychology and Desegregation Research,'.American Psychologist, 16 (1961) 105-112. ”cordon w. Allport, "Prejudice: A Problem in Psychological and Social Causation,” Toward A General Theo of Action, eds. Talcott Parsons and Edward.A. Shils (New York: Harper & Row, Inc., 1951), p. 381. Rokeac sore interest objects acros attitudes tow attitude towa retarded the been a failur encountered w ticated attem Single measm interpretatic attitude and social behavj attitudesnor He belieVed t cognitively j degrees of in is differenti to be pifilt llsh Sci %, pp. #9 we shall never be able to deal adequately with the problem of consistency and incon- sistency in individual behavior. Nor shall we discover those conditions that are known to arouse tolerant, and to weaken intolerant, modes of response. Rokeach stated that attitude theorists have generally been more interested in the theory and measurement of attitudes toward objects across situations than in the theory and measurement of attitudes toward situations across objects. The divorcing of the attitude toward situation from the attitude toward object severely retarded the growth of attitude theory. The consequences have been a failure to recognize that an attitude object is always encountered within some situation. There have also been unsophis- ticated attempts to accurately predict behavior on the basis of a single measure of the attitude toward the object, and unjustified interpretations and conclusions about the inconsistency between attitude and behavior. Rokeach postulated that an individual's social behavior must always be mediated by at least two types of attitudes--one activitated by the object, the other by the situation. He believed that attitude toward object and toward situation will cognitively interact with one another and will have differing degrees of importance with respect to one another. Resulting behavior is differentially influenced by the two sets of attitudes.5u suMilton Rokeach, "The Nature of Attitudes,” a working paper to be published in the International Encyclopedia of the Social, Sciences, pp. 6-7 and ll-12. The Second P The In is Bluner. scientific r neet his thr Aconcept mu class of em; clearly this It must enab of objects t “V the usag criteria wer 10 A: P 50 The Second Position: Attitude vs. SitBéEEEB. The most outspoken representative of the second position is Blumer. He criticized the use of the concept attitude in scientific research.55 The usage of the concept attitude did not meet his three requirements for concepts in empirical science.56 A concept must point clearLy to the individual instances of the class of empirical objects to which it refers. It must distinguish clearly this class of objects from other related classes of objects. It must enable the development of cumulative knowledge of the class of objects to which it refers. Reasons that were offered to eXplain why the usage of the concept attitude did not meet the above criteria were: 1. An attitude is not perceived directly but must be pieced together through the process of inference. 2. Attitudes should be distinguishable from opinions, beliefs, etc. 3. Attitudes are defined as a tendency or prediSposition to act. The overwhelming majority of studies tell nothing of the relation of attitude to action. A serious weakness in Blumer's criticism of attitude is the substitute unit of analysis he offered in place of attitude. It has the same weakness he credited to the concept attitude. Blumer preferred the social act as the unit of analysis of human behavior. He posited an intervening process that exists between the tendency to act and the action. The intervening process is built up in the 55Herbert Blumer, QAttitudes and the Social Act." Social 2.299.195» 3 (1955). 59-65. 56Ibid., p. 59. flow of fl 1 53; puts him the self conce in empirical s t is extremel self. It has self as a com of the class c Blumer concept but aj we“ behavio: eflphasized a I ideology and t imPOI‘tant poi relations sin of the Pelati attitudes whi most importer. prejudice to Blumer his diSCOUten I the attitude 5? State Her“: “03‘“??? Ame 51 flow of gglf_interaction. Blumer's dependence on the use of the term §3l§_puts him in the same position as those using the concept attitude. The self concept does not meet his three requirements of a concept in empirical science. The self is difficult to isolate or perceive. It is extremely difficult to delineate between the self and non— self. It has not been clearly demonstrated that the usage of the self as a concept has led to the development of cumulative knowledge of the class of objects to which it refers. Blumer was not only critical of attitude as a scientific concept but also of its particular usage in studies of attitude and overt behavior toward minority groups§firIn his review Blumer emphasized a number of different points. The influence of ideology and major public events on social scientists was an important point. The major trend in theory and research on race relations since World war II stemmed from the belief that the nature of the relations between racial groups results from the feelings and attitudes which the groups have toward one another. Prejudice is the most important attitude, and therefore, it is important to study prejudice to understand discrimination. Blumer strongly criticized the above trend. He expressed his discontent for the prejudice-discrimination axis (prejudice as the attitude and discrimination as the overt behavior). He was also strongly critical of the research and theory on the California F- 57Herbert‘Blumer, ”Research on Race Relations in the united fizatzz ofomerica,' International Social Science Journal,10 (1958) 3 7- scale and simz‘ research stem: personality d the inconsist' members of mi 52 58 scale and similar instruments. Blumer seriously questioned the research stemming from the idea that racial prejudice is due to personality deficiency. After reviewing some of the literature on the inconsistency between attitude and overt behavior toward members of minority groups, he stated:59 The crucial issue raised by such research studies is that of 'situation' versus ‘attitude' in the explanation of behavior. Several specialists in race relations have been led to the position that the social setting of action instead of the racial attitudes of the participants is the prime determinant of behavior. He went on to state:60 Because of the failure to fill in the crucial gap between an attitude and its overt expres- sion the vast outpouring of research on racial attitudes and feelings has not yet tied in with, or thrown light on, the actual behavior of racial groups in their complex association. Lacking a verified connection with concepts of actual racial association, the large body of findings on racial attitudes stands apart with- out having contributed to an integral body of racial theory. Throughout the article Blumer made several important points in his critical assessment. There were, however, shortcomings. Blumer conceived of the issue in terms of situation vs. attitude instead of situation and attitude. Re preferred to criticize a 581n a much shorter article, ”Intergroup Relations vs. Prejudice: Pertinent Theory for the Study of Social Change," Social Problems, # (1956) 173-176, Arnold M. Rose takes a very similar position to Blumer on these two points. sgBlumer, ”Research on Race Relations in the United States of America,“ op. cit., p. “27. 60Ibid., p. 434. nonistic su‘nstitut Ar actual it should he 'comnissi character terms of actual be PI‘Edoming unfavoral 0f the u: unfavom the Jews in the U: and Mosle 1950's. 53 monistic theory (attitudes cause discrimination) and propose a substitute monistic theory (social situations cause discrimination). Another criticism of Blumer is his strong emphasis on studying actual intergroup behavior instead of intergroup attitudes. He should have made a distinction between acts of ”omission” and "commission”. Interracial relations in the United States can be characterized by a lack of contact and interaction, at least, in terms of the frequency, duration, priority and intensity of the actual behavior. Both favorable and unfavorable acts of omission predominate. Although Negroes have been the victims of overt unfavorable acts, the frequency, duration, priority and.intensity of the unfavorable acts is slight in comparison to some other recent unfavorable acts toward minority groups, e.g., the treatment of the Jews in Germany during world war II, the treatment of Blacks in the Union of South Africa and the mutual aggression of Hindus and Mbslems toward one another during the late l9h0's and early 1950's. Blumer's theoretical approach appeared in another article.61 His main thesis was that race prejudice exists basically in "a sense of group position” rather than in a set of feelings which members of one racial group have toward the members of another. A scheme of racial identification is necessary as a framework for racial prejudice. The identification involves the formation of an image of a conception of one's own racial group and of another 61Herbert Blumer, “Race Prejudice As A Sense of Group Position,” Pacific Sociological Review, I (1958) 3-7. racial group ' attitudes and research in r which given r There are fou (l) a feeling race is intri prietary cla: (Lt) a fear an on the peroga The Se 0f the indivjl a common OI'ie feelings and r‘ficducecl to it the members C lies in a fej f(mus is on J to individua: 54 racial group in terms of the relationship of the groups. The attitudes and feelings used in the major trend of theory and research in race prejudice are actually a resultant of the way in which given racial groups conceive of themselves and others.62 There are four basic types of feeling in a sense of group position: (1) a feeling of superiority; (2) a feeling that the subordinate race is intrinsically different and alien; (3) a feeling of pro— prietary claim to certain areas of privilege and advantage; and (h) a fear and suspicion that the subordinate race harbors designs on the perogatives of the dominant race. The sense of group position, however, transcends the feelings of the individual members of the dominant group, giving such members a common orientation that is not otherwise to be found in separate feelings and views.63 It is a general orientation and cannot be reduced to individual feelings and beliefs. It acts as a norm for the members of the dominant group. The source of race prejudice lies in a felt challenge to this sense of group position. ‘The focus is on the position of group to group rather than individual to individual. The major criticism of this short article of Blumer's is the misplaced concreteness. Few authors would disagree with his discussion of the development and nature of prejudice as it is related to groupings of people in society. However, when Blumer 621mm. p. 3. 631bid., p. 4. st; “f_‘ int 55 states that the sense of group position transcends and cannot be found in separate feelings and views, he is entering the realm of the mystical or superorganic. The relationship of part—whole phenomena in the social sciences remains a complex and open question in the philosophy of science.6’+ By not attempting to come to grips with this important issue, Blumer deprived his readers of the opportunity for making a favorable assessment of his theoretical approach. Blumer may have felt compelled to stress the transcend- ental qualities of the sense of group position because of his aversion to attitude and other acquired behavioral disposition conceptualizations. His attempt led to even more perplexing problems. Moreover, he used concepts such as feelings to describe individuals' orientations to behavior. It is difficult to make a sharp, meaningful distinction between a concept like feeling which Blumer preferred and a concept like attitude which he criticized. Other writers have stated similar, but less radical positions. DeFleur and westie have written theoretical articles and completed research on the relationship between attitude and overt behavior.65 They stated that there is the I'fallacy of expected correspondence," i.e., expectation of correspondence between various universes or categories of reSponses toward the attitude stimulus. There are 6“Fora discussion of the dimensions of this problem, see Ernest Nagel, The Structure of Science (New York: Harcourt, Brace & world, Inc., 1961 , pp. 535-5h6 and Abraham Kaplan, The Conduct of In ui (San Francisco, California: Chandler Publishing Company, 19%35. Pp. 80-82. 65Melvin L. DeFleur and Frank R.'Westie, iAttitude As A Scientific Concept,“ Social Forces, 42 (December 1963) 25-28. situatic situatic response The was the prin i.e., ir to anotl'. expected a situat exhibit Althoug} authors 0f Conte Stimuli their pa 56 situational factors such as group norms, roles, definitions of situations and other social constraints which materially mediate responses in situations involving either attitudes or overt action. The reason for the lack of correSpondence is the fact that one of the primary characteristics of modern society is normative conflict, i.e., inconsistent demands on the individual from one group situation to another.66 One type of correspondence that can be legitimately expected is correspondence between the “normative requirements" of a situation, and the kind of behavior a given individual will I exhibit in that situation regardless of past behavior or disposition.67 Although the prescribed norms may definitely influence behavior, the authors overstated their case. It is unrealistic to posit a theory of contemporariby, assuming that the individual can only react to stimuli present in the ongoing activity and ignore the relevance of their past exporiences.’ .Raab and Lipset also considered the social situation as the 66The idea of normative conflict was stated more precisely by Dietrich C. Reitzes. He thought that sociologists could make the best contributions to the field of race relations by a systematic use of two basic concepts--institution and organization. The theoretical roots of the sociological perspective were planted in Wirth's theory on an urbanized way of life, i.e., individual behavior becomes segmented due to the complexity of institutions and organi- zations he encounters. The author stated that in order to understand, predict and control the behavior of individuals the crucial point was not how the person acted before, or in other situations or how he said he would act, but rather how the situation was defined by relevant organizations and his personal relationships to these organi- zations, i.e., which organizations or which definition he accepts or rejects. Idetrich C. Reitzes, IInstitutional Structure and Race Relations," Phylon (Spring 1959) 98—66. primary deter were strongl} mental and me as 'a little waiting for ‘ providing onl inconsistenc; 57 primary determinant of both attitude and overt behavior.68 They were strongly opposed to Allport's definition of attitude as a mental and neural state of readiness. They labeled such conceptions as 'a little mental package tucked away in a corner of the brain, waiting for the proper stimulus to bring it to life."69 After providing one of the most thorough reviews of the literature on inconsistency, they stated:70 Thus, the mass of modern evidence runs counter to the 'attitude—first' fallacy which olds that prejudice is a lurking state of mind that spills over into overt behavior . . . . Actually, there emerges an understanding that the key to prejudice must be found outside the realm of attitude-behavior relationships. The evidence has demonstrated how both attitudes and behavior are affected by the social frame of reference in which they occur . . . . Perhaps then, the most effective and workable approach to understanding the phenomenon of prejudice is through an investi- gation of the kinds of social situations which give rise to and sustain prejudiced behavior and attitudes. Unfortunately the argument that Raab and Lipset presented is dependent on a single definition of attitude on which they placed distorted emphasis. As acquired predispositions to behavior, attitudes need not be defined as ”little mental packages tucked away in a corner of the brain“, or as I'a lurking state of mind that Spills over into overt behavior”. There is nothing sinister or mysterious in the functioning of attitudes. It would be a terrifying exPerience, 68Earl Raab and Seymour Hartin Lipset, ”The Prejudiced Society,“ American Race Relations Today, ed. Earl Raab (Garden City, New Jersey: Ebubleday and Company, Inc., 1962), pp. 29-55. 69Ibid.. p. 53. 7OIbid. , pp. tang. however, to bs’ the capabilitl' associating tE appropriate m: A quot I researcher' s 58 however, to be forced to take part in ongoing activities without the capability of recalling previous orientations to behavior, associating them with present stimuli and responding in the appropriate manner. A quote from Simpson and'Yinger adequately summarizes this researcher's appraisal of those who favor the second position:71 Some students in the field have suggested that individual prejudice has little to do with intergroup relations, that these vary with changes in the social structure, not with changes in individual attitudes. we support this emphasis on sociological factors as a strategic matter, and would only caution against swinging the pendulum too far in that direction. Summary of the Theoretical Significance of the Social Situation One of the striking characteristics of the literature is the widespread emphasis on the necessity of taking into account the social situation in analyzing the relationship between attitude and overt behavior. The question remains, however, as to how helpful is the recommendation. Generally, the variable social situation is couched in vague endorsements. In 1943 Green criticized the conceptualization of social situations. Many of his criticisms hold true today. The concept social situation continues to be a poor tool for analysis. It is seldom defined experimentally and operationally. Used as a blanket concept, it may include social institutions, social relationships, patterns of social interaction, a lifetime of experiences or 71Simpson and linger, op. cit., p. 522. elementary beh experimental To a greater to which he w social situat rather than d determine whj. necessary to eXpectations .10 current ev extent to Whi more Susceptj bAVe parthi; ani have four a lay Of life 01‘ life in W which the out O. 59 momentary behavioral phenomena. Even if precise operational and experimental definitions are provided, a serious problem remains. To a greater or lesser extent, the individual creates the situations to which he will react. He is not a passive agent. Most structured social situations also provide a latitude of behavioral expression rather than determining one single course of action. In order to determine which cluster of expressions will be chosen it becomes necessary to obtain information on past experiences, individual expectations beyond the immediate situation, as well as reaction to current events. Another source of variation in reaponse is the extent to which individuals are maleable. Some people may be much more susceptible to manipulation than others. There are people who have participated in many different segments of segmented culture and have found it difficult to integrate a code of behavior and a way of life from the many conflicting codes of behavior and ways of life in which they have participated. Finally, the entities which the outside observer calls social situations may not exist for 72 the participants. The referents for the observer may, for the participants, merely be momentary contacts on a time—space matrix which have different meanings and are used for different purposes.73 Thus, it is relatively unsafe to assume common affects of social situations on participants. There are many possible sources 72This particular point is supported by Howard and Tracy Kendler and Hyman in Chein et al., op. cit., pp. 27-28 and 38-39. 73Arnold w. Green, "The Social Situation in Personality Theory," American Sociological Review; 9 (1943) 388-393. of variatior as a variab] a wide range consensus tk is helpful i systematical social situa Factors that In th narrowly def characterist 60 of variation. Moreover, the concept itself, is difficult to define as a variable because of its usage as a blanket concept to cover a wide range of social contacts. Although there is a widely based consensus that knowledge of characteristics of social situations is helpful in predicting behavior, few attempts have been made to systematically isolate and precisely measure characteristics in social situations that affect the definition of the situation. Eactors that Affect the Expression of Attitudes In this study the concept definition of the situation was narrowly defined in order to overcome the ambiguity caused by characteristics in the social situation that influence an individual's definition of the situation. The concept is defined as the subjective interpretation of what the appropriate behavior to exhibit Fshould be” in the testing situation. Two characteristics were assumed to influence the definition of appropriate behavior in the testing situation. The first was the normative expectations involved in the role that subjects play as research respondents. The second was the subcultural normative expectations of college students to express liberal racial attitudes. There has been.a renewed interest in the testing situation as a form of social behavior. Lenski and Leggett reported that more is involved in the research interview than gathering data. A rela- tionship between respondent and interviewer develops. The behavior of each is influenced by perceptions of the situation and the appropriate psychology factors th agreement of interac role eXpec A remarka‘ play his Orne conc rather t1: ment is a an exper: 61 7h Orne addressed his concern to the social appropriate social norms. psychology of the psychological eXperiment. He outlined situational factors that are important in testing situations. There is explicit agreement on the part of the subject to participate in a special form of interaction known as ”taking part in the experiment”. The mutual role expectations of the researcher and subject are well understood. A remarkable characteristic is the extent to which the subject will play his role and place himself under the control of the experimentor.75 Orne concluded that the subject must be recognized as an active rather than a passive participant, and that the psychological experi- ment is a special form of interaction and that subjects' behavior in an experiment is a function of the totality of the situation.76 Cicourel and Back §t_§l, have outlined game theory models to account for the behavior of respondents in testing situations. The authors stressed the fact that respondents adopt certain roles and play the game according to the appropriate normative expectations they perceive to be relevant.77 Each of the above authors was concerned with the fact that role playing according to the perceived 7”Gerhard E. Lenski and John C. Leggett, "Caste, Class and Deference in the Research Interview,” American Journal of Sociology, 65 (1960) 463. 7Sl‘i’artin T. Orne, ”On the Social Psychology of the Psychological Experiment,- American Psychologists, 17 (1962) 777. 761bid. , p. 783. 77Aaron V. Cicourel, method and Measurement in Sociolo (New York: The Free Press of Glencoe, 19355, pp. 203-207 and Kurt‘w. Back at al., "The Subject Role in Small Group Experiments,” Social Forces, HE (1965) 181-187. . normative ex: between varia Resea: racial attit: patterns by leave the do! reliable ind. suggest that sophisticati runs counter Altho racial Preju- in the expre I members . Ir United State moms Of the major hypotw 62 normative expectations has an effect on hypothesized relationships between variables. Research on the normative expectation of expressing liberal racial attitudes is limited. In reviewing a study in voting patterns by social class, vander Zanden concluded that the results leave the door open to questions as to whether attitude studies are reliable indicies of racial and segregationist sentiment. The data suggest that better-educated groups because of intellectual sophistication are reluctant to state prejudiced sentiment since it runs counter to the norms of.American democratic beliefs.78 Although Pettigrew found personality factors related to racial prejudice, he also emphasized the important effects of norms in the expression of negative attitudes toward minority group members. In both South Africa and the Southern region of the United States unusually prejudiced attitudes reflect the dominant norms of the white society.79 The author concluded by stating a majorhypothesis:80 In areas with historically embedded traditions of racial intolerance, exter- nalizing personality factors underlying prejudice remain important, but socio- cultural factors are unusually crucial and account for the heightened racial hostility. ~ 78James W; vander Zanden, Race Relations_in Transition (New York: Random House, Inc., 1965), pp. 111+..117. 79Pettigrew, "Personality and Sociocultural Factors in Intergroup.Attitudes: A Cross—National Comparison,“ op. cit., Pp. 35'390 BOIbid. , p. 40. The corollary historically personality 1 sociocultural heightened f2 Linn': that he cons: favorable at‘ of attitude ‘ major eXplan: attending a ' being politi I toward the N aP‘pl‘oval but Education.”-E the college cam-y over i stronger 001: 0f the Colle light of 0th the mslflfts university 8 1n intense; 63 The corollary of Pettigrew‘s hypothesis would be, ”In areas without historically imbedded traditions of racial intolerance, externalizing personality factors underlying prejudice remain important, but sociocultural factors are unusually crucial and account for heightened favorable expressions toward members of minority groups." Linnis study of students at the University of Wisconsin found what he considered to be discrepancies between the expression of favorable attitudes and overt behavior toward Negroes. The expression of attitude was more favorable than the expression of behavior. His major explanation was the cultural milieu of the college students attending a large midwestern university which had a reputation for being politically and racially liberal. He stated, “Liberal attitudes toward the Negro are, in most circles, not only criteria for social approval but a sign of intellectual maturity--a sign of a ’liberal' education."81 However, conforming to the normative expectations in the college community of expressing liberal attitudes does not carry over into patterns of direct interracial contact because of stronger conflicting normative expectations both within and outside of the college community.82 Linn's explanation is reasonable in light of other indirect evidence that is available. In Chapter I the results of Harris' representative sample of American college and university students was reported. The figure on willingness to engage in interracial activities were fantastically high. For'example, 85 81Linn, op. cit., p. 359. 821bid., pp. 360-36u. percent ap: said they ‘ marriage. interraciaf The degree of and attitu commitment setting in C{nestions tOWEI‘d the relatively meallized The mspor re313011883 the neceSE beha‘Tior 4: In the reSea] definite 1 In the sec I“539011893 nomatiVe expectatic commitment \ 8 3A 64 percent approved of belonging to the same social club, 47 percent said they would date a Negro and 36 percent approved of inter; marriage.83 These figures do not align with the observations of interracial behavior on college campuses. The above two characteristics can systematically alter the degree of relationship between each of the antecedent variables and attitude, and the degree of relationship between attitude and commitment, and attitude and overt behavior. The usual research setting in this area generally consists of a researcher asking questions for which the responses imply no future courses of action toward the object of the questions. In this type of situation it is relatively easy to elicit responses that reflect conformity to idealized patterns of behavior suggested by the subcultural norms. The respondent is more likely to give idealistic patterns of reSponses to attitude statements because he is not made aware of the necessity to compromise his ideals with actual plans of behavior toward the object of the attitudes. In contrast to the usual research setting is one in which the researcher asks questions for which the reSponses imply definite future courses of action toward the object of the questions. In the second research setting it becomes difficult to elicit responses that are thought to be appropriate according to the normative eXpectations of the respondent's role and subcultural expectations. When.a respondent is asked questions about his commitment to future behavior he is forced to respond in terms 83Above p. 7. of more real: is a discrep‘ he actually ideals to me In th situation, a appropriate situation in variable tha independent Elationship bGha‘fior. T the situatio I related to t In °Pder to ram“? Pla 65 of more realistic orientations toward future behavior. If there is a discrepanqy between what he perceives he ought to do and what he actually plans to do in the future, he must compromise his ideals to meet the contingencies of the situation. In the research setting subjects orient themselves to the situation, ascertain their interest and proceed to carry out the appropriate patterns of action. In essence the definition of the situation in the usual testing situations is considered as a variable that systematically reduces the consistency among the independent variables and the dependent variable attitude, and the relationships between attitude and commitment and attitude and overt behavior. The fourth independent variable is the definition of the situation. The fourth and fifth propositions which.are related to the fourth independent variable are: The greater the extent to which attitudes are expressed in a research setting that implies future involvement with the object of the attitude, the greater the relation- ship between each of the antecedent variables and attitude . The greater the extent to which attitudes are expressed in a research setting that implies future involvement with the object of the attitude, the greater the relation- ship between attitude and commitment and attitude and overt behavior. In order to test the fourth and fifth propositions students were randomly placed in different experimental groups. The experi- 66 mental conditions consisted of altering the order of appearance of the commitment and attitude scales. When the commitment scale appeared before the attitude scale the expression of attitudes were made in a research setting that implied future involvement with the object of the attitude. The experimental conditions are described in greater detail in Chapter IV. Attitudes Toward Members of a Minority Group Despite the previous proposition that stated the relation- ships between attitude and the other two manifestations of overt behavior are dependent upon the definition of the situation, attitude is considered an important determinant of overt behavior. It is a basic part of the Specific derivation from the general propositions cited in Chapter 11. Factors in the testing situation can reduce but not destroy the relationship between attitude and overt behavior. As an intervening variable attitude is influenced hy antecedent variables and the definition of the situation. The independent variables mold orientations to behavior toward a class of objects. The orientations are manifested as consistent response patterns. In the area of minority group relations the studies of attitude as an independent variable affecting other manifestations of overt behavior.are not voluminous. A large number of studies focus on attitude as a dependent variable, or describe patterns of association with personality and social characteristics without attempting to posit a causal relationship. Two studies that tested the proposition--attitude is a determinant of overt behavior-— reported significant relationships. 67 Wilner,'walkley and Cook designed a comprehensive study of segregated and integrated public housing involving a total of 1,200 interviews. They develOped a composite index to measure initial attitudes toward minority groups. When physical proximity was controlled, this index was positively associated with self-reports of the extent of interracial contact.8u The study of DeFleur and westie has already been described in detail.85 Using chi—square, they found a significant relationship between attitude and expressed commitment to take part in overt behavior.86 A number of studies reported that there was no significant relationship between attitude and overt behavior toward members of a minority group. The authors of these studies offered several reasons for the inconsistency. In the two studies of public accommodations proprietors and managers were interviewed. LaPiere and Kutner EEJEL, reported that conflicting attitudes are present in situations of face-to-face contact.87 Conflict was resolved in favor of good business relations rather than expressing prejudice toward members of a minority group which resulted in inconsistent 8Minniel M. Wilmer, Rosabelle P. Walkley and Stuart W. Cook, "Residential Proximity and Intergroup Relations in Public Housing Projects,” Journal of Social Issues, 8 (1952) 58. 85Above pp. h3.#4. 86DeFleur, ”Verbal.Attitudes and Overt.Acts: An Experiment on the Salience of Attitudes,'o op. cit., p. 671. 87Richard T. LaPiere, littitude vs. Actions,” Social Forces, 13 (193M) 236-237 and Bernard Kutner, Carol'Wilkins and Penny'Yarrow, "Verbal Attitudes and Overt Behavior Involving Racial Prejudice,” Journal of Abnormal and Social ngchology, h? (1952) 652. 68 behavior. In their study of a rural community Brookover and Holland found that adults consistently expressed a set of pervasive and highly unfavorable sentiments toward members of minority groups but the attitudes did not carry over into active forms of hostile behavior. Behavior toward members of minority groups was considered to be a function of particular situations.88 Saenger and Gilbert reported inconsistency between the buying patterns of New York shoppers and their attitudes toward Negro clerks. Conflicting attitudes and the social setting were used to explain the inconsistency. If buying a particular object was important, it outweighed negative attitudes. The social setting was in tolerant department stores where Negro clerks were employed. It was suggested that patterns of behavior were adapted to the situation.89 Killian studied 150 southern white laborers who lived in the Near West Side of Chicago and who had highly unfavorable attitudes toward Negroes who lived in the same area. The laborers did not act out their attitudes for the following reasons: the lack of leadership, organization and consensus in the anomic neighborhood; the absence of legal sanctions for segregation; and the exaggerated conception of Negro strength and unity as contrasted against the 88Wilbur Brookover and John Holland, 'An Inquiry Into the Meaning of Minority Group Attitude Expressions,” Anerican Sociological Renew, 17 (1952) 197-198. . 89Gerhart H. Saenger and Emily Gilbert, l'Customer Reactions to the Integration of Negro Sales Personnel," International Journal of minion and Attitude Research, 1+ (1950) 72-7 . 69 perceived white weakness and disunity.9o' Fishman studied whites in a neighborhood where the ethnic composition was changing. He found that those with the most unfavorable attitudes tended to stay in integrated sections of the neighborhood. The variable used to eXplain the inconsistency was social status. White non-manual workers had greater opportunity to move than manual workers, and they moved despite their favorable attitudes.91 Two studies were designed to test the relationship between attitudes and overt behavior in ''conformity" experiments, i.e., experiments modeled after Sherif's and Asch's ”classic" designs. Both studies predicted different patterns of behavior when the judgments of subjects were confirmed by a confederate who was a member of a minority group rather than a white student. Bray attempted to explain the inconsistency by stating that attitude is never elicited alone. The social situation provides stimuli for the expression of additional personality characteristics and attitudes.92 Malof and Lott explained the inconsistent behavior by criticizing the Adornb g§_§l, theoretical framework from which their prediction was derived. They preferred Rokeach's approach to interracial behavior in which belief congruence is regarded as a more important determinant than 90Lewis M. Killian, “The Adjustment of Southern White Migrants to Northern Urban Norms,I Social Forces, 32 (1953) 69. 91Joshua A. Fishman, ”Some Social and Psychological Determin- ants of Intergroup Relations in Changing Neighborhoods: An Intro- duction to the Bridgeview Study,‘ Social Forces, 40 (1961) #5. 92Douglas W} Bray, "The Prediction of Behavior from Two Attitude Scales,” Journal of.Abnormaljand Social Psychology, 45 (1950) 81.82. 7O racial or ethnic characteristics in the acceptance of others.93 The authors also mentioned that situational characteristics are important in determining how attitude will be expressed in overt behavior}!+ Mann found significant relationships among the cognitive, affective and behavior components of racial prejudice for Negro and white graduate students. The relationships were, however, insignificant for the whites when they were analyzed separately. The author did not attempt to interpret the findings.95 Linn's study has already been mentioned.96 Students had more favorable attitude than overt behavioral expressions. This finding was explained by pointing out that students had normative support for expressing favorable attitudes, but lacked support for becoming involved in overt behavior with Negroes.97 In one of the more recent and original articles on social attitudes, Campbell has criticized the frequent interpretations of inconsistency between attitude and other manifestations of behavior. From previous research in psychology on the-phenomena of I'cheating" 93Milton Rokeach, The Open and Closed Mind (New York: Basic Books, Inc., 1960), p. 153. 9“Milton Malof and Albert Lott, ”Ethnocentrism and the Acceptance of Negro Support in a Group Situation," Journal of Abnormal and Social szchology, 65 (1962) 257-258. 95John H. Mann, ”The Relationship Between Cognitive Affective and Behavior.A3pects of Racial Prejudice,“ Journal of Social Psychology, 49 (1959) 226-227. 96Above pp. Mans. 97Linn, op. cit., pp. 358—36“. 71 he developed the idea of attitude expression within a hierarchy of situational threshold differences. The studies on cheating reported that the extent of cheating was strongly related to the opportunity to cheat as well as the proclivity. Campbell felt that the opportunity to express attitudes in overt behavior was also related to the opportunity structure.98 He stated that the litera- ture on inconsistency between attitude and overt behavior often confuses correlational inconsistenqy with situational threshold differences, and thus exaggerates the degree of inconsistency. For example, the inconsistency that LaPiere reported would have been demonstrated if those who refused face-to-face (.h percent) accepted by questionnaire, or if those who accepted by questionnaire (7.5 percent) refused face.to-face. No report of such cases occurred. In other words the questionnaire provided a low threshold for the expression of negative attitudes. However, the face-to-face contact provided a higher threshold for the expression of negative behavior. This second condition inhibited the expression of negative attitudes. Campbell also criticized Minard's study of white miners' behavior toward Negro miners. He stated that Minard reported no instances of true inconsistency, i.e., being friendly in town and hostile in the nines.99 In fact, Campbell concluded that the two items, mine and town correlated perfectly, using any index not biased by uneven item marginals. In short what Campbell said is that researchers 98Campbell, op. cit., pp. 157.159. 99Fora closer inSpection of Minard's research see R.D. Minard, op. cit., pp. 29-44. 72 should not expect to find a one-to-one relationship between attitude and overt behavior because situational thresholds can vary.100 Using Campbell's outline as a model there would be incon- sisteney between attitude and overt behavior in two types of socio- cultural environments. The first would be in an environment where there was a low threshold for favorable attitude expression and a high threshold for favorable behavioral expression when subjects‘ exhibited favorable behavior and unfavorable attitudes. For example, white college students would be inconsistent if they expressed unfavorable attitudes and favorable overt behavior toward Negro students. The second type of environment would be one in which there was a low threshold for unfavorable attitude expression and a high threshold for unfavorable behavioral expression when subjects exhibited unfavorable behavior and favorable attitudes. ' For example, white southerners would be inconsistent if they expressed favorable attitudes and unfavorable overt behavior toward Negroes. When Campbell's criteria are used to evaluate the reported inconsisteney in the above articles, the interpretation of the results changes. No evidence of inconsisteney was reported in the LaPiere, Kntner §t_§l,, Brookover and Holland, Killian, Saenger and Gilbert and Malof and Lott articles. Linn found two inconsistent students out of 31+ and DeFleur and Westie found 5 students out of 1&6. me to the research design and the data presented the inconsistency could not be determined in the Fishman article. Bray reported a _ 1000ampbell, op. cit., pp. 160.161. 73 negative correlation of -.149 between attitudes toward Negroes and the number of times the white students agreed with a Negro confed- erate in the conformity experiment. Mann reported a negative correlation of -.54 between the affective component of attitude and a measure of overt prejudice for white students. The number of people who were unfavorable in their attitudes and favorable in their behavior could not be determined. Part of the confusion on the nature of the relationship between attitude and other manifestations of overt behavior stemms from the confusing interpretations. Many authors disproved that attitude was the necessary and sufficient cause of overt behavior. However, in the field of social peychology it is relatively easy to disprove that a single determinant is the necessary and sufficient condition of a resultant. Moreover, it was often assumed that the relationship could be tested by assigning equal or absolute weights to units of attitude and overt behavior. The existence and direction of threshold differences were ignored to disprove the hypothesis. Due to Campbell's insights and the re-examination of the literature, attitude was considered to be an intervening variable that influenced other manifestations of overt behavior. The sixth proposition is: The greater the degree of favorable attitude toward members of a minority group, the greater the favorable commitment and overt behavior toward members of the minority group. 7h Favorable attitudes were considered as a contingently necessary condition for other manifestations of overt behavior, i.e., it is probably true that whenever there is an expression of favorable overt behavior, there is always favorable attitude, however, favorable attitudes do not always indicate favorable overt behavior. The differences in the low threshold for favorable attitude and the high threshold for favorable expressions of overt behavior was also considered. Scores on the measures of attitude, commitment and overt behavior were not given equal or absolute values. The degree of relationship was tested by comparing the relative intensity of the degree of favorability on the measures of attitude, commitment and overt behavior. Commitment to Become Involved in Possible Activities With Members of.A Minority Group ' Due to the anticipated differences between the expression of attitude and overt behavior, an additional variable was included. The commitment threshold was thought to be closer to the overt behavior threshold than the attitude threshold. Committing oneself to involve- ment in possible interracial behavior is similar to definite acts oriented directly toward members of a minority group (overt behavior). The decisions imply definite consequences for the subject. Responses to attitude statements are made in situations that are relatively free of definite consequences. 75 The literature on the variable commitment is sparse. Definitions stress different points. Etzioni considered commitment as positive involvement. It refers to the positive cathetic- evaluative orientation of an actor to an object, characterized in terms of intensity and direction. Under the heading of moral involve- ment he defines two types of commitment. They differ in their foci of orientation and in the structural conditions under which they develop. Pure moral commitment is based on internalization of norms and identification with authority. Social commitment is dependent upon sensitivity to pressures of primary groups and their members. The author's discussion of commitment focuses on orienta- tions of members toward an organization.101 Goffman in his well developed style presents a structural definitionzloz A concept that is often employed in the discussion of roles is that of commitment. I propose to restrict this term to questions of impersonally enforced structural arrange- ments. An individual becomes committed to something when, because of the fixed and interdependent character of many institu- tional arrangements, his doing or being this something irrevocably conditions other important possibilities in his life, forcing him to take courses of action, causing other persons to build up their activity on the basis of his continuing in his current under- takings, and rendering him vulnerable to unanticipated consequences of these under- takings. He thus becomes locked into a lolAmitai Etzioni, gigomparative Analysis of Complex Organiza- tions (New York: The Free Press of Glencoe, Inc., 1 1 , pp. 8-11. 102Erving Goffman, Encounters (Indianapolis, Indiana: Bobbs- Merrill Company, Inc., 1961), pp. 88-89. 76 position and coerced into living up to the promises and sacrifices built into it. Typically, a person will become deeply committed only to a role he regu- larly performs, and it is left to gallants, one-shot gamblers, and the foolhardy to become committed to a role they do not perform regularly. Kurt Lewin's theory and teaching Spawned many ideas related to commitment. In his work in I'group dynamics", field theory was used to explain changing patterns of behavior. Participation in groups was considered a powerful force on individual behavior. He characterized the behavior of individuals in groups as being in a state of quasi—stationary social equilibria. The degree of stability or change was dependent upon the forces directed at the group and behavior within groups. Change occurred when decisions were made to alter behavior. Research on leadership training, changing food habits, work production, criminality, alcoholism and prejudice indicated that it was usually easier to change individuals formed into a group than to change any one of them separately.103 Lewin's conceptualization of commitment is similar to what Etzioni defined as social commitment. He was also cognizant of what Goffman calls structural pressures. As a student of Lewin, Pelz defined commitment as the degree to which a decision is indicated publicly. She designed a study to measure the effects of discussion, decision, commitment and 1oBKIurt Lewin, I'Frontiers in Group Dynamics,” 22:_E¥E°' pp. 227“2350 77 consensus in group decisions. The study involved volunteering and taking part in behavioral-science experiments. Testing hypotheses she found that: (1) group discussion was not a more effective induce- ment to action than the lecture method or no persuasion attempt; (2) a more public commitment or indication of a decision was not more effective in influencing overt behavior than a less public one; and (3) the process of making a decision and the degree to which group consensus is obtained and perceived were alone capable of generating differences as large as those reported in the classic experiments of Lewin's co-workers.1ou Brehm and Cohen attempted to extend Festinger's theory of cognitive dissonance by adding the variable commitment. They were interested in attitude-discrepant behavior. They defined commitment as the degree of voluntary choice to engage in a pattern of behavior. In their experimental studies the authors found a greater degree of attitude-discrepant behavior under experimental conditions of volun- tary choice. Investigation of the role of commitment in the theory of cognitive dissonance is considered to be an important contribution. It aids in the specification of psychological implication, and thus, the determination of what is consonant and what is disonant. It also aids in the specification of the ways in which a person may try to reduce dissonance.lo5 83 10“Edith Bennett Pelz, ”Some Factors in Group Decision,” in Readings in Social Psychology,.op. cit., pp. 212-219- 105 Jack W. Brehm and Arthur Cohen, lorations in Co nitive Dissonance (New York: John Paley and Sons, Inc., 15325, pp. 8-9. 195 and 217. I 5‘." 78 Becker has developed a definition of commitment than can be applied to a wide range of theoretical considerations. Commit- ment is defined as being engaged in consistent lines of activity. The activities persist over time. They are seen by the actor as directed toward the same goal. Other feasible alternatives are rejected.106 In elaborating on his definition Becker stated that there were three major elements of commitment:107 l. The individual is in a position in which his decision with regard to some particular line of action has consequences for other interests and activities not necessarily related to it. 2. The individual places himself in that position through his own prior actions. 3. The committed person must be aware that he has made a side bet acted in such a way to implicate other interests and must recognize that his decision in this case will have ramifications beyond the side bet. In this study a modification of Etzioni's definition will be used. He defined it as positive involvement, i.e., cathetic- evaluative orientation of an actor to an object characterized in terms of intensity and direction. This particular definition is similar to definitions of attitude. What distinguishes commitment is the decision to engage in a particular line of action. In.making a choice the individual selects one pattern of activity with.its corresponding consequences. The particular*alternative chosen is considered to have either the most favorable or the least unfavorable 106Howard 8. Becker, "Notes on the Concept of Commitment,” American Journal of Sociology, 66 (1960) 33. 1071mm. pp. 35-36. 79 consequences. Therefore, commitment is defined as making a decision that will determine a particular pattern of action that will have consequences for the individual. Commitment is neither a completely voluntary or determined act. The individual does choose among alternatives in light of the consequences he perceives. However, the alternatives and consequences are influenced by the social milieu in which the commitment is expressed. A comparison may help to illustrate the meaning of commitment. One instrument that has been frequently used to measure attitudes toward members of minority groups is the social distance scale. Essentially, the social distance scale presents to subjects a number of stimulus statements asking them to indicate the degree to which they would interact with members of minority groups, e.g., from close kinship by marriage to excluding from country. The state- ments imply hypothetical involvement with members of minorities. However, a social distance scale does not elicit responses that measure actual commitment to become involved with members of a minority. No definite decisions are made with respect to future behavior. Commitment is considered as an intervening variable. The antecedent variables influence the choice of alternatives. In turn commitment influences patterns of behavior that are directed toward the object of the commitment. The relationship between commitment and the antecedent variables and commitment and overt behavior is not the same as the relationship of attitude to the antecedent variables 80 and overt behavior. Attitude is strongly influenced by the defini- tion of the situation, i.e., attitude reSponses are influenced by the normative expectations of research respondents and the subcultural normative expectations of what is the proper responses to give to statements about a particular class of objects. These two factors reduce the extent of relationship between attitude and the other variables. Commitment reaponses are not as strongly influenced by the characteristics of the testing situation. Decisions to commit oneself to patterns of action outside the testing situation are made at higher threshold level that is closely related to overt behavior. Therefore, the seventh.and eighth propositions are: The extent of the relationship between commitment and the antecedent variables .will be greater than the extent of the relationship between attitude and the antecedent variables. The extent of the relationship between commitment and overt behavior will be greater than the extent of the relation- ship between attitude and overt behavior. In order to account for the variable commitment the deriva- tion of the general proposition should be modified as follows: B = Com (antecedent variables); Att (definition of the situation and antecedent variables) where B = observable behavior directed toward a class of objects Com = commitment toward the class of objects Att = attitude toward the class of objects. 81 Overt Behavior Directed Toward Members of a Minority Group The major dependent variable is overt behavior. It is defined as acts directed toward the object of attitude and commitment. Attitude and commitment responses can also be considered overt behavior, i.e., observable patterns of action. In this context, however, overt behavior is given a more restricted classification. Overt behavior reaponses involve interaction with the object of the attitude. Overt behavior is the major form of behavior to be explained and predicted. The derivation from the general proposition in Chapter II suggested a possible way of developing an explanation of patterns of overt behavior toward members of a minority group. The behavior is analyzable in terms of a series of probabilistic explana- tions in which.a series of variables are related to the dependent variable. The variables influencing overt behavior are considered to be some of the contingently necessary conditions rather than sufficient conditions. The relationships between the antecedent and intervening variables are analyzable in terms of probabilistic explanations of their interrelationships. Theoretical gypotheses Derived from the Propositions From the eight propositions a number of theoretical hypotheses can be inferred. The first three propositions stated that each antecedent variable would be related to manifestations of behavior. Three hypotheses were derived directly from.each proposition. Proposition I stated: 82 Those people who have a high level of authoritarianism as measured by the F- scale will express less favorable attitude, commitment and overt action toward members of minority groups. The hypotheses are: l. The higher the degree of authoritarianism, the less favorable the attitude toward members of a minority group. 2. The higher the degree of authoritarianism, the less favorable the commitment toward members of a minority group. 3. The higher the degree of authoritarianism, the less favorable the overt behavior with members of a minority group. Proposition II stated: If the costs of avoiding interaction are low and if there are available alternative sources of reward, the more varied the intergroup contact, the greater the degree of favorable attitude, commitment and overt behavior toward members of a minority g mup.108 The hypotheses are: k. The greater the extent of past intergroup contact, the more favorable the attitude toward members of a minority group. 108The qualifications of Proposition II will not be mentioned in the theoretical hypotheses. The sample of college students controls for the qualifications. In the college community the costs of avoiding interaction are low and there are available alternative sources of reward. 83 5. The greater the extent of past intergroup contact, the more favorable the commitment toward members of a minority group. 6. The greater the extent of past intergroup contact, the more favorable the overt behavior with members of a minority group. Proposition III stated: The greater the extent to which significant others are perceived to have favorable attitudes and actions toward members of minority groups, the more favorable the attitude, commitment and overt behavior toward members of a minority group. The hypotheses are: 7. The greater the degree of perceived support from significant others, the more favorable the attitude toward members of a minority group. 8. The greater the degree of perceived support from significant others, the more favorable the commitment toward members of a minority group. 9. The greater the degree of perceived support from significant others, the more favorable the overt behavior with.members of a minority group. Each of the antecedent variables are important characteristics of the individual. Although each of the variables can be analyti- cally distinguished, each person has some measure of the three characteristics. Since each of the antecedents are related to the same manifestations of behavior and since each person retains some value of each of the antecedents, it follows that in a college community the three antecedents should be related to each other. Therefore, it was hypothesized: 10. ll. 12. 84 The higher the degree of authoritarianism, the less the degree of past intergroup contact. The higher the degree of authoritarianism, the less the degree of perceived support from significant others. The greater the extent of past intergroup contact, the greater the degree of perceived support from significant others. Since each of the antecedent variables are related to each other, the combined influence of each of the three antecedent variables should be related to the three manifestations of behavior. 13. 14. 15. Therefore, it was hypothesized: The lower the degree of authoritarianism, the greater the extent of past intergroup contact, and the greater the degree of perceived support from significant others, the more favorable the attitude toward members of a minority group. The lower the degree of authoritarianism, the greater the extent of past intergroup contact, and the greater the degree of perceived support from significant others, the more favorable the commitment toward members of a minority group. The lower the degree of authoritarianism, the greater the extent of past intergroup contact, and the greater the degree of perceived support from significant others, the more favorable the overt behavior with members of a minority group. Propositions IV and V stated that the research setting in which the attitudes were expressed influenced the relationship between attitudes and the other variables. The research setting‘ influenced the definition of the situation. Five hypotheses were derived directly from the two propositions. Proposition IV stated: 85 The extent to which attitudes are expressed in a research setting that implies future involvement with the object of the attitude, the greater the relationship between each of the antecedent variables and attitudes. The hypotheses are: 16. 17. 18. The extent to which attitudes are expressed in a research setting that implies future involvement with the object of the attitude, the greater the inverse relationship between the degree of favorable attitude toward members of a minority group and the degree of authoritarianism. The extent to which attitudes are expressed in a research setting that implies future involvement with the object of the attitude, the greater the relationship between the degree of favorable attitude toward members of a minority group and the extent of past intergroup contact. The extent to which attitudes are expressed in.a research setting that implies future involvement with the object of the attitude, the greater the relationship between the degree of favorable attitude toward members of a minority group and the degree of perceived support from significant others. Proposition V stated: The extent to which attitudes are expressed in a research setting that implies future involve- ment with the object of the attitude, the greater the relationship between attitude and commitment and attitude and overt behavior. The hypotheses are: 19. The extent to which attitudes are expressed in a research setting that implies future involvement with the object of the attitude, the greater the relationships between the degree of favorable attitude and commitment toward members of a minority group. 20. 86 The extent to which attitudes are expressed in a research setting that implies future involvement with the object of the attitude, the greater the relationship between the degree of favorable attitude and overt behavior toward members of a minority group. Despite some evidence to the contrary, Preposition VI stated that attitude was an independent variable that influenced other manifestations of behavior. Proposition VI stated: The greater the degree of favorable attitude toward members of a minority group, the greater the favorable commitment and overt behavior toward members of the minority group. The hypotheses that were derived from the proposition are: 21. 22. The greater the degree of favorable attitude toward members of a minority group, the greater the degree of favorable commitment toward members of a minority group. The greater the degree of favorable attitude toward members of a minority group, the greater the degree of favorable overt behavior toward members of a minority group. The last two propositions were statements about the relation- ship between commitment and the other variables in the study. Four hypotheses were derived from the two propositions. Proposition VII stated: The extent of the relationship between commitment and the antecedent variables will be greater than the extent of the relationship between attitude and the antecedent variables. The hypotheses arer 87 23. The inverse relationship between commitment and the authoritarianism will be greater than the inverse relationship between attitude toward members of a minority group and authoritarianism. 24. The relationship between commitment and past inter- group contact with members of a minority group will be greater than the relationship between attitude toward members of a minority group and past intergroup contact. 25. The relationship between commitment and perceived support from significant others will be greater than the relationship between attitude toward members of a minority group and perceived support from significant others. Proposition VIII stated: The extent of the relationship between commitment and overt behavior will be greater than the extent of the relation- ship between attitude and overt behavior. The hypothesis is: 26. The relationship between commitment and overt behavior will be greater than the relationship between attitude toward members of a minority group and overt behavior toward members of a minority group. ' - Summary In this chapter the rationale and evidence for developing specific propositions and hypotheses were presented. Each of the variables in the propositions was defined and the literature relevant to each of the variables was reviewed. Eight prOpositions and twenty—six hypotheses were developed. The development of the specific attributes of the general proposition was designed to establish clarification of the 88 relationShip between attitude and overt behavior toward members of a minority group. Clarification was attempted through three refinements of previous research. The first was the multivariate approach to the problem. Four independent variables were related to three manifestations of overt behavior. Attitude and commitment were considered intervening variables. Analytically, attitude and commitment intervene between both the antecedent and concurrent conditions and overt behavior. They intervene in the sense that they can be considered as both independent and dependent variables depending upon the relationships that are examined. The second refinement was the attempt to determine the effects of the testing situation on the expression of attitude. This second refinement involved the manipulation of the order of appearance of questions designed to measure commitment. It was assumed that subjects would define the situation somewhat differently if they were asked to commit themselves before being asked their attitudes than viceaversa. In one type of experimental situation students were asked to respond to attitude statements with no awareness that they would be aSked to what extent they were willing to interact with Negro students. In the other type of experimental condition students were asked to respond to attitude statements after they had been asked to what extent they were willing to interact with Negro students. In the first type of experimental condition the usual definition of the situation in a testing environment was assumed to influence the 89 attitude statements, and thereby, alter the relationship between attitude and the antecedent variables and between attitude and the other manifestations of overt behavior. The experimental conditions will be described in detail in the following chapter. The third refinement was to develop the variable of social psychological commitment and compare the relative strength of commitment and attitude as explanatory variables. CHAPTER IV RESEARCH DESIGN AND PROCEDURES Introduction In order to test the propositions and hypotheses a verifi- cation study was designed. This chapter describes the design and procedures that were developed. Before gathering the data for the major test of the theory an extensive pilot study was completed.1 Decisions on particular procedures to use in the major study were dependent on the results of the pilot study. Information on the pilot study will be presented jointly with information on the major study. This information will be used as evidence to justify the decisions that were made in gathering data for the major study. Experimental Design Both the pilot and major studies employed experimental designs. In the pilot study there were two treatment groups. In both groups students were asked to respond to the authoritarianism (Modified F- scale), past intergroup contact (Past Exp-scale) and perceived 1The opportunity for doing the pilot study was the result of a suggestion by.A. O. Heller. During spring term.l964 he tahght a graduate course on the theory and measurement of social attitudes. He thought one of the best pedagogical devices was to have students become involved in.a research project. He suggested that the pilot study he used as a class project. Thirteen graduate students, who were taldng the course, were organized into a research team that gathered and helped to evaluate the data. 90 91 support from.eignificant other (PSSO-scale) scales. Each treat- ment group was differentiated by the order of the presentation of the commitment scale (Com-scale). In experimental condition I the Com-scale preceded the attitude scale (Att-scale). In experi- mental condition II the Att-scale preceded the Com-scale. The Com-scale implied future interaction with Negroes. The order of presentation was used as an experimental stimulus to systematically influence the manner in which attitudes toward Negroes would be expressed. The responses to the two scales in the different order of presentation was designed to alter the student's definition of the situation (D of S). In the major study it was decided that the effects of the experimental design could be analyzed.more thoroughly if four treatment groups were used instead of two. The order of presenta- tion of the antecedent variables was the same as the pilot study. Experimental variation was limited to the ordering of the Com—.and Att-scales. The four treatment groups were: 1. The Att-scale before the Com-scale 2. The Com-scale before the Att-scale 3. The Att-scale without the Com-scale #. The Com-scale without the Att-scale Two additional treatment groups were added to the major'study to examine the relationships between attitude, commitment and overt behavior (Overt-scale) in more detail. Four treatment groups provided information to evaluate the interaction between commit- ment and attitude. when students were asked to respond to both scales they were expected to be consistent in their response 92 patterns, i.e., responses to the first of the two scales would . influence the reSponses to the second scale. The two additional treatment groups provided information to evaluate the relationships between the variable that was second (when both attitude and commit- ment were measured) and the other variables, and they provided information on the relationships between the intervening variables and overt behavior when each variable was measured by itself. The different treatment groups do not systematically affect all the variables. The first three antecedent variables preceded the experimental stimuli and therefore, their interrelationships were not influenced. In treatments (2) and (4) the relationships between commitment and the antecedent variables were expected to be the same. In treatments (1) and (3) the relationships between attitude and the antecedent variables were expected to be the same. The relationship between attitude and overt behavior in treatments (1) and (3) and the relationship between commitment and overt behavior in treatments (2) and (4) were not expected to be signifi- cantly different. The only relationships that were expected to change were the relationships between the variable that was second (when both attitude and commitment were measured) and its relation- ship to the other variables. In other words, the commitment scale in treatment (1) was presented in a different experimental condition than the commitment scale in treatments (2) and (4). The attitude scale in treatment (2) was presented in a different experimental condition than the attitude scale in treatments (1) and (3). 93 Sample The pilot study data was gathered during the spring term of 1964c A small systematic sample of Michigan State University undergraduates was obtained from the university's Data Processing Center. The university was not considered solely as an institution of higher learning. It was analyzed as a community, i.e., a functioning social system. Within this community students live and interact with other people directly and indirectly involved with the academic institution. The criteria for sampling the students were selected in order to sample those most likely to be participants in the university community. In the pilot study 46 students selected were full-time students who lived in the Lansing—East Lansing metropolitan area. Freshmen were excluded because they were relatively new arrivals on campus, not as familiar with the prevalent attitudes and sanctions governing the patterns of social relationships. Foreign students were excluded because their familiarity with interracial activities in this country was considered to be either limited or viewed from a distinctive perspective. Part-time students were excluded because of their often minimal contact with students outside of the classroom. This last criterion of exclusion also applied to students who lived out- side of the Lansing-East Lansing metropolitan area. The data for the major study were gathered during winter term 1965. In the pilot study a sample was selected from the total ‘university population and the students were interviewed in their 94 places of residence. In the major study four undergraduate courses in sociology, ”Introduction to Social Psychology", were used. Originally it was planned to draw a random sample from the university population. The major reason for the change in procedure was the tentative nature of research funds. The awarding of a federal grant from the National Institute of Mental Health was uncertain. Moreover, the university policy on the usage of funds by a pre- doctoral fellow was not defined in the same way as the official manual from the.Pub1ic Health Service. It was extremely doubtful that monies would be made available to hire and train.interviewers. Since the captive audience of undergraduates in the classroom setting did not impose the problem of hiring and training interb viewers, the major data were gathered in the classroom setting. The sample was composed of students who were enrolled in one of the four classes of “Introductory Social Psychology". Plans were made to sample only three out of the four classes. The data from the three classes were gathered on January 22, 1965. This was followed by six.small group discussions held on January 24, 25.and 26. {A date early in the term was selected in order that attendance at the small group discussions would not be impeded by students having to study for exams. The number of students who completed the questionnaire was 224. The expected.number was 300. In order to guarantee adequate representation in each of the four treatment groups it was decided to obtain information from the fourth class. The questionnaire was administered to the fourth class two weeks after the original administration. The total number 95 was increased to 276. D16 to the late date in the term and the difficulties inherent in organizing a series of small group discussions, the fourth class was not asked to participate in the small group discussions. Thus, data on the Overt-scale is available only for three classes. There are three reasons why the introductory courses in social psycholog were selected. First, in the introductory course there was a greater likelihood of obtaining a better cross-section of college sophomores, juniors and seniors than in advanced specialized courses. The second reason was that freshmen were not admitted into the courses. The research was designed to study students already socialized in the college community. Freshmen had only been on campus four months. In that short period of time they might not have been able to develop attitudes and patterns of interaction toward Negroes in the college community. The third reason was to involve students in a social psychological experi- ment that would be relevant to the course content the students were studying. In all four classes, the topic I'Theory and Measurement of Social Attitudes” was part of the course outline. In all four classes the tOpic of social attitudes had not been presented at the time the questionnaire was distributed. After the students had the opportunity to take part in all phases of the experiment, the researcher was invited to the four classes to explain the study. Both the instructors of the classes and the researcher preferred giving the students a thorough briefing of the research objectives. It was felt that students are often asked to participate in experi- 96 ments, but are seldom given the opportunity to become familiar with the objectives of the research. In contrast to the usual procedures this research project served as a learning device for the students. Out of the total of 276 students, 13 were.excluded from the analysis. Three did not complete one or more of the scales. Ten students did not meet the criteria of being U.S. citizens, residents of the Lansing-East Lansing metropolitan area and full-time students carrying at least 12 credits. Thirtyetwo students who completed the questionnaire but refused to identify themselves were retained in the analysis. Randomization Procedures The questionnaires for treatments (1) and (2) were distributed at a ratio of 5:2 with the questionnaires for treatments (3) and (4). The first two treatment groups contained measures of both attitude and commitment. Treatment groups (1) and (2) were of major importance for testing propositions IV, V, VII and VIII. In order to guarantee an adequate distribution of scores a larger proportion of students were given questionnaires for treatments (1) and (2). The four treatment groups had 92, 93, 39 and 39 students respectively. The questionnaires were randomly distributed in each of the four classes. Each student had the same probability of falling into any one of the four treatment groups as any other student. In order to verify the procedures of randomly distributing students 97 to treatment groups, the KrushaleWallis OneeWay Analysis of Variance test was applied to the rank order distribution of a series of variables.2 It was hypothesized that there would be no significant differences on the rank order of the variables for the treatment groups, except for the expression of attitude. The .05 level of significance was used to determine whether students in the four treatment groups were significantly different. A number of control variables were used to test the randomization of students into treatment groups. The variables were: sex, age, region of country where the home town was located, present residence, number of credits taken winter term, grade point average, occupa- tion of the father or major wage earner in family and religious affiliation.3 For one of the above variables there was a significant difference among treatment groups (H value = 9.830). There was a significantly wider geographical distribution of the students' home towns in Treatment (4). The geographical distribution of the students was computed with the scores of each of the six major scales. The variables were: authoritarianism, past intergroup contact, perceived support from significant others, commitment, attitude and overt behavior. It was not found to be significantly related 2For a description of the one-way analysis of variance test, see Sidney Siegel, Non rametric Statistics (New”Tork: MbGraweHill Book Company, Inc., 19 , pp. 1 -193. 3This information was gathered on the last page of the questionnaire. For closer inspection see Appendix A. 98 to any of the six variables. Thus, it was concluded that the difference in geographical distribution in treatment (4) did not significantly affect the six major variables. Each of the six major variables was correlated with the treatment groups. The KruskalAWallis test was used to test for significant differences in the rank ordering of each variable by treatment groups. One of the six variables was found to be significantly related to the treatment groups (H value = 9.057). The attitude scores were significantly lower for treatment (2) (i = 178.8) than for treatments (1) and (3) (i's = 185.21and 189.9). This difference, however, was predicted. The D of S was expected to be different in treatment (2). The Att-scale followed the Com—scale. The degree of favorability of the Att- scale was expected to be lower in the research setting that implied future involvement with the object of the attitude. Thus, the analysis of both background variables and the major variables revealed randomization procedures were effective. Operational Definitions, Reliability and validity of the Scales A sample questionnaire for treatment (1) has been provided in.Appendix.A. Each scale is labeled. The only difference between the questionnaires for the four’treatment groups is the order of presentation of the attitude and commitment scales. In order to determine the reliability of the scales the technique of split- half reliability was used. Using a table of random numbers the items were subdivided into two groups. The total scores of the two 99 sets of items were correlated.by the product moment correlation coefficient. Using the Spearman-Brown propheoy formula, the reliability of the total scales was estimatedfp The method selected to estimate the validity of the scales was construct validity.5 The internal consistency and the predictive ability of the scales were examined. The method used to determine the internal consistenoy of the scales was the correlation of item scores with the total score. The relationship between the item and total score is an indicator of the discriminating power of the item, i.e., the item can be used to distinguish between subjects who have different attitudes. When items are significantly related to total scores, they are considered to be valid items. Product-moment and point bi- serial correlation coefficients were computed for the relationship between item and total scores.6 The predictive validity of the scales was determined by the extent to which the predictions stemming from the propositions were accurate. If the test of relationships between variables were relatively strong and in the predicted direction the scales were considered to be valid indicators of the theoretical constructs. “J. D. Guilford, P chometric Methods (New York: McGraw-Hill Book Company, Inc., 1936 , pp. 18 21. 5Lee J. Cronbach and Paul E. Meehl, ”Construct Validity in JPeychological Tests,‘ ngchological Bulletin, 52 (1955) 281-302. 6Henry E. Garrett, Statistics in P cholo and Education (How York: Longmans, Green.and Company, 1959), pp. 134—150 and 380- 384. 100 The California F-scale One of the most extensive research projects in the field of social psychology was The Authoritarian Personalit .7 Adorno §t_§l, developed the California F-scale to measure the personality syndrome of Iauthoritarianism“. Rather than being part of the authors' theoretical orientation at the beginning of their research, authori- tarianism slowly emerged as a result of empirical observation and statistical analysis.8 The F-scale had two purposes. The first was to construct a disguised instrument to measure prejudice without mentioning the names of any specific minority group. The second was to develop a measure of anti—democratic trends of the personality (a fascistic outlook on life).9 The scale was designed to tap nine characteristics believed to be associated with anti- democratic beliefs and prejudice. They were: conventionalism; authoritarian submission; authoritarian aggression; anti-intraception; superstition and stereotyping; preoccupation with power and ”tough- ness'; destructiveness and cynicism; projectivity; and sex (an exaggerated concern with sexual ”goings-on“). A series of items 10 were developed to measure each of the above characteristics. The final form of the scale (forms 40 and 45) consisted of 29 items. 7Adorno et a1. op. cit., pp. 222-224. 8John Madge, The Origins of Scientific Sociology, op. cit., Po 379. gAdorno et al., op. cit., pp. 222-224. loIbid. . pp. 224-2Lu. 101 The authors felt that the scale did measure the authoritarian syndrome. The items were interrelated in the sense that each item was significantly correlated with the scale as a whole.11 In this study one major modification was made to overcome one of the criticisms of the F-scale. In the scale all items are phrased so that an agreeing response is taken to indicate authori- tarianism. Scales designed in this manner have been criticized on the grounds that the theoretical variable cannot be separated from the response set to acquiesce, i.e., the tendenoy to agree with any opinionated or doctrinaire statement about human affairs. A large body of literature has developed around the concept of response set. One point of agreement is that response sets are more likely to 12 The F-scale has occur when items are difficult or ambiguous. been regarded as an ambiguous scale.13 In order to measure the characteristics of authoritarianism without the contamination of the response set to acquiesce a modified form of the 29 item California F-scale was used. The major modification was to provide 15 items that were opposite in meaning to the original statements of the F-scale. Fourteen items remained unchanged. Bass developed what he called llIbid. , pp. 261-26 2. lelLlen E. Edwards, "Social Desirability or Acquiescense in the MMPI? A Case Study with the S.D. Scale,” Journal of Abnormal and Social chholog, 63 (1961) 351-359. Lee-J. Cronbach, “Further Evidence of Response Sets and Test Design, " Education and Psycholo- ggcal Measurement, 10 (1950) 3-31. 13Chapman and Campbell, op. cit., p. 158. 102 the G-scale. Each statement of the scale was opposite in meaning to the original statement. For'example, two original items on the F-scale were, ”People can be divided into two distinct classes, the weak and the strong,‘ and "Homosexuals are hardly better than criminals and ought to be severely punished”. Bass changed these items to read, IPeople cannot be divided into two distinct classes, the weak and the strong,” and ”Homosexuals are not criminals and should not be punished". The author tested the validity of his G-scale on.a sample of students. He found that all items were not ranked equally. A four point scale was used to measure the degree of 'oppositeness". .An item was completely opposite in meaning if it had a mean rating of 4.00. Bass found that items had mean scores ranging from 3.99 to 1.86.1“ The 15 items ranked most opposite in meaning in Bass' G-scale (all had mean ranking above 3.67) were combined with the 14 items of the F-scale. This modifi— cation is defined as the modified F-scale, corrected to reduce the response set to acquiesce. A 6-point forced choice set of response categories was used for each item. The responses were: strongly agree (+3); agree (+2); agree in part (+1); disagree in part (-1); disagree (-2); and strongly disagree (-3). Omissions or indicators of uncertainty were scored as zero. All possible item scores were transformed to values ranging from 1 to 7. The range of possible scores on the scale was 29 to 203. A high score indicated a high level of authoritarianism. luBernard N; Bass, QAuthoritarianism or.Acquiescence,' Journal of Abnormal and Social Psychology, 50 (1955) 617-620. 103 Using the technique of split-half reliability, Adorno 933$. found that the average reliability coefficient for Froms 40 and 45 was .90, their range .81 to .97.15 In the pilot stuchr the corres— ponding reliability was .52. It was decided to continue to use the modified F-scale because of the higher levels of reliability reported in the literature. In the major study the reliability of the scale we increased to .74. A series of techniques have been used to estimate the validity of the original F-scale. It we found to correlate with Ethnocentrism, Politico-Economic Conservatism and XenOphobia scales.16 The analysis of the internal consistency of the F-scale has been attempted by using different techniques which resulted in different conclusions. It has already been reported that Adorno 9113;. found the item scores significantly correlated with the total scores.17 Christie and Garcia and O'Neil and Levinson using factor analytic techniques, detected empirical clusters of personality character- istics that were similar to the theoretical structure of the F syndrome.18 found his results like those of Christie and Jahoda demonstrated Camilleri using the centroid method of factor analysis that the F-scale was not unidimensional. He also found the ”Adams et al., op. cit., p. 257. 16Delbert c. Miller, Handbook of Research Desi n and Social Mpgsurement (New York: mvid McKay Compazv, Inc., 1955), p. 301. 17See above p. 100. 18John Madge, op. cit., p. 422. 104 factors were not related in a meaningful fashion when the items were grouped into the categories of the original researchers. Camilleri concluded that the F-scale was not a good measure of the authoritarian personality theoretical structure. He also stated that the original 29 item F-scale did not provide item scores that could be summated and correlated with other behavioral variables. The same total score could result from a variety.of reSponse patterns.19 In the pilot study the product moment coefficients between item and total scores for the modified F-scale were used as a measure of internal consistency. Only 11 out of 29 items yielded significant correlations in the predicted direction. This item analysis did not indicate good discriminating items nor unidimen- sionality. Table 1 reports the correlations for the major study. Four out of the 29 items were not significantly related to the total scores. These items were: 7. Some of the goings-on in this country, even in places where people might least expect it, are tame compared to the wild sex life of the Greeks and Romans. lb. The artist and the professor'are much more important to society than the businessman and the manufacturer. 16. Those people who were in positions of authority before Genmany entered World war II should not have been used to keep order'and prevent chaos after the war. 19Santo F. Camilleri, “A Factor Analysis of the F-Scale,’ Social Forces, 3? (1959) 322-323. - TABLE 1 105 RESULES OF THE ITEMJTOTAL SCORE.ANALISIS OF THE MODIFIED F-SGALE Standard a Item Mean Deviation Pearson r Significance 1. 2.77 1.86 .35 p. 4 .05 2. 3.19 1.63 .14 p. < .05 3. 4.38 1.97 .34 p. < .05 4. 1.98 1.26 .41 p. <: .05 5. 2.69 1.48 .43 p. < .05 6. 4.35 1.70 .35 p. 4 .05 7. 4.49 1.56 .00 not significant 8. 2.73 1.70 .23 p. <1 .05 9. 2.57 1.49 .36 p. 4 .05 10. 3.92 1078 037 p. < 005 110 2.89 lei-"9 .21. p. < 005 12. 3.06 1.71 .46 p. 4 .05 13. 3.81 1.73 .61 p. < .05 14. 5.10 1.44 .10 not significant 15. 2.82 1.32 .36 p. < .05 16. 4.44 1.52 -.03 not significant 17. 2.66 1.56 .44 p. <1 .05 18. 3.40 1.71 .12 not significant 19. 4.01 1.69 .34 p. 4.05 20. 6.04 1.09 .18 p. 4 .05 2].. 3.19 1.78 .42 p. 4 .05 22. 2.69 1.61 .31 p. <: .05 23. 3091‘" 1073 053 p. < .05 21+. 3075 2.08 .36 p. < 005 25c ”.13 1.66 cm p0 < 005 as. 3.49 1.60 .21» p. < .05 270 30% 1.82 .2“ p. < .05 280 3059 108“ .38 p. < 005 290 “.90 1.61 .22 p. < 005 8‘N = 263. A correlation of .121 is necessary for the relationship to be significant at the .05 level of significance. 106 18. weakness and difficulties can hold us back; sheer will power is not enough to overcome difficulties. The first three of the four items can be characterized as being exceedingly ambiguous, i.e., making it difficult for the students to give meaningful responses. .The fourth item approached significance. Although the results of the item analysis for the modified F-scale were not impressive for the pilot study, the results of the test of validity were promising. Previous studies have demonstrated that the level of authoritarianism was related to nominal religious group affiliation. Catholics have been found to score higher on the F-sca1e than Protestants and Jews.20 The explanation of these findings is that the religious ideology of Catholics is more authori- tarian than the religious ideology of Protestants and Jews. In the pilot study the mean scores on the modified F-scale for the Catholics, Protestants and Jews were: .i = 84.3 (N = 9): i = 79.2 (N = 29): and i = 77.3 (N = 6) respectively. Using the Krushal- wallis test of one-way analysis of variance, the differences between ranks of the three groups were computed. The H—value was 5.97 in the predicted direction. This value for df = 2 was significant beyond the .10 level and.approached the critical value (5.99) at the .05 level of significance. Thus, the scale was considered to indicated construct validity. 20Milton Rokeach, op_. cit., pp. 109.119. 107 Further evidence of the validity of the modified F-scale was available in the test of proposition I which stated that there would be significant inverse relationships between authoritarianism with attitude, commitment and overt behavior toward members of a minority group. In order to determine whether the variables were significantly related, Kendall's rank correlation coefficient (Tau) was computed.21 The degree of association between the commitment and authoritarianism (r. -.30) and attitude and authoritarianism (r. -.31) were significant at the .10 level (N = 46). The degree of association between overt behavior and authoritarianism (r. -.23) was significant at the .20 level (N = 46). Although the extent of the associations were not great, they were in the predicted direction and did indicate that the proposition was probably true. Therefore, it was decided to retain the variable for the major study. The Past Intergroup Contact Scale In order to measure the degree of past intergroup contact, a scale (Past Exp-scale) was designed to measure the extent and type of past experience white college students had with Negro college students. A twenty item scale was used in the pilot study. Eight out of the 20 items were, however, designed to elicit responses concerning the same type of previous intergroup contact-ta1king five minutes or more with a Negro student. The items differed ”Siege. mu pp- 213-223. 108 according to the topic that had been discussed, e.g., Sports, dating, etc. The research setting was changed from an interview to a classroom setting. In order that the students would have the opportunity to complete the questionnaire within a 45 minute time- span, it was decided in the major study to reduce the eight items to one general item leaving a 13 item scale. The Past Exp-scale measured varying degrees of intimate contact. The degree of intimacy varied from having worked with a Negro to having gone to a social activity with a Negro of the opposite sex. The projected time interval into the past varied. It was reasoned that some forms of intergroup contact had a likelihood of occurring more frequently than other forms, e.g., the possibility of becoming acquainted with a Negro in a class appeared less likely than the possibility of eating at the same table with a Negro. Examples of the questions are: l. Daring the last year did you ever sit next to a Negro in class whom you got to know? 2. During the last 30 days have you sat down to eat at the same table with a Negro? 3. Do amr of your close friends have close friends who are Negro? The subject was asked to reapond Eyes” or “no" to each item. .A des' response was scored as 3 and a ”no. reaponse was scored as l. Omissions, question marks or other indicators of doubt were scored as 2. The possible total scores ranged from 13 to 39. 109 .After responding to the past exp-scale students were asked to express whether their past experiences were favorable or unfavor- able. Separate questions were used for each reSponse. After'the question asking if they had favorable past experiences, they were asked to describe from one to three of these past experiences. The same procedure was used after the question on unfavorable past experiences. The descriptions of the past experiences were ranked as being either strongly favorable (l), slightly favorable (2), no response (3), slightly unfavorable (4), and strongly unfavorable ( 5). The total scores for both questions ranged from 4 to 8. It was found that the past exp—scale was correlated with the favorable evaluation of past experience (r —.45) but did not correlate with unfavorable evaluation of past experience (r -.O4). When the past exp-scale was correlated with the combined favorable and unfavorable evaluation of past experience the correlation drOpped to r -.36. The Tau value of the relationship between the favorable evaluation of past experience and the modified F-scale and the favor- able evaluation of past experience and the PSSO—scale were less than the past exp-scale correlation with the same antecedent variables (Tau values +.08 and -.24 compared to -.14 and +.36). As a result of the analysis of the favorable and unfavorable evaluations of past experience, it was decided that the past exp—scale was a better measure of past intergroup contact. The above analysis revealed the past exp-scale was correlated with the favorable evaluation of past experience and revealed the past exp—scale was more sensitive to 110 the relationships with other variables than the favorable evalua- tion of past experience. The Split-half reliability for the past exp-scale in the pilot study was .86. In the major study it was .77. The greater reliability of the pilot study was probably due to the larger number of items in the scale. The internal consistency scale in the pilot study was good. Eighteen out of 20 items were signifi- cantly correlated with the total scores. Table 2 reports the item- total score correlations for the major study. Every item was significantly correlated with the total scores. TABLE 2 RESULTS OF THE ITEM-TOTAL SCORE.ANALISIS FOR THE EAST INTERGROUP CONTACT SCALE Standard a Item Mean Deviation Pearson r Significance l. 1.96 1.00 .43 p. ‘5 .05 2. 1.78 .98 .60 p. “L .05 3. 2.15 .99 .56 p. 4 .05 4. 1.81 .98 .35 p. 14L .05 5. 2.39 .92 .59 P. <: .05 6. 2.17 .99 .64 p. ¢: .05 70 2066 075 on? p. 4 005 8. 2.55 .84 .61 p. ¢: .05 9. 2.87 .48 .39 p. ‘1 .05 10. 1.51 .87 .50 p. 4 .05 11. 2.88 .47 .33 p. < .05 12. 2.75 .66 .32 p. 4 .05 _lj. 1.90 1.00 .60 p. < .05 an = 263. A correlation of .121 is necessary for the relationship to be significant at the .05 level of significance. lll Supporting evidence for the predictive validity of the past exp-scale was available in the test of pr0position II. It stated past intergroup contact would be significantly related to attitude, commitment and overt behavior. Using Kendall's Tau the degree of association between attitude and past intergroup contact (r = .29) and commitment and past intergroup contact (r = .25) were significant at the .10 level and the .20 level, respectively (N = 46). The association between past intergroup contact and overt behavior (r = .10) was significant at the .40 level (N = 46). The measures of association were in the predicted direction and indicated that the proposition was probably true. Thus, it was decided that past intergroup contact was successfully operationalized. The Perceived Suppgrt of Sigpificant Others Scale In order to measure perceived support of significant others, a scale was designed to measure the degree that significant others would participate in interracial activities. A 54 item scale was developed for the pilot study. There were 9 stimulus questions, each having 6 possible significant others associated with each question. The type of interracial activity varied from contri- buting $1.00 to help finance the activities of the 'Congress of racial Equality, ' the ”Student Non-Violent Coordinating Cormdttee, I and the ”National Association for the Advancement of the Colored O People" to agreeing to protest against segregated housing in one's home town. The six classes of significant others were: 112 (1) close friend; (2) roommate; (3) parents; (4) pastor or religious leader; (5) faculty member and (6) other. For each of the 54 items the responses were Kyes', “maybe” or ”no”. The respective scores for the responses were 1, 2 and 3. .The method of selecting total scores was to compute the percentage of des” responses from the total number of responses that were answered. Since primary interest was focused on the degree of positive support, the “yes" responses were given the value of l and the other responses scores of zero. All of the students did not consider each of the six classes of significant others as important. In some cases the parents were deceased, or the roommate was evaluated negatively. Thus, many students did not answer all 54 items. In order to make the response patterns comparable the scores were obtained by determining the percentage of Eyes” responses from the total number of responses that were answered. The range of possible total scores was .00 to 1.00. In the major study the same stimulus questions were used. However, 5 classes of significant others were used instead of 6. In order to allow'the students to answer the complete question- naire within a 45 minute period, the “other“ response was eliminated from the major study. Inspection of the response patterns of the pilot study also indicated that some students had difficulty in responding. They did not consider a religious leader or faculty member as a significant other. Others wanted to distinguish between friends of the same sex and friends of the opposite sex. 113 Married students did not know whether the response roommate applied to their situation. Therefore, new response categories were selected for the major study. They were: (1) closest friend of the same sex; (2) closest friend of the opposite sex; (3) parents; (4) roommate (or husband and wife); and (5) some older person whom the students respected, e.g., professor, minister, uncle, etc. The students were asked to specify the older person. The split- half reliability of the PSSO-scale in the pilot study was .90. In the major study the corresponding reliability was .96. The reliability was probably increased by removing the ambiguity in the responses. In the pilot study 17 out of the 54 items were not significantly related to the total score, indicating that a relatively large proportion of items did not have good discriminating power. When the response patterns were improved the interval con- sistenqy of the scale increased. Table 3 reports the itematotal score correlations. Every item was significantly correlated with the total score in the major study. The predictive validity of the pilot study PSSO-scale was estimated by correlating the PSSO.scale with attitude, commitment and overt behavior. The corresponding Tau's were .26, .41 and .30 (N = 46). Attitude, commitment and overt behavior were significantly related to the PSSOascale at the .20, .05 and .10 levels of signifi- cance. The measures of association were in the predicted direction and indicated that the proposition was probably true. Thus, it was decided that perceived support of significant others was success- fully operationalized. 114 TABLE 3 RESULTS OF THE ITEM-TOTAL SCORE ANALYSIS FOR THE PERCEIVED SUPPORT FROM SIGNIFICANT OTHEIE SCALE Standard Point Bi- Item Mean Deviation Serial r"‘l Significance 1. .73 .44 .51 p. c .05 2. .65 .48 .46 p. < .05 3. .40 .49 .43 p. < .05 4. .58 .49 .44 p. 4 .05 50 .83 037 035 p0 < 005 6. .37 .48 .59 p. 4 .05 70 029 .145 052 p. A 005 8. .21 .41 .44 p. 4 .05 90 .24 .43 .51 p. < .05 10. .50 .50 .52 p. 4 .05 ll. .36 .48 .56 p. < .05 12. .29 en's 053 p0 < 005 13. .13 .34 .38 p. 4 .05 14. .22 J42 .52 p. < .05 15. .1le .50 .52 p. 4 .05 16. .38 .49 .57 p. < .05 170 0% .144 053 p. < 005 18. .27 .44 .53 p. < .05 19. .22 .41 .53 p. < .05 20. .46 .50 .48 p. 4 .05 21. .84 .37 .49 p. < .05 22. O74 0144 051 p. ‘005 23. .53 .50 .48 p. 4 .05 24. .65 .48 .44 p. 4 .05 25. .83 .38 .173 p0 ‘ 005 fig .77 0&2 014’9 p0 < 005 270 070 0% oil's Po < 005 28. 051 050 0% p. < .05 290 .64 .178 .40 p. < 005 30. 079 on]. .41 p. < .05 310 0% cm 0146 p0 < 005 32. .20 .40 .43 p. < 005 33. .14 .35 .24 p. < .05 31‘". 016 .37 036 p0 4 005 35. .36 .48 .35 p. < .05 o 025 cm 050 p. < 005 370 022 042 .55 p. < 005 38. .09 029 em I). < 005 39. 015 036 ofl p. < 005 115 Table 3 cont. 40. .40 .49 .49 p. .<: .05 1+1. .32 .47 .54 p. 4. .05 42. .27 .44 .54 p. 4 .05 43. .21 .41 .47 p. <1 .05 44. .21 .40 .52 p. < .05 45. .46 .46 .50 p. 4 .05 3N = 263. .A correlation of .121 is necessary for the relationship to be significant at the .05 level of significance. Definition of the Situation There was not direct measurement of the independent variable definition of the situation. Its existence was judged by the effects on the two intervening variables when the experimental conditions were manipulated. In the pilot study one interview situation (Experimental Condition I) consisted of students being asked to respond to attitude statements before they were aware of the questions asking them to commit themselves to interracial activities. In the other interview situation (Experimental Condition II) the order of appearance of statements was reversed. In experimental condition I it was expected that students would respond to the attitude items in terms of the normative expectations they perceived to be related to playing the role of a 'good' research respondent and the role of the "liberal” college students. In experimental condition II the D of S was expected to be quite 116 different. Students were asked to respond to attitude statements after they had committed themselves to varying degrees of possible interaction with Negroes. The interview situation no longer retained its “play-like” characteristics.22 Decisions were already made that involved patterns of possible interaction outside of the testing situation. It was expected that students would be consistent with their commitment when they expressed their attitudes. Moreover, it was expected that attitudes would be more highly associated with authoritarianism, past intergroup contact, perceived support from significant others, commitment and overt behavior in experimental condition II. In order to test the effects of the two different D of 8's Tau's were computed for the relationship between the five variables and attitude. The respective Tau values in experimental condition I (N = 22) were: authoritarianism, -.l4; past intergroup contact, .35; perceived support from significant others, .01; commitment, .36; and overt behavior, .24. In experi- mental condition II (N 24) the Tau's for the same relationships were: -.37. .26, .48, .59 and .63. The direction of the prediction was accurate in four out of five relationships. The average difference between the pairs of Tau's for the four relationships 22Back at al., op. cit., p. 183. The authors draw the analogy between participating in.a research experiment and.p1ay. They out- line Caillois' five features of play. It is separate, uncertain, unproductive, free and is governed by rules and.make—believe. Roger Caillois, Man, Play and Games (Glencoe, Illinois: The Free Press, 1961). pp. 9.10. 117 in the predicted direction was .33. Using a conservative estimate of significance the Tau's for’attitude and overt behavior in the two experimental conditions were not found to be significantly different.23 The Attitude Scale In order to measure attitudes toward Negroes a scale was designed to measure the degree of favorable orientation toward Negroes. A 32:item scale was developed.2n .A number of suggestions were followed in determining the content and structure of the items. The scale was labeled QAttitudes Toward Negroes on the MSU Campus! in order to impress upon students the nature of the questions. Following the suggestion of DeFleur and Westie the object of the scale was specified in detail. This procedure was used so that the students would not consider the object of the attitude items as being ambiguous. They did not have to respond to items that referred to Negroes in general. They were responding to items that referred to fellow students about whom they had an.opportunity to form some judgment either through direct or indirect contact. 23Maurice a. Kendall. sank Correlation Methods (London: Charles Griffin and Company, Limited, 1948), p. 53. The author stated that if the standard error was Z Tan -Tau2, then the difference is not significant. The test for 31 ficance is a conservative test using the maximum variance for an estimate of standard error. .Appendix B provides addition statistical information on all the Tans computed in each treatment group. 2“Due to a typographical error in the response categories one of the items had to be excluded in the major study. Item 18 could not be used. 118 Edwards' summary of fourteen criteria for constructing items was taken into consideration but not followed rigidly.25 Examples of items are: 1. I think there are Negroes qualified to be class president. 2. I would feel extremely uncomfortable dancing with a Negro student. 3. I would prefer sharing living quarters with any white rather than with a Negro student. The items covered a variety of experiences on campus, e.g., dating, student government, living and eating together, athletics, academic abilities, militancy, etc. Thirteen items expressed favorable attitudes toward Negro students and 18 expressed unfavorable attitudes. In order for students to be consistent, they had to both.agree and disagree with items. The same items were used in both studies. A 7-point forced choice set of responses were used for each item. They were phrased in the same manner as the modified F-scale responses. The range of possible scores was 31 to 217. The estimated split-half reliability of the pilot study was r .91 and in the major study it was .89. In the pilot study 28 out of the 32 items were significantly correlated with the total scores. Three of the 4 remaining items approached significance. The pilot study data was encouraging. Table 4 reports the correlation between item and total scores for the major study; Every item was significantly correlated with the total score. In the pilot study evidence of the predictive validity 25Edwards, op. cit., p. 13. 119 TABLE 4 RESULTS OF THE ITEM-TOTAL SCORE ANAIISIS FOR THE ATTITUDE SCALE fis‘tandard a Item Mean Deviation Pearson r Sigpificance 1. 6.30 .82 .60 p. < .05 2. 6.16 1.08 .32 p. A .05 3. 6.50 .94 .44 p. 4 .05 4. 6.29 1.01 .70 p. 4 .05 50 4e72 1088 .68 I30 4 .05 6. 5.75 1.12 055 p0 < .05 70 5.61 1.28 .62 p. < .05 8. 4.01 1.82 .72 p. 4 .05 90 3.81 1099 056 pe 4.05 10. 6.11 1.12 .46 p. < .05 11. 6.26 1.06 .66 p. . < .05 12. 5.86 1.13 .74 p. 4 .05 13. 6.06 1.07 .56 p. < .05 14. 6.06 1.19 .70 p. 4 .05 15. 5.91 1.32 .71 p. < .05 16. 6.58 .70 .63 p. 4 .05 17- 6.50 .59 .54 p. 4 .05 19. 5.97 1.33 .74 p. < .05 20. 5.18 1.43 .64 p. < .05 21. 5.42 1.50 .62 p. < .05 22. 6.15 .99 .66 p. < .05 23. 6.01 1.16 .54 p. 4 .05 24. 6.30 .93 .65 p. < .05 25. 6.04 .96 .57 p. < .05 26. 3.50 1.96 .58 p. < .05 27. 4.81 1.88 .57 p. < .05 28. 6.33 .81 .67 p. < .05 29. 4.95 1.67 .26 p. < .05 30. 6.38 .77 .66 p. < .05 31. 5.06 1.54 .46 p. < .05 32. 6.04 1.09 .41 p. 4 .05 33. 6.45 .63 .Jl _JDL 4 .05 3N = 224. A correlation of .132 is necessary for the relationship to be significant at the .05 level of significance. 120 of the attitude scale was obtained by testing preposition VI. It stated that attitude would be significantly related to commitment and overt behavior. Using Kendall's Tau the correlation between attitude and commitment was .47. The correlation between.attitude land overt behavior was .42. Both were significant beyond the .05 level (N = 46). The measures of association were in the predicted direction and indicated that the proposition was probably true. Thus, it was decided that attitude was successfully operationalized. The Commitment Scale In order to measure the variable commitment a 10 item scale was develOped. It was unrealistic to create a longer scale. The longer the scale, the more the students would have become skeptical of the manifest function of the commitment items. Questions were designed that implied taking part in interracial activities with Negro students. The introduction to the scale stated that the questions involved possible future experience with Negroes on campus. It went on to state, “If programs or activities could be set up to help improve interracial understanding on campus would you consider participating?” The students were asked to give a general response which was not included in the scale. Following this introduction the students were asked to gggm;t_ themselves to nine different forms of possible activities. If they committed themselves to any of the nine activities, they were asked to give their phone number. This last item was included in the scale to reinforce the idea of the possibility of interaction. Care was 121 taken to construct items that would appear realistic to the students. The items varied in the extent of commitment to engage in interracial activities.2'6 Examples of items are: l. would.you agree to participate in a small group discussion on the topic of white students' social relations with Negroes on campus? 2. would you agree to go to coffee or lunch with a mixed racial group of students to talk about interracial problems on campus? 3. would you agree to protest against segregated housing in East Lansing with Negro students? The students were given three choices of responses on the nine items--!yes', ”maybe“ and "no“. The des' was given the weight of one, "maybe", two and ”no”, three. Primary interest was in the degree of positive commitment. Therefore, it was decided to score the ”yes” responses as one and the remaining responses as zero. The range of possible scores was 0 to 10. The same scale was used in both the pilot and major studies. The split-half reliability of the pilot study was .82. In the major study the corresponding reliability was .89. In both the pilot and major studies every item was correlated with the total scores. Table 5 reports the results of the item-total score analysis for the major study. The pilot study data revealed that the Com-scale was a valid predictor of overt behavior. The Tau rank correlation between commitment and overt behavior was .42 (N = 46). 26The items of the PSSOBscale were modified and used as the stimulus questions in the Com—scale. The modification involved limiting the object of the items to Negro students. 0 122 The relationship was significant beyond the .05 level in the predicted direction. Thus, it was decided that the construct commitment had been successfully operationalized. TABLE 5 RESULTS OF THE ITEM-TOTAL SCORE ANAHSIS FOR THE CQ‘IMITMENT SCALE Standard Point Bi... Item Mean fiDeviation serial 13 Significance l. .63 .48 .70 p. < .05 2. .36 .48 .67 p. <: .05 3. .32 .47 .63 p. 41.05 1+. .63 0% 075 p0 <‘ 005 5. .53 ~50 .75 P- 4 ~05 6. .2. 041 055 Po < 005 70 037 .158 .73 p. < .05 8. .37 0‘48 .71 p. < 005 9. .% .44 .54 p. < .05 10. féj .48 JL p. 4 ,0: a N = 224. A correlation of .132 is necessary for the relationship to be significant at the .05 level of significance. The Overt Behavior Scale In order to measure the degree of overt behavior a scale was developed to measure behavior that was congruent with attitude and commitment toward Negro students. The scale consisted of six ~~ distinct manifestations of behavior. The behavior measured ms the extent of willingness to become involved in small group discussions with members of the campus chapter of the NAACP. The small group discussions were defined as orientation sessions on what students at MSU were doing to improve race relations. The NAACP chapter was selected because it was the only active student interracial 123 organization on campus during winter term 1965. This organization represented a heterogeneous group of students, e.g., newcomers to those who advocated strong tactics of militanqy. At the end of the questionnaires students were asked if they would be willing to participate in small group discussions. If they said yes, they were asked if they would be willing to attend one of the discussions scheduled for three different days at both 4:00 p.m. and 8:00 p.m. The students were asked to indicate the time and day they wished to attend the discussions. During the weekend following the administration of the questionnaire students were contacted by telephone and asked if they still planned to attend the small group discussions. This third step was designed to discover if students actually planned to attend the small group discussions and to provide a reminder of the times and place. At the six small group discussions every student was asked to give his name. This information was used to identify those who came with theirquestionnaire data. At the beginning of each meeting the researcher defined himself as a member of the NAACP. He introduced representatives of the student NAACP chapter who had previously volunteered to lead the small group discussions.27 Each of the leaders had previously met with the researcher to discuss a general outline for the discussions. .At each session the campus historywof the organization was discussed, particular areas of 27The representatives were among the most active members of the organization.. Mbst of them had held positions of leadership within the organization. .At one meeting there were three former presidents of the organization and additional officers. 124 discrimination on and off-campus were cited, and future activities on the campus were brought to the students' attention. Students participated in lively conversation with the members of the NAACP. At the end of the small group discussions students were asked if they would be willing to sign up to participate in any of the following activities planned for winter term. 1. 2. 3. 4. 5. 6. Help work on a project in Missippi that emphasized remedial education and developing leadership. The project was Sponsored by the All-University Government. Attend the pley "The Man Called Nigger' that was being presented during winter term to help raise funds for the project designed to send MSU students to Mississippi. ‘work on a committee that was involved in trying to reach a solution to the off-campus housing problem for Negro and foreign students. Help work on a campus newsletter designed to cover student involvement in local and national civil rights activities. Wbrk on a fund raising campaign to help finance civil rights organizations. Become a member of the NAACP. The scale measuring congruent behavior had six discrete points from refusing to participate in the testing situation to actually signing up to participate in future activities beyond the small group discussions. The range was 0 to 5. The standard textbook procedures for analyzing scales were difficult to apply to the overt-scale. A split—half reliability test was inadeeuate for analyzing five dishotomous items. Inspection of the scale revealed very few inconsistencies in the degree of involve- 125 ment. The time ordering of responses that increased in the degree of involvement in interracial activities reduced the possibility of inconsistency. In the pilot study three responses were out of order in a total of 230 recorded reaponses for 46 students. In the major study seven responses were out of order in,a total of 1100 recorded responses for 220 students. Evaluating the results in terms of scalogram analysis procedures revealed a coefficient of reproducability of .99. Thus, the scale had a high level of internal consisteney. Table 6 reports the results of the item-total score analysis. All items were significantly correlated with the total score. The overt-scale was the dependent variable that two of the three independent and the intervening variables predicted. Both the attitude and commitment scales were found to be correlated with overt behavior beyond the .05 level of significance. Therefore, it was decided that overt behavior was successfully operationalized. TABLE 6 RESULTS Q“ THE ITEM-TOTAL SCORE ANAHSIS FOR THE OVERT BEHAVIOH.SCALE 1' Standard Point Bi- ‘ Item Me§n_ Ibviation serial ra Significance lo 036 0% 0 P0 < '05 2. .30 .46 .88 p. 4 .05 3° 019 039 086' p. < .05 4. .11 .31 .81 p. 4 .05 50 .10 .30 .78 h‘ .05_ 8‘N = 220. .A correlation of .133 is necessary for the relationship to be significant at the .05 level of significance; 126 Reliability of Research Procedures Helen Peak has stated that it is misleading to Speak of the reliability of a test with the implication that reliability is a property only of the instrument itself, for the error observed is the result of variation in the whole complex of determinants of the measured event. QAny statement about what a test measures and how reliably the measurement is made, must be accompanied by informa- tion regarding the conditions under which the statement is true.28 In this study numerous attempts were made to control sources of unreliability. The administrator of the questionnaire was constant in three of the four classes. A competent graduate student volunteered to administer the instrument to the three classes who were given the opportunity to respond to the overt-scale. The research director did not take part in the administering of these questionnaires. He was scheduled to take part in the small group discussions. It was felt his presence in both situations would affect the small group discussions. The graduate student was instructed on the purposes of the research, the research procedures and the type of questions students would probably_ask and he was given a set of responses to handle the questions. No significant abnormalities occurred in the testing situation. One question that was frequently asked was, 2afielen Peak, “Problems of Objective Observation,“ Research Methods in the Behavioral Sciences, eds. Leon Festinger+and Daniel Katz New York: Holt, Rinehart and Winston, 1953), p. 293. 127 ”IS there actually going to be small group discussions?” The administrator assured the students that the discussions were actually planned and that positive reSponses to the questions involved more than taking part in a classroom experiment. In the fourth class the data were gathered two weeks after the first three classes. No small group discussions were planned. The director of the research project administered the questionnaire. The reSponse procedures were clearly explained to the students. The rSSponse alternatives were well Spaced so that reSponses would not be confused by students or coders. Each questionnaire provided column numbers in the right hand margin for coders to place the number correSponding to the reSponses. This procedure facilitated the task of transposing numbers from the questionnaire to IBM punch card sheets. Three coders were trained on how to use a code book. A twenty per cent sample of the questionnaires revealed one systematic error, which was corrected, and random errors that comprised less than three percent of all the rSSponses that were coded. Trained university personnel punched and verified the IBM cards.29 A computer programmer was hired to program the data according to programs available in the Computer Center Library. 29Only one error out of all the numbers that were punched on cards was discovered. This error caused a loss of about three weeks work on the analysis. It was an alphabetic instead of a numeric punch. The computer program that was used was not designed to handle alphabetic punches, and thus, the computer refused to process the data until the error was located. 128 The means and standard deviations of each of the Six scales were checked and found to correSpond to the calculations of the computor. Conclusions The most convincing evidence for reliability is the genenal consistency between the pilot study results and the major study results. There is no exact method for comparing the two sets of data. Eifferent testing conditions, relatively small N in the pilot study, different time periods in which the data were gathered and different students all contributed to sources of variability. However, there are remarkable comparisons.30 In general, the pre- dicted direction of the hypotheses was consistent. There were A similar'patterns of responses to both sets of scales. Although the data were not reported the average scores on the scales were compar- able, e.g., the Scores on the attitude scale were highly favorable in both groups of data. This evidence suggested that the results were both stable and reproducible. One exception to the similarity between the pilot and major studies was the extent of overt participation. In the pilot study 25 percent of the students attended the small group discussions and in the major study only 10 percent attended. The pilot study was conducted under conditions involving interaction with the interviewer in a two person-group setting and the data were gathered during the 3oAppendixB provides comparable statistical information. 129 Spring term when the weather was good. In contrast the major study was conducted under conditions of group administration and the small group discussions were scheduled during a period when an ice storm covered the Lansing area. The differences in the two studies systematically reduced the turn out at the small group discussions in the major study. The validity of the operationalization of the scales has been demonstrated by the degree of internal consistency and the predicted relationships between variables. The presentation of the findings in Chapter V will provide further criteria on which to evaluate the predictive validity of the scales. This chapter has been devoted to answering one major question, fHow good is the data?” The answer to the question is relative to the strictness of the criteria emplqyed in evaluation. The ideal would be an isomorphic relationship between the theory and the observations, measurements and the mathematics employed in the quantification of the variables. .At best, the data indicate a develOped representation of the theoretical perspective and not perfect correSpondence between theory and methodology. When this study is compared to previous research there are a number of improvements. For'example, the F-Scale was modified to overcome the acquiescent response set. The attitude scale presented objects to subjects that were not ambiguous. .A scale to measure commitment was developed and the overt-scale included 130 direct interaction with members of a minority group outside of the testing situation. Within the restrictions of limited time, money and personnel, the methodological procedures revealed both reliable and valid data. Thus, the methodological procedures are considered to be an accurate derivation of the theoretical framework. CHAPTER V ANALYSIS OF THE RESULTS Introduction This chapter will be divided into two major sections. The first section reports the results of the statistical analysis of the hypotheses.1 The second section is devoted to interpreting the results. The statistical test of thezull hypotheses is used as the primary source of evidence to evaluate the theoretical proposi- tions and hypotheses. A decision was generally made to accept the theoretical hypotheses if the statistical tests did reveal signifi- cant relationships. However, decisions to accept or reject hypotheses were not solely dependent on the statistical test. All a significance test pg£_§g_can give is a probability statement about obtaining a certain result if the given hypothesis is true. The actual decision to accept or reject the test of hypotheses is dependent on additional factors that are not part of the formal lKendall's Tau was consistently used as the statistical test of the relationship between variables. It was decided to remain with a statistic that provided a uniform test of the propositions. Appendix C describes how the statistic was computed. Appendix.B provides additional statistical information that was not used in the direct test of the hypotheses. Product—moment correlation coefficients, Kendall's Tan and Kruskal and Goodman's Gamma were computed for all the relationsgips between scales in each of the four treatment groups in the major study and the two treatment groups in the pilot study. 131 132 mechanics of a statistical test. The criteria used to reach decisions will be covered in more detail in the second section. Analysis of Propositions and Hypotheses Proposition I The first proposition stated that there would be an inverse relationship between the level of authoritarianism and the degree of favorable attitude, commitment and overt action toward members of minority groups. This pr0position was tested by three research hypotheses: 1. There will be an inverse relationship between the modified F-scale and the att-scale. 2. There will be an inverse relationship between the modified F-scale and com-scale. 3. There will be an inverse relationship between the modified F-scale and the overt-scale. Tables 7, 8 and 9 report the results for the three research hypotheses. The Tau's for the three hypotheses were -.29. -.19 and -.18. Each Tau was Significant beyond the .05 level. Thus, proposition I was statistically confirmed. Proposition II The second proposition stated that there would be a positive relationship between the extent of past intergroup contact and the degree of favorable attitude, commitment and overt behavior toward members of a minority group. This proposition was tested by three research hypotheses: 133 TABLE 7 THE RELATIONSHIP emu AUTHORITARIANISM AND ATTITUDE FOR.TREATMENTS (1). (2).AND (3) Authoritarian- Attitude Scores ism 90 100 110 120 130 140 150 160 170 180 190 200 210 Total Scores_:99 109 119 129 139 149 159 169 179 189 190 209 Z 1_l 40-49 1 1 50-59 6069 l l 2 70-79 1 2 l 1 2 7 80-89 1 1 4 4 7 8 25 90-99 1 1 1 4 5 9 5 8 5 39 100-109 1 1 2 4 8 9 15 7 8 7 62 110.119 2 3 3 5 12 6 11 11 4 57 120-129 1 1 1 5 1 4 4 4 1 22 130-139 1 2 2 1 1 7 140-149 150-159 1 1 2 Total 1 1 3 5 8 41; 33 24 47 33 33 25 224 Tau = -.29 Standard error = .045 P.< .05 134 TABLE 8 THE RELATIONSHIP BETWEEN AUTHORITARIANISM AND COMMITMENT FOR TREATMENTS (1). (2) AND (4) £2310 ritarian- Commitment Scores Scores 0 l 2 ,3 4 _5_ 6 Z__ 8 9 10 Total_ 60-69 2 2 70—79 2 1 2 1 1 1 1 9 80.89 5 2 2 1 3 1 3 2 4 23 90-99 4 1 3 6 4 5 6 1 6 3 2 41 100-109 13 8 3 3 2 8 3 5 3 9 4 61 110.119 11 7 3 11 7 6 4 5 3 57 1m4w 3 3 3 2 3 4 2 1 3 a 130.139 1 2 2 5 140-149 150.159 1 1 2 Total 32 26 14 27 18 3o__;z_14 18 15_ 13. 224 Tau = -.19 Standard error = .045 P.‘<: .05 THE RELATIONSHIP BETWEEN AUTHORITARIANISM AND OVERI‘ 135 TABLE 9 BEHAVIOR FOR ALL TREATMENTS Authoritarianism Overt Behavior Scores Scores 0 l 2 3__ 4 5 Total 60—69 1 1 2 70-79 4 l l l 7 80.89 14 3 1 3 5 26 90-99 21 9 3 6 39 100-109 37 3 6 5 1 6 58 110-119 36 7 7 3 2 1 55 120-129 17 2 2 2 23 130-139 7 7 140-149 150-159 2 2 Total 138 16 26 17 3 20 220_w Tau = #618 Standard error = .046 P. < .05 136 4. There will be a positive relationship between the past exp-scale and the att-scale. 5. There will be a positive relationship between the past exp-scale and the com-scale. 6. There will be a positive relationship between the past exp—scale and the overt-scale. Tables 10, 11 and 12 report the results of these three research hypotheses. The Tau was significant beyond the .05 level. The Tau's for the three hypotheses were .38. .38 and .27 respectively. Thus, proposition II was statistically confirmed. Proposition III The third proposition stated there would be a positive relationship between the extent of perceived support from Signifi- cant others, favorable attitude, commitment and overt action toward members of a minority group. This proposition was tested by the following three research hypotheses: 7. There will be a positive relationship between the PSSO—scale and the att-scale. 8. There will be a positive relationship between the PSSOascale and the com—scale. 9. There will be a positive relationship between the PSSO—scale and the overt-scale. Tables 13, 14 and 15 present the results for these three research hypotheses. The Tau's for the three hypotheses were .37. .50 and .39 reSpectively. Each Tau was significant beyond the .05 level. Thus, proposition III was statistically confirmed. 137 TABLE 10 THE REIATIONSHIP EETwEEN PAST INTERGROUP comer AND ATTITUDE FOR TREATMENTS (1). (2) AND (3) PaSt Inte r_ Attitude Scores group Contact 90 100 110 120 130 140 150 160 170 180 190 200 210 Total Scores _99 109 119_ 129 139 149 159 162_g179' 182, 122 202 220 13 2 1 2 . 1 '6 15 1 1 1 3 2 4 4 3 3 1 23 i7 1 2 4 4 5 1 1 1 20 19 1 2 2 7 4 6 8 2 1 33 21 2 2 7 2 5 4 1 4 27 23 1 4 3 9 6 6 1 3o 25 3 12 4 5 4 28 27 1 1 3 2 3 4 18 29 '1 1 3 2 2 5 2 16 31 1 2 4 4 ll 33 1 1 2 2 6 35 1 2 1 4 37 39 1 1 2 Total 1 1 L A J 8. _Jl_1_ 33 24 47 33 33 25 224 Tau = .38 Standard error = .045 P. < .05 138 TABLE 11 THE RELATIONSHIP BETWEEN PAST INTERGROUP CONTACT.AND COMMITMENT FOR TREATMENTS (1), (2) AND (4) Past éggzg' Commitment Scores Contact Scores 0 l 2 3 4 ,_5 6 7A 8 9 10 Total__ 39 l l 37 1 l 2 35 1 2 1 4 33 l l l 1 l l 6 31 1 1 ' 2 2 4 2 1 13 29 2 ' 1 2 2 2 1 1 2 13 27 1 2 2 2 1 2 2 3 4 19 25 4 l 4 l 7 1 2 4 2 3 29 23 2 3 l 5 3 9 l l 1 l 27 21 5 5 1 7 2 4 1 1 3 29 19 6 3 4 3 4 4 2 1 3 3o 17 2 3 l 3 4 5 2 2 1 23 15 13 2 2 2 3 1 23 13 3 1 l l 6 Total 432. 26 14 27 18 _30 _17 14 18 15 13 224 Tau = .38 Standard error = .045 P. < .05 TABLE 12 139 THE RELATIONSHIP BETWEEN PAST INTERGRGJP CONTACT AND OVERI‘ BEHAVIOR FOR ALL TREATMENTS Past Inter- Group Contact Overt Behavior Scores Scores 0 l 2 3 4 5 Total 13 5 l 6 15 23 l 1 25 17 16 1 2 1 1 21 19 20 3 4 3 1 1 32 21 18 3 2 3 1 1 28 23 21 4 4 l 1 31 25 14 1 6 2 1 3 27 27 8 1 2 5 16 29 8 1 2 13 31 2 2 4 2 12 33 2 2 4 35 1 1 2 4 37 39 l 1 Total 138 16 26 17 j 20 220 Tau = .27 Standard error = .046 P. < .05 140 TABLE 13 THE REIATIONSHIP REerEN PERCEIVED SUPPORT FROM SIGNIFICANT OTHERS AND ATTITUDE FOR TREATMENTS (1). (2) AND (3) Perceived §32§0rt Attitude Scores Significant Others 90 100 110 120 130 140 150 160 170 180 190 200 210 Total Scores g_9971097119 129 139 149 159 169 179 189_199 209 220 .000-.099 l l l 2 2 7 2 l6 .100-.199 1 1 2 3 3 4 1 15 .200—.299 l l l 5 8 9 6 3 3 37 .300-.399 l l 2 1 2 5 7 7 9 2 2 39 .400-.499 1 3 6 1 11 9 4 5 40 .500-.599 1 4 3 10 4 11 l 34 .600-.699 2 1 1 3 7 2 16 .700-.799 l 2 4 l 4 6 l8 .800-.899 1 l 1 3 .900-.999 l 5 6 Total ;_l__ l 3 5 _8__l_l 33 24 47 33 33 25 224 Tau = .37 Standard error = .045 P.< .05 141 TABLE 14 THE RELATIONSHIP BETWEEN PERCEIVED SUPPORT ERG! SIGNIFICANT OTHERS AND CQMITMENT FOR TREATMENTS (1), (2) AND (4) Perceived :Egiort Commitment Scores Significant Others Scores 0_:l_32 33, 4 _5 6 7 8 .9 10 Total .000-.099 ' ll 1 2 2 16 .100-.199 7 2 2 5 1 1 18 .200-.299 7 6 1 3 4 4 2 1 2 30 .300-o399 5 8 2 11 3 5 l 1 2 2 4O .400-.499 6 6 4 5 5 6 3 5 2 2 44 .500-.599 2 l 2 4 7 3 4 2 l 3 30 .600-.699 2 1 5 2 2 4 1 1 18 .700-.799 1 2 2 3 6 4 18 .800-.899 1 1 1 2 5 .900-.999 1 1 l 3 Total 32 26 14 27 1330 17 14 18 1; 33 225; Tau = .50 3 Standard error = .045 P. < .05 142 TABLE 15 THE RELATIONSHIP BETWEEN PERCEIVED SUPPORT FRO! SIGNIFICANT OTHERS.AND OVERT BEHAVIOR.FORHALL TREATMENT GROUPS gerceived ngport Overt Behavior Scores Significant Others Scores 0 1 2 <3 4 45_ Total .000-.099 18 18 .100-.199 l6 1 17 .200-.299 27 2 3 3 35 .300-.399 27 2 4 l l 2 37 .400-.499 25 4 9 1 4 43 .500-.599 15 6 4 l 5 32 .600-.699 6 4 1 2 l 2 16 .700-.799 3 2 1 5 4 15 .800-.899 l 1 2 4 .900-.999 l l l 3 Total 138 16 26 17 3 ._2O 220 Tau = .39 Standard error = .046 P. 4 .05 143 The tenth, eleventh and twelfth theoretical hypotheses were derived indirectly from prepositions I, II and III. Each of the three antecedent variables was related to the three dependent variables. Each one represents a distinguishable characteristic of the individual. Since each of the individual's antecedent variables are related to the same dependent variables they were expected to be related to each other. The three theoretical hypotheses were tested by the following three research hypotheses: 10. There will be an inverse relationship between the modified F-scale and the past exp-scale. 11. There will be an inverse relationship between the modified F-scale and the PSSO-scale. 12. There will be a positive relationship between the past exp-scale and the PSSOascale. Tables 16, 17 and 18 present the results for these three research hypotheses. The Tau's for the three hypotheses were -.l4, -.l5 and .39 reSpectively. Each Tau was significant beyond the .05 level. Thus, theoretical hypotheses 10, 11 and 12 were statistically confirmed. The thirteenth, fourteenth and fifteenth theoretical hypotheses were also derived from propositions I, II and III. These three theoretical hypotheses state that the combined effects of the three antecedent variables would Significantly affect each of the three dependent variables. The three theoretical hypotheses . were tested by the following three research hypotheses: 144 TABLE 16 THE RELATIONSHIP BETWEEN AUTHORITARIANISM AND PAST INTERGROUP CONTACT FOR ALL TREATMENT GROUPS *1 Authoritar- Past Intergroup Contact Scores ianism Scores 39 37 35 33 31 22 22 25 23 21 12 12 15 13 Total 40.49 1 1 50-59 60.69 1 1 2 70-79 1 1 3 1 1 1 1 9 80-89 112443415221131 90—99 1 445685744 47 100.109 1 143568981267271 110.119 2246119118112'66 120-129 11221533441 27 130-139 1 1 2 l l 1 7 140-149 150—159 - 1 1 2 Total 2 1 433,616 22 34 333336 26 26 z 263 Tau = —.14 A Standard error = .041 P. < .05 145 TABLE 17 THE RELATIONSHIP BETWEEN AUTHORITARIANISM AND PERCEIVED SUPPORT FROM SIGNIFICANT OTHERS FOR ALL TREATMENT GROUPS ——__—-— PerceiVed Support Fzgm SignifiEant Others Authori- tarianism .000 .100 .200 .300 .400 .500 .600 .700 .800 .900 Total Scores .099 .199 .299 .399 .499 .599 .699 J99 .899 .999 40-49 1 1 50-59 60.69 2 2 70-79 1 1 1 3 2 l 9 80-89 1 2 3 4 8 3 1 7 l l 31 90-99 2 1 6 10 10 10 2 3 1 2 47 100-109 6 8 12 10 6 9 6 9 3 2 71 110-119 7 7 11 12 16 8 4 1 66 120-129 2 1 3 4 8 4 4 1 27 130.139 1 1 2 1 1 1 7 140-149 150-159 1 1 2 Total 19 20 _39 42 52;_39 19 22 5 6 263 Tau = -.15 Standard error = .041 P. 4 .05 146 TABLE 18 THE RELATIONSHIP BETWEEN PAST INTERGROUP CONTACT AND PERCEIVED SUPPORT FRQA SIGNIFICANT OTHERS FOR ALL TREATMENT GRCIJPS fiizrgmup Perceived Support From Siggfioant Others Contact .000 .100 .200 .300 .400 .500 .600 .700 .800 .900 Total Scores .099 .199 .299 .399 .499 .599 .699 J99 .899 .999 39 1 1 2 37 1 1 35 3 1 -4 33 1 2 2 1 1 7 . 31 1 3 4 4 4 16 29 2 4 3 1 4 16 27 3 3 7 4 2 2 22 25 2 1 8 12 6 2 1 34 23 4 9 5 5 5 3 2 33 21 l 4 5 7 2 8 3 2 1 33 19 5 4 8 6 8 2 3 36 l7 3 3 3 3 6 6 1 1 26 15 8 2 6 7 2 l 26 13 2 1 2 1 1 7 Total 19 20 39 42 52 39 19 22 6 263 Tau = .36 Standard error = .041 P. 4 .05 147 13. The lower the modified F—scale, the higher the past exp—scale and the higher the PSSOascale, then the higher the com-scale. 14. The lower the modified F-scale the higher the past exp-scale and the higher the PSSOascale, then the higher the att-scale. 15. The lower the modified F-Scale, the higher the past exp-scale and the higher the PSSOascale, then the higher the overt-scale. In order to test these hypotheses each of the three antecedent variables was divided at its mean. The values of the modified F- scale were reversed in order to make the high scores consistent with the other two scales. Students were placed into one of four cate- gories. They could score above the mean on all three variables, (3); score above the mean on two variables, (2); score above the mean on one variable, (1); or they could score below the mean on all three variables, (0). Thus, there were four groups of students. The rank ordering of the dependent variables were compared for the four groups of students. Tables 19, 20 and 21 report the results for the three research hypotheses. Tau's for the three hypotheses were .48, .42 and .36 reSpectively. Thus, theoretical hypotheses l4, l5 and 16 were statistically confirmed. The three tables also reveal the marked cumulative affects of the three antecedent variables on each of the three dependent variables. The means for each of the dependent variables in the marginals of the table present a graphic demonstration of the marked differences between the four groups of students. The mean scores increase in the predicted direction. 148 TABLE 19 THE RELATIONSHIP BETWEEN THE COMBINED AFFECTS OF AUTHORITARIANISM, PAST INTERGRGIP CONTACT AND PERCEIVED SUPPORT FOR SIGNIFICANT OTHERS ON CCMI‘ETMENT Combined Commitment Scores Commit- Effects ment Scores__ 0 1 2._3> 4 ,5_ 6 73 8 9 10 Totg; Mean <_fi (0)--- 158497222 49 2.18 (1) + - - 16 ll 7 10 6 12 4 4 4 3 77 3.25 (2)++- 16382118577 1 59 5.27 (3) +++ l 3 5 3 3 7 5 12 39 7.64 Total 32 26 14 27 18 30 17 14 18 15 13 224 4131 11 Tau = .48 3 Standard error = .045 P.<.05 TABLE 20 THE RELATIONSHIP BETWEEN THE COMBINED AFFECTS (F AUTHOIETARIANISM, PAST INTERGRGJP CONTACT AND PERCEIVED SUPPORT FOR SIGNIFICANT OTHERS ON ATTITUDFS Tombined Attitude Scores Atti- Effects 90 100 110 120 1""3"0 145—150" 1' 6'0 170 180 190 200 210 To- tude Score-EL i9 102 112 IQ 132 122 152 162 122 182 122 202 220 tal Mean (0)--- 2 3 4 710 7 7 8 1 49 168.5 (1)+- 1 1 1 2 4 3 18 10 15 12 7 6 80 177.6 (2)++- 1 5 3 17 10 9 8 53 192.0 (3)+++ 4 8 3 16 11 42 200.7 Total 1_ 1 3 J 8 13.1;33 24 47 33 33 25224 183.3 Tau = .42 Standard error = .045 P. < .05 149 TABLE 21 THE RELATIONSHIP BETWEEN THE COMBINED AFFECTS W AUTHORITARIANISM, PAST INTERGRCUP CONTACT AND PERCEIVED SUPPORT OF SIGNIFICANT OTHERS ON OVERT BEAVIOR Combined Overt Behavior Scores Overt Effects Behavior Scores __ O l 2 3, 4 ,_5 Tot§l_ Mean (0) - - - 42 2 5 3 1 53 .47 (l) + - - 57 5 7 3 3 75 .57 (2) + + - 29 8 5 5 2 5 54 1.22 (3) + + + 10 1 9 6 12 38 2.55 Total 138 16 26 17 j 20 220 . _1. 05 Tau = .36 - Standard error = .046 P. < .05 Proposition IV The fourth proposition stated that the greater the extent to which attitudes are expressed in a research setting that implies future involvement with the object of the attitude, the greater the relationship between each of the antecedent variables and attitude toward members of minority groups. The experimental setting was hypothesized to influence the subjects' definition of the situa- tion. In order to test this proposition the three following research hypotheses were tested: 16. There will be a greater inverse relationship between the att-scale and modified F-Scale in experimental condition II than in experimental condition I. ‘ 150 17. There will be a greater positive relationship between att—scale and past exp-scale in experimental condition II than in experimental condition I. 18. There will be a greater positive relationship between att-scale and PSSOascale in experimental condition II than in experimental condition I. Experimental condition II was treatment (2). In treatment (2) the com-scale preceded the att-scale. In experimental condition I obtained in treatments (1) and (3) the att-scale preceded the com— scale. In experimental condition I it was believed that students would reSpond much more favorably to the att-scale and that the degree of association between attitude scores and scores on the antecedent variables would be lessened. The means were different and in the direction predicted. The mean score in treatments (1) and (3) was 186.6. In treatment (2) the mean was 178.8. Tables 22, 23 and 24 report the results for the three research hypotheses. In Table 22 the Tau's for the modified F-scale in experimental condition II and experimental condition I were —.29 and -.30 reSpectively. In Table 23 the reSpective Tau's for the past exp-scale were .31 and .46. In Table 24 the reSpective Tau's for the PSSO-Scale were .29 and .49. The differences in the Tau's were actually in the opposite direction than predicted. Experimental condition I (att-scale before com-scale) had higher correlations than experimental condition II. Thus, proposition IV and theoretical hypotheses l6, l7 and 18 were not statistically confirmed. 151 TABLE 22 THE RELATIONSHIP BETWEEN ATTITUDE AND AUTHORITARIANISM UNDER TWO EXPERIMENTAL CONDITIONS Authoritarian- ism 90 100 110 120 130 140 150 160 17 180 190 200 210 Total Scores 99 109, 119 £29, 139. 122 159 169 179 189_ 199 209 220 Experimental Condition I--Treatments (l) and (3) 40-49 1 1 50-59 60-69 1 1 2 70-79 1 1 1 2 5 80-89 1 4 1 2 6 14 90-99 1 l 3 1 4 3 6 4 23 100-109 1 2 l 6 10 4 4 5 33 110.119 1 2 1 3 6 3 9 6 4 35 120-129 2 1 1 4 2 1 11 130.139 1 2 2 1 6 140-149 150-159 1 1 Total 1 1 1 2 3 5 15 13 30 20 20 20 131 Experimental Condition II-Treatment (2) 40-49 50-59 60-69 70-79 1 1 2 80.89 1 3 5 2 11 90-99 1 l 4 5 2 2 1 16 100-109 1 2 2 7 3 5 3 4 2 29 110-119 1 1 2 2 6 3 2 5 22 120.129 1 1 1 3 3 2 11 130-139 1 1 140-149 150-159 1 1 _Total 2 3 5 6 l8_> ll l7 13 13 _5 _93 Experimental Condition I Tau = -.30 Experimental Condition II Tau = -.29 Standard error = .140 P. 7 .05 152 TABLE 23 THE RELATIONSHIP BETWEEN.ATTITUDE.AND PAST INTERGROUP CONTACT UNDER TWO EXPERIMENTAL CONDITIONS ‘Pkast Inter- Attitude Scores Group Contact 90 100 110 120 130' 140 150 160 170 180 190 200 210 Total Scores ‘99 109 119 £139 149 159 169 179 189 199 209 220 EXperimental Condition I -- Treatments (1) and (3) |""‘ .1:- 13 15 1 1 l 2 1 17 1 19 1 21 23 25 27 29 31 33 l 35 37 39 Totall 112351513302020 Experimental Condition II--Treatment (2) 14 12 18 19 19 NHl—‘H mumwm i-‘N NHNNNH Huxlxoutowm «F'w-F-‘VH H N?H?#HHH H HmkHNwHwHI—J l—‘ mwmxloo‘d'o N O 131 13 15 1 17 1 19 1 21 23 25 27 l 1 29 31 33 35 1 37 39 1 __ijal 2 7 13 IL 18 13 __12 J Experimental Condition I Tau Experimental Condition II Tau Standard error =‘.140 P.7 .05 l-‘N 3 1 1 HNI—I l-' N-F'NH Pu l—‘t—‘le-‘NI?NI—' H H lr-‘l-‘NF" F'i—‘NNH NNN b) I-" HNNl—‘NI—‘N |-' l—‘r—‘H 8 PE l—’ H muunoxoo oomooxom u nN 14's.. 153 TABLE 2“ THE RELATIONSHIP BETWEEN ATTITUDE AND PERCEIVED SUPPORT FROM SIGNIFICANT OTHERS UNDER TWO EXPERIMENTAL CONDITIONS Perceived Support From Attitude Score 8 Signi fi- cant Others 90 100 110 120 130 140 150 160 170 180 190 200 210 Total Scores 99 129 119 129 139 149 159 16L179 189 122 202 220 Experimental Condition I-u-Treatments (l) and (3) . 000-. 099 1 1 l l 1+ 2 10 .100-. 199 2 2 1 4 9 . 200... 299 1 1 1 3 5 8 a 1 2 26 . 300-. 399 1 1 3 u 5 1 15 .400-.499 l 3 5 6 3 5 23 .500-.599 2 l 6 2 8 l 20 .600—.699 1 1 3 1+ 2 11 .700-.799 2 1 2 5 10 .800-.899 1 1 2 .900-0999 ' 5 5 Total 1 1 1 2 3 5 15 13 30 20 20 20 131 Experimental Condition II--Treatment (2) .000-.099 l 2 3 6 .100—.199 1 1 1 2 1 6 . 200-. 299 1 2 3 1 2 2 1 12 . 300... 399 1 2 1 1 u L» 3 1+ 1 1 22 .400-.l+99 1 2 3 1 6 3 1 17 .500-.599 1 2 2 4 2 3 11» ;600-.699 1 l 3 5 .700-.799 l 3 1 2 2 9 .800-.899 1 l .900-.999 l 1 Total 2 23 15 6 18 ll 12, 13 132 5. 93 Experimental Condition I Tau = .49 Experimental Condition II Tau = .29 Standard error = .1140 P. > .05 154 Proposition V PrOposition V stated that the greater the extent to which attitudes are expressed in a research setting that implies future involvement with the object of the attitude, the greater the relationship between attitude and commitment and attitude and overt behavior. In order to test this proposition the two following research hypotheses were tested: 19. There will be a greater positive relationship between the att-scale and com-scales in experimental condition II than in eXperimental condition I. 20. There will be a greater positive relationship between the att-scale and overt-scales in experimental condition II than in eXperimental condition I. Tables 25 and 26 report the results for the twc research hypotheses. In Table 25 the Tau's for experimental condition II and experimental condition I were .49 and .54, respectively. In Table 26 the reSpective Tau's were .35 and .40. The relationships were in the direction opposite than predicted. Thus, proposition V and theoretical hypotheses 19 and 20 were not statistically confirmed. These findings come as somewhat of a surprise. In the pilot study there were significant differences in the predicted direction for hypotheses l9 and 20. Table 27 reports a comparison of the findings from the pilot and major studies for the relationships between attitude and commitment and attitude and overt behavior. The contrast between the two studies is remarkable. The experimental effects were quite different in the two studies. An.explanation for these differences will be presented in section two. 155 TABLE 25 THE RELATIONSHIP BETWEEN ATTITUDE AND CGIMITMENT UNDER TWO EXPERIMENTAL CONDITIONS Attitude Commitment Scores Scores 0 l 23 4 5 6 2 8i 10 Total Experimental Condition I—-Treatment (1) only 90-99 1 1 100.109 110.119 1 1 120-129 130-139 1 1 uw4A9 1 1 2 150-159 2 1 1 1 5 160.169 3 2 2 2 1 10 170-179 1 5 1 1 1 9 180-189 1 1 1 7 3 6 1 3 1 24 190-199 1 1 1 2 2 u 1 3 1 1 17 200-209 1 1 1 2 1 2 3 2 13 210-220 1 1 3 2 1 1 9 Total 1211 713 814 7 8 7 1 4 92 Experimental Condition II—-Treatment (2) 90-99 100-109 110-119 120-129 2 2 130-139 3 3 1%4M 21 1 1 5 150-159 2 2 1 1 6 160-169 5 n 2 3 1 1 1 1 18 170-179 1 1 2 2 1 2 2 11 180-189 2 1 5 4 1 1 1 2 17 190.199 1 1 3 3 1 1 1 1 1 13 200-210 a 3 2 u 13 210-220 1 1 3 5 Total 1612498126568j J3 Experimental Condition I Tau = .54 Experimental Condition II Tau = .49 Standard error = .128 P.7.05 156 TABLE 26 THE RELATIONSHIP BETWEEN ATTITUDE AND OVERT BEHAVIOR UNDER TWO EXPERIMENTAL CONDITIONS Overt Attitude Scores Behavior 90 100 110 120 130 145 150 160 170 180 190 200 210 Total Scores 99 109 119 129 D9 142 159 169 179 ]._89 199 209 220 Experimental Condition I--Treatments (l) and (3) o 1 1 1 2 3 2 13 8 17 5 6 4 63 1 1 l 3 5 1 11 2 2 2 3 5 3 15 3 l 1 3 2 3 l 11 u 1 1 l 3 5 2 6 8 Total 1 1 1'. 2 3 3 15 12 26 l6 17 14 111 ' Experimental Condition II-Treatment (2) o 2 2 5 5 15 7 7 9 5 1 58 l 1 1 1 3 2 1 3 l 5 3 l 2 1 4 4 5 2 2 3 1 8 Total 2 2 5 6 _15 L l; 11 ___11 4 J8 Experimental Condition I Tau = .40 Experimental Condition II Tau = .35 Standard error = .15 P.> .05 157 TABLE27 A COMPARISON OF FINDINGS FROM THE PILOT STUIK AND MAJOR STUDY FOR THE RELATIONSHIP BETWEEN ATTITUDE AND CQdMITMENT AND ATTITUDE AND OVERT BEHAVIOR Pilot Stugz Experimental Condition I. (attitude scale before commitment scale) Commitment Overt behavior Attitude Tau = .36 Tau = .24 N = 22 N = 22 Experimental Condition II . (commitment scale before attitude scale) Commitment Overt behavior Attitude Tau = . 59 Tau == .63 N = 24 N = 24 Ma or Stu Experimental Con tion I. (attitude scale before commitment scale) Commitment Overt behavior Attitude Tau = .54 Tau = .35 N = 92 N = 111 Experimental Condition II. (commitment scale before attitude scale) Commitment Overt behavior Attitude Tau = .49 Tau == .40 N = 93 N = 78 Proposition VI Proposition VI states that there will be a positive rela- tionship between attitude and commitment and attitude and overt behavior. This proposition was tested by the following two research hypotheses: 158 21. There will be a positive relationship between att-scale and the com—scale. 22. There will be a positive relationship between the att-scale and the overt-scale. Tables 28 and 29 present the results for these two research hypotheses. The Tau's for the two hypotheses were .49 and .39, respectively. Both Tau's were significant beyond the .05 level. Thus, preposition VI and hypotheses 21 and 22 were statistically confirmed. Proposition VII The seventh proposition stated that there would be a stronger relationship between commitment and the antecedent variables than between attitude and the antecedent variables. In order to test this proposition the three following research hypotheses were tested: 23. There will be a greater'inverse relationship between the com-scale and modified Pascale than between the att-scale and the modified F-scale. 24. There will be a greater positive relationship between the com-scale and the past exp—scale than between the att—scale and the past exp-scale. 25. There will be a greater positive relationship between the comhscale and the PSSObscale than between the att-scale and the PSSObscale. In order to test the hypotheses only students in treatments (1) and (2) were used in the analysis. Students in these treatments responded to all three scales. Sets of correlations between scales for the same individuals were compared to discover if the correlations 159 TABLE 28 THE RELATIONSHIP BETWEEN ATTITUDE AND CQ‘IMITMENT FOR TREATMENTS (1) AND (2) Attitude Commitment Scores Scores 0 l 2 3 4 _:5_ 6 Z_ 8g__9 1—19, Totalw_ 90-99 1 1 100-109 110-119 1 1 120-129 2 2 130-139 4 4 140-149 3 1 1 1 l 7 150-159 4 3 1 1 1 1 11 160-169 8 6 4 5 2 1 1 1 28 170-179 2 6 3 3 1 1 2 2 20 180.189 1 3 2 7 3 11 5 4 2 1 2 41 190—199 1 2 l 5 5 5 2 4 1 1 1 28 200-209 1 1 1 1 6 1 2 6 2 5 26 210-220 1 1 1 1 3 2 4 3 16 _Tgt§1__ 28 23 gL1_,221 16 26 13 1 1 lI‘ 18 Tau = .49 Standard error = .049 P. ¢£L.O5 160 TABLE 29 THE RELATIONSHIP BETWEEN.ATTITUDE.AND OVERT BEHAVIOR FOR TREATMENTS (1), (2) AND (3) Overt Attitude Scores Behavior 90 100 110 120 130 146 150 160'175 180 190 200 210 Total Scores #99 109 119 129 139 149 159_169_179 189 199 299 220 O 1 1 3 4 8 7 28 15 24 14 11 5 121 1 1 1 3 6 2 1 l4 2 1 2 5 4 5 3 20 3 1 1 4 2 5 2 15 4 l l l 3 5 2 2 5 7 16 Total 1_1 1 3_ 4 8 9‘ 30 21 _39. 27 28 18 189 Tau = .39 Standard error = .049 R4 .05 were significantly different. Tables 30, 31 and 32 reportthe results of the three hypotheses. In Table 30 it may be seen that the Tau's for the com—scale and the modified F-scale and the att- scale and the modified F-scale were -.17 and -.27 respectively. In Table 31 the Tau's of com- and att-scales with the past exp- scale were . 35 and .34. In Table 32 the Tau's of com. and att-scales with the PSSO-scale were .48 and .36. Two of the differences in Tau's were in the predicted direction. The relationship of commitment and attitude with authoritarianism was in the direction Opposite than pmdicted. The differences between the three sets of Tau's were not statistically significant. Thus, proposition VII and hypotheses 23, 24 and 25 were not statistically confirmed. 161 TABLE 30 131E RELATIONSHIPS BETWEEN AUTHORITARIANISM AND CWTMENT AND AUTHORITARIANISM AND ATTITUDE FOR STUDENTS IN TREATMENTS (1) AND (2) Aztigfiigm Commitment Scores Scores 0 1 2 j 4 _5 6 J 8 j #10 Total 40-49 50-59 60-69 2 2 70-79 2 l 1 1 l 1 7 80-89 4 l 2 l 3 1 3 2 17 90-99 4 3 5 4 4 3 l 4 3 2 33 100-109 11 8 l 3 2 7 2 5 3 6 4 52 110-119 9 7 3 8 6 6 4 4 l 48 120-129 3 2 3 1 2 3 2 l 2 19 130-139 1 2 2 5 140-149 150-159 1 l 2 Total 28 23 11 22 16 26 13 13 13 9 11 185 Authori- Attitude Scores tarianism 90 100 110 120 130 140 150 160 170 180 190 200 210 Total Scores 22 102 112 122 132 142 152 162 122 182 122 202 220 40-49 50-59 60-69 1 1 2 70-79 1 2 l 1 2 7 80-89 1 2 3 7 4 17 90-99 1 l l 3 4 9 5 5 4 33 100-109 1 l 2 4 7 7 14 6 6 4 52 110-119 1 2 3 5 ll 5 8 10 3 48 120-129 1 1 l 5 1 4 2 3 1 19 130-139 1 2 1 1 5 140-149 150-159 1 1 2 Total 1 l 2 4 7 11 28 20 41 28 26 16 185 Authoritarianism--Commitment Tau = -.17 Authoritarianism--Attitude Tau = -.31 Standard error = .142 9.7.05 162 TABLE 31 RELATIONSHIPS BETWEEN PAST INTERGROUP CONTACT AND COIMITMENT AND PAST INTERGROUP CONTACT AND ATTITUDE FOR TREATMENTS (1) AND (2) Past Intergroup Commitment Scores Contact Scores 0 l 2 3 If; 6 7 8 9 10 Total 39 1 '7 1 37 l 1 35 1 2 3 33 1 1 1 1 1 5 31 1 1 2 2 2 1 9 29 2 1 2 2 2 1 l l 12 28 1 2 l 1 2 1 3 4 15 27 4 1 3 1 6 l 2 2 1 2 23 25 1 l 23 2 2 1 5 2 8 l l l 23 21 4 3 1 7 l 3 l l 2 23 19 6 3 3 2 3 4 2 1 3 27 17 2 2 l 1 4 4 1 1 1 17 16 11 2 1 2 3 l 20 13 2 1 1 1 5 Total 28 23 11 22 16 26 13 13 13 9 11 185 Past Intergroup Attitude Scores Contact 90 100 110 120 130 140 150 160 170 180 190 200 210 Total Scores 9L109 119 129_139 142159 169 17L1§9 122 209 220 1 39 l 37 _ 1 l 35 2 1 3 33 l 1 2 1 5 31 1 2 4 2 9 29 l 1 3 2 2 2 1 12 28 1 1 2 3 2 3 3 15 27 2 11 4 3 3 23 25 1 l 23 1 4 1 7 4 5 l 23 21 2 2 6 2 4 3 1 3 23 19 1 2 2 6 4 5 6 1 27 17 1 2 3 3 5 1 1 1 17 16 1 l 2 2 4 3 3 3 1 20 13 1 1 2 1 ~5 Total 1 1 2 4 75 11__28 20 41 28 26 16 18 Past Intergroup Contact--Commitment Tau = .35 Past Intergroup Contact-Attitude Tau = .34 Standard error = .138 P. 7 .05 163 TABLE 32 THE RELATIONSHIPS BETWEEN PERCEIVED SUPPORT FRCM SIGNIFICANT OTHERS AND COMMITMENT AND PERCEIVED SUPPORT FROM SIGNIFICANT OTHERS AND ATTITUIES FOR TREATMENTS (1) AND (2) __P—érceived Support From Significant C°mmltment Scores Others Scores 0 l 2 3 4 5 6 7 8 j _19 Total .000-.099 9 1 2 1 13 .100-.199 5 2 1 4 1 13 .200-.299 7 5 1 3 4 4 2 1 1 28 .300-.399 5 7 1 11 2 5 l 2 2 36 .400-.499 5 5 l 5 4 5 3 2 2 32 .500-.599 2 1 4 5 3 5 1 l 3 25 .600-.699 2 l 2 4 1 2 4 1 17 .700-.799 1 2 2 1 2 4 3 15 .800-.899 l 1 l 3 .900-.999 l 1 1 3 Total 28 23 11 22 16 26 13 13 13 9 11 185 Perceived Su ort From Si 911.152“ flatitude Scores Others 90 100 110 120 130 140 150 160 170 180 190 200 210 Total Scores 22 102 112 122 132 142 152 162 122 182 122 202 220 .000-.099 l l 2 2 5 2 13 .100-.199 1 1 1 3 2 4 1 l3 .200-.299 1 1 l 4 6 8 5 3 29 .300-.399 1 2 1 2 5 7 6 7 3 2 36 .400-.499 1 3 5 l 9 7 3 3 32 9500-.599 l 5 2 7 3 7 l 26 .600—.699 2 1 1 3 6 2 l5 .700-.799 1 1 4 1 4 4 15 .800-.899 1 1 1 3 .900-.999 1 2 3 Total 1 2 1 4 7 11 29 20 40 2L 29 14 185 Perceived Sfiport From Significant OthersuConnnitment Tau Perceived Support From Significant Othe rs-Attitude Tau . Standard error = .133 R7 .05 815 II II 164 Proposition VIII The eighth proposition stated that there would be a stronger relationship between commitment and overt behavior than between attitude and overt behavior. In order to test this proposition the following research hypothesis was tested: 26. There will be a greater positive relationship between the com-scale and overt-scale than between the att-scale and the overt-scale. Table 33 reports the results for the hypothesis. The Tau for the com-scale and overt-scale was .58. The Tau for att-scale and overt- scale was .40. The difference between Taul -Tau2 was in the predicted direction. The difference, however, was not statistically significant. Thus, proposition VIII and hypothesis 26 was not statistically confirmed. Interpretation of the Results Tecisions to confirm or reject hypotheses are not entirely dependent upon significance tests. Snedecor'has aptly stated this point in reference to chi-square, however, his comments can be applied to other statistical tests. He stated:2 Some people adapt a rather slavish attitude toward tests of significance, rejecting the [null] hypothesis if chi-square is more than 3.841 and accepting it if chi-square is less. This indicates inadequate appreciation of the nature of the information acquired by sampling. 2George W3 Snedecor, Statistical Methods (Ames, Iowa: The Iowa State College Press, 1948), p. 23. 165 TABLE 33 THE RELATIONSHIPS BETWEEN OVERT BEHAVIOR AND CCTMTMENT AND OVERT BEHAVIOR AND ATTITUDE FOR TREATMENTS (1) AND (2) gzzitior' Commitment Scores Scores 0 l 2 3 4 L 6 7 8 9 _;_9_ Total 0 27 19 9 18 8 15 4 2 l 1 104 1 1 2 2 2 l 1 9 2 1 1 3 3 3 1 3 15 3 3 2 2 3 2 1 13 4 1 2 3 5 1 3 2 7 13 Total 27 20 10 21 12 22 11 9 9 5 11 157 Overt Attitude Scores Behavior 90 100 110 120 HOMO 180 190 200 210 Total Scores 10 ll 1 1 l 1 16 1 18 1 20 220 0 1 1 2 .3. 7 7 25 14 20 12 9 3 104 1 2 4' 2 1 9 2 1 5 4 3 2 15 .3 1 4 2 5 1 13 4 1 1 1 3 5 2 2 4 5 13 Total 1 12379251734232312 152 Overt Behavior--Commitment Tau = . 58 Overt Behavior-Attitude Tau = .40 Standard error = .133 19.7.05 166 A sample furnishes evidence not proof. This evidence is to be added to that already accumulated from experience and reports of other research. Usually, also, there is collateral information accruing during the process of the experiment. It is the investigator's responsibility to integrate all this evidence and to reach a decision. He cannot evade this responsibility merely by citing a value of chi-square. In this study decisions to tentatively accept or reject hypotheses were dependent upon: additional information on the data gathering procedures, comparable statistical findings, previous research experience and an assessment of the theoretical constructs and their operationalization. The research hypotheses testing the first proposition were concerned with the relationship between the modified F-scale and the three dependent variables. Although statistically significant, the extent of the relationShips were not great. The modified F-scale was also not highly correlated with the other two antecedent variables. The measure of association for the relationShip between the PSSO-scale and the past exp-scale was more than twice as great as the relation- ship of the modified F-scale with these two antecedent variables. The past exp-scale and the PSSOascale measure characteristics that are theoretically related to characteristics of authoritarianism. Students in a sample drawn from a Northnn college who were highly authoritarian would not be expected to have significant others willing to engage in activities with Negroes and would not be expected to have had frequent contact with Negroes. The data 167 indicate, however, that the modified F-scale is not highly associated with the PSSO-scale and the past exp-scale. The findings are somewhat surprising. Since 1950 the variable authoritarianism and its operationalization have played a central role in research on prejudice. The continued use of this variable in the present research on prejudice can be questioned. The past exp-scale and the PSSOascale were found to be better predictors of attitudes, commitment and overt behavior. .Although proposition I was statistically confirmed it appears to add little to knowledge in this area of research. Thus, the decision to accept preposition I remains most tentative. The evidence to support propositions II and III is more convincing. The pattern of relationships reveal that both the past exp-scale and the PSSO-scale are related to each other and to the three dependent variables. The PSSOhscale is a E555 predictor of the dependent variables than the past exp-scale. As an antecedent variable it is more strongly related to the com- and overt-scales than either the past exp-scale or the modified F-scale and is asgood a predictor of the att-scale. The three antecedent variables do have cumulative affect on the three dependent variables. Hypotheses 13, 14 and 15 were confirmed. These three hypotheses predicted that there would be significant relationships between the cumulative affect and attitude, commitment and overt behavior. The Tau's were significant. 168 There were also marked differences in the means of attitude, commit- ment and overt behavior. These results illustrate the advantage of using a multivariate approach to predict dependent variables. One unresolved question is the weighting that should be given to each of the antecedent variables when they are grouped together. In the test of the above hypotheses the antecedents were assumed to have equal weight. The findings on the modified F-scale suggest that its relative importance should not be considered equal to that of the past eXp- and PSSOascales. Proposition VI stated that there would be a positive relation- ship between the degree of favorable attitude and favorable commit- ment and favorable attitude and overt behavior. Moderately strong statistical relationships were discovered. These relationships exist across treatment groups (see Appendix B). The relationships are not perfect, i.e., the att-scale scores do not eXplain all the variation on the com- or overt-scales. It is stronger, or equal to, any of the three antecedent variables relationship to overt behavior. Thus, knowing the intensity of a person's attitude is helpful in predicting his behavior toward the object of the attitude. Propositions IV, V, VII and'VIII were not statistically confirmed. One of the major reasons was that these propositions were more difficult to prove. The hypotheses for propositions I, II, III and VI were easier to prove. The statistical test for the null hypothesis demonstrated the extent of relationship. The test of the hypothesis in propositions IV, V, VII and VIII did not use the same test for the null hypothesis. The tests were 169 used to determine whether there were significant differences between sets of relationships each of which was already found to be statistically significant. Moreover, the statistical test for Tau1 -Tau2 is a conservative test. The standard error in the denominator is determined by estimating the maximum variance. For example, in testing hypothesis 26 the Tau for the com-sca1e and the overt-scale was .58. The Tau for the att-scale and overt-scale was .40. They were not found to be significantly different. The same product—moment correlation coefficients for the two sets of rela- tionships were .71 and .42 respectively. Using a test designed to estimate the extent of difference for two independent variables on the same dependent variable in the same population, the difference between rzy -rzx was found to be significant beyond the .025 level.3 In this particular test of a hypothesis it appears that the signifi- cance of the difference between Taul -Tau2 is difficult to precisely determine. This same difficulty does not appear in all the tests that were not statistically significant. In some instances the sets of relationships were in the opposite direction than predicted and in other cases the Tau's were only slightly different. Proposition IV stated that the relationship between attitude and the three antecedent variables would be influenced by the defini- tion of the situation. Proposition V stated that the relationship 3Fora discussion of the statistic see Helen M. walker and Joseph Lev, Statistical Inference (New York: Holt, Rinehart and Winston, 1953 , pp. 2 -257. 170 between attitude and commitment and attitude and overt behavior would be affected by the definition of the situation. The fact that the data did not support the corresponding hypotheses 16, 17, 18, 19 and 20 was not anticipated. The pilot study data indicated that the correlations between the att-scale and com-scale and the att-scale and overt-scale were significantly greater when the com- scale preceded the att-scale in the interview; The reSpective Tau's were .59 and .63 (N = 2”) compared to .36 and .24 (N = 22). Moreover, the antecedent variables were more strongly related to the att-scale when the com-scale preceded the att-scale.except for past experience. The Tau's for experimental condition II were -.37 (modified F-scale), .26 (past exp-scale) and .48 (PSSObscale) compared to -.14, .35 and .01 in experimental condition I. In the pilot study there were only two treatment groups. In the major study both treatments (1) and (3) were considered as experimental condition I in testing four of the five major hypotheses. Inapection of the measures of association for individual treatment groups in.Appendxi B reveals that treatment (3) does contribute slightly to the magnitude of the unanticipated findings. The effect of treatment (3), however, does not account for all the differences. When treatments (1) and (2) are compared separately the same pattern of relationships exists. Thus, the inclusion of treatment (3) data with treatment (1) data does not explain the unanticipated findings. 171 Characteristics of the data suggest that the decision to accept the null hypothesis should not be made. Three related sources of evidence are important. First, the mean scores on the attitude scale were lower in experimental condition II. This finding indicated that placing the com—scale before the att-scale did have the predicted effect on the expression of attitudes. Second, the relationships between sets of variables in the two experimental conditions were in the opposite direction than predicted. The modified F-scale is the one exception. It is a less sensitive instrument, and, therefore, it should not be expected to indicate patterns of relationships. The fact that the findings are opposite to that predicted suggests there may have been some factor other than chance causing the results. One explanation might be that the effects of experimental condition I affected the students more uniformly than experimental condition II. In experimental condition I students gave highly favorable responses to the att—scale, which resulted in moderate correlations between the att-scale and the five other variables. In experimental condition II some of the students may have also given highly favorable scores and others responding to the experimental stimulus may have significantly decreased their degree of favorability, i.e., they were consistent with their low levels of commitment. For example, if two students in experimental condition I had only moderate PSSO-scale scores and both responded veny favorably to the attitude statements, and two students in experimental condition II had only moderate PSSO-scale 172 scores and one responded very favorably to the attitude statements and the other did not, then the relationship between the two variables in experimental condition I would be greater than in experimental condition II. The third source of evidence that is related to the decision of not accepting the null hypothesis is the pilot study data. In the pilot study the results were in the predicted direction. The relationships between the att-scale with two of the three antecedent variables, commitment and overt behavior were greater in experimental condition II. The major reason for the differences between the pilot study and the major study may be the two different research settings in which the data were gathered. In the pilot study students were interviewed in their places of residence. The interviewer asked the questions and the students responded. In the major study questionnaires were distributed to classes of students. After initial instructions, students reSponded to the questions as they were phrased in the questionnaire. The form of interaction was more intense in the interview situation than in the classroom setting. The interviewers direct interaction with the students in a small two person setting could have had a greater*inf1uence on the student's responses. When confronted directly to respond to the attitude statements students may have felt greater social pressures. 173 The nature of the effects of social pressures on the eXpression of attitudes has not been adequately resolved. Edwards has argued that the method of direct questioning is inadequate when individuals are reluctant to publicly express their attitudes on controversial issues. He statedzu "Only when the social atmosphere is free from felt or’actual pressures toward conformity might we expect to obtain evidence about a person's attitudes hy means of direct questioning.“ Edwards believed that the most accurate expression of attitudes occurs in conditions of anonymity, e.g., the administration of the questionnaire in the classroom. In contrast to Edwards, Hyman took a different position. He was con- cerned with the inconsistency that was reported between attitude and overt behavior toward members of a minority group. He thought that many inconsistencies between attitude and behavior were due to the inconsistencies in the interpretations the researcher gives to attitude measurement and its relation to behavior. He thought pri- vate attitudes revealed under test conditions might not be expressed in the more normal situations of everyday life. Reasons are offered _ for the discrepancy. One is the social setting. The average testing situation is different from the normal situation of coercive social forces in which attitudes are expressed. Many testing situations provide anonymity. The respondent under such circumstances can ”Allan L. Edwards, op. cit., p. 3. 174 express himself without concern with the consequences of the expres- sion and can reSpond with various levels of reflection and concern for completing the task. In contrast to the areana of daily life, the issues in the testing situation are unloaded of their emotional overtones. He stated that if our aim is to predict a given kind of behavior in a given setting, we should design tests so that they incorporate the fundamental aspects of the setting into the tests.5 Hyman's comments on anonymity suggested that he would prefer an interview situation in which the social.pressures to conform to normative expectations would be stronger than in a classroom setting. Both authors are only partially correct. The data from both the pilot and major studies reveal that the interview (pilot study) was both the "best" and ”worst" way for collecting information on the relationship between attitude and overt behavior. In experi- mental condition I (att-scale before com-scale) the Tau for the relationship between attitude and overt behavior was .2#. In experi- mental condition II (com—scale before the att-scale) it was .63. The respective Tau's for the major study were .40 and .35. This data suggest that the more effective way to predict behavior from attitude is to measure attitude in an interpersonal setting, after having asked respondents to commit themselves to various degrees of interracial activities. The data also indicate that in experimental SChein.et al., op. cit., pp. 38-u0. 175 condition II in the pilot study the antecedents and commitment are strongly associated with attitude in this testing environment.6 The above is one interpretation for the non-significant relationships for'hypotheses in propositions IV and V. It relies heavily on the data gathered in the pilot study. The interpretation may be wrong. The data may be the result of some unknown source of error. Although it sounds trite, this area could benefit from future research. At this stage it may be incorrect to accept the hypotheses of no difference that were used to test prepositions IV and V. PrOposition VII stated that the extent of relationships between commitment and the antecedent variables would be greater than the relationships between attitude and the antecedent variables. Proposition VIII stated that the extent of relationship between commitment and overt behavior would be greater than the relation- ship between attitude and overt behavior. Hypotheses 23, 21+, 25 and 26 were used to test these two propositions and they were not statistically sigiifiwt. With respect to the antecedent variables it is interesting to note that the relationship between commitment and authoritarianism (-.l7) is less than the relationship between attitude and authoritarianism (-.27). However, authoritarianism is the poorest predictor of the three manifestations of overt behavior. The antecedent that is the best predictor of the manifestations of 6See Appendix B. 176 overt behavior is the PSSO-scale. The com-scale relationShip to the PSSO—scale (.48) was greater than the relationship between the att-scale and the PSSO-scale (.36). The pattern of relationships suggests that the relationships between the antecedent variables and commitment are stronger than the relationships between antecedent variables and attitude when the antecedent variables are better predictors of behavior. Thus, although the hypotheses of no different relating to proposition VII were not rejected, the findings indicate that the proposition may be true when antecedent variables are strongly associated with manifestations of overt behavior. The difference between commitment and overt behavior (.58) and attitude and overt behavior (.110) was not significant when Kendall's test of significance was used. The difference was in the predicted direction. It should also be noted that in the separate treatment groups the relationships between commitment and overt behavior were greater than the relationships between attitude and overt behavior (see Appendix B). This pattern of relationships suggests that proposition VIII might be true Two major factors that make it difficult to prove the preposition statistically are: the conservative test for Tau and the crude measure of overt behavior. Although the overt-scale does indicate the extent to which students were willing to engage in interaction with Negroes, it is not a very sensitive instrument. The majority of students had the same score-.zero. They were 177 unwilling to become involved with Negroes. Apparently, acts of omission predominate both inside and outside of the testing situation. How to develop a more sensitive instrument to measure white's actions toward Negroes remains an open question. Summary After examining the predicted relationships between variables a number of decisions were made. PrOpositions II, III and VI were considered to be true. Both the statistical tests of hypotheses were confirmed and the patterns of relationships were consistent. Although statistically significant, proposition I was considered to be tentative. The extent of the relationships between the modified F-scale and the antecedent variables was slight. Higher correla- tions were expected. Moreover, the extent of relationship between the modified F-scale and the att—scale was larger than the extent of relationship between the modified F-scale and com-scale. The com—scale was, however, a better predictor of overt behavior than the att-scale. The main point in evaluating the outcome of the research in terms of statistics is the pattern of results is as important as the precise level of significance achieved by a particular set of differences. The modified F-scale was found to be signi- ficantly related to the antecedent and dependent variables but the pattern of relationships was unimpressive. 178 Prepositions IV, V, VII and VIII were more difficult to confirm due to the conservative statistical test. There is conflicting evidence in the data for propositions IV and V. The pilot study data supported the predictions. The major study data did not. The experimental conditions in the major study were not effective, i.e., they did not systematically alter the student's definition of the situation. The interwiew situation in the pilot study did have the hypothesized affect on attitude. The data for prepositions VII and VIII were inconclusive. Although not statistically significant, the patterns of relation- ship between the com-scale and the antecedent variables revealed that the com-scale relationships to antecedents is greater than the att-scale, when the antecedents are moderately associated with overt behavior. The com-scale is also a better predictor of overt behavior than the att-scale, but the difference between the Tau's did not result in statistically significant differences. Using Hotelling's test for the effects of two independent variables on the same dependent variable for the same population the differ- ences were found to be significant. The theoretical guideline that was used to develop the specific prepositions stated: B = Com (antecedent variables); Att (definition of the situation and antecedent variables). The variables commitment, attitude, perceived support from significant others, past intergroup contact and authoritarianism 179 were defined as contingently necessary conditions for overt behavior. The definition of the situation was considered a contributory condition that affected the expression of attitude and was not considered a necessary condition for overt behavior. A simple way to test the various components of the general proposition was to compare each of the antecedent and independent variables with overt behavior in a two-hy-two table. The independ- ent variables were divided at the mean. The dependent variable-- overt behavior--was not divided at its mean. It was thought a more meaningful dividing point would be an indication of willingness to interact with Negro students outside of the testing situation. Therefore, everyone who had a score of 3 or more was defined as being definitely willing to engage in interracial activities. These subjects were students who said they would come to the small group discussions when they were contacted after having responded to the questionnaire. This dividing point distinguished between those who only expressed willingness in the testing situation from those who remained interested in coming to the small group discussion. As contingently necessary conditions of overt behavior one cell in a two-hy-two table was expected to have a very low frequency. This cell would contain those students who were less favorable on the antecedent and intervening variables and who were more favorable on the overt-scale (they had scores of 3 or more on the overt—scale). Tables 34, 35, 36, 37 and 38 report the distribution in two-hy-two tables. 180 TABLE 34 THE RELATIONSHIP BETWEEN AUTHORITARIANISM AND OVERT BEHAVIOR Expressed Willingness Outside the Testing Situation to Interact with Negroes . No Yes Total Authoritarianism Below i 80 29 109 Above i 100 ll . 111 Total 180 #0 220 TABLE 35 THE RELATIONSHIP BETWEEN PAST INTERGROUP CONTACT AND OVERT BEHAVIOR. Expressed Willingness Outside the Testing Situation to Interact with Negroes No Yes ' .Total Past Intergroup Below i 98 1N 112 Contact Above i 82 26 108 Total 180 “0 220 181 TABLE 36 THE RELATIONSHIP BETWEEN PERCEIVED SUPPORT FROM SIGNIFICANT OTHERS AND OVERT BEHAVIOR EXpressed Willingness Outside the Testing Situation to Interact with Negroes _ No Yes Total Perceived Below i 128 10 138 Support From - Significant Above x 52 30 82 Others TOtal _ 180 (+0 229— TABLE 37 THE RELATIONSHIP BETWEEN ATTITUIE AND OVERT BEHAVIOR Expressed Willingness Outside the Testing Situation to Interact with Negroes No Yes Total Attitude Below 56 81 6 87 Above 3? 74 28 102 Total 155 j“ 182 182 TABLE 38 THE RELATIONSHIP BETWEEN comnmmm AND ovmu BEHAVIOR Expressed Willingness Outside the Testing Situation to Interact with Negroes No Yes Total Commitment Below 56 98 4 102 Above i 54 32 86 Total 152 36 188 The data indicate very few cases of inconsistency. If students were actually willing to take part in overt behavior, they had those characteristics of the antecedent and intervening variables that were hypothesized to be related to favorable overt action toward Negroes. The largest percentage of inconsistent cases was for past intergroup contact (6.4%) and the smallest percentage was for commitment (2;1%). Thus, the antecedent and intervening variables are considered to be contingently necessary but not sufficient conditions for willingness to engage in interaction with Negro students. Conclusion The analysis of the eight propositions reveal that there was supporting evidence for propositions I, II, III, IV, V, VI, VII and VIII. The continued use of the variable authoritarianism within the theoretical structure is questionable. Its strength of 183 association with other variables was slight. Its pattern of relationships reveals that it does not fit the pattern of relation- ships that was expected. Prepositions II, III and VI can be accepted as true. The data revealed that both the statistical test and the pattern of relationships were consistent with the predictions. The patterns of relationships for propositions IV and V reveal that the propositions may be true when tested in experimental situations that can successfully manipulate the definition of the situation. Preposition VII may be true when antecedent variables are strongly associated with overt behavior. The patterns of relationShip suggest that preposition VIII is probably true, although the stringent test of significance does not reveal statistically significant differences. The analysis of the components of the general proposition indicated that the three antecedent variables and the two intervening variables were contingently necessary conditions for willingness to engage in overt behavior with Negroes outside of the testing situation. CHAPTER VI SUMMARY AND CONCLUSIONS Overview This study was designed to investigate the relationship between attitude and overt behavior. The task was not as simple as it first appeared. Research on white's attitudes and actions toward members of minority groups reported inconsistencies. Research designs were previously developed to disprove a monistic relationship. The majority of studies diSproved that attitude was the necessary and sufficient condition for particular patterns of overt behavior. This was often the extent of the argument. Other determinants of overt behavior were used to describe why incon- sistencies existed but the determinants were not measured and/or controlled. In a few studies additional variables were measured and/or controlled to explain why patterns of overt hehavior were not dependent upon attitude. Examples of this type of research are studies hy DeFleur and westie, Linn, Bray'and.Pettigrew; Although valuable, these studies were limited in scope. One additional determinant was selected to demonstrate that it was either a better predictor of overt behavior or a major factor that caused the inconsisteney between attitude and overt behavior. 184 185 Following the advice of Suchman, Dean and Williams, it was decided to study the problem with what the authors called the complex-adequate hypothesis. It outlines several interrelated variables, all of which are necessary conditions for the consequent condition to be predicted and/or explained, but none of which alone is a sufficient condition. The advantage of this approach is that it attempts an optimal decision between the rigor of the simple, analytic abstract hypothesis and the complexity of concrete events.1 A middle-range theory was develOped to explain the degree of consistency between attitude and overt behavior. One general preposition in the field of social psychology that had frequently been used to explain patterns of overt behavior was: behavior’is a function of both individually acquired characteristics and the social context in which the behavior is expressed. A specific derivation of this preposition was culled from the literature. In its final form the preposition stated: B = Com (antecedent characteristics of the individual); Att (definition of the situa- tion and antecedent characteristics of the individual). It was used as a guideline to examine white students' attitudes and overt behavior toward Negro students. Three independent variables (antecedent characteristics) were derived from the literature in social psychology. They A were: authoritarianism as measured hy the modified F—scale, 1See above pp. 20-21.. 186 past intergroup contact with Negroes, and perceived support of significant others for engaging in interaction with Negroes. The fourth independent variable was the definition of the situation (concurrent condition in the social situation). Two factors that influenced the definition of the situation were postulated. They were the normative expectations of being a research reSpondent and the normative expectations to express liberal racial attitudes in a college environment. Experimental conditions were designed to provide stimuli that would systematically alter the definition of the situation. The conditions consisted of altering the order of presentation of the commitment and.attitude scales. These two variables were defined as intervening determinants between the four independent variables and overt behavior. Eight propositions and twenty-six hypotheseS'were developed to explain the resultant overt behavior and the relationships between determinants. In order to test the theory a verification study was designed. White students at.MHchigan State University were used as the sample for both the pilot and major studies. In the pilot study a systematic sample of sephomores, juniors and seniors were interviewed in their places of residence. In the major study students in four‘intro- ductory classes of social psychology were administered a question- naire in the classroom. Data were analyzed for 46 students in the pilot study and 263 students in the major study. 187 Scales were designed to measure the five determinants of overt behavior and they were found to be both reliable and valid. The eXperimental stimuli in the pilot study had the predicted effect on the expression of attitudes. Small group discussions designed to bring Negro and white students together to improve interracial relations in the campus community were used as the measure of overt behavior. Although it was a crude measure of overt behavior, it was found to be a valid resultant of the determinants. Therefore, the operationalization of the deter- minants and resultant were successful. Major Findings The general proposition of the study was found to be true. The three antecedent and the two intervening variables were found to ”approach” necessary conditions for expressed willingness to take part in interaction with Negroes, i.e., they were not absolute necessary conditions but contingent necessary conditions. The antecedent variables were found to be interrelated and the combined influence on overt behavior was greater than the individual effects in two out of three cases. The combined influence was greater than the individual influence of authori- tarianism and past intergroup contact, but it was not as great as perceived support from significant others. The antecedents also influenced the intervening variables which were the better predictors of overt behavior. These findings suggest the utility of a multivariate approach. 188 As a contributory condition influencing the expression of attitude, the definition of the situation was not found to be an important determinant of attitude in the major study. The controlled stimuli in one experimental condition did not make the expression of attitude more consistent with commitment, overt behavior and the antecedent variables. Propositions IV and V stated that the definition of the testing situation was an important contributory condition. They were not relegated to the limbo of unconfirmed theoretical statements. The pilot study data suggested that the propositions might be true in a carefully controlled experimental environment. As a determinant of overt behavior, attitude was both a consistent and inconsistent predictor. The determination of consistency or inconsisteney depends in part upon the definition of inconsistency and the evidence that is marShalled to prove the argument. Campbell is correct in his criticism of researchers who have demonstrated the inconsistency between attitude and overt behavior. Generally the researcher's argument is that attitudes are inconsistent with overt behavior if attitude and overt behavior are not isomorphic. Using this argument the data in the major study was very inconsistent with overt behavior. Only one out of 189 students who responded to both the att- and overt-scales, had a total score on the att-scale that indicated a negative attitude toward Negro students. However, only 19 of 189 the 188 students who had positive attitudes attended the small group discussions that were presented as an opportunity to improve interracial relations. The above findings do indicate that there is not an isomorphic relationship but they do not prove that the relationship is inconsistent. Campbell asserted that isomorphic relationships do not exist between different situational thresholds. The eXpression of favorable or unfavorable attitudes in the ”play environment" of a testing situation is quite different than the eXpression of favorable or unfavorable acts directly toward Negroes. In this study the data revealed that the expression of attitude was consistent with the expression of overt behavior. The intensity of attitude was found to be significantly related to the intensity of commitment and overt behavior. The intensity of attitude was also found to be significantly related to the intensity of the antecedent variables. Moreover, only 6 cases of true inconsistency, i.e., willingness to interact with Negroes outside the testing situation and scoring below the mean on the att-scale, were found in 189 cases. As an explanatory variable attitude was a contingently necessary condition and a relatively good predictor of overt behavior. Thus, attitude was not incon- sistent with overt behavior. In the major study the best single determinant of overt behavior was commitment. The data indicated the strongest relation- ship and the least inconsistency. Propositions VII and VIII were 190 designed to demonstrate that commitment was both more strongly related to the antecedents and overt behavior than attitude. Due partially to the stringent test for Tau1 -Tau2, the propositions were not confirmed. However, the pattern of relationships suggested that the proposition might be true if the antecedent variables are moderately related to overt behavior and if a less stringent test is used to determine the significance of the data. The development of the middle-range theory had several advantages. The multivariate approach did reveal patterns of relationships that could not have been determined by analyzing isolated hypotheses. For example, each of the variables in simple bivariate distributions could not have been compared and evaluated in terms of the other determinants that were related to the variables. The attempt to relate the outcome of a single bie variate analysis to a larger body of theory would have been difficult. The relationship of each additional variable to attitude and overt behavior would be uncertain. Little evidence would be available to determine how a particular hypothesis was deduced or what followed from it. This research provided information to fill an important research gap. The majority of studies on attitude have not attempted to measure the overt behavior correlate of attitude. Attitude is either considered primarily as a dependent variable or no causal relationship is predicted. Those studies that 191 have attempted to measure overt behavior have had difficulty in establishing a valid operational definition. Indirect measures of overt behavior have often been used. They attempt to measure some limited form of commitment to interaction with the object of the attitude. In this research careful consideration was given to the development of both a commitment and overt action measure. The commitment scale items were structured to imply the definite possibility of future interaction with Negroes. In order to measure overt behavior a series of small group discussions were organized. These discussions were designed to improve interracial understanding. The extent of wanting to become involved in improving interracial understanding was used as a measure of overt behavior. ,Although these discussions were difficult to arrange, they did reveal information that was worth the investment of time and resources. The data did provide some evidence to evaluate the national trend of expressing more favorable attitudes toward Negroes. Like the findings of the studies reported in "Chapter I“ the data reveal that favorable attitudinal responses are made to statements that imply little or no direct contact and imply no involvement in planned actions designed to bring about a change in interracial relations. Although there has been a major shift from predominately unfavorable to favorable attitudes, this Shift does not indicate that whites have significantly changed their patterns of overt behavior toward Negroes. 192 In this particular content area it may be useful to consider favorable attitudinal responses as symbol acts. Favorable reSponses paraphrase the perceived values and norms of a society or community and are not the sufficient condition for perceived plans for action. The degree of the intensity of the favorable responses, however, does reveal information that can be used to predict degrees of favorable commitment and overt behavior. Limitations The major limitation of this study was the sample that was used to verify the theory. In the pilot study a small systematic sample of sophomores, juniors and seniors were selected. Four introductory classes in social psychology were selected for the major study. The empirical findings can only be generalized to similar Michigan State University white college students. Previous research, however, indicates that similar results were obtained on one or more of the hypotheses in the theoretical framework. DeFleur and.Westie found a consistent relationship between attitude and commitment. Linn found that students were affected by the liberal norms on racial topics in a college environ- ment. He also found that students tended to have more favorable attitudinal eXpressions than overt behavioral eXpressions. In a national sample of whites, Harris and Brink found that whites with previous social contact had more favorable attitudes toward Negroes than those who did not have previous social contact. 193 Zetterberg suggested a principle to use in extrapolation 2 ”It is more beyond the sc0pe of empirical data. He stated: probable that a hypothesis holds true outside the population on which it has been confirmed than that the contrary of the hypothesis holds true in the new population.” While this principle may not be correct for a single hypothesis, it probably does hold for a series of hypotheses interrelated through prepositions. The probability of a single hypothesis being confirmed because of chance differences or a systematic bias is much greater than it is for a series of hypotheses. It is probably true that the sets of relationships described in this study extend to non-student populations. The relative degree of favorable attitude, commitment, overt behavior, past intergroup contact and perceived support from.significant others is probably much less for a heterogeneous population. Members of communities in various regions of the country have more invested in the institutionalized norms that directly or indirectly contributed to the development of unfavorable manifestations of behavior. These norms restrict the range of favorable influence to which the relatively “unattached" students are exposed. Although the degree of favorability may be much less, the patterns of relationship would not be eXpected to contradict those discovered in this research. 3 ZZetterberg, @. cit., p. 128. 19h Implications for’Future Research This research attempted to establish closure on an old but persistent problem in the field of social psychology, however, the findings suggest possibilities for future research. Different patterns of relationships were discovered in the interview and classroom research settings. Students were more subject to the experimental controls in face—to-face interaction than in the classroom. One possible explanation is students had less anonymity and were exposed to what they perceived to be social pressures in the interview situations. The contrasting findings in the pilot and major studies indicate that the factors that affect the definition of the testing situation need to be carefully examined. In the major study the experimental stimuli were not carefully controlled. Students were given initial instruction to answer the questions in their order of appearance. However, some students were observed paging through the questionnaires before answering the questions, and in a few cases they discussed either their reSponse patterns or the questions with fellow classmates. One way to maintain the experimental controls would be to distribute the questionnaire in various sections so that the subjects could not be aware of what followed a particular scale. The interaction between students could be controlled if more emphasis was placed on the ”scientific nature' of the experiment, stressing the importance of obtaining independent observations. 195 Additional data on the experimental conditions in the interview situation are needed. The N sizes in the pilot study were small. Further evidence would help to evaluate the importance of the definition of the situation on the expression of attitudes in an interview setting. Although the general theoretical orientation is considered to be applicable for a non-student population it needs to be tested. Major difficulties exist in the operationalization of commitment and overt behavior for a community study. One possibility would be a community that had an.active program developed to improve interracial understanding, e.g., visiting days promoted by church groups or dialogues sponsored by the Southern Christian Leadership Conference. A sample of respondents could be interviewed prior to some publicized activity. The results of the interview could be compared with.the actual participation. One crucial factor would be the necessity of establishing the close cooperation of those Negro and white leaders who directed the campaign. Before beginning another study it would be helpful to develop a weighting system for each of the determinants. The data revealed that they should not be assigned equal importance as predictors of overt behavior. Commitment appears to account for more variation on the expression of overt behavior than attitude. The precise amount cannot be determined because the 196 variance cannot be adequately computed and the extent of inter- action between determinants is not known. The data does suggest that the better predictors are those questions that are more directly related to the pattern of behavior under consideration. EXploring different ranking criteria might lead to a composite index that would be the most efficient predictor. Further internal analysis of the scales measuring the antecedent and intervening variables may reveal important sube scales that have theoretical significance. For example, the total score for the perceived support of significant others was used throughout the analysis. The support from peers may have added significance for some students and not others. It is conceivable that there are full-time students living within the college community who positively evaluate only the attitudes and actions of significant others outside of the university setting. These students may be less favorable in their attitudes and actions. Another area of future exploration is the relationship between commitment, attitude and overt behavior in other content areas. It was found that whites' attitudes toward members of a minority group was not as good a predictor of overt behavior as commitment. It would be interesting to discover if this same pattern of relationship holds true for different attitudes. 197 Traditional techniques for measuring attitudes may measure hypothetical orientations to behavior. Reaponses to attitude statements may be made without taking into consideration important contingencies surrounding possible plans for engaging in behavior. In contrast commitment involves making a decision with regard to some particular line of action that has consequences for other interests and activities. The decision is dependent not only on the antecedent characteristics of the respondents and their attitudes, but it is also based on the perceived contingencies surrounding the consequences of behavior. Attitude scales may only measure what respondents feel they ”ought“ to express and not what they actually would plan to do if forced to make a decision. Particular content areas may exhibit a greater or'lesser correSpondence between attitude, commitment and overt behavior than manifestations of behavior of whites toward minority group members, e.g., attitude toward occupational achievement or religious institutions and beliefs. The theoretical significance of commitment has not been adequately explored in the field of social psychology. The lack of investigation may be the oversight of the professionally disinterested or alienated social scientists who lack commitment to patterns of behavior outside of their'professional role. Another possibility is the difficulty of its operationalization. 198 A third difficulty is ethical considerations. Is it ethical to have respondents commit themselves to socially defined unethical behavior? Is it ethical to measure commitment without providing respondents the opportunity to carry out their decisions? Nevertheless, the researcher interested in predicting various patterns of overt behavior may find that a measure of commitment is the best predictor of the overt behavior under investigation. The range and relative intensity of various factors that affect the decision to become involved in a particular pattern of behavior with perceived consequences need to be carefully examined. One important variable that was found to be significantly related to commitment was the degree of perceived support from significant others. APPENDIX A A Survey of Student Attitudes and.Actions Toward Negroes Supervisor: Dr. Santo F. Camilleri Assistant: Mr. Jim Fendrich 355—6637 Student Number Schedule Number I / 0°15- This survey is interested in your attitudes and actions. Only those directly involved in the research project have access to your name. They will ngt.reveal your name to anyone who is not directly involved. To begin with, we would like to have your responses to a number of Opinions which many segments of the general public feel to be important socially and personally. The best answer to each statement is your personal opinion. The statements cover many different and opposing points of view. You may find yourself agreeing strongly: with some of the statements and disagreeing just as strongly with others, and perhaps uncertain about other statements. You can be sure that many people feel the same as you do. Please answer all the statements. There are six possible responses to all of the statements. The responses are: Strongly.Agree ='SA Agree = A Agree in Part =.A-P Disagree in Part = D-P Disagree Strongly Disagree = SD Please answer every statement by circling the response with which you agree. [The Modified F-Scale] 1. Some day it will probably be shown that astrology can Col. 4 explain a lot of things. SA A A-P D4? D SD +3 +2 +1 -1 -2 “3 199 200 2. When a person has a problem or worry, it is best for him to think about doing something about it, not be distracted by more cheerful things. SA A A—P D—P D SD -3 -2 -1 +1 +2 +3 3. Science has its place, but there are many important things that can never possibly be understood by the human mind. SA A A-P D-P D SD +3 +2 +1 -1 -2 '3 4. Sex criminals such as those who rape and attack child— ren should be treated as sick people, not publicly whipped or worse. SA A A-P D-P D SD -3 -2 -1 +1 +2 +3 5. Homosexuals are not criminals and should not be punished. SA A A-P D-P D SD -3 -2 -1 +1 +2 +3 6. Nowadays more and more people are prying into matters that should remain personal and private. SA A A-P D-P D SD +3 +2 +1 —1 -2 -3 Some of the goings-on in this country, even in places where peOple might least expect it, are tame compared to the wild sex life of the Greeks and Romans. 7. SA A A—P D-P D SD --3 -2 -1 +1 +2 +3 8. No one is born with an urge to jump from high places. SA A A—P D-P D SD -3 -2 --1 +1 +2 +3 9. Wars and social troubles may someday be ended by wisdom and education, not by an earthquake or flood that will destroy the whole world. SA A A-P _3 —2 -l D-P D SD +1 +2 +3 Col. 5 Col. 6 Col. 7 Col. 8 Col. 9 Col . 10 C01 . ll Col. 12 ll.‘ 10. 11. 12. 13. 14. 15. 16. 17. 18. 201 If people would talk less and work more, everybody would be better off. SA A A-P D-P D SD +3 +2 +1 -1 -2 '3 Familiarity does 323 breed contempt. SA A A-P D-P D SD —3 -2 -1 +1 +2 +3 There is hardly anything lower than a person who does not feel a great love, gratitude, and reSpect for his parents . SA A A-P D—P D SD +3 +2 +1 -1 -2 -3 Obedience and reSpect for authority are the most important virtues children should learn. SA A A-P D-P D SD +3 +2 +1 -1 -2 '3 The artist and the professor are much more important to society than the businessman and the manufacturer. SA A A-P D-P D SD -3 -2 -1 +1 +2 +3 An insult to one's honor should always be punished. SA A A-P D-P D SD +3 +2 +1 -1 -2 '3 Those people who were in positions of authority before Germany entered world war II should not have been used to keep order and prevent chaos after the war. SA A A—P D4? D SD -3 -2 -1 +1 +2 +3 Most of our social problems would be solved if we could somehow get rid of the immoral, crooked, and feeble-minded people. SA A A-P D-P D SD +3 +2 +1 -1 -2 ’3 'Weaknesses and difficulties can hold us back; sheer will power is not enough to overcome difficulties. SA A A-P D-P D SD -3 -2 ”-1 +1 +2 +3 Col. 13 Col. 14 C01. 15 Col. 16 \j’ Col. 1? Col. 18 Col. 19 C01. 20 Col. 21 19. 20. 22. 23. 24. 25. 202 What this country needs most, even more than laws and political programs, is a few courageous, tireless, devoted leaders in whom the people can put their faith. SA A A—P D-P D SD +3 +2 +1 -1 -2 -3 No one ever learned anything really important through suffering. SA A A-P D-P D SD -3 -2 -1 +1 +2 +3 No sane, normal, decent person could ever think of hurting a close friend or relative. SA A A-P D-P D SD +3 +2 +1 -1 -2 -3 People cannot be divided into two distinct classes, the weak and the strong. SA A A-P D-P D SD -3 -2 -1 +1 +2 +3 What youth needs today is strict discipline, rugged determination, and the will to work and fight for family and country. SA A A-P D-P D SD +3 +2 +1 -1 -2 -3 No person should have complete faith in some super- natural power whose decisions he obeys without question. SA A A-P D-P D SD -3 -2 —1 +1 +2 +3 Young people sometimes get rebellious ideas, but as they grow up they ought to get over them and settle down. SA A A-P D-P D SD +3 +2 +1 -1 -2 -3 A person who has bad manners, habits and breeding can hardly eXpect to get along with decent people. SA A A-P D-P D SD +3 +2 +1 -1 -2 ’3 Col. 22 Col. 23 Col. 2“ Col. 25 Col. 26 Col. 27 Col. 28 Col. 29 203 27. A person does not have to worry about catching an infection or disease just because many different kinds of people move around and mix together a great deal nowadays. SA A A-P D-P D SD -3 -2 -1 +1 +2 +3 28. Most people don't realize how much our lives are controlled by plots hatched in secret places. SA A A-P D-P D SD +3 +2 +1 -1 -2 -3 29. Human nature being what it is, universal peace will come about eventually. SA A A-P DuP D SD ~3 -2 -1 +1 +2 +3 (PLEASE CHECK TO SEE IF YOU HAVE ANSWERED ALL OF THE RESPONSES) Col. 30 Col. 31 Col. 32 0013. 33- 35 This next task is different. about your past experience with Negroes. 204 [The Past ESCperience Scale] statements by circling Yes or No. l. 7. 10. During the last year did you ever sit next to a Negro in class whom you got to know? Yes No Do any of your close friends have close friends who are Negro? Yes No During the last 30 days have you sat down to eat at the same table with a Negro? Yes No During the past year did you work with a Negro? Yes No During the past year did you study with a Negro? Yes No Do you have close friends of the same sex who are Negro? Yes No Do you have close friends of the opposite sex who are Negro? Yes No In the past year have you gone to a social activity with a Negro of the same sex? Yes No In the past year have you gone to a social activity with a Negro of the opposite sex? Yes No During the past 30 days have you spent 5 minutes or more talking with a Negro student? Yes No Col. Col. Col. C01. C01. Col. Col. C01. C01. Col. we would like some factual information Please answer all the 37 38 39 41 #2 43 45 ll. 12. 13. 14. 15. 16. 17. 205 Have you ever taken part in any demonstration spon- sored by a Negro group? Yes No Do you know the name of the president of the campus NAACP? Yes No In the last year have you participated in a small social gathering that was mixed racially? Yes No Would you say that your past experiences on campus with Negroes have been favorable? Yes No Could you briefly describe some of these favorable experiences. First eXperience: Second eXperience: Third experience: would you say that your past experiences on campus with Negroes have been unfavorable? Yes No Could you briefly describe some of these unfavorable experiences. First experience: Second experience: Third experience: Col. 47 C01. #8 Col. 49 Cole. 50- 51 Col. 52 Col. 53 C01. 5“ Col. 55 C018. 56— 57 Col. 58 Col. 59 Col. 60 Col. 61 C013. 62+ 65 (PLEASE CHECK TO SEE IF YOU HAVE ANSWERED ALL OF THE RESPONSES) 206 [The Perceived Support for Significant Other Scale] The following are some statements about participation in inter- racial activities. know might do if they were asked to participate. responses to each question. circling either Yes-—Maybe——No. 1. Would the following people you know agree to go to coffee or lunch with a mixed racial group to talk about interracial problems? a) b) C) d) 8) Your closest friend of the same sex Your closest friend of the opposite sex Your parents Your roommate (or husband or wife) Some older person whom you reSpect, e.g., professor, minister, uncle, etc., please Specify would the following people you know agree to have a Negro as a neighbor? a) b) o) d) Your closest friend of the same sex Your closest friend of the opposite sex Your parents Your roommate (or husband or wife) Some older person whom you respect, e.g., professor, minister, uncle, etc., please specify Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes 1 You are now asked to indicate what people you There are five Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe 2 Please answer each statement by Col. Col. Col. Col. Col. Col. Col. Col. Col. Col. 10 11 12 13 14 207 would the following people you know agree to spend a weekend at the home of a Negro if the Negro invited them? a) Your closest friend of the same sex b) Your closest friend of the opposite sex 0) Your parents d) Your roommate (or husband or wife) 9) Some older person whom you respect, e.g., professor, minister, uncle, etc., please specify would the following people you know agree to invite a Negro to spend a weekend at their home? a) Your closest friend of the same sex b) Your closest friend of the opposite sex c) Your parents d) Your roommate (or husband or wife) e) Some older person whom you respect, e.g., professor, minister, uncle, etc., please specify Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes 1 Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe 2 Col. Col. C01. C01. Col. Col. Col. Col. Col. Col. 15 20 24 5. 208 would the following people you know agree to participate in a small group discussion on the topic of Whites' social relations with Negroes? a) Your closest friend of the same sex b) Your closest friend of the opposite sex 0) Your parents d) Your roommate (or husband or wife) e) Some older person whom you respect, e.g., professor, minister, uncle, etc., please specify would the following people you know agree to attend a lecture or conference on the topic of Whites' social relations with Negroes? a) Your closest friend of the same sex b) Your closest friend of the opposite sex c) Your parents d) Your roommate (or husband or wife) e) Some older person whom you reSpect, e.g., professor, minister, uncle, etc., please Specify Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes 1 Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe 2 Col. Col. Col. Col. Col. Col. C01. C01. Col. Col. 25 27 30 31 32 33 3h 209 Would the following people you know agree to protest against segregated housing in their home town? a) Your closest friend of the same sex b) Your closest friend of the opposite sex 0) Your parents d) Your roommate (or husband or wife) e) Some older person whom you respect, e.g., professor, minister, uncle, etc., please Specify would the following people you know agree to attend a regular meeting of a chapter of the NAACP? a) Your closest friend of the same sex b) Your closest friend of the opposite sex c) Your parents d) Your roommate (or husband or wife) 9) Some older person whom.you reSpect, e.g., professor, minister, uncle, etc., please specify Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes 1 Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe 2 Col. Col. Col. Col. Col. Col. Col. Col. ,Col. ‘Col. 35 37 38 39 40 41 42 “3 210 9. Would the following people you know agree to contribute $1.00 to help finance the activities of a Negro action group such as (SNCC) (NAACP) (CORE)? a) Your closest friend of the same sex Yes Maybe No Col. 45 1 2 3 b) Your closest friend of the Yes Maybe No Col. 46 opposite sex 1 2 3 c) Your parents Yes Maybe No Col. 47 l 2 3 d) Your roommate (or husband or wife) Yes Maybe No Col. 48 l 2 3 e) Some older person whom you respect, Yes Maybe No Col. 49 e.g., professor, minister, uncle, l 2 3 etc., please Specify Cole. 50- 52 (PLEASE CHECK TO SEE IF YOU HAVE.ANSWERED.ALL OF THE RESPONSES) 211 [The Attitude_Sca1e] Attitudes Toward Negroes on the M.S.U. Campus Use the Same ReSponses as Indicated on the lstgpage! Please answer each of the statements by circling the response with which.you agree. 1. I think there are Negroes qualified to be class presidents. SA A AeP D-P D SD +3 +2 +1 —1 -2 -3 2. Negro students all look alike. SA A AeP DeP D SD -3 -2 -1 +1 +2 +3 3. I think research would show that Negroes definitely' are only capable of getting poorer grades than white students. SA A A-P DeP D SD ,3 -2 e1 +1 +2 +3 4. I wouldn't want Negroes in positions of responsible student leadership on campus. SA A A—P D-P D so -3 -2 .1 +1 +2 +3 5. I wouldn't mind at all if I lived in an area that was integrated. SA A AeP DeP D SD +3 +2 +1 -1 -2 -3 6. I find some Negroes attractive. SA A A-P D—P D SD +3 +2 +1 -1 -2 -3 7. Negroes on campus want too much. SA A AeP D-P D SD +3 -2 -1 +1 +2 +3 8. I would like to go on a double date with a Negro couple. SA A A-P 9.? D SD +3 +2 +1 .1 -2 -3 Col. Col. Col. C01. C01. C01. C01. C01. 10 11 12 9. 10. 11. 12. 13. 14. 15. 16. 17. 212 I would feel extremely uncomfortable dancing with a Negro student. SA A AeP D-P D SD -3 -2 -1 +1 +2 +3 Negroes are better in Sports because they come from more primitive backgrounds. SA A AeP D-P D SD -3 -2 -1 +1 +2 +3 Eating at the same table with a Negro wouldn't bother me. SA A AMP DeP D SD +3 +2 +1 _1 -2 -3 It would be a good experience to get to know more Negroes on campus. SA A A-P DmP D SD +3 +2 +1 -1 -2 -3 Negro students don't take care of their personal hygiene. SA A A—P DeP D SD :3 e2 -1 +1 +2 +3 I'd hate to be seen walking across campus alone with a Negro. SA A AeP D-P D SD -3 -2 -1 +1 +2 +3 Negroes should stick to themselves. SA A A-P DeP D SD -3 ~2 -1 +1 +2 +3 Any white student is better than a Negro student. SA A An? D—P D SD +3 e2 .1 +1 +2 +3 No one forgets so easily as a Negro student. SA A A-P DeP D SD -3 -2 -1 +1 +2 +3 C01. C01. 001. Col. C01. 001. Col. Col. Col. 17 18 19 20 l8. 19. 20. 22. 23. 24. 25. 27. 213 When given a chance Negroes can do just as well in school as anyone else. SA A A-P DeP D SD +3 +2 e+1 —1 +2. +3 I wouldn't want to see a Negro president of.AUSG. SA A A—P D-P D SD +3 -2 -1 +1 +2 +3 I would attend a small group discussion to become more informed about NegroeWhite relations on campus. SA A A-P DeP D so +3 +2 +1 -1 -2 -3 Only unprincipled students would go on an interracial date. SA A AmP D-P D SD -3 -2 -1 +1 +2 +3 I would like to see Negroes get equal treatment in all areas of campus life. SA A A-P D-P D SD +3 +2 +1 -1 -2 -3 Negroes want the same things out of life that I do. SA A AeP 0.? D SD +3 +2 +1 -1 -2 -3 The more Negroes come to MSU the lower the standards get. SA A An? DeP D SD _3 -2 -1 +1 +2 +3 I wouldn't mind working with Negroes on some campus project. SA A AeP DeP D SD +3 +2 +1 -1 -2 .3 I hate to see a white and Negro going steady together. SA A A-P D-P D SD -3 -2 -1 +1 +2 +3 I would prefer sharing living quarters with any white rather than with a Negro student. SA A A-P D-P D SD -3 -2 -1 +1 +2 +3 Col. ,4/ Col. Col. C01. C01. Col. 0010 Col. Col. Col. 22 23 24 25 30 31 280 30. 31. 32. 33. 214 I think the only thing that Negroes can contribute to campus life is better athletics. SA A An? D-P D SD -3 -2 _1 +1 +2 +3 The only way that Negro students can obtain full equality on campus is through the help of white students. SA A AeP DeP D SD +3 +2 +1 -1 -2 -3 The more Negro professors we get on campus the lower will be the quality of teaching. SA A A-P DeP D SD -3 -2 -1 +1 +2 +3 The reason why Negroes want fraternities and sororities of their own is that they want to stay be themselves. SA A As? DeP D SD -3 -2 -1 +1 +2 +3 I think there are Negroes on campus who will be more successful in the future than I will. SA A AeP D-P D SD +3 +2 +1 -1 e2 —3 Some Negro students are smarter than I am. SA A AeP DeP D SD +3 +2 +1 .1 -2 -3 Col. 32 Col. 33 Col. 34 V Col. 35 (PLEASE CHECK TO SEE IF YOU HAVE ANSWERED ALL OF THE RESPONSES) 215 [The Commitment Scale] Here is a different kind of question involving your possible future experience with Negroes on campus. If programs or activities could be set up to help improve inter- racial understanding on campus would you consider participating? Please check your responses. Yes No Maybe Col. 60 A If called would you commit yourself to participate in any of the following possible activities. Please circle your response. 1. WOuld you agree to go to coffee or lunch with a mixed racial group of students to talk about interracial problems on campus? 2. Would you agree to Spend a weekend at the home of a Negro attending M.S.U. if he or she invited you? 3. ‘Wbuld you agree to invite a Negro at M.S.U. to spend a weekend at your home? 4. Wbuld you agree to participate in a small group discussion on the topic of white students' social relations with Negroes on campus? 5. ‘Would you agree to attend a conference on the tOpic of white students' social relations with Negroes on campus? 6. 'Wbuld you agree to protest against segregated housing in East Lansing with Negro students? 7. would you agree to attend a meeting of the Campus Chapter of the NAACP? Yes Maybe No 1 Yes Yes Yes Yes Yes Yes 1 2 Maybe Maybe Maybe Maybe Maybe Maybe 2 3 Col. 61 _ v" Col. 62 Col. 63 Col. 64 Col. 65 V/ Col. 66 L/ Col. 67 216 8. would you if asked agree to contri— Yes Maybe No bute $1.00 to help finance the 1 2 activities of a Negro action group such as (SNCC) (NAACP) (CORE)? 3 9. would you agree to work on an inte- Yes Maybe No grated committee for collecting funds l 2 to rebuild churches in Mississippi? If you have answered Yes or Maybe to any of the above questions, please give your phone number. 3 Col. 68 Col. 69 L/ Cols. 70- 71 Col. 72 I. «3‘ l - 5‘ .I . 217 [The Overt Behavior Scale] 1. Some students are planning small group discussions on inter- racial relation§_with members of the Campus Chapter of the NAACP. The discussion would be an orientation to what students on campus are doing to improve race relations. ‘Wbuld.you agree to attend one of these meetings? Yes No 2. If you have answered Yes would you please indicate which one of the following meetings you will attend. The meetings will be held in Room 319 Berkey Hall on: Sunday, January 2hth at 4:00 p.m. and 8:00 p.m. Monday, January 25th at 4:00 p.m. and 8:00 p.m. Tuesday, January 26th at 4:00 p.m. and 8:00 p.m. Day Time Col. 73 Col. 74 Col. 75 (Questions asked after questionnaire distribution by telephone) 3. "Do you still plan to attend one of the small group discussions?” Yes Maybe No A. Actual attendance at the small group discussion. 1. student attended 2. student did not attend 5. Students who attended further committed themselves to participate in interracial activities. 1. students signed up to participate in interracial activities 2. students did not sign up to participate in interracial activities Col. 76 C01. 77 Col. 78 218 we would now like some personal information. Name: Sex: Male Female Are you a U.S. citizen? Yes No Where is your home town? City State Where are you now residing? On campus E. Lansing Lansing Other than Lansing, please specify ’ How old are you? Are you single ? or married ? What is your major in school? —(Please be specific) How many credits are you taking this term? ___0-6 ;__7-9 _10-12 _13-15 _16 or more What is your all-university grade point average? What campus organizations do you belong to? l. 2. 3. What is the occupation of’your father or the major wage- earner in.your family of origin? (Please be Specific) What church do you belong to? Is there any general or specific comments you would care to make about the questions that were asked? THANK YOU FORNYOUR COOPERATION! Col. Col. Col. C01. C01. Col. Col. Col. C01. C01. Col. Col. Col. Col. Col. Col. 41 42 1+3 45 47 5O 51 52 53 54 55 APPENDIX B Pilot Study Definitions 1 = Modified F-scale 2 = Past EXperience scale 3 = Perceived Support of Significant Others scale 4 = Attitude scale 5 = Commitment scale 6 = Overt Behavior scale Treatment (1) Attitude before Treatment (2) Commitment before Commitment Attitude N = 232 N = :¥+ vari— Pearson vari- Pearson ables r Tau Gamma _f ables r Tau Gamma 1-2 -.20 -.08 -.11 1-2 -.22 -.16 -.20 1-3 .07 .13 .19 1-3 -.52 -.44 -.59 1-4 -.29 —.23 -.34 1.4 —.47 -.37 -.49 1-5 -.16 .14 .19 1-5 -.52 -.43 -.55 2~3 .20 .19 .22 2—3 .11 .ll .12 2.4 .44 .35 .40 2.4 .32 .26 .30 2-5 .35 .27 .29 2.5 .34 —.25 -.27 3-4 .05 .01 .01 3-4 .55 .48 .58 3-5 .19 .18 .20 3-5 .75 .62 .73 4-5 .37 . 36 .41 4.5 .72 .59 .69 1-6 -.13 -.14 -.21 1-6 -.38 «.33 e.44 2.6 .02 .01 .02 2.6 .33 .30 .34 3-6 .12 .09 .12 3-6 .67 .60 .69 4-6 .33 .24 .31 4.6 .73 .63 .74 15:6 .28 .23, .27_ 5-6 AZ;__ .58 .68 220 Major Study Treatment (1) Attitude before Treatment (2) Commitment before Commitment Attitude = 92 N = 93 Vari- Pearson —'Va§i- Pearson _ables r Tau Gamma ables r Tau Gamma 1-2 -.14 -.ll -.13 1-2 -.11 -.08 -.09 1-3 -.10 -.08 -.09 1—3 -.27 -.19 -.22 1.4 -.30 -.28 -.33 1-4 -.34 -.29 -.35 1-5 -.24 -.l6 -.19 1-5 -.24 -.18 -.21 2-3 .46 .32 .35 2.3 .39 .26 .28 2-4 .40 .31 .35 2.4 .40 .31 .34 2-5 .53 .41 .46 2.5 .52 .38 .41 3-).“ '57 045 049 3.1+ .40 .29 031 3-5 .61 .45 .47 3-5 .67 .51 .54 4-5 .65 .54 .61 4-5 .61 .49 .54 :: 79 N = 78 1-6 -.19 -.08 -.11 1-6 -.26 -.21 -.36 246 .24 .17 .23 2-6 .39 .30 .46 3-6 .49 .39 .53 3.6 .41 .35 .57 4-6 .46 .44 .60 4-6 .37 .35 .56 5.6 .69____ .57 _ .73 5-6 .21 .60 .BL Treatment (3) Attitude only N = 39 variables Pearson r —F7Tau Gamma 1-2 -.52 —.30 -.35 1-3 -.37 -.21 -.23 1-4 -.tu+ -.33 -.38 2-3 .79 .59 ~63 '2-4 .72 .58 .64 3.4 .62 .44 .47 = 32 1-6 -.30 -.21 -.28 2.6 .57 .41 .52 3-6 .52 .48 .63 4.6 .41 .30 .40 221 Treatment (4) Commitment only N = 39 Variables Pearson r Tau Gamma_ 1-2 -021 -016 -019 1-3 -.31 -.22 -.24 1'5 -031 -021" -027 2-3 .62 .144 .47 2.5 069 .53 058 3'5 073 056 .60 N = 31 1-6 -.38 -.L+1 -.55 2.6 .34 .29 .38 3-6 .52 CM .58 _1 5-6 .5: .51 .65 APPENDIX C In order to obtain the probability for the relationship between two variables it was necessary to have two different types of information. The first was Kendall's Tau corrected for tied scores. The second was the standard deviation. The Tau's were computed by the Michigan State University 3600 computor. The program for obtaining Taub from a bivariate frequency distri- bution was corrected for ties. Kendall has derived a rather forbidding—looking calculation for computing the sampling variance of S when ties exist. The formula is:1 no: - 1) (2N + 5) - Enjmj - 1) (313+ 5) - Enkmk - 1) <2nk + 5) j k 18 r.- * # nj(nj-1) (nj-z) anmk-l) (nk-Z) + h— F 900 (N ~33 (N - 2) l Enjhj - 1) Erik (nk - l) + j k 2017‘???- 1) 1william L. Hays, Statistics for P cholo sts (New York: Holt, Rinehart and Winston, l9 , p. . 222 223 The above formula was not used. The reason was the original computations were for Tau and not S. In order to use Kendall's correction factor all the S values would have had to be computed by hand along with the sampling variance of S. The probability associated with the distributions of the data were computed by a formula provided by Siegel. The formula was:2 Z: I‘ 2(2N + 5) 9N N - 1 There was the possibility of contributing to Type I error using Siegel's formula. The standard error estimate does not take ties into consideration. Comparisons were made by the computing of the probability by the two different methods. The relationship between the modified F-scale and the overt-scale in Treatment (4) was selected. The N was 32. These were 17 ties in one column and Tau was .21. This data were the most likely to be affected by Siegel's formula. Using Kendall's method 2 was 1.69. Using Siegel's method the 2 was 2.25. Both were significant beyond.the .05 level. Therefore, Siegel's formula did not contribute significantly to type I error. ZSiegel, op. cit., BIBLIOGRAPHY Books Adorno, T. W. et al. The Agthoritarianr Personality. New York: Harper and Brothers, 1950. Allport, Gordon. The Nature of Prejudice. New York: Addison- Wesley, Inc., 19 . Bettelheim, Bruno and Janowitz, Morris. Social Change and Prejudice. New York: The Free Press of Glencoe, 1964. Brehm, Jack W. and Cohen, Arthur. $1orations in Cognitive Dissonance. New York: John iley 8: Sons, Inc., 19 2. Brink, William and Harris, Louis. The Negro Revolution in America. New York: Simon and Schuster, 19 . Cicourel, Aaron V. Method and Measurement in Sociolog. New York: The Free Press of Glencoe, 1964. Coles, Robert. The Dese re tion of Southern Schools: A P chiatric St 91. New York: Anti-Defamation League of B'nai B'rith, 1963. Deutsch, Morton and Collins, Mary E. Interracial Housing. Minneapolis, Minnesota: University of Minnesota Press, 1951. Edwards, Allan L. Techniques of Attitude Scale Construction. New York: Appleton-Century-Crofts, 1957. Etzioni, Amitai. A Comrative Analysis of Complex Organizations. New York: The Free Press of Glencoe, Inc., 19 1. " Garrett, Henry E. Statistics in P cholo and Education. New York: Longmans, Green am Comparv, 1959. Goffman, Erving. Encounters. Indianapolis, Indiana: Bobbs-Merrill Compam', Inc., 1961. Gordon, Milton M. Assimilation in American Life. New York: Mord University Press, 1964. ~ 224 225 Guilford, J. D. Echometric Methods. New York: McGraw-Hill Book Company, 1610., 193 o Hays, William L. Statistics for Lsycholoasts. New York: Holt, Rinehart & Winston, l9 . Homans, George C. The Human Group. New York: Harcourt, Brace and World, Inc., 1950. Kaplan, Abraham. The Conduct of Inquiry. San Francisco, California: Chandler Publishing Comparw, 1964. Kendall, Maurice G. Rank Correlation Methods. London: Charles Griffin and Comparw, Limited, 1948. Lewin, Kurt. Field meorLinjgcial Science. Edited by Dorwin Cartwright. New York: Harper 8: Row, 1951. Madge, John. The Tools of the Social Science. Garden City, New York: Doubleday & Company, Inc., 19 5. . The Origins of Scientific Socigggv. New York: The Free Press of Glencoe, 1962. Mead, George Herbert. Mind, Self and Society. Edited by Charles W. Morris. Chicago, Illinois: The University of Chicago Press, 1959 . Merton, Robert K. Social Theom and Social Structure. Glencoe, Illinois: The Free Press, 1963. Miller, Delbert C. Handbook of Research lbsi sign and Social Measurement. New York: mvid McKay Comparw, Inc., 1965. ' Nagel, Ernest. The Structure of Science. New York: Harcourt, Brace and World, Inc., 19 l. Newcomb, Theodore M. Social ngcholog. lst ed. New York: Henry Holt & Comparw, Inc., 1950. , Turner, Ralph H. and Converse, Philip E. Social Pachology. 2nd ed. New York: Holt, Rinehart 8: Winston, Inc., 19 5. . The Acggintance Process. New York: Holt, Rinehart and Winston, Inc., 1961. 226 Rokeach, Milton. The Open and Closed Mind. New York: Basic Books, Inc., 1960. Sellitz, Claire et a1. Research Methods in Social Relations. New York: Henry Holt 8: Company, Inc., 19760. Sherif, Muzafer. Intergroup Conflict and Cooperation: The Robbers Cave meriment. Norman, Oklahoma: University of Oklahoma Books Exchange, 1961. , and Sherif, Carolyn W. Reference Grows. New York: Harper 8: Row, Inc., 1964. Shibutani, Tamotsu. SocietLand Personality. Englewood Cliffs, New Jersey: Prentice-Hall, Inc., 1961. Siegel, Sidney. Nogarametric Statistics. New York: McGraw-Hill Book Company, Inc., 1956? Simpson, George Eaton and Yinger, J. Milton. Racial and Cultural Minorities. New York: Harper .9. Row, Inc., 1965. 2nd ed. Snedecor, George W. Statistical Methods. Ames, Iowa: The Iowa State College Press, 19 . Suchman, Edward A., Dean, John P. and Williams, Robin M., Jr. Desegregation: Some Eggpositions and Research Su stions. New York: Anti-Defamation League of B'nai B'rith, 19%. Vander Zanden, James W. Moe Relations in Transition. New York: Random House, Inc., 196 5. Walker, Helen M. and Lev, Joseph. Statistical Inference. New York: Holt, Rinehart &Winston, 1953. r i Wilmer, mniel M. et al. Human Relations in IErracial Housin . Minneapolis, Minnesota: University of Minnesota ss, 1955. Zetterberg, Hans L. (h Theory and Verification in Sociolog. Totowa, New Jersey: The Bedminister Press, 19 3. Articles and Periodicals Allport, Gordon. 'Prejudice: A Problem in Psychological and Social Causation,” Toward A General Theory of Action. Edited by Talcott Parsons and Edward A. Shils. New York: Harper and ROW, Inc., 1951! PP. 365-387. 227 Back, Kurt'W. et al. ”The Subject Role in Small Group EXperiments,' Social Forces, 44 (1965), 181-187. - Bass, Bernard M. QAuthoritarianism or.Acquiescence,' Journal of .Abnormal and Social Psychology, 50 (1960), 616-623. Becker, Howard S. I'Notes on the Concept of Commitment,'.American Journal of Sociology, 66 (1960), 32-40. . Blumer, Herbert. WAttitudes and the Socia1.Act,' Social Problems, 3 (1955). 59-65. . . "Race Prejudice As A Sense of Group Position," Pacific Sociological Review, 1 (Spring 1958). 3-7. ~ . ”Research on Race Relations in the United States of figsrica,’ International Social Science Journal, 10 (1958). 3—447. . Bogardus, Emory S. ”Racial Instance Changes in the United States During the Past Thirty'Years," Sociolo and Social Research, 42 (November-December 1958), 127-135. Bury, Douglas W} "The Prediction of Behavior From Two Attitude Scales,” Journal of Abnormal and Social ngchology, 45 (1950), 64-84. Brookover,‘Wilbur and Holland, John. fiAn Inquiry Into the Meaning of Minority Group Attitude Expressions,“ American Sociological Review, 17 (April 1952), 196-202. BPOPhY. I. M. “The Luxury of Anti-Negro Prejudice,” Public Opinion Quarterly, IX (1946), 456.466. Camilleri, Santo F. !A Factor'Analysis of the F-scale,” Social Forces. 3? (1959). 316-323- ' Campbell, [bnald T. 'Socia1.Attitudes and Other.Acquired Behavioral Dispositions,” ngchology: A Stugy of a Science. Edited by Sigmund Koch.. New York: McGraw-Hill Book Company. 1963. pp. 94-172. Campbell, Ernest Q. “On Desegregation and.Matters Sociological,“ Phylon, (Summer 1961), 135-145. . ”Campus 65,' Newsweek Magazine, (March 22, 1965), 43-63. 228 Chapman, Loren J. and Campbell, Donald T. ”The Effect of Acquiescence Response-Set Upon Relationships Among the F-Scale, Ethnocen- trism and Intelligence," Sociometry, 22 (June 1959), 153-161. Charters, W. W., Jr. and Newcomb, Theodore M. "Some Attitudinal Effects of Eacperimentally Increased Salience of a Membership Group,‘I Readings in Social Psycholog. Edited by Eleanor E. Maccoby, Theodore M. Newcomb and Eugene L. Hartley. 3rd ed. New York: Holt, Rinehart & Winston, Inc., 1958, pp. 276-281. Chein, Isidor, Deutsch, Morton and Hyman, Herbert and Jahoda, Marie eds. ”Consistency and Inconsistency in Intergroup Relations," Journal of Social Issues, Vol. V, No. 3 (1949). 2-61. . Christie, Richard. "Authoritarianism Ila-examined,“ Continuities in Social Research: Studies in the ScoEe and Method of‘the Authoritarian Personality. Edi by Richard Christie and Marie Jahoda. New York: The Free Press of Glenc'oe, 1954, pp. 123-196. Cottrell, Leonard 5., Jr. ”The Analysis of Situational Fields in Social Psycholog," American SocioloE._ca1 Review, 9 (1943), 370-387. . Cronbach, L. J. "Further Evidence of Response Sets and Test Design,” Education and Psychological Measurement, 10 (1950), 3—31. DeFleur, Melvin L. and Westie, Frank A. ”Attitude as a Scientific Concept,’ Social Forces, 42 (1963), 17—31. . ”Verbal Attitudes and Overt Acts: An Ewcperiment on the Salience of Attitudes," American Sociological Review, 23 (December 1958), 667-673. Deutsch, Morton. "Field Theory in Social Psychology," Handbook of Social Psycholog. Edited by Gazdner Lindsey. . Reading, Massachusetts: Addison-Wesley, Inc., 1954, pp. 181-222. Edwards, Allen E. ”Social Desirability or Acquiescence in the MMPI? A Case Studywith the S.D. Scale,” Journal of Abnormal and Social ngcholog, 63 (1961), 351-359. Fishman, Joshua A. ”Some Social and Psychological Determinants of Intergroup Relations in Changing Neighborhoods: An Introduction to the Bridgeview Study," Social Forces, 40 (October 1961), 2.51. . Green, Arnold W. “The Social Situation in Personality Theory," American Sociolcgical Review, 9 (1943), 388-393. 229 Harding, J. and Hogrefe, R. ”Attitudes of White Department Store Eruployees Toward Negro Co-workers,“ Journal of Social Issues, 8 (1952), No. 1, 18-28. Himmelstrand, Ulf. 'Verbal Attitudes and Behavior: A Paradigm for The Study of Message Transmission and Transformation,” Public Opinion Quarterly, XXIV (Summer 1960), 224-250. Hogrefe, Russell, Evans, Mary C. and Chein Isidor. ”The Effects on Intergroup Attitudes of Participation in an Interracial Play Center,” reported in The Handbook of Social_Psycholog[. Edited by Gardner Lindzey. Reading, Massachusetts: Addison- Wesley, Inc., 1954, pp. 1021-1061. Hyman, Herbert H. and Sheatsley, Paul B. ”Attitudes Toward Desegregation," Scientific American, 211 (July 1964), 16-23. . I'The Authoritarian Personality-4 Methodological Critique, " Continuities in Social Research: Studies in the Scogg and Method—of the Authoritarian Personality. Edited by Richard Christie and Marie Jahoda. New York: The Free Press of Glencoe, 1954, pp. 50-122. Killian, Lewis M. 'The Adjustment of Southern White Migrants to £252th Urban Norms,I Social Forces, 32 (October 1953). 9. Kutner, Bernard, Wilkins, Carol and Yarrow, Penm. l'Verbal Attitudes and Overt Behavior Involving Racial Prejudice,“ Journal of Abnormal and Social ngcholog, 47 (1952), 649-652. Lambert, Richard D. and Bressler, Marvin. “The Sensitive-Area Complex: A Contribution to the Theory of Guided Culture Contact,’ American Journal of Sociolm, 60 (1955), 583-592. LaPiere, Richard T. "Attitudes vs. Actions,“ Social Forces, 13 (December 1934), 230-237. Lenski, Gerhard E. and Leggett, John C. ”Caste, Class and Deference in the Research Interview,“ American Journal of Sociology, 65 (1960), 463-467. Linn, Lawrence S. "Verbal Attitudes and Overt Behavior: A Study of Racial Iliscrimination,’I Social Forces, 45 (1965), 353-364. Lohman, Joseph P. and Reitzes, Dietrich c. "Deliberately Organized Groups and Racial Behavior,“ American Sociological Review, 19 (1954). 382-344. . 230 MacKenzie, Barbara K. ”The Importance of Contact in Determining Attitudes Toward.Negroes,' Journal of Abnormal and Social Echologr, 43 (1948), 417-441. Malof, Milton and Lott, Albert. ”Ethnocentrism and the Acceptance of Negro Support in a Group Situation,“ Journal of Abnormal and Social Psycholggy, 65 (October 1962), 254-258. Mann, John H. “The Relationship Between Cognitive,.Affective and Behavior ASpects of Racial Prejudice,“ Journal of Social -£§ychology, 49 (1959), 223-228. . Martin, James G. and westie, Frank R. “The Tolerant Personality,“ American Sociological Review, 24.(1959), 524-525. - Minard, R. D. ”Race Relations in the Pocahontas Coal Field,” Journal of Social Issues, Vol. 8, No. 1 (1952), 29-44. Miyamoto, Frank and Dornbush, Sanford M. ‘A Test of the Interactionist Hypotheses of Self-Conception," American Journal of Sociology, LXI (1956). 399-403. . Mussen, Paul H. ”Some Personality and Social Factors Related to Changes in.Chi1dren's Attitudes Toward Negroes," Journal of Abnormal and Social ngchology, 45 (July 1950), 423-441. Newcomb, Theodore M. FAttitude Levelopment as a Function of Reference Groups,“ Readings in Social Psychology. Edited by Eleanor C. Maccoby, Theodore M. Newcomb and Eugene L. Hartley. 3rd ed. New York: Holt, Rinehart and‘Winston, Inc” 1958. pp- 265-275. . "The Prediction of Interpersonal Attraction,“ American ngchologist, 11 (November 1956), 575-586. Orne, Martin T. "On the Social Psychology of the Psychological Experiment,” American ngchologist, 17 (1962). 776-783- Palmore, Erdman B. “The Introduction of Negroes into White Departments,I Human Organization, 14 (1955), 27-28. Peak, Helen. ”Problems of Objective Observation," Rgsearch_Methods in the Behavioral Sciences. Edited by Leon Festifiger and Daniel Katz. New York: Holt, Rinehart & Winston, 1953, pp. 243-299. Pearlin, Leonard I. ”Shifting Group Attachments and.Attitudes Toward Negroes,” Social Forces, 33 (1954), 47-50. 231 Pettigrew, Thomas F. ”Personality and Sociocultural Factors in Intergroup Attitudes: A Cross-National Comparison," Journal of Conflict Resolution, 2 (March 1958), 29-42. . ”Social Psychology and Desegregation Research," American'Psychologist, 16 (1961), 105-112. Photiadas, John D. and Biggar, Jeanne. ”Religiosity, Education and Ethnic Distance," American Journal of Sociology, LXVII, No. 6, (May 1962), 666-673. Raab, Earl and Lipset, Seymour Martin. “The Prejudiced Society,” American Race R§_1ations Toda . Edited by Earl Raab. . Garden City, New Jersey: Doubleday and Comparv, Inc.. 1962. pp. 29-55. Reitzes, Dietrich C. “Institutional Structure and Race Relations,” Phylon, First Quarter (Spring 1959), 48-66. . Rokeach, Milton. "The Nature of Attitudes,” a working paper to be published in the International Engyclopedia of_the Social Sciences, pp. 1-1 . Rose, Arnold M. "Intergroup Relations vs. Prejudice: Pertinent Theory forthe Study of Social Change," Social Problems, Saenger, Gerhart H. and Gilbert, Emily. ”Customer Reactions to the Integration of Negro Sales Personnel," Public Opinion Quarterly. 4 (1950). No. 1. 57-76. Sanford, Nevitt. "The Approach of the Authoritarian Personality,- ngcholog of Personality. Edited by J. L. McCary. New York: Steiner, Ivan D. and Johnson, Homer H. I'Authoritarianism and Conformity," Sociometry, 26 (March 1963), 21.-34. Stewart, Don and Hault, Thomas. ”A Social-Psychological Theory of the 'Authoritarian Personality'," Amerigag Journal of Sociology. 65 (1959). 274-279. Stouffer, Samuel A. et a1. ”Negro Infantry Platoons in White Companies,” Readings in Social Pjgcholog. Edited by Eleanor E. Maccoby, Theodore M. Newcomb a Eugene L. Hartley. New York: Holt, Rinehart a. Winston, Inc., 1958, pp. 596-601. 232 watson, Jeanne. “Some Social and Psychological Situations Related to Change in Attitude,“ Human Relations, 3 (1950), 15-56. Wilner, D. M., Walkley, Rosabelle P. and Cook, 5. w. ”Residential Proximity and Intergroup Relations in Public Housing firoéects,’ Journal of Social Issues, Vol. 8, No. 1 (1952). 5" 90 Yarrow, Marian et a1. ”Acquisition of New Norms: A Study of Racial Desegregation in Interpersonal Dynamism in a Desegregation Process,I Journal of Social Issues. Edited by Marian Radke Yarrow. Vol. 14 (1958), 1-63. Yinger, Milton J. "Research Implications of a Field View of Personality,” American Journal of Sociology, LXIV (1963), 580-592. Papers Haller, A. 0. EA Point of View on Social Psychology.' A working paper presented to Alpha Kappa Bolts of Michigan State University, October 25, 1963, pp. 1-21. . ISome Principles of Attitudes and Behavior,“ a working paper, 1960, pp. 1-8. UNV.L I IBRQRIES ‘ WIIWIWIWIWll 757605 (DEB MICHIGAN STRTE WW"WWII“ 312930