' MORAL CONFORMITY IN OPEN AND CLOSED GROUPS: A COMPARATIVE STUDY OF MORAL DECISION - MAKING AMONG MORMON, CATHOLIC AND PUBLIC SCHOOL CHILDREN Dissertation for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY A. GARY SHEPHERD 1976 This is to certify that the thesis entitled MORAL CONFORMITY IN OPEN AND CLOSED GROUPS: A COMPARATIVE STUDY OF MORAL DECISION-MAKING AMONG MORMON, CATHOLIC AND PUBLIC SCHOOL CHILDREN presented by A. Gary Shepherd has been accepted towards fulfillment of the requirements for PhoDo degree in SOCiOlOgy 7a..“ 2%— / Major professor Date /?M/’7‘ 0-7639 .41“:- WW M: Q C" vp-" fi\~ -..‘. E “V...“ V ‘ ”1"\ '- ~ 'V-«u v-‘lot. ABSTRACT MORAL CONFORMITY IN OPEN AND CLOSED GROUPS: A COMPARATIVE STUDY OF MORAL DECISION-MAKING AMONG MORMON, CATHOLIC AND PUBLIC SCHOOL CHILDREN BY A. Gary Shepherd At its most general level, this study has been concerned.with investigating the impact of social structure on socialization of conformity to moral standards. The social structural feature of concern here is degree of system aperture-closure. Moral standards are represented by presumed adult conventions regarding good and bad behavior. Conformity is analyzed in terms of responsiveness to influence source, i.e., self standards vs. parent stan- dards vs. peer standards. Socialization processes take on an inferred status (since they are not actually measured in this study). serving theoretically as the link between open- closed system constraints and degree of conformity to moral standards. At a more specific level, this study represents a critical response to a line of research initiated by Urie Bronfenbrenner, Edward Devereux and other investigators at ”U‘ V Invp~ .S ACa‘A:‘.‘H . t “I . n F F A L‘ “a le ‘. 0..“ CF "-31.23 "Vlo Q br': . :" :v: -~.‘c ba‘.“ “E 'a.,:!- ‘ za“ fl, I--.:.; S ‘5.9‘ ‘n i svsuggq‘g &.L Rafi/Ce. “e 96:- .. .- I‘E 5932‘? to J I‘. “£291,335 2 ‘51. . S ‘ «\u‘o': oper-‘Clt E“: z I J .- A. Gary Shepherd Cornell University. The Cornell studies have examined moral preferences of school children in a number of coun— tries (ranging from the United States to the Soviety Union). and have concluded that children reared in collective socialization settings tend to adhere to more acceptable moral standards than do children reared in individualized socialization settings. The present study attempts to extend the implications of the Cornell studies by (1) exam- ining and defining more carefully the larger social and cultural contexts in which moral socialization occurs, and (2) refining the conceptualization and measurement of moral preference. The general social variable focused on in this study is the degree to which social systems differ with regard to "aperture-closure." Three basic dimensional sets for defining open-closed systems were extracted from the litera« ture. These dimensions include: system "linkage" (i.e., degree of exchange and interdependence between systems), system "consensus“ (i.e., with regard to beliefs, norms, and values), and system “conformity-deviance sanctions.“ Of these basic dimensions, one--the dimension of "consensus"—- was selected as being most amenable to measurement in the present study. An empirical indicator of "consensus" was developed (the "Index of System Engagement," or ISE) and applied to the groups being studied. Groups from which samples were drawn for testing 'were intended to represent, on an §_priori basis, social I t trawl-"WE . fiu.¢eoi‘d. “ t I ' ‘ Q ‘ :unqq' ft!“ 579' .u-v'o. V““u“’“ 13:2: tze H- I ' t f Mug-uh. 1v- »eruuuso V gees ‘ a :n‘}“'P :‘L‘Cl h “sob abaov' u. 'AA A. \ n t ....:s .3: Mar? . I ..~ I“ ' u.‘ W!- A"; "I “:‘V'v. -5- "'7fl~ ' ‘J .:'._‘C.‘! . n "1 vu b‘ud a a H :‘(ECOTSE' *I V . mi." YO!" A: C: v _ ‘4‘in “‘t~e.’ :‘QI ‘ ' "kite to c'” ‘ s.‘ 5* IIEre I; :. .~~"‘L -. at; '-(,h a7 “A n ‘ s ‘ :valt; h It! A. Gary Shepherd systems that could be expected to vary along the open-closed continuum. These groups were Mormon, Catholic and Public School children. The ISE proved to be useful in distin- guishing the Mormon and Catholic samples in predicted ways (necessary information for the ISE was not available for Public School children). Thus, significantly higher ISE scores for Mormon children (controlling on sex and social class) supported the a priori designation of Mormons as representing a relatively more closed system than Catholics. Reconceptualization of relevant moral issues involved differentiating categories of moral responses made available to children in the previous Cornell studies. Critical reading of the original Moral Dilemmas Test (MDT) suggested that two basic types of underlying moral dimen- sions were being measured. The first type of moral response consisted of situations wherein the child is required to resist peer pressures to behave in "anti-social" ways. Potential positive or negative adult sanctions, depending upon the response, are implicit in the dilemma situations.) The present study contends that the socially approved response in this kind of situation (from an adult perspec- tive) represents a "negative-passive" morality. The second type of moral response reflected in the MDT involved situa— tions that require performance of an act that affirms a moral principle and implies a benefit for the recipient of the act. This kind of moral response is called, in the preSent study, "positive-active" morality. SS!!! v ouch, : ‘\ : ”'2 I' A y . Mal: “(2.8 .‘ I ' :ua: 0‘6 0'._‘ '“'.. b“ ‘.I- I a up..‘ ‘ "F .‘ ~v:o .L'I'e-Pa SS W “: Iea‘ .., uh. .d‘ “w I" 2'- I -u :. ¢S:er azia . -. ”tongs.“ ~ Nb'..b.La‘ 3; 's \ § ‘ o “’3’ I2} ,; . Si" "if; " See +1. . 5.. ~. ‘1 s :v .=:“"L.' I A '4‘ ‘- “‘E. ‘o 5". \“s ‘e ""s h ‘ fi .. I~ \:S‘A\ VJ “ Q ..§es Cf t ix. :32!“ C set S; I“ I h “‘3‘. d‘ual v n !_ bor.f: '21..“ "\an ‘V " raft, H A. Gary Shepherd Since positive-active decisions appear to be morally more fundamental than negative-passive decisions, and since the original MDT consisted of a larger number of negative-passive dilemma items, additional positive—active dilemmas were created to supplement the old items. All of the MDT items (both old and new) were subjected to factor and cluster analyses. These analyses provided support for the theoretical distinctions by producing item clusters that could consistently be identified as negative—passive vs. positive-active. In addition, a third dimension was also discernable from a sub-set of the original MDT items; this consisted of dilemmas that emphasized a certain achievement orientation. Arguments were presented that diminished the moral importance of achievement orientation and thus ques« tioned the desirability of including such.items in the MDT. Children's responses to the MDT were analyzed across three experimental conditions: (1) Base-—children were told that no one would see their answers to the dilemma items; (2) Adult—-children were told that their parents would see their answers to the dilemma items; and (3) Peer—-children were told that their classmates would see their answers to the dilemma items. Although.statistical significance was not always obtained, comparisons between :responses of the three samples did provide suggestive support for several hypotheses. For instance, when only the cxriginal, conceptually undifferentiated MDT results were examined, rankings were in the predicted direction: on ctr-all cc"' u... . I. . 331.3365 were III;':'G' ' “‘0I$984 we.e “4‘:v‘vy:‘ tL' ““‘Q‘NS A: Size cvergall .:.‘ A;:: V IA UH... ‘6..e‘ Nb! )5“ r "1‘ . new“: ‘eSnL h'61: Vt. . rec-He‘lltlr‘.‘ Vual st .. Vflfih‘l BL; \ 9 v‘ _ "‘- ,. revue b“‘¢’u.e. 2“. t. ‘7 Y“ . vol cvw‘ lta-l 1;:4. M9!) were a that ‘ 9 “‘5~Enaards k h l Skisa CL‘ ‘ ‘ “‘eQrer “it. “.‘h . ' . IE PASSI‘ 1‘“ '15‘35 he “res . ‘ SIS a: “'eSe :b “:~i§ ‘ y “5 hp '4 :.‘ . ‘rr‘ ‘yt "e. u , ~ \‘\ ‘ :1 t S: “x \‘L the ‘ n watt ‘L s C.“ ‘ "‘E A. Gary Shepherd over-all conformity to conventional moral standards, Mormon children were highest, Catholics next, and Public School children were lowest. When the basic moral dimensions underlying the modified MDT were analyzed separately, the same over-all rankings were obtained, but with some impor- tant differences. First, isolating negative-passive MDT items resulted in Mormons being even more conforming to conventional standards compared to Catholic and Public School children. When, however, responses to positive- active MDT items were isolated, Catholic and Public School children were at least as likely to conform to conventional moral standards under the Base and Peer conditions as were Mormon children. These differences in degree of conformity to the two basic moral dimensions are interpreted as sup- porting the notion that closed social systems will place greater stress on sanctioning the behavioral outcomes of negative-passive moral situations. With regard to positive— active moral standards, the supported notion is that closed systems will have only a relatively perfunctory concern that differs little from positive-active concerns manifested in more open systems. These conclusions find modest additional support from the correlations between ISE and MDT scores. Although tme strength of these associations turned out to be dis- appointingly small, some encouraging evidence can be gleaned from the patterning of association. For instance, Mormon correlations between ISE and negative-passive MDT scores IKE CCCSLSZET. azrcss sex .3: ""11 ‘flFA‘iooi um“. VVc.“ .. I. I. ‘ ».~~. ‘ . .s ”I 3.9 ' ':h 5“ are adrz“ “ ‘::. ..‘|“:S & a. H‘ “t s) A 5... V e‘:“1‘ ' \bs 0:: ;F «‘3' 55.. ‘v A. Gary Shepherd were consistently positive and higher than for Catholics across sex and social class controls for the three experi- mental conditions. On the other hand, when correlations between ISE and positive—active scores were examined, increases were found in the incidence of negative relation- ships, as well as a consistent decrease (sometimes reversal) in Mormon-Catholic differences across sex and social class controls for the three experimental conditions. Ambiguous and/or negative findings of this study are generally characterized by the "smallness" of many of the predicted differences between samples and the blurring of some of these differences when general patterns are broken down into component parts for subanalyses. Weak- nesses in the instruments and procedures employed in the study are examined as potential causes for some of the ambiguous results. Suggestions for remedying flaws of measurement, design and sampling are detailed. Also sug— gested are additional new measuring instruments and pro- cedures that should help improve the interpretation of results obtained in any further replications. Such repli— cations would help to place the study of comparative moral behavior more firmly within a social matrix. m... C‘ mu 0 K. MORAL CONFORMITY IN OPEN AND CLOSED GROUPS: COMPARATIVE STUDY OF MORAL DECISION-MAKING AMONG MORMON, CATHOLIC AND PUBLIC SCHOOL CHILDREN BY A. Gary Shepherd A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Sociology 1976 A ACKNOWLEDGEMENTS This dissertation has had a relatively lengthy career, beginning with ideas generated in the winter of 1973 and ending in final revisions of the manuscript during the present winter. A number of peoplew-too many, in fact, to mention them all here-~have contributed necessary assis- tance at various points during the interval from inception to culmination. However, public expression of my apprecia- tion is especially due to the following individuals and organizations: To Sr. Thomas Aquinas and Mr. Al Dalimonte, super- intendents of the Lansing Diocesian Parochial Schools, and to the principals of various of these schools; to President Don R. Earl and various bishops in the Liberty, Valverda, Olympus and Holiday stakes of the Church of Jesus Christ of Latter Day Saints; to Dr. Jack Mawdsley and Dr. Garth Errington, superintendents of the Battle Creek Public School District, and to the principals of various of these schools: for permission to use classes of children willing to parti- cipate in my study. ii 2: . 1 LEE 53.7.65, ; T0 La; H’v~ ; 'Q‘Cr.e Shez‘u . .. "'31.. FaC ‘ t ‘ 1 s.‘ . ‘ '51”; . C a. has. “-4". o“ -' and a. s“ \\.: ‘ -r To Mr. Alvin Shepherd, Dr. Steven Burton and Karen Graham for helping to arrange contacts with Mormon bishops. in the various stakes from which MIA and Primary classes were sampled in the study. To Lauren Snow Shepherd, Susan Shepherd, Mrs. Marjorie Shepherd, Peggy Snow and Dr. Darryl Thomander for assisting in the administration of the MDT and other ques— tionnaire forms to the children used as subjects. To Bill Brown and Dr. Jack Hunter for assistance in setting up the PACKAGE computer routines and for helping to interpret the output of these routines. Also to the Com- puter Institute for Social Science Research at Michigan State University for providing the computer time necessary to carry out the data analysis. To Gordon Shepherd and Dr. Frank Camilleri for insightful ideas and suggestions concerning the construction and limited meaning of several of my measures and for their help in evaluating some of my data results. To Lauren Snow Shepherd for typing portions of the rough draft (a real labor of love, since she is even a worse typist than I), Elizabeth.Halstead for her careful proof- reading, and Irene Orr for her professional typing of the final manuscript. To my dissertation committee members, Dr. Denton Morrison, Dr. William Faunce, and Dr. Eugene Jacobson for their tolerance and encouragement. iii a and f . I ”lo 4 V" V A .‘te‘svn' II to I o I" ‘1) ‘1 U "1 . . I. , .. :he sec; 1 C .. ‘ ':"‘\na‘ ""‘w-LCCical ‘ ‘. ‘NL... ‘. vi-A...‘ .10: ‘2 3n" n: "‘ “VA. U ‘ 5“. v . b‘l‘:‘ “‘Q‘ “E? ‘ ‘ ‘ as 3 § (In "his ”he. And finally, to Dr. Bernard Finifter, my committee chairperson, who easily stands out from among all who have helped. Dr. Finifter brings together an unusual command of the social psychological literature, finely honed methodological skills, and a creative mind. This happy combination has provided me with a constant and generous source of provoking ideas and technical evaluation. Dr. Finifter has always left room for disagreement 0n particular issues but, at the same time, has always insisted that my own thinking reflect the same high standards of thoroughness and critical analysis that he demands of himself. Far beyond the exercise of doing an acceptable piece of disser- tation research, this is what a graduate education should be all about. iv :fi' A' 'V 3‘ F“ ~ A azulsziif C... 9 n N:. 3? ‘H‘?’ ~ \ a. .- ~ 23". OF FYI-9", ,_ ices; 'Wn- \‘ I KC. 0? REPE‘:‘ ‘ g p I.“ ‘ ‘- " ' M...e‘ 0 on ~::~ p- “O~ U K F ‘n L. “h ’4 (l) (0 “Vin TABLE OF CONTENTS ACKNOWLEDGEMENTS . . . . . . . . . . . LIST OF TABLES . . . . . . . . . . . . LIST OF FIGURES . . . . . . . . . . . . LIST OF APPENDICES . . . . . . . . . . . Chapter I. II. THE RESEARCH PROBLEM . . . . . . . THE Introduction . . . . . . Objectives of the Present Study . Review of Relevant Literature . . Summary of Research Findings . . Some Limitations of the Cornell Studies and Elaboration of the Present Study's Objectives . . . . . . Chapter I--Footnotes . . . . . . THEORETICAL FRAMEWORK . . . . . . Introduction . . . . . . . . . Open—Closed Systems . . . . . General Definitions for Identifying Open-Closed Systems . . . . . . Selection of Open-Closed Dimension for Elaboration in the Present Study . Some Basic Moral Dimensions: Negative vs. Positive, Passive vs. Active . Negative-Positive, Passive-Active As Structural Properties of Moral Situations . . . . . . Statement of the General Hypothesis . Modification of the General Hypothesis Chapter II--Footnotes . . . . . . Page . ii . ix . xii -xiii . l . l . 4 O 6 O 17 . 21 . 28 . 31 . 31 . 31 . 45 . 48 . 50 . 59 . 62 O 62 . 66 lagter 0" on. O 3’ ‘3’“ v:-: » *M—ou DC? D‘MV,‘ ”’0. A ..s ‘fi‘vvv ' a "‘ IV... (I: r) .1 Cl”. ~ o-a :71 l '1 F") ’ t i ('1 r—I U) :1 (I) r) p: 1 Cr .PI Chapter Page III. METHODS AND PROCEDURES, PART 1: TEST POPULATIONS AND SAMPLES . . . . . . . . 70 Introduction . . . . . . . . . . . 70 RATIONALE FOR SELECTION OF COMPARATIVE POPULATIONS . . . . . . . . . . . . 70 The Mormons . . . . . . . . . . . 70 Catholics . . . . . . . . . . . . 76 U.S., Cross-Sectional . . . . . . . . 77 DESCRIPTION OF SAMPLES . . . . . . . . . 78 Mormon Children . . . . . . . . . . 78 Catholic Children . . . . . . . . . 82 Public School Children . . . . . . . 84 Chapter III--Footnotes . . . . . . . 88 IV. METHODS AND PROCEDURES, PART 2: THE MORAL DILEMMAS TEST . . . . . Introduction . . . . . . . . . . . 92 THE ORIGINAL MDT . . . . . . . . . . . 92 General Content and Format Description of Test Items . . . . . . . . . . 92 Scoring of Test Items . . . . . . . . 93 Test Conditions and Experimental Design . . 94 Underlying Dimensions of the MDT Items . . 99 Empirical Support Reported for the A. Priori Dimensionality of the MDT Items . . . . . . . . . . . . 100 Critique of the Original MDT . . . . . 102 MODIFICATION OF THE MDT . . . . . . . . 103 Description of New Items Costructed to Supplement the Original MDT . . . . . 103 Empirical Evaluation of the Expanded MDT (Original Plus New Items) Based on Results from the Mormon, Catholic and Public School Samples . . . . . . 105 Chapter IV-—Footnotes . . . . . . . . 129 vi Chapter page V. METHODS AND PROCEDURES, PART 3: THE INDEX OF SYSTEM ENGAGEMENT . . . . . . . . . 132 Introduction . . . . . '. . . . . . 132 General Description of the Index of System Engagement (ISE) . . . . . . 132 Validating and Scaling the Dimensions of System Engagement . . . . . . . . 135 Combining Subdimensions Into a Composite Index of System Engagement . . . . . 147 Chapter V--Footnotes . . . . . . . . 155 VI. FORMULATING AND TESTING RESEARCH HYPOTHESES . . 158 Introduction . . . . . . . . . . . 158 HYPOTHESES INVOLVING THE INDEPENDENT VARIABLE . 158 Hypothesis I: System Engagement . . . . 159 Test of Hypothesis I . . . . . . . . 160 Discussion of Results for Hypothesis I . . 161 Elaboration of Hypothesis I . . . . . . 166 HYPOTHESES INVOLVING THE DEPENDENT VARIABLE . . 173 Hypothesis II: MDT Scores . . . . . . 174 Elaboration of Hypothesis II . . . . . 174 Results for the Original ("Undifferen— tiated“) MDT . . . . . . . . . . 176 MDT Results for the Negative-Passive Dimension . . . . . . . . . . . 190 Hypothesis for the Positive-Active Dimension . . . . . . . . . 199 MDT Results for the Positive-Active Dimension . . . . . . . . . . . 199 Hypothesis for the Achievement Dimension . 208 MDT Results for the Achievement Dimension . 211 SYSTEM ENGAGEMENT AND MORAL CONFORMITY: EXPLICITLY HYPOTHESIZING RELATIONSHIPS BETWEEN THE INDEPENDENT AND DEPENDENT VARIABLES . . . . . . . . . . . . 218 Hypothesis III . . . . . . . . . . 218 Results for SE and Negative—Passive MDT Scores . . . . . . . . . . . 220 Results for SE and Positive-Active MDT Scores . . . . . . . . . . 225 Results for SE and Achievement MDT Scores . 227 Chapter VI--Footnotes . . . . . . . . 229 vii fl .6 carer D-’ F"P’ I... LAN... . s H y A {I lb ,.I H" h b ,I .- A t‘ L 4 '1 . ‘F 'H . dg~~‘ ‘G‘sz ".v u; , Cflhapter \IJII. EilflBLIOGRAPHY Introduction . Positive Findings Moral Conformity Negative-Ambiguous Findings Deficiencies of the Present Recommended Improvements Final Assessment Chapter VII——Footnotes viii CONCLUSIONS AND RECOMMENDATIONS nonwoooo (D (0 cosmooo Page 230 230 231 236 242 246 253 259 260 321 .0 ‘1' a: . LIST OF TABLES Table Page 1. Cross—Cultural Comparisons of Childrens' Mean MDT Scores as Reported in Bronfenbrenner (1975:485) . . . . . . . 7 2. Sex and Social Class Characteristics for Mormon Sample . . . . . . . . . . . 82 3. Sex and Social Class Characteristics for Catholic Sample . . . . . . . . . . 85 4. Sex and Social Class Characteristics of Public School Sample . . . . . . . . . 87 5. Mean MDT Scores Under Base Condition Only, All Groups (Mormon, Catholic, Public School) Combined . . . . . . . . . . 107 6. Mean MDT Under Base Condition Only, Mormon and Public School Subanalyses . . . . . . 108 7. Major MDT Item Clusters Emerging from Cluster Analysis. All Sample Scores Combined, Experimental Conditions Ignored . . . . . 110 8. Intercorrelations for A Priori Clusters, Original Cornell MDT Items.‘ Samples Combined, Test Conditions Ignored. Cluster Reliability Alphas in Diagonal . . . 113 9. Factor Analysis of Expanded MDT, Samples Combined, Test Conditions Ignored . . . . 114 10. Separate Factor Analysis Results for Each MDT Form, Samples Combined (N = 487), Test Conditions Ignored . . . . . . . . . 118 ix Tale H. 'Eult Cor RF Us the 3- Inter Pcs 8a: Neg Di: A. D L02 (5. S2 3' I T L ‘nLe‘PC Table Page 11. ”Multidimensioanultiform" Matrix, Samples Combined (N = 487), Test Conditions Ignored (Alpha Reliabilities in Parenu theses Main Diagonal) . . . . . . . . 123 12. Intercorrelations Between Negative-Passive and Positive-Active Dimension Scores. Within Same Form vs. Different Form Comparisons . . 125 13. Intercorrelations Between Achievement by Negative-Passive and Positive-Active Dimensions Across MDT Forms . . . . . . 126 14. Analysis Results for Weak MDT Items . . . . 128 15. Correlations Between Parents' Attendance and Membership . . . . . . . . . . . 137 16. Time/Activity Combinations . . . . . . . 137 17. Correlation Between Relatives and "Closeness" to Respondent‘s Family . . . . . . . . 139 18. Membership/Closeness Scoring Combinations for "Relative" Dimension . . . . . . . . . 141 19. Intercorrelations Between Categories of Friends Rated as Church Members . . . . . 143 20. Correlations Between Neighbors Membership and Contact with Respondents‘ Family . . . . . 145 21. Scoring Summary for Family Home Evening/ Confession Variables . . . . . . . . . 148 22. Minimum-Maximum Church Activity Scores . . . 149 23. Intercorrelations Among SE Subdimensions . . . 150 24. Factor Analysis of SE Subdimensions (Yielding Loadings on Only One Factor) . . . . . . 151 25. Standardization of Rank Scores for SE Subdimensions . . . . . . . . . . . 152 26. Intercorrelations Between Original SE Dimension Rank Scores and Standardized SE Dimension Rank Scores . . . . . . . 153 Table 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. Mean SE Dimension Scores' . . . . . . . Mean SE Dimension Scores, Controlling for Social Class. . . . . . . . . . . Mean SE Scores, Controlling for Sex~ . . . Mean MDT (Original Items Only) for Mormon, Catholic, and Public School Children . . Mean Differences Between Test Conditions for Mormon, Catholic and Public School Children . . . . . . . . . . . . Mean Negative-Passive MDT Scores for Mormon, Catholic and Public School Children . . . Mean Differences Between Test Conditions on Negative—Passive MD‘s for Mormon, Catholic and Public School Children . . . . . . Mean Positive-Active MDT Scores for Mormon, Catholic and Public School Children . . . Mean Differences Between Test Conditions on PositivesActive MD‘s for Mormon, Catholic and Public School Children . . . . . . Mean Achievement MDT Scores for Mormon, Catholic and Public School Children . . . Mean Differences Between Test Conditions on Achievement MD's for Mormon, Catholic and Public School Children . . . . . . Correlations Between SE and Negative-Passive MDT Scores for Mormon and Catholic Children 0 O O O I I O O O O O 0 Correlations Between SE and Positive-Active MDT Scores for Mormons and Catholic Children 0 O O O O O O O O O O 0 Correlations Between SE and Achievement MDT Scores for Mormon and Catholic Children . xi Page 163 168 172 177 182 191 196 200 205 212 216 221 226 228 K) o " ‘.. Ufa?“ Sc: L... Figure 1. LIST OF FIGURES Page Graphic Presentation of Mean ORIGINAL MDT Scores for Mormon, Catholic and Public SChOOl Children 0 O O O O O O O O O 317 Graphic Presentation of Mean NEGATIVE-» PASSIVE MDT Scores for Mormon, Catholic and Public School Children . . . . . . . 318 Graphic Presentation of Mean POSITIVEvACTIVE MDT Scores for Mormon, Catholic and Public School Children . . . . . . . . . . 319 Graphic Presentation of Mean ACHIEVEMENT MDT Scores for Mormon, Catholic and Public SChOOl Children I O O O O O O O O O 32 0 xii N. '5‘ If.) "' Appendix I. A. II. III. Iv. LIST OF APPENDICES Page Forms X, Y, and Z of the Moral Dilemmas Test (MDT); B. Instructions for Administration of the MDT; C. Back— ground Questionnaire for Construction of the Index of System Engagement (ISE) . . 262 Suggested Supplementary Measures: A. Peer Group Closure Scale; B. Parental MDT Evaluation Questionnaire . . . . . . . 303 Correspondence: A. Letter from Presiding Bishopric, Church of Jesus Christ of Latter Day Saints (Mormons); B. Letter from Edward Devereux . . . . . . . . 309 Factor/Cluster Analysis of Combined MDT Items Pooled Across All Subjects (N = 487); B. Analysis of Variance for MDT Scores. Comparisons for Mormon, Catholic and Public School Children, Controlling on Sex and Social Class; C. Graphic Pre— sentation,of Mean MDT Scores Across Base, Adult and.Peer Conditions. Comparisons Between Mormon, Catholic and Public School Children . . . . . . . . . . 311 xiii CHAPTER I THE RESEARCH PROBLEM Introduction Ultimate concern with "moral" issues is inherent in the subject matter of the behavioral sciences. For instance, moral issues implicitly underlie most of the specific substantive t0pics, grouped under the broad cone ceptual canopy of socialization, with socialization rightfully regarded as constituting the core of social psychological inquiry (Clausen, 1968:3). Within the realm of early socialization concerns, psychology proper focuses on the consequences of specific child-rearing practices and the growth of cognitive and emotional capa— cities during that period when the child is learning to distinguish "right" from "wrong," culminating in a more— or-less coherent system of values and moral beliefs. Sociology proper reflects this same focus at the more general conceptual levels of social control and deviance. Ideally, the over-all child-rearing patterns typical of Sfimen social groups are examined to see how these prac- tices, by eventually producing conforming group members, serve to re ‘ A 'fl;ys boa». . I.) 'ri (I) «) . ,- (3 .L‘ (3 r r v '1 - , . serve to reinforce the existing normative and moral orders. There exists a vast and sprawling literature dealing with the moral aspects of socialization. It appears that the greater weight of this literature favors a psychological perspective. Psychologists have made explicit their interest in moral questions, have formu- lated theoretical and empirical problems that reflect this explicit interest, and have thus produced many valuable insights concerning the psychology of moral activity. These insights are generally classifiable under three broad areas of psychological emphasis: (1) Social learn- ing--most often associated with modeling and imitation of "preferred" conduct paradigms, exemplified by Bandura (1969); (2) Cognition--paradigms of moral judgement and the acquisition of moral concepts, exemplified by Piaget (1932) and Kohlberg (1963); and (3) Internalized "feeling" states--conscience or super—ego based guilt paradigms, exemplified by Freud (1930).1 Contemporary sociology, however, has shown less interest in stressing the direct moral implications of socialization theory and research (Inkeles, 1968; Maccoby, 1971). While discussion of social values and normative influence has always been central to sociology, efforts are not usually made to explicitly distinguish moral values from other kinds of values.2 Thus, specifically rmual values and their social antecedents tend to remain insured in as: as clear. iith the exI A.“ Q :. SQ'JA Cfias . rati2; tech t;":1 ‘ my iitt‘. I 9"!” “"9539 e3: ism-tion The I “fish it, 3323;: re I.“ we; '1' :~ Up, in It: “My tc '9 . “mills: ki p...“ 4 4‘ '1‘} . s s “e S “a "“36" ‘ GIT-v“ \_“ne 18 S‘-. . “13.3,”; Ia“ iE‘v obscured in much sociological analysis rather than stand out as clearly recognized variables in their own right. With the exception of studies that have attempted to link social class to differences in parental values, child- rearing techniques, and environmental pressures,3 rela- tively little actual research has been addressed to the influence exerted by social structural variables on the acquisition and expression of moral values. The present study is a response to this lack. Through it, I have attempted to examine one potentially important relationship between the social organization of a given group, the particular socialization techniques that are likely to reflect the organization of the group, and the moral consequences for children who are exposed to particular kinds of socialization experiences within their group.4 The social structural variable that I am most concerned with in this study is the degree to which.a group is either “open" or "closed" to the introduction and/or expression of alternative norms and values. The socialization experience that I am most interested is the degree to which children are exposed to a consis— tent set of values and norms. And the moral behavior that I examine is the degree to which children conform to group standards of right and wrong when under pressure to deviate. The general hypothesis that has guided this study has been, simply, that the more "closed" the social structure of a group, the more likely there will be con- formity to the moral standards of that group. Objectives of the PresenE_Study The research antecedents of my study are largely found in the work either conducted or inspired by Urie Bronfenbrenner (1967, 1970a, 1970b) and Edward Devereux (1970, 1972) of Cornell University. These Cornell studies imply much concerning relationships between social structure and moral behavior of the young by making cross—cultural comparisons of the degree of manifest congruence that exists between adult and children's standards of behavior. Provo- cative cultural differences have been found, particularly by contrasting moral decision-making patterns of Russian and American children. Peer group norms in the Soviet Union conform amazingly to idealized adult standards, while peer group norms in the 0.8. tend to conflict with idealized adult standards. One of the important variables that has been proposed by Bronfenbrenner to account for these differ- ences is the degree to which childrens' major socialization experiences are structured, consistent and coordinated. This initial line of research has been carried out elsewhere by others. However, apart from varying cultural settings, these studies have been more in the spirit of virtual replications, and therefore have not evolved new conceptualization of the problem, nor refined measurement " “e va'i (U :ua .. “ I ul:¢.3.l°u C A: 6",‘1 -.S: as carrot 11; iii-i. not or.‘ '. 'IA ' V “o .‘.H r r: .. ‘ u .. I If‘r o y . I v D ) r 4. a T I \:’. . “=5 .‘s'e ~ ‘ Q P: .5 sue C: ’2 i v w‘ \ I s v b"ah ce~v€ \4 C "I 9 “V“; n WC 1 VOINO: a..a‘ C . .‘CV I... ‘:_~D:' . .5s , y’ar‘. I «a 1|:" I ‘ Q\ "“5 0‘ n. ‘ . 0. ; .If‘s I s as. , \r: I: . ”‘51s , ‘ Q ~E ‘- u‘:- “35:41 y; ‘ S~Iae ‘ k I x ‘g \‘ ‘ \_r\ \_I of the variables involved, nor offered much critical evaluation of surface findings. Although the establishment of stable empirical findings is, of course, important, one cannot help but eventually ask what such findings may mean, not only in and of themselves, but beyond themselves in their implications for other issues. For instance, may we validly infer that, on the whole, children of particular social groups are in some sense "more moral" than children of other groups? What are the actual moral dimensions that are to be compared? To what degree might these dimensions differ from conven- tional conformity? What do the relevant socialization differences tell us about the larger systems which produce them? Do system differences imply the Operation of general variables that potentially may be applied to the analysis of "moral socialization" at many social levels, from small groups, through sub—cultures to nationestates? The study represented is an attempt to deal with some of these questions through a type of extended or "systematic” replication of the Cornell projects.5 In con— ducting this systematic replication, I have dealt with the same basic research problem and employed the same basic research strategies. I have utilized the original test instruments and attempted to sample from the same age group of<flfildren. However, I have also drawn samples from new populations, representing different kinds of subvcultures. I have attempted to refine and‘add to the original test List?" "5 al.- “. I :explored ' N‘eza i~=‘ I .l 1" CV“- ‘ Q “Fay- \ . OF R :N'U‘W t. t . M""flI'U-Q "in."‘“;tv l ‘ . “e u. :3 . C} v. '4' Le. \‘CS 0 1T. the :xz‘ . ":07. ‘1'- ‘J‘ it“. £r‘1 ”' v~‘ instruments, as well as introduce some new measures of unexplored variables. And, most importantly, I have tried systematically to conceptualize those aspects of the problem that appear to link social structure to moral conformity in children. More specifically, my study has three major goals: (1) to identify and measure general social-structural variables that lead to distinctive socialization eXperi— ences for children in different social systems; (2) to relate socialization differences to distinctive patterns of moral conformity; and (3) to modify the view of comparative moral behavior implied in the Cornell studies and, in the process, refine the instrument which has been used to measure moral orientation. Before elaborating the inten« tions of the present study, however, it will be helpful to summarize briefly the Cornell research that has been carried out to date. A condensed presentation of cross—cultural findings is given in Table 1.6 Review of Relevant Literature Bronfenbrenner (1962) initiated this line of inquiry with an article detailing his field observations of the intensive "character education" program that oper- ates in the Soviet school system.7 The crux of this pro- gram involved an overriding stress on collective respon- sibility for living up to a consistently presented set of social standards. .From the beginning elementary school Table 1 Cross Cultural Comparisons of Childrens‘ Mean MDT Scores as Reported In Bronfenbrenner (1975:485) Mean Across Country N Base Adult Peer Conditions U.S.S.R. 1. Boarding 188 13.82 15.62 15.04 14.83 2. Day 165 11.81 12.49 12.32 12.20 3. Hungarya 13.28 15.17 13.74 14.06 4. Czechoslovakiaa 10.36 10.38 7.64 9.46 5. Polanda 6.94 7.60 3.90 6.14 6. Japana 3.77 4.62 2.90 3.76 7. Canadaa 3.58 4.27 0.91 2.92 8. West Germanya 1.79 4.43 2.26 2.83 9. Great Britainb 274 3.71 3.12 1.05 2.63 10. U.S.A. 158 2.43 2.96 1.27 2.22 Israel 400 11. Kibbutza 2.26 1.80 0.62 1.56 12. Citya 2.77 1.22 0.52 1.50 13. Hollanda 1.27 2.10 0.16 1.18 14. Scotlanda 1.31 1.77 — 1.89 0.40 15. Switzerlanda'c — 1.59 — 0.76 — 3.91 - 2.09 aSample size not available. b Bronfenbrenner table. Beloff and Patton (1971) results; not included in original cSwiss children were, on the average, one year older than children in other samples. . . ‘ "'n-:.e. _ z'vbosg ‘ "‘ b w u-.,“(‘" “ ‘Iouu" twy‘, ‘ ‘ tents , ‘4 I. I: II ' i“ a, Ahe a-Ag.‘ . »I\\,latl:‘~ x. .. 5‘:I ' ' 70‘s! ‘1",674 4 “ N: I“ days, children are placed into various organized groups or ”collectives" that are set up to Operate according to specified adult standards of behavior. The collectives then quickly become the focal points for rewards and punish- ments. For instance, each child is a member of a series of "nested" collectives-~his row, classroom, school, school district, etc. Competition among collectives—-row with row, classroom with classroom, etc.——provide superordinate goals that foster cohesiveness while avoiding “in—group" vs. "out-group" antagonisms. Rewards and punishments are directed at various groups rather than at individuals. This, arrangement createsaistrong teamwork ethic and transforms the peer group into a powerful secondary source of adult influence. Indeed, according to Bronfenbrenner (1962:556) it is not long before the children's collective virtually ”surpasses the family as the principle agent of socializa— tion." Armed with explicitly developed adult criteria for evaluating performance and conduct, the collective is even- tually given responsibility for regulating many of its own affairs. Procedures encouraged for accomplishing this regulation include reporting observed peer deviancy to the group and public self-criticism of personal faults. The cbservable consequence of this controlled peer socialization is that Soviet children appear to be unusually well behaved and conforming to established adult norms. ‘AI : . .v...'::"lt1'v ‘ § . an «Ends “t VA“? A V ‘M F» If: “9.: ‘N. (l) \ P0 ‘ C (1‘ .l v- I U ’ ‘aq‘ts w in: C‘ass :CR‘ H In . ‘heS F0 ‘\ ‘\ n O . in z... ‘ b. ‘lC+; {‘35: 1 ° A :'\~ :C‘ \‘165 ‘ 7 0‘1 These initial field observations were put to an empirical test several years later after Bronfenbrenner and his associates had deve10ped an instrument (the "Moral Dilemmas Test") and procedure for measuring differences in conformity to peer vs. adult sources of influence in various moral situations (Bronfenbrenner, 1967). The Moral Dilemmas Test (hereafter referred to as the MDT) consists of a number of short stories which always pose the same underlying dilemma for respondents: whether to go along with friends who are urging and/or engaging in some behavior that would not be approved by adults, or to resist the pressure from friends and instead conform to the standard that would be approved by adults (see Appendix I for the full set of actual moral dilemma stories). These stories are presented under different experimental manipulations, including telling respondents that (1) no one will see their answers ("Base" condition), (2) parents will see their answers ("Adult" condition), and (3) classmates ‘will see their answers (“Peer" condi- tion). These experimental conditions, and the MDT itself, are discussed more fully in Chapter III. Bronfenbrenner administered the MDT to school (fluldren in both the U.S.S.R. (Moscow) and U.S. (Ithaca, New York). Results were dramatic and consistent with smedictions (see summary in Table 1, lines 1 and 2 vs. line 10). Soviet children averaged much higher scores under all experimental conditions, indicating a pervasive ———V aiult crier. scare $19.1; ' ’ 1 venom... ‘ ' b reaazness . ‘Iu ' 0 €118.25. '1- 3C'iiet C): i 1 I 'cp k. a” \Ujs1‘2ed 1 .‘1:H‘ k, A t“ dawn 1 ‘ngh‘ -.~b~ed a 5;. 4“ ”.C 4‘ Lb \ Q ‘ k 86le .‘L “3:: ““4: l.~re ‘ In," ‘\ usEd (WM, “\ \F .5 \.‘e ‘Qh u§cep1 q 10 adult orientation to moral standards (i.e., a high mean score signifies a reported readiness to conform to con- ventional, adult-approved standards. A low score indicates readiness to engage in mischievous activity being urged by friends in the dilemma story). This orientation was most impressively evident under the Peer condition; when Soviet children thought their classmates would know their responses, scores actually increased over those previously obtained in the neutral Base condition. In stark contrast, U.S. children's scores dipped to their lowest level when informed that peers would see all choices made in resolving the dilemmas. Girls were consistently more adult-oriented than boys under all test conditions for both groups. The U.S.S.R. portion of this study was later repli- cated (Bronfenbrenner, 1970), because Bronfenbrenner recognized that the previous sample of Soviet children had been drawn exclusively from boarding schools. This fact suggested a potentially important intervening variable, since boarding school children are virtually cut off from the day-to-day influences of home and family, and are thus all the more intensively exposed to the molding impact of the Soviet educational system. Bronfenbrenner reasoned that where the number of primary socialization agents is reduced ("monistic“ socialization), the child will be much nmre susceptible to consistent socialization influences. 0n the other hand, if the child is exposed to multiple Socialization agents (”pluralistic“ socialization): ‘ n . “send. “V e EEC €Xp€c§ ,. salt 301‘: $15 the f ‘ . 335“ ll Bronfenbrenner infers that influence will become fragmented and expectations will tend to be conflicting. Therefore, by sampling from regular Soviet day schools, in which children Spend only part of the day while the rest of their time is occupied with family and neighborhood friends, Bronfenbrenner expected to find somewhat less adult-oriented morality than revealed in the boarding schools. Nevertheless, adult conformity was still expected to be considerably higher than was true for the earlier U.S. sample. These predictions, as seen in Table 1, were borne out. This demonstrated the importance of degree of exposure to homogeneous socialization experi- ences beyond just the operation of the children's collec- tive. Similar investigations have been carried out in other cultures. Without actually presenting the data, Devereux, Bronfenbrenner and ROgers (1969) refer to pre- liminary results of MDT testing among children in England. According to their summary, English children were consis- tently less "adult oriented" than even American children. Within the English sample, boys were reported to be gener- ally more yielding than girls to peer pressure to violate adult norms. This sex difference is, of course, consistent vfixh the findings in both New York and Moscow studies. Responding to this tentative report, Beloff and Patton (1970) administered the MDT, under the usual experimental conditions, to a larger sample of British school 0:11. ..'.A' FRI-A ‘ :évuy‘“..:u . “is 6):“ng It... .“ “ 'h. 12 school children. At issue was a more conclusive test of the assertion that socially desirable response tendencies among British children would be relatively low, with pronounced peer pressure in opposition to adult standards. This expectation was partly based on the observation of an apparently high societal value placed on the function“ ing of independent peer group cultures (i.e., free from adult intervention) in Great Britain. Mean MDT scores showed that British children are indeed strongly affected by peer pressure along "anti— social" lines, but actually no more so than U.S. children. In fact, "Base" and “Adult" scores pg; s2 were a little higher than for their U.S. counterparts. However, belief that parents would audit test results did not lead to greater shift to adult conformity (when compared to scores initially obtained under the Base condition) for British boys and actually decreased scores for British girls. One particularly important outcome of Bronfenbrenner's Soviet testing was repeated here: Adult orientation of children attending boarding schools was uniformly greater than that of regular day school students. Also, as before, girls' scores were always higher than boys‘. MDT scores have also been obtained from a sample of Canadian school children (Mason, 1972). Not surpris- ingly, given presumably similar socialization environments, scores under each of the three test conditions are very cflose to those achieved by U.S. children in the initial :ww 5 “RV “.401. .w‘ e 1; - 35113.". j J sccres '12:. 303‘s shows; "'7" :sass were :‘ n I‘ . ‘g. bCHC‘§~ W| 13 Bronfenbrenner study. One difference that does emerge is a slight tendency for Canadian boys to decrease their scores under the Adult condition, whereas, of course, U.S. boys showed a slight increase in their scores. Canadian girls were consistently higher than Canadian boys under all conditions. Results of MDT testing have also been reported for Israeli children (Rim and Seidenross, 1971). In a non— Kibbutz sample, it was found that "anti-social" peer influence--for both boys and girls--was even stronger than for U.S. and British children. As in Great Britain (but to a greater degree) the Adult condition actually served to decrease conformity to conventional standards. Another surprising outcome was the fact that Israeli girls' scores were uniformly lggg£_than the boys' scores. These findings are difficult to interpret because Rim and Seidenross provide so little concrete information about their sample and no discussion at all of the social and cultural variables operating in Israeli socialization (urban, non-Kibbutz) that led them to hypothesize that “in a conflicting situation, with pressure from peers vs. adults, children in Israel would be more responsive to the pressure of their age groups (p. 36).” A separate study of Israeli children, very recently published by Bronfenbrenner and others (Bronfenbrenner, §£_al., 1975) lends some clarification to questions raised by Rim and Seidenross, as well as introduces new issues ~:e direc this study :0 the 30: 553285 for 53.362018 | 576873, t? ‘ 351 C02: :5 {Eflect Thus, Rich; 7:;- W2: .2 va'. ‘ .. :36 inch: Ti" ; ”My 1 ._ .. .‘sEaSurES h 5m ‘43}; and 1 ‘L \E in?!” “93.: 5‘... Midis 14 more directly relevant for the present investigation. In this study, Kibbutz and non-Kibbutz childrens' responses to the MDT were obtained and compared.8 The over-all scores for bgth_groups of children were low and quite comparable to the results obtained by Rim and Seidenross. However, the pattern of responses under the three experi- mental conditions for girls in each group was interpreted to reflect differences in collective vs. family upbringing. Thus, Kibbutz girls consistently scored higher than Kibbutz boys across test conditions, achieving their high— est score under the Adult condition. This pattern was just reversed for non-Kibbutz children, as Rim and Seidenross had previously found. In other words, girls raised under collective conditions were more "likely to shift their responses in the direction of conventional standards when threatened with social exposure (p. 484).“ Several speculations are offered as to why collec- tive experiences should influence girls but not boys. These include citation of experimental evidence that girls nay simply be more compliant with external conformity pressures (which are more prevalent in a collective set— ting), and the fact that Kibbutz boys are more exposed to the influence of both parents than are Kibbutz girls, thus providing boys with a more “pluralistic" socialization experience. Still, the overall conclusion is that the Israeli Kibbutz is a "less pure example of collective uPbringing than the Soviet Boarding School (p. 485)." B: esting 00:; and resgon; w . 1““! only“. .‘_ in}: ' th“tlon ( L 353:6 . ind in :r~.. l e;s‘dre f,- .K. pH ttQ' .. Condit 3.“. 'a 0,: the 1 6 9st: . . ‘nu n “i“ ‘ ”5110.1. 37“, . "953m r-l T! :C‘m. . .3 1h be: 558-13 acts I 15 Bronfenbrenner and his colleagues also make inter- esting comparisons between Kibbutz children's reSponses and responses already obtained from children in the Soviet Union. The magnitude of scores under each experimental condition shows Soviet children to be overwhelmingly more adult-oriented than Kibbutz children. These over-all differences are explained as a function of adherence to a morally strict and authoritatively imposed Communist ideology for the Soviets, vs. an implicit value orientation in the Kibbutz that favors a certain degree of autonomy and self-assertion. On the other hand, comparisons of scoring pattern§_ revealed that Israeli and Soviet children both tend to respond in the same way to pressure from peers as to pressure from adults. In other words, MDT scores under the Peer condition were not much.different from scores obtained under the.Adult condition--a result rarely reported from testing in other cultural settings. It is suggested that this similar outcome is due to the existence of single, pervasive national commitments affecting both the old and young in both countries such that social pressure from peers acts in the same direction as social pressure from adults. In the Soviet Union this national commitment is again identified as the relentless development of Communist ideals; in Israel, as simple national survival.9 Several investigations have been carried out that employed the MDT without including experimental manipulations (64., th. .1 53! 111516. v.” (""':r y 'd‘ “nut- 2'16 child: I 1 hence"? w ' r g. n? .6311 n t E» v \“fin Ch‘ ‘ ‘, y 3;: €321” hes L 4,: 5’3“; i" ’ M. ' 33; '1]! ‘ Rat ‘ n I inc “ 33‘: l6 (e.g., the Adult and Peer conditions). Devereux (1972), for instance, administered a shortened version of the MDT (under the Base condition only) to samples of U.S. and German children. By varying the absence-presence of the childrens' teacher during the testing, Devereux was able to show that the German children-dwho were presumably reared in a more authoritarian and adult controlled environment--were significantly likely to be more conform- ing to adult standards when the teacher was present. U.S. children, on the other hand, were not significantly affected in their conformity by either the absence or presence of their teacher during testing. These findings have at least two implications of interest here. The first implication is substantive: children who must live in strictly regulated, “authority- bounded" mOral environments may simply comply with ”correct" behavior when authority cues are present. This tendency would then naturally be at the expense of developing an autonomous sense of right and wrong. For German children, authority cues were provided by adults. But for Soviet children, within the childrenfs collective, peers serve an analogous function since the peer group becomes an extension of adult expectations. The second implication is methodological and involves the administration of the MDT in various cultural settings; i.e., in authoritarian groups it is possible that, even when teachers are absent from the room, the ;:ese:ce 0 I V ‘N rn fir» 'V‘aou A..- 9 in; it . d 1‘ h K.::y“‘. t «‘6‘. F \ .:\2 ‘q ‘tszaV-h * h b 1.6“ 3:." 1‘. u.‘ , \ {ES’ 1 I? ) r‘v:e¥. b . E: ‘l W C‘J\A 0“- ‘H a! 33.4; ‘5 A 0f a“ {NJ ‘II \=\‘ ‘ A ‘M 1? presence of adult experimenters may bias the children toward more adult—oriented responses. Again, this becomes problematic in assessing results Of Soviet testing but not, apparently, for assessing middle class U.S. results. Finally, Beloff and Temperley (1972) correlated MDT scores (again obtained only under the Base condition) of Scottish children with sociometric peer ratings. They found a consistently negative relationship between popular- ity and adult moral orientation for both boys and girls. At least within Scottish peer groups, then, the most attractive and influential members are precisely those children who stand most in Opposition to conventional adult norms. In passing, it seems reasonable to speculate that, in the Soviet childrenk;collective, this relationship would likely be just the reverse. Again, this should be the case since the peer group becomes a vehicle for imposing adult norms, and popularity within the peer group is importantly tied to exemplifying these norms. Summary of Research Findings One Of the major implications coming out of the research reviewed above is that, in a number Of Western countries, various societal values and norms regarding "proper" behavior are rather idiosyncratically presented to children. Relative lack Of a consistent and articulated code Of "proper” behavior is often thought to be partly a reflection of modern, urbanized society in which.an 1 increasing 265:3! GtCl LTtC the c EEEt‘or p by‘ . — r _ - - — ——' I‘ Q“ “l. h ‘- Lh‘ldr ‘lh‘s R . :7~._‘ “t‘n We for « )0 3'6; NIH ‘rh . L slnhl. 3 1 LI {Eta \: H t“ 18 increasingly broad range of institutional agencies supplant the traditional socialization role of the family (e.g., day care centers, schools, recreational programs, the mass media, etc.). These incursions of non-family institutions into the socialization process are not themselves, however, a matter of rational planning and coordination. Consequently, norms and values transmitted in these institutional contexts tend to be vague, situationally specific, and sometimes mutually contradictory. A related feature of modern, urban society in the west, according to Bronfenbrenner, is an apparent decrease in the integration of youth and adult realms of activities. concerns and responsibilities. Largely cut off from reward- ing ties to adult society, American children and youth develop strong identities within the age-segregated world of peer groups. To Bronfenbrenner and his associates, the evidence seems clear that, within these peer groups, norma- tive standards of behavior often move in directions opposed to adult expectations. However, there are other cultures and societies in which children's peer group norms appear to manifest very little conflict with adult expectations. In particular, as we have seen, Bronfenbrenner perceives and presents evidence for such a situation existing in the U.S.S.R. Although the Soviet Union is itself an industrial and increasingly urban society, the socialization of children there appears to be deeply affected by institutionally coordinated ani fie promulgatic 303% the young 935178 channelil 0f adult stem I i 50Viet child. s d the Child ' S wor] soup is itself one AlthOUg Man? ”vs th l9 coordinated and ideologically based programs which unify the promulgation and acceptance of adult values and norms among the young. This is partially accomplished by pur- posive channeling of peer group sanctions into the service of adult standards via such organized social devices as the children's collective. The likelihood, then, that the Soviet child's own standards of conduct will differ markedly from adult standards is diminished, since (1) the child's world Of group pressures and expectations is programmed to coincide with an adult world-view, and (2) the adult world-view incorporated into the child's peer group is itself a comparatively uniform and articulated one. Although Bronfenbrenner refrains from saying so, in many ways the ”children’s collective" may be viewed as a microcosm of the ideal state of affairs desired in Soviet society at large. While moral consensus in the U.S.S.R. is, of course, far from absolute, basic values that are deemed important are carefully infused by the state into many levels of social functioning. For instance, De George (1969) notes the impact Of Officially sanctioned programs of moral emphasis in areas of life ranging from factory work to Soviet science and even to the Soviet legal system. .According to De George, ”Communist Morality'l is seen by .Soviet leaders as a possible instrument to replace rule (of force and eventually perhaps even formal rule.of law through the individual ‘8 internal control of himself, in accord with cer' tion and discigl violate any of :- Reglacing l: extent alrez. 'corrades c:r offenses are peers. The the moral c: criminal C05. include publ - - - etc. ( 20 accord with certain norms, along with spontaneous correc- tion and disciplining by his fellow citizens should he violate any Of these norms. Thus, Replacing law (some) by morality has to some extent already been implemented in the Soviet 'comrades courts.‘ In these courts minor offenses are informally tried . . . by one's peers. The offenses are Often violations of the moral code not specifically covered by the criminal code . . . and the . . . penalties include public apology, social reprimands . . . etc. (p. 98). Notwithstanding the vision of various Utopian ideals, it is difficult to imagine that a system Of formal laws and Official legal sanctions could actually come to be totally supplanted by universal internalization Of "cor— rect" standards, even within a closely controlled politi- cal state such as the Soviet Union. However, it must be recognized that, in the U.S.S.R. (and apparently to an even greater extent in the Peoples Republic Of China),11 this is a goal that can be seriously considered and even successfully implemented, on a modest scale, within some sectors of social life. Both Of these facts appear to point to certain structural features of the Soviet system which facilitate, and perhaps even require, external imposition of moral consensus. In contrast, the relative diversity Of values and diffuseness Of socialization experiences that exist in the U.S. and other Western societies suggest somewhat opposite structural features. My attempt to explicate and generalize these system ___fi differences is i :ore fully in C': Sore Limitations 3113 £13 ratLCI‘. Antone?- associates have collective level in different cul 125137319 the ki L . ‘ that are cond“C 1' :0: eXaiple , Bro: coordinated SOCi exposed to. part in t- t .e SChOOlSo tzoetrination t school. .eetio of 0“ "wt .“Q p it be poss lies ' Exxstin s ft “16 o the r L) Elm. ‘\ 8V 21 differences is alluded to directly below and then taken up more fully in Chapter II. Some Limitations Of the Cornell Studies andElaboration Of‘the Present Study's Objectives Although Bronfenbrenner, Devereux, and their associates have been concerned primarily with comparing collective levels Of moral decision making among children in different cultures and societies, they do not go far in analyzing the kind Of social features and arrangements that are conducive for particular socialization outcomes. For example, Bronfenbrenner contrasts the intentional and coordinated socialization experiences the Soviet child is exposed to, particularly through the children's collective in the schools, with the casual and laissez-faire moral indoctrination that typifies U.S. childrenfs experience in school. This comparison is both interesting and important enough, but other questions remain. For example, moral laissez-faire vs. moral collectivization is an Obvious reflection of contrasting political and social structures. WOuld it be possible, then (let alone desirable), given these existing structures, for either country to success— fully adapt the general socialization techniques employed by the other? Indeed, by turning the issue around, the general system of socialization that has in fact evolved in these or any identify and ar. That is, we 51“.: system processe. tively move bac that generates J . ' ”‘Nn A‘O":u' «Ye incre? serial mits thy; (3M3, “H.\, ugh even 1' the“ is Certair a . .fi t 181 “‘5 p‘t ential if. 22 in these or any other social system should help us to identify and analyze important group characteristics. That is, we should be able to examine specific social system processes, such as socialization, and then induc— tively move back toward characterizing the larger system that generates such processes. The utility of this procedure increases when we are dealing with more obscure social units than those represented by nation-states (although even in the case of presumably well known states there is certainly ample room left for more careful des- cription and comparison of differences and similarities). Phrasing our questions in these terms should also lead us to consider the usefulness of Przeworksi and Teune's (1970) dictum that the ultimate goal of comparative research is to substitute names of variables for the proper names of social systems: When we find that societies differ with regard to a particular characteristic, we can ask what it is about these societies that causes this differ— ence. If the factor first considered does not answer the question satisfactorily, it is possible to consider other factors, gradually replacing the notion that 'nations differ' by statements formu- lated in terms of specific variables (pp. 29-30). A major objective of my study, then, is to identify general social variables that contribute to the differen- tial organization of socialization experiences for the young. One such variable that I have paid particular attention to is the degree of system "aperture-closure." The potential importance of aperture-closure as astructural d: gaseous social: Em U.S.S.R. ‘15 tion process (" M . .gntln“ «11:1, IEfEV NE at 3:1 ir EifS‘s‘e " I Mtlc atte r L 0 Ocial “a? be en made t «“53 0c M & nEigk‘bC 23 a structural dimension is suggested by the relatively homo- geneous socialization process ("closed") in countries like the U.S.S.R. vs. the relatively heterogeneous socializa- tion process ("open") in countries like the U.S.A. Again, although I wish to conceptualize aperture—closure as a continuum, reference to the U.S.A. and U.S.S.R. in this preliminary context is not intended to establish either absolute poles or preferred units of analysis, but simply to make use of two fairly clear examples of contrasting system types for which a variety of information and data are already available. Definition and discussion of the Open-closed variable is taken up in Chapter II. A second socia1-structural variable of importance for studying socialization of moral behavior is level of social class. General formulations about the development and expression of moral orientations have too often been based only on middle class samples. This criticism is applicable to the Cornell studies and their off—shoots, most of which have attempted to characterize the broad moral orientation of youth in the U.S., Soviet Union, Great Britain, Israel, and other countries without paying systematic attention to differences that might be attribu- table to social class levels.12 Apparently, some efforts have been made to include school children from different types of neighborhoods in order to improve the over—all representativeness of samples. However, most samples have still wound up having a "pretty heavy middle class bias" (see letter from Edward Devereux, Appendix III). Devere working class bcth U.S. and cues were eith ‘° surprising. the mtmducti Create a great: 521:1; working a 'H‘q; . . “‘5 finding is I . CL Evidence th 24 Devereux (1972) did make some rough comparisons of working class and middle class moral decision making among both U.S. and West German children when authority-related cues were either absent or present in the testing situation. Not surprisingly, social class differences did emerge: the introduction of authority-related cues appeared to create a greater shift toward conformity with adult norms among working class than among middle class children. This finding is, of course, congruent with the large body of evidence that links lower and working-class home environments to a general pattern of arbitrary, authori- tarian and punitive disciplinary techniques which, in turn, tend to mold children who "externalize" values rather than ”internalize" them (see, for example, Miller and Swanson, 1960; Bronfenbrenner, 1961). Additional reasons often advanced for class differences in value orientation include emphasis in the home on obedience and conformity rather than on self-direction, which is taken as a consequence of typical working class occupational structures (Kohn, 1969; Berkowitz, 1966, 1967). Greater peer and environmental pressure experienced by lower class children to violate conventional norms has also been suggested (Psathas, 1957; Wilson, 1959). While it seems evident that social class exerts an independent and powerful influence on the development of moral orientations, the strategy I have adopted for this study is to introduce social class primarily as a control variable, so tr. regresenting d; indez-endent in; is, I want to s closure after i. been taken into Critical has“ conceptuai 25 variable, so that I can more clearly gauge the general, independent impact of “open—closed" social systems. That is, I want to see if differences remain among populations representing different degrees of system aperture- closure after intra—system social class differences have been taken into account. Critical questions can also be raised about the basic conceptualization and measurement of moral orienta- tion in the Cornell studies. Bronfenbrenner, Devereux, and others working with the same problem often appear uncritically to equate moral behavior with conformity to prevailing societal norms. This restricted vieWpoint is further limited by employing a measure (the MDT), which emphasizes negative and passive moral expressions, i.e., behavior is counted as pro-social and, therefore, "moral,“ to the degree that a norm is not violated or that tempta— tion to do wrong is resisted. Stress on moral prohibitions and degree of compliance with rules has both characterized much other research and been criticized before for the one-sided view that it fosters (Breznitz and Kugelmass, 1968). Aronfreed (1963), whose own work has included investigation of moral responses to transgression, concludes that Most of the psychological research on moral behavior has focused on its prohibitive and punitive components. Yet people are obviously moral in a broader sense than that of merely avoiding reacting to transgressions . . . It is clear that the reinforcements which.originally define transgressions and their consequences may be verY ‘31: to be expe: (pp. 437-43 A relat that the actual the ET appears 5313:; general cent of moral v: tat underlie t": 1:!- l .6... to be as c: an instrument U m- - ...e orientatii MY include 0‘: tion ~ ' SmPathy, 1&"n ‘WVUESS of 7?. films-vb 3 & complain is , ‘ Pittel and u 26 be very different from those which define actions to be experienced as rewarding or praiseworthy (pp. 437-438). A related shortcoming of the Cornell approach is that the actual content of the "moral dilemmas" used in the MDT appears inadequate as a sample of the theoretical domain generally understood to be encompassed by the con- cept of moral values. For instance, the moral principles that underlie the MDT items are few in number and do not seem to be as crucial in import as one would hope for in an instrument that purports to be a general measure of moral value orientation. These principles, as measured by the MDT, only include obedience to parents' wishes, achievement motiva- tion, sympathy, rule adherence and rule enforcement. This narrowness of measurement scope and content is a common enough complaint about research in the area of moral values. As Pittel and Mendelsohn (1966) point out in their general review of moral measurements: Many tests sample only a small number of moral or ethical areas . . . thus limiting their generality. The content typically sampled is based on categories of conventional morality or on the author's theoretical preoccupation. Many dimensions of behavior which are poten- tially morally salient are thereby excluded (p. 33). If the MDT may be criticized on grounds that the moral dimensions which make it up represent too constricted a range and lack sufficient salience within the range that is encompassed, then we must call into question the kind of generalizing that seems to be implied by some of the Cornell studies. That is that children from certain groups tend to be 'more it just specific: test this not] I33334 concern D p C: Rafi 27 to be ”more moral" than children from other groups, not just specifically but in general.13 In order to adequately test this notion, it is necessary to expand the focus of moral concern to a correspondingly general level. Therefore, another extension of previous research that I am attempting in the present study is to supplement the MDT with additional new items designed to (1) require positive and active responses to moral situations, i.e., the moral actor must do "good" rather than simply avoid "wrong-doing," and (2) reflect a broader range of moral values such as justice, altruism, sharing, keeping promises, etc. Elaboration of these objectives constitutes the second major conceptual focus of Chapter II. CHAPTER I--FOOTNOTES lSee Reger Brown (1965) for a very helpful review and integration of these three themes.- zRoger Brown (1965:404) makes this same point with the reminder that while "mOral values carry a sense of absolute obligation, of 'ought' or 'should,‘ . . . other values [i.e., non-moral] carry only a sense of desirability, of 'I like' or 'I want.’ Aesthetic values, economic values, the values of success and of health generate preferences and yearnings rather than categorical imperatives.” 3Comprehensive reviews of the literature on rela- tionships between social class and issues related to the study of moral values have been carried out by Bronfen- brenner (1958) and Hess (1970). 4The theme of culture, socialization practices and personality consequences has been much more widely inves- tigated, particularly by anthropologists. One recent sociological study of culture and personality that has .some relevance for this dissertation is the attempt by Gonzalez and Tamayo (1974) to assess the extent that "open" societies produce "open-minded" persons and "closed" societies produce "closed-minded" persons. Their usage of Open-closed society is confined to informal and intuitive categories while my own use of this same variable attempts a more systematic and measurable conception of Open—closed social differences. SFinifter (1973) has developed a very useful typology of replication strategies wherein the major generic distinctions are identified as being "virtual" replications, ”systematic" replications, and "psuedo" replications. These replication modes are represented as intersecting with the type of data base used to produce a specification of sub strategies. By Finifter‘s scheme, my study represents the sub strategy of "inferential derivation"--the intersection of an attempt to systemati- cally replicate with an independent sample. According to Finifter, "the inferential derivation replicator is free to construct a line of reasoning that only indirectly and 28 nferentially 5fee to choos: Which ti .3eor' J lSthat the It 533" how they 11‘ “(5985 bet'n 2.34:. .S to test; SMA‘Y he is {g uata .3" Nada, Polar to I73)? howled - :e literati ‘nre 531.1001 SparaFiC t°ws rd tit} 29 inferentially represents the initial findings . . . is free to choose or design a study to include any variables which theory indicates are relevant. The main requirement is that the replicator who uses this strategy explicitly show how there are reasonable conceptual or empirical linkages between his derivations, the new hypotheses he plans to test, and the specific findings of the reference study he is replicating." 6Data reported in this table for Hungary, Czech03* lovakia, Poland, Japan, Holland and Switzerland have not, to my knowledge, been formally published and discussed in the literature. 7This program is more fully described in Bronfen- brenner's book, Two Worlds of Childhood; the U.S. and U.S.S.R. (1970). F 8Using confession as an indicator of moral orienta- tion, Luria,et al. (1963) have already demonstrated the greater sensitivity of Kibbutz children to transgression situations when compared to Israeli children who live at home with their parents. 9The notion that moral socialization may be affected by a general social characteristic such as "national commit- ment” points in the same direction as questions that I have posed in the present study. My attempt to arrive at a more systematic understanding of importantly involved social characteristics is taken up in the next chapter. 10Bronfenbrenner does not really elaborate, in his published reports, concerning the kinds of arrangements and preparations made in advance of his testing in Soviet schools. Without information to the contrary, we are led to wonder if students might not have been affected by prior knowledge that an American scientist would be coming to involve them in an experiment. Even if no prior knowledge of the experiment existed among students, the context of the actual testing would surely make clear the propaganda value of impressing American researchers with "correct" responses. 11Sidel in Women and Child Care in China (1972), presents a fairly recent and'informative observational report of the massive moral and political indoctrination program that permeates all phases of childrearingin China today. 12Bronfenbrenner is certainly not generally insensi- tive to the importance of social class factors. Indeed, his earlier work on social class differences in child- rearing practices is well known (1958). Nevertheless, consideration in much of the Bronfenbrenner were difficult fficially a u should still 1: those parents a distinction found to be 'rr. l3For trenner makes q 1" Af-"serican so: , 3O consideration of social class differences is not apparent in much of the actual research done with the MDT. In Bronfenbrenner's case, it may be that class distinctions were difficult to make in the Soviet Union, since it is officially a "classless'I society. However this may be, it should still be possible to distinguish between children whose parents are, say, bureaucrats vs. factory workers, a distinction that Kohn (1969) and Berkowitz (1966) have found to be important in the differential development of value orientations. 13For instance, in Two WOrlds of_ghildhood, Bronfen- brenner makes it clear that he views his findings as point- ing to "disruptive trends in the process of socialization in American society (p. v)" which are "of sufficient gravity to require some programmatic action (p. viii).” He then sharply contrasts this state of affairs with the effective- ness of Soviet upbringing methods. The empirical support for these assertions is found in comparing Soviet and American childrens' MDT scores. By implication, then, Soviet children are considered "more moral" than American children. Elsewhere, Bronfenbrenner (1968:203) explicitly equates higher MDT scores with a "greater claim to virtuous behavior.” CHAPTER II THE THEORETICAL FRAMEWORK Introduction In this Chapter, I will attempt to sort out and discuss the conceptual issues that have been central to my investigation of conformity to moral standards. Chief among these issues are (l) establishment of "open" vs. "closed” systems as a theoretically general and important social-structural variable in the socialization of moral conformity; (2) clarification of the underlying moral dimensions which appear to be at issue in the testing con— text established by Bronfenbrenner and his associates; and (3) linkage of aperture—closure to dimensions of moral conformity through the development of theoretical hypo- theses. gpen:Closed Systems The original Bronfenbrenner studies can serve as a point of departure for developing the theoretical framework for the present study. As we have seen, Bronfenbrenner points to specific features of socialization in the U.S.S.R. 31 to account for :nfomity bet'n these include: involvement of younger age or: and important eI 0: group forces i” 3‘1: ing group gram rather th 5L1: rewards , et 8. the name of M4 . 32 to account for the large differences in degree of adult conformity between American and Soviet children. Some of these include: use of social models through large scale involvement of older children and adults to work with younger age groups; "character education" as an explicit and important expectation of the school system; employment of group forces in reinforcing desired behaviors within an enduring group context (i.e., peer and self-criticism, group rather than individual competition, group punishment and rewards, etc.); and the assignment of responsibilities in the name of superordinate goals of the classroom, school and community. One outcome of these and related socialization strategies is the prevention of value conflict. Thus, because Soviet children are confronted with less divergence in moral outlook in their interactions with other persons in family, school and peer groups, their conformity to a more consistent set of standards is naturally increased. This explanation is given added credence by both Bronfenbren- ner's (1970) and Beloff and Patten's (1970) finding that adult conformity is greater for boarding school children than for day school children, whether in the Soviet Union or in Great Britain. To account for this difference, one simply extends the original notion concerning imposed consistency of standards:. boarding school children are exposed to fewer socializing agents while day school children are eueshed in a . increases the : tations. Now, t; this analysis Within a broad: 34:31 ConcePt P‘IPose is the This concept ha ccznitive SYSte fie“been by R: {"CC’SnitiVe sin ti Schweder l ( ”‘9" 1961). an the . Se th eOr belief 33 enmeshed in a multiple socialization setting, which increases the potential for exposure to conflicting expec- tations. Now, the major theoretical task here is to place this analysis of differences in socialization settings within a broader, structural framework. One general struc- tural concept that suggests itself as useful for this purpose is the notion of "Open" vs. "closed“ systems. This concept has been applied more at the level of individual cognitive systems as illustrated in the cognate approaches developed by Rokeach ("open-closed mindednessr 1960), Bieri ("cognitive simplicity—complexity," 1959), and Harvey, Hunt and Schroeder ("abstract vs. concrete cognitive function- ing," 1961). Although specifics vary, a common focus for all these theories is how values, ideas, attitudes, and beliefs become patterned into coherent and interrelated systems, and the degree to which such systems, once con— structed, are conducive and receptive to the input of new and variegated information from outside the system. For instance, the "closed-minded" individual tends to maintain the homogeneity of an already fixed set of beliefs and assumptions by rejecting or distorting new and different cognitions. In contrast, the "open-minded" individual tends to be more receptive to new cognitions. at least for purposes of evaluation, and hence more flexible in modifying, enlarging and reorganizing his belief system when deemed appropriate. The no individuals wt. from their env collective or | the way they 5* bitenee. the ; “Vidal fine bound to author tioned group is h' n elrned raiecti. to such SOC-1&1 Exclusin mew: “mouse wi di~ Q'Se‘nination Fina1er close is analogOuS ‘ As Wi‘ ire -, . 'l‘l’ldual s o 34 The mode of psychological integration preferred by individuals when confronted by diverse cognitive inputs from their environment poses a clear analogy for analyzing collective or social integration achieved by groups in the way they structure their normative systems. For instance, the rigid belief system of the closed-minded individual finds its counterpart in groups that are tightly bound to authoritarian ideology, sacred traditions, unques- tioned group myths, pervasive taboos, etc. The closed- mdnded rejection of unfamiliar cognitions may be likened to such social phenomena as ethnocentrism, xenophobia, exclusive membership requirements, prohibition of social intercourse with outsiders, as well as the prohibition of dissemination and/or practice of alien customs, etc. Finally, closed-minded distortion of unfamiliar cognitions is analogous to out-group stereotyping. As with the cognitive system of open-minded individuals, the social system of open groups reverses the above tendencies by virtue of its greater flexibility and adaptiveness. Thus, open social systems would be characterized by institutionalized procedures for making normative and ideological changes, allowance for freer social exchange with non-system members, greater permitted diversity of opinion and behavior, etc. From this rough analogy, it may be argued (as indeed, Lauman and Schuman,l967, have already done) that a closed-open conceptualization of social structure has inch in comon Folk-Urban, Ge: hindein‘s Rec? | 'Hechanical' s: is said to be a! or honOgeneity differences be: J an 'Organic“ ice mutual depe Esterose neous L Open-c; tplications t'. r“, r«6~5 Odern COT; \ 35 much in common with such classical community dimensions as Folk-Urban, Gemeinschaft-Gesellschaft, and especially Durkheim's Mechanical vs. Organic solidarity. In a "Mechanical" social structure, for instance, integration is said to be achieved through the fundamental alikeness or homogeneity of its units. Thus, it is "closed" to differences between units. On the other hand, integration in an "Organic" social structure is brought about through the mutual dependence upon one another of dissimilar or heterogeneous units. Thus it is “open" to differences between units. Open-closed group differences have fundamental moral implications that are most apparent in contrasting small, pre-modern communities with large, modern communities. For instance we can contrast the transmission of moral values in “simple" vs. "complex" societies: "In a tight knit and narrow community, all the influences bearing upon individuals are united in pressure towards social conformity," thus drastically reducing the incidence of moral deviancy and ensuring social unity (Bull, l969:3). Or, as Ossowska (1970:43) sums it up, “In folk societies, moral order pre- vails, the binding together of men consists in common conceptions as to what is right, in common ideals, in common convictions as to the good life." All of this stands out vividly against the social and moral constraints Oper- ating in complex, urban groups wherein "a technical order Predominates over a moral order . . . the bonds that coordinate the not rest on Shi ”LL .awer from mu: (Cssowska. 197C Nadel's self-r gulatior. teen-closed sys ,cm. «nation are tecause of thei :::iation, ince ir“. A . tea to inext 36 coordinate the activities of men in the technical order do not rest on shared conviction as to the good life . . . rather from mutual usefulness, deliberate coercion . . . (Ossowska, l970:43)." Nadel's (1971) discussion of social control and self-regulation also lends substance to a conception of open-closed systems. Most of Nadel's examples of self- regulation are drawn from "primitive" societies where, because of their small scale and lack of internal differ— entiation, incentives to conform to normative standards are linked to inextricably related social roles. Thus, "the public standards of conduct are affected only because every man is both a family head and a potential rank-holder, a farmer and a possible candidate for Priestship (Nadel, l971:8)."- . However, a system of social control based on self— regulation need not be confined to simple societies, as we learned from DeGeorge's discussion of "comrade courts" in the Soviet Union, and as Silver (1966) makes clear in analyzing the orthodoxy requisites for role and status allocation that developed in theAmerican South. Finally, as Nadel (1971) himself points out: Highly complex societies, too, exploit it whenever they are organized on 'totalitarian' lines, e.g., - when social promotion of any kind is impossible unless the candidate professes the 'right' kind of religion or political conviction or lives accord- ing to approved standards of morality (p. 9). Of course, even in complex but non-totalitarian societies, self-regulation based on conformity to approved series of condu: social control socialization ; tinction to be closure represe iistinctions i." relative Openne eatination of Its the establ The deg 37 modes of conduct and thought is a critical element of social control and one of the basic functions of the socialization process in any group. However, the dis— tinction to be made here is one of degree, since aperture- closure represents a continuum. One of the important distinctions implied in this continuum concerns the relative Openness of groups in at least permitting a fair examination of ideas and modes of conduct that depart from the established norms (vs. out-of-hand-rejection). The degree to which normative deviations are tolerated in a group, the degree to which such tolerated deviance is privately or publically acknowledged, and the range of deviant behaviors that may be engaged in without social stricture are the foci of Ryan and Strauss‘s (1954) conception of "loose—tight" social structure. Their notion has several apparent similarities with the conceptualiza— tion of open-closed systems that I am proposing here. Thus, a "loosely structured" society is characterized by social norms which explicitly admit a wide range of alternative channels of conformity. Behavior which.never~ theless exceeds the variety of acknowledged normative Jbounds is still broadly tolerated. Finally, the values of (group organization, formality, permanence and solidarity rare underdevelOped such that group roles are subordinate to individual ends. I All t? toward the 'ti cox-e to reverE new tive patt with enphaSiS and iuties in Sperli of closed socil the thrust of harmony, and co being fundanen Lie conflicts 38 All this is in marked contrast with societies toward the "tight" end of the continuum which increasingly come to reverse these same characteristics. In general, normative patterns will be clearly and narrowly marked with emphasis upon the observation of reciprocal rights and duties in specified situations.l Sperlich (1971) discusses the intolerance aspect of closed societies from a political point of view. Since the thrust of closed social systems is toward order, harmony,and consensus, Sperlich views such systems as being fundamentally anti-political, having little use for the conflicts implicit in political activities. According to Sperlich, the touchstones by which Open and closed sys- tems can be identified politically are, "regulation of con- flict (no more than necessary) and suppression of conflict (as much as possible). No society is perfectly closed, but a society will move toward one or the other of these two types to the degree that it suppresses or welcomes and regulates diversity and conflict (p. 187)." Sperlich's discussion is, in turn, admittedly much indebted to Popper's (1966) earlier, more Philosophical work, The Open Society and Its Enemies. A case in point is Popper‘s attack on Plato's political theory which, according to POpper, provides the prototype for the "closed society": [Plato] certainly believed in . . ..a general historical tendency towards corruption, and in the possibility that we may stop further. corruption in the political field by arresting all political change_. . . He tries to realize it by esta the evils degenerate which is f ‘ is the bes of the 501: the arrest" Mainta,I time and in a V obviously no e5 accomplishing l of some minima‘ general applice elated set of ings, impulses amp members In clo eliplicit on th 39 it by establishment of a state which is free from the evils of all other states because it does not degenerate, because it does not change. The state which is free from evil of change and corruption is the best, the perfect state. It is the state of the Golden Age which knew no change. It is the arrested state (p. 21). Maintainance of even a "working" consensus over time and in a world fraught with complex issues is obviously no easy matter. Various social mechanisms for accomplishing this are well known and include establishment of some minimal ideological bases for conduct that have both general applicability and broad appeal such that an arti- culated set of guiding principles can “resonate" with feel— ings, impulses, predispositions and circumstances common to group members (Tomkins, 1965). In closed societies, ideology becomes increasingly explicit.on the one hand, while increasingly employed to simplify and/or distort past and present problems on the other hand. In less developed closed societies, "magic" and myth making Operate to reinforce group taboos and the prevailing worldview (Bergson, 1935). The function of myth is no less important in more deveIOped, or modern, closed societies in which public Opinion is manipulated, and ideologically supportive myths are promulgated through centralized control of information channels (Silver, 1966). In contrast, open societies are characterized by the permissible operation of rational critiques that challenge group myths- (Bergson, 1935). The eff gropaganda contI lecitinacy atta illustrated in totes, he Cor: I riardian of Sov Laership . Tr.- Ls given a moral l5:;er\’i8(alrs E1176 fits into the t: hoped" that when “to“: sail and nece: vases may help ' Eatisf mi“ in tread moral ac lfibnh “firefly forms 33.3%: slide, and at 40 The effectiveness of ideological coherence and prOpaganda control are heightened by a sense of moral legitimacy attached to system agencies. This is well illustrated in the Soviet Union where, as DeGeorge (1969) notes, the Communist party is recognized as the “guide and guardian of Soviet morality" as well as providing political leadership. Therefore, Party decisions "carry moral as _ well as political authority . . . Morality is thus a means of developing patriotic consciousness and strengthening the authority of the party (1969: 109)." Even routine labor is given a moral quality through appeals to Communist tenets: "Supervisors are encouraged to show workers how their work fits into the total scheme for producing Communism; it is hoped that when they understand how their work is socially useful and necessary, the moral satisfaction of doing their tasks may help to make up for the absence of creative satisfaction in the task itself (1969: 110)." Given wide— spread moral acceptance, public opinion can then be "properly formed by techniques of mass education and propaganda and . . . relied upon to continue the education of individuals. Public opinion is to correct the individual and to help him analyze and improve his conduct in the light of the collective. The criticism of the collective is to stimulate and reinforce individual self-criticism (l969:ll4)." Where moral legitimacy is not accorded to the instituted authority structure Simmel (1964) reminds us of the paradox societies' thl ”he conflictfu not itself dif Where the (I is strong zation, it associatic:I contents a: these grog; centralizai perogative Unitary f0: «ht What is in 1 Piano menon I is The Effect ’55 Vitaliz cohesion. fEEIings Q bets, All 41 of the paradoxical tendency towards closure of "Secret societies" that Spring up in opposition to the regime. The conflictful relationship between the two systems is not itself difficult to apprehend: Where the overall aim of the general society is strong (particularly political) centrali- zation, it is antagonistic to all Special associations, quite irrespective of their contents and purposes. Simply by being units, these groups compete with the principle of centralization which alone wishes to have the perogative of fusing individuals into a unitary form (p. 375). But what is ironic, as Nisbet (1966) elaborates upon the phenomenon, is that The effect of conflict, of persecution, can be as vitalizing to the secret society's sense of internal freedom as it can to its sense of cohesion. Secret societies often combine feelings of freedom on the part of their mem— here. All that makes the members wish to with? draw from the felt oppressions and frustrations of the larger society causes them to wish to intensify the feeling of oneness within the secret society. Such intensification commonly leads also to centralization and rigor of authority within. Gradually the monolithic character of the small society comes to seem a very sign of the members' release from the tyrannies and corruptions of the outer society. In the totalism of its power, the secret society feels, not despotism, but a new form of freedom, one in which all may participate, one freighted with redemptive mission. The whole history of religious sects and revolu- tionary movements is illustrative of this (pp. 163-64).2 Of course, even in presumably closed societies (particularly in those that are also large and complex), we find discrepancies between the ideal and the actual and variation in acceptance and adherence to "preferre “ group standards. St: in the differer 55? school sturl these discrepa: ties. What a;: the ever Preset in the SOCiety )3! ‘; ' “‘\‘ate SOCIE mileagues’ and as a faithful Pressures 6X8“ If Con! mmment. t) “he standards l ”in “we group S . at! _ ~C EXamlned t 42 standards. Such discrepancies are illustrated,for example, in the difference in moral conformity between boarding and day school students in the Soviet Union. And, as we know, these discrepancies are much magnified in more open socie» ties. What appears to be crucial for the individual, given the ever present likelihood of at least some value conflict in the society as a whole, is the degree to which the immediate social environment (e.g., family, friends, colleagues, and other "significant others") actually serves as a faithful and accurate mediator of the normative pressures exerted by the larger social system. If conflict remains even within the immediate social environment. then individuals are likely to fall back upon the standards espoused by the most salient of their refer~ ence groups. A case in point is provided by Rosen (1955) who examined the immediate antecedents of Jewish adolescent attitudes toward the use of kosher meat-an issue that was a source of conflict in the Jewish community being studied. Attitudes about kosher meat--both positive and negative--were found to be highly related to the position taken by either family or peers, depending upon which one was rated as the most important reference group (measured by observation and sociometric reports). In another study relevant to the moral impact of reference groups, Johnson and Neal (1970) found that voluntary participation in pro- social activities (e.g., community action prbgrams and the like) was significantly related to perceived reference activitY- A more closed 9‘0“? pl» acholedges a m theory aP rajor distincti systems, met-bet unchanging. Wht state of const (having affini said to derive is follows. First affected by d 43 group norms, whether supportive or rejecting of such activity. A more systematic and abstract treatment of open vs. closed group phenomena is presented by Ziller (1965) who acknowledges a debt, by broad analogy, to the general sys- tems theory approach developed by Bertalanffy (1950). The major distinction suggested by Ziller is that in closed systems, membership composition tends to remain stable and unchanging, whereas membership in open systems is in a state of constant flux. Several important consequences (having affinity with points already discussed above) are said to derive from this difference, and are summarized as follows. First, normative structures and decision making are affected by development of different time perspectives. In closed groups, relationships are constant, long lasting, interlinked,and thus established traditions come to prevail. In open groups, however, relationships are transitory: "the future is indistinct,“ and thus actions are governed by expediency demands of the present. Interpersonal sanctions are more difficult to maintain in open groups. For instance, "in open groups, either a or b may leave the group volun- tarily . . . [which] tends to reduce the power aspects of the relationship since the length of time is reduced over *which.either person is able to impose sanctions on the other (p. 169)." Reciprocity norms in open groups, there“ fore, tend to function between the individual and group as a group rat}: be the case in Second, I severe "equili‘: such factors as ~ . - etc. On- =°Pin9 with th Elaberation of “.4“ . Snags. .GIIOVJP 44 as a group rather than between individuals per se, as would be the case in more closed systems. Second, Open groups are confronted with a more severe "equilibrium“ problem than closed groups, due to such factors as changing leadership at inopportune times, . . . etc. One characteristic Open group tendency for coping with these kinds of disruptive forces is an elaboration of formal role structure and secondary relation- ships: "Group members react to the position to which the new member has been assigned rather than to his charac- teristics as a unique individual (p. 167).“ Third, because of the continuous flow of members, Ziller affirms that in contrast to closed groups, open groups have an expanded frame of reference. As alternate modes of belief and participation become available in the environment external to the open group, non-conformity to group standards is perceived as being less important in its consequences. But in closed groups, "members learn only a definite course of ideas, opinions, beliefs, and values. They have little opportunity to listen to many ideologies and learn different beliefs (p. 168).“ There— fore, the consequences of non-conformity to group standards are perceived as much more threatening in closed groups. General Definii Sven-Closed 5y.» Al thou Ifbership f 1:] is that it tre. generalizabili sub-systems wi mite opposite whole (e.g., a exclusive club Ofsanization i: 3PPIOéCh coinc: altific‘ately will I haVe Love llWhich be 2”ing. 45 General Definitions for Identifying Qpen—Closed Systems 4Although somewhat limited by focusing primarily on membership flux, one of the virtues of Ziller's formulation is that it treats aperture—closure as a variable that can theoretically be applied to any social grouping. This generalizability is particularly useful in distinguishing sub-systems within larger systems--when the part may have quite opposite aperture-closure characteristics than the whole (e.g., a religious sect in a secular society, an exclusive club in a socially mobile society, a military organization in a democratic nation, and so on). This approach coincides well with my own intention in develop- ing a general conception of open-closed systems that ultimately will be useful for understanding moral conformity. I have attempted to assimilate those ideas reviewed above which best appear to serve this purpose and, in so doing, have arrived at three general dimensions which I consider to be fundamental in defining open-closed social systems. These three general dimensions are: (1) The degree to which a particular social system is linked to other social systems within the larger social order; (2) the degree of consensus among individual units of a system concerning issues that need to be resolved by that system; and.(3) the degree to which a system is capable of sanctioning d8\ normative patté Each of ism into more below. Note 1:} its descriptior It may be assurr 513.3111 reversir. A. Wie 535951 to the l. 2. 3. 3- Wit 46 sanctioning deviance from, or conformity to, established normative patterns. Each of these broad characterizations may be broken down into more specific definitional points which I present below. Note that since open-closed represents a continuum, the descriptions are phrased in terms of relative closure. It may be assumed that relative aperture can be defined by simply reversing the directionality of the statements. A. With regard to Linkage, a social system is closed to the degree that 1. little contact occurs between that system and other inclusive systems 2. when contact does occur with other systems, access to and exchange between the system in question and different systems are made difficult 3. that system is independent of other systems in regulating its own major social func- tions (e.g., political, educational, economic, etc.) B. With regard to Consensus, a social system is closed to the degree that 1. within member networks of significant reference groups, unanimity exists in defining acceptable behaviors and beliefs 2. normative definitions are shared by all potential reference groups within that system 3. a common, system—generated ideology exists which gives direction, coherence, support and substance to normative definitions 4. the ideology is explicitly formulated and officially legitimized 5. the flow of information is controlled, and communication channels are subservient to ideological expression C. Wit; is closed to th 1. 47 C. With regard to Deviance and Conformity, a system is closed to the degree that 1. sanctions available to that system for rewarding desirable behavior and punishing undesirable behavior are both powerful and cover a wide range of possible expressions 2. sanctions—~and their individual consequences-- are highly related to other areas of system functioning and control 3. sanctions and controls are accorded a high degree of legitimacy 4. the governing structure which administers formal sanctions and controls (and in whose name informal sanctions are justified) is authoritarian and centralized The above definition has the virtue of bringing together a number of related conceptual elements that have previously been found to be important. It is obvious, however, that each sub-dimension is itself in need cf fur— ther refinement, clarification of terms, and operational definitions, just as it is obvious that not all of these sub-dimensions could be easily comprehended by a single empirical research project. Indeed, some of the sub—dimensions point to specific variables that, taken singly, would be very difficult to measure empirically. For instance, we would no doubt be justified in regarding the political structure of the Soviet system as being authoritarian and centralized, but how much more so than that of, say, the U.S.? Or, if we are deal- ing with less inclusive systems, it would prove no easy task to empirically gauge the degree of difficulty of "access and exchange" between say, a Jewish population and the larger malty in ad the lar; eases. diffe rits of co: m“ ex;erier Selection of “23553535 n Present Stuc‘ \ Cert ks and \ kt. “h are 5+ referellce c “lne the EF-‘m: . “9311129 '>‘ . ‘V‘I‘lal vs 1“ tral' lb, 48 community in New York City compared to a Jewish population and the larger community of Moscow. In these kinds of cases, differences may be assumed to exist between the two units of comparison, but largely at a gross observational and experiential level. Selection of Open-Closed Dimension of “Consensus“ for Elaboration in the Present Study Certain sub—dimensions defined above, however, may be more readily and intensively pursued in empirical study. These are some of the sub-dimensions that define the general system property of "consensus.“ The aspects of Open-closed systems that come under scrutiny here are the degree of intentional patterning of norms and values into a coherent moral system, and the degree to which this system is sup- ported by an extended network of groups of people who inter- act with each other mainly on the basis of these shared norms and values. The larger the number of norms and values that are shared, and the larger the number of potential reference groups that share them, the more "closed" we may define the system to be. Laumann (1968) emphasizes a similar approach in examining open—closed friendship systems, or what he calls "radial vs. interlocking associational networks." In general, a radial or open associational system exists when the acquaintances of an individual do not interact among themselves. In an interlocking or closed system, however, an individual's friends also know and interact with one zest, corn: greater soc 31:219.. the "man o :3 + unity. 5“. C '::Savn 49 another. As associational networks become more interlock- ing, they are characterized by increasing intimacy, commit- ment, commonly held definitions of the situation, and thus greater social influence over given individuals located within the network. The antecedents of these theoretical expectations are, of course, well known for small group research on cohesiveness. For the purposes of this study, then, "Consensus" is the open-closed dimension that receives primary attention. As will be shown later in Chapter III, "Consensus" is measured by a type of network analysis that takes into account the amount of contact among actors in a system who are assumed to share a set of common standards. The system actors of interest in the present context are: the child, who is pre- sumably being socialized within the system, and the major agents of socialization for the child, including parents, teachers, relatives, siblings, neighbors, and the child's own friends. Even though the measurement focus of this study will be on the "Consensus" dimension, it must not be supposed that the “Linkage" and "Sanction" dimensions become somehow reduced in actual importance for the par- ticular systems being examined. Since the effects of these other dimensions on real life systems cannot be controlled for, their impact must be assumed, even though it is not directly measured. Therefore, I have tried to take into account these dimensions of open-closed systems by select- ing systems for comparison in which differences between all of the di: .‘tre accurate sub-system w. iistinguishab “st to def inl 113: is to de: Pct-ulation sag aitached to t. 3.95"?“ ' “de33 I 1 citations a 50 all of the dimensions may be assumed on an a;priori basis. More accurately stated, pOpulations representing separate sub—systems were first chosen for study that appeared to be distinguishable by the three general criteria that I have used to define open—closed systems. Once chosen, my inten- tion is to demonstrate an empirical difference between population samples that derives from the specific criteria attached to the dimension of system consensus. Population descriptions, specific sampling procedures, and sample limitations are discussed in Chapter III. Some Basic Moral Dimensions: Negative vs.fiPositive, Passive vs. Activefifi It now remains for me to clarify the moral dimen- sions that appear to be most relevant in the context of open-closed social systems and to suggest what I think the relationship may be between these particular social and moral dimensions. In so doing, however, it is certainly not my intention to attempt an extensive inquiry into the nature of "morality," the "good life," the “virtuous char- acter,” etc. Instead, I intend merely to elaborate the theoretical meaning of the two kinds of moral elements that I believe underlie the dilemma situations that Bronfen- brenner and his associates have developed. I have identi— fied these two elements above as Negative-Passive moral orientation and Positive-Active moral orientation, and have crith tension. In szctrastin tions. Ne 5539 their is altruis 3:35 of r khan; Aval 1 ‘v. ‘5‘ act tha "‘m‘ "419 the hive ref 51 .have criticized the MDT for over-emphasizing the former dimension. In general, Negative vs. Positive is constituted by contrasting motivational subcomponents of moral orienta- tions. Negative refers to instrumental motivations that have -their basis in fear of punishment. Positive refers to altruistic motivations that are based on empathy and norms of reciprocity. Passive vs. Active represent opposite behavioral subcomponents of moral orientations. Passive refers to either absence of behavior (as in refraining from an act that breaks a rule) or reactive behavior (as in umving the individual away from transgression of a rule). Active refers to affirmative behavior that attempts to exPress and concretely apply a general moral principle. Although there are four possible combinations of motiva— tiOnal and behavioral elements according to this scheme, I Will be concerned here primarily with the Negative-Passive and Positive-Active combinations, since these are the combinations that appear most clearly in the MDT.3 Durkheim. These terms have recognizable echoes in Durkheim's (1965) famous analysis of those ritual elements which together constitute the Sacred community, e.g., the Negative cult and the Positive cult. The Negative cult consists of rites of abstention that forbid certain behaviors in order to keep separate the Sacred and Profane Spheres of life: A whole 9 this stat Since the and to ke ing upon abstentio pose to 9 formed by scribe ce themselve so they a is CC:;. 3 338). Extra Slit, then, i Essa-59 Profa iSbad (intro: 8 ~13: Engaginc Thus, “w and “t .85}, Stage fC Rich leads 1 Effects, reli 1t to HCQ giou: conSt: flee , film heart ffdsht . 52 oup of rites has the object of realizing of separation which is essential. r function is to prevent undue mixings p one of these two domains from encroach- he other, they are only able to impose s or negative acts. Therefore, we pro- ve the name negative cult to the system these Special rites. They do not pre- tain acts to the faithful, but confine to forbidding certain ways of acting; 1 take the form of interdictions, or as y said by ethnographers, of taboos (p. olating from the point of view of the Negative is good (moral) to avoid profane activity, e activity--at least in the religious context-- 1), since it contaminates and prevents one in sacred activity, which.is good (moral). in addition to maintaining boundaries between d," the Negative cult also serves as a prepara— the actual expression of sacred activities, to the function of the Positive cult: he importance of the Negative cult may ough it may indirectly have positive t does not contain its reason for in itself; it introduces one to the life, but it supposes this more than utes it. If it orders the worshipper cm the profane world, it is to bring to the sacred world. Men have never at their duties towards religious ht be reduced to a Simple abstinence ommerce; they have always believed that d positive and bilateral relations with gious forces], whose regulation and on is the function of a group of ritual To this Special system of rites we ame of Positive cult (p. 366). im identifies several distinct classes of the Positive cult, that constitute sacred behavzor and placu exgressio .53 behavior (e.g., sacrificial, imitative, representative and piacular). Each of these has concrete meaning and expression in Specific religious activities, most clearly in primitive, totemic societies to which Durkheim resorted for illustrations. Generally Speaking, however, all of these rituals touch upon the reciprocal nature of man's relationship with the Sacred and, indeed, upon the recip- rocal relationships of men with one another which Durkheim viewed as the ultimate basis for all religious and moral beliefs. To engage in sacred activity is to define what these relationships should be and to "uphold and reaffirm . . . the collective sentiments . . . and ideas" which maintain the unity of the group (pp. 474—75). Thus, in their extrapolated, modern guise, we find certain aspects of the Positive cult reflected in such normative “virtues" as justice, love, self-abnegation, loyalty, etc.4 Negative-Positive and Passive-Active dimensions are at least implicit in many other of the innumerable concep- tual forays that have been made into the bramble of moral definitions.5 In the more recent social psychological literature, two views of morality that are difficult to ignore, keeping these dimensions in mind, are the tremen- dously influential ideas of Piaget (1932) and the equally impactful contemporaneous ideas of Kohlberg (1963). The present study is not concerned, EE£.§EJ with the twin problems that identify both Piaget and Kohlberg‘s work, e.g., individual moral development and moral reasoning. fl r Heu'ertheless, ' :‘nild via the ticated and co ml consider sccially cruci fiat, cognitis issgnonynous . . v ~Qb‘ at? + I. One italysis and < ‘azicas staoes 331131058 fllm a» ,. *" aresmrahl' Pi o a e “Li-Pier; o Lent .54 Nevertheless, their investigations of moral growth in the child via the gradual acquisition of increasingly sophis- ticated and complex moral percepts have pointed to several moral considerations that appear to be psychologically and socially crucial. That is, both Piaget and Kohlberg assert that, cognitively Speaking, the development of moral ideas is synonymous with a natural movement toward increasing maturity. One implication of this assumption is that ana lysis and descriptive charting of moral reasoning at Var ious stages of development should be helpful in uncover- 1119 those fundamental dimensions of moral principles that ate presumably universal. Piaget. Of the two, Piaget's conception of morality is simpler; only two general stages of development are identified, these being the well-known stages of "heteronomy" (-"subject to another's law") and "autonomy" ("subject to One's own law"). Heteronomous characteristics that are of particular interest for evaluating the moral meaning of the MDT irlclude the sacredness and inflexibility imputed to authoritatively posited rules, concern for scrupulously °beying such rules, and the notion of expiatory punishment when the rules are broken. There appears to exist at least an implicit connection between these heteronomous charac- teristics and the Negative-Passive combination that I Perceive underlying many of the MDT items. There is, most 55 : emphasis on specific rules emmanating from Rules of conduct at this level tend to l prohibited acts, and moral indignation and/or :ly a function of the magnitude of consequences 'om engaging in such acts. Lgh heteronomy, as conceived by Piaget, more than I have mentioned here, it is at n to suggest that heteronomous morality, in .3 other characteristics, has a strong ten- :gative-Passive in orientation. The behavioral L orientation (Passive) is to avoid violating .y defined rules. The motivational basis of .on (Negative) is largely to avoid the con- punishment that are inextricably linked to .nt characteristics of the "Autonomous" stage inferential in their implications for a re moral orientation. Nevertheless, as I am, these relevant characteristics include: :ciprocity, equality and cooperation, empathic 'iew of rules as relative rather than absolute, .vioral motives rather than consequences; and For restitution, when acknowledged rules are rather than expiatory punishment. 2 aspects of autonomous morality are thus Le operation of reciprocity norms and the cooperative exchanges. Interaction systems 56 based on mutual trust and mutual respect require doing things for one's fellows, so that the cooperative enter- prise may be maintained (recall the analogous function of Durkheim's Positive cult). Empathy and the development of role-taking Skills are also important in this regard, because they lead to anticipation of, and active efforts to fulfill, peer needs. Even the preference for restitution when broken rules are involved implies an active moral sense rather than the passivity of the heteronomous malfac- tor being "justly punished" for misdeeds. Positive elements of autonomous morality are also Visible, chiefly in the emphasis placed upon delineating the intentionality behind an act rather than preoccupation with externally imposed rules that either proscribe or Prescribe the act. Indeed, the flexibility with which rules are themselves viewed and handled implies an addi- tiOneal positive consequence, namely that in place of innumerable lists of concrete rules, one becomes increas- ing 13; inclined to guide his moral conduct in terms of general, overarching principles that are internally validated and thus not directly contingent upon fear of punishment or h°Pe of reward. Kohlberg. Actually, this last mentioned emphasis on general, guiding principles remains only a potential deVelopment to be inferred from Piaget‘s description of antOnomous morality. It is, of course, Kohlberg who most __‘_1 1 insist cgposec 32:51 (3 isjust. Just lies hone. Just; print verse P93? There tions oblige anothe for re moral . reason is cal; 57 insistently elevates the notion of general principles (as opposed to specific "virtues”) to the highest pinnacle of moral deve10pment. The most general of these principles is justice: .Justice is not a concrete rule of action, such as lies behind virtues like honesty . . . To be honest means don't steal, don't lie, don't cheat. Justice is not a set of rules, it is'a moral principle . . . a mode of choosing which is uni— versal, a rule of choosing which we want all people to adopt always in all situations . . . There are exceptions to rules . . . but not excep- tions to principles. A moral obligation is an obligation to respect the right or claim of another person. There is only one principle basis for resolving claims: justice or equality . . . A moral principle is not only a rule of action but a reason for action. As a reason for action, justice is called respect for persons (Kohlberg, 1970: 69-70). . The Positive—Active element implicit in Piaget‘s mOral autonomy is clearly indicated above by Kohlberg. Matllre moral choices are made on the basis of an internalized Prfiirlciple of justice which requires that actions be taken Whiczh insure both the rights and intrinsic worth of other peli'sons. Kohlberg‘s own formulation of growth toward moral matlurity is more complex than Piaget's and involves six "stages." These stages are not a simple elaboration of Piaget's scheme, with heteronomy and autonomy remaining as poles while the ground between is more finely differentiated. Instead, as Hoffman (1970:277) notes, there is an internal mixing of these dimensions such that "elements of Piaget's hetaxonomy can be found in Kohlberg's stages 1 to 4 and I “‘1‘: elements 6.‘ 5 stages, 1'1 and the Pa within the Po: orientatior MEGS an Negative-pa hsfiumentaj a'Pre-moral SegatiW‘Pas me 'right mistic neec needs °f 0the an incl'Pient 58 elentents of Piaget's autonomy in Kohlberg‘s stages 2 to 6-” Similarly, upon examination of Kohlberg's six stages, it is evident that the Negative—Positive dimension and the Passive-Active dimension are also variously combined within these stages . For instance, stage I ("punishment and obedience orientation"), because of its preoccupation with external Cormands and avoidance of punishment, clearly reflects a Negative-Passive dimension. However, the next stage (“naive instrumental hedonism") , which is considered to still be at a 'pre-moral" level, suggests the operation of both He-9ative«-Passive and Positive-Active combinations. Thus, while ”right action“ is largely defined in terms of meeting e9C>istic needs, it is also occasionally so defined when n$86.3 of others are met as well. There is, at stage 2, an incipient sense of value relativity, egalitarianism and "orientation to exchange and reciprocity" (at least in a Pragmatic way, i.e., "You scratch my back and I'll serratch yours,"' Turiel, 1973:235) . Stage 3 ("good-boy/girl morality“) , with its stress on maintaining good relations with others by helping and tfiring to please and its definition of ”goodness“ in terms of conventional virtues, primarily reflects a Positive- AC“'tive dimension. However, stage 4 (“authority and social order maintainance') , in turn, appears to give greater Weight to Negative—Passive orientations having to do with 59 "duty" and respecting authority, designated rules and " earned rights . " Stage 5 (“legalistic contract and democratic law") implies both a Negative-Passive concern with binding obli— ga tions ensuing from arbitrary (albeit consensual) laws and agreements, and a Positive-Active emphasis upon the "Possibility of changing law in terms of rational considera- tions of social utility" (Turiel, 1973:236). Finally, at Stage 6 ("principles of conscience"), Positive-Active ele- mellts are prominent in moral choice and action, as indi- ca‘ted earlier . Ifleiative-Positive, Passive-Active ptructural Pgerties of _°ral Situations For purposes of my analysis, the principal theore— tical insight to be gained from the preceding discussion of Durkheim, Kohlberg and Piaget is the inference that Positive- Ne-gative and ActivesPassive moral dimensions do not neces— sarily represent substantive values p_e__1; s2, but rather are inherent structural properties of all moral situations and mOral actions, from the morally simple to the morally com- plex. This would seem to be particularly true for the behavioral dimension of Active vs. Passive. Thus, an indivi- dufil does not himself typically value. his Passive response tel'ldencies over his Active response tendencies, or vice versa (although moral theorists may do so). .Rather, it is often the case that, given a particular kind of moral choice 60 situation that does involve substantive values and stan- dards, one simply acts in ways that can then be described as Passive or Active. For instance, the situation at hand may simply involve a clear-cut prohibition with no other moral issues at stake: one either breaks a rule or does not, and the potential consequences for self, should the rule either be obeyed or violated, are known. In such a situation, one can only be "moral" (at least in conventional terms) by liehaving in a Negative-Passive way, the morally mature along with the morally immature, because the situation does not call out for more advanced moral percepts. The motivational component of Negative vs. Positive is o in actual situations, more ambiguous to determine. There do exist, after all, substantive altruistic values (5-. e., clustering around the notion of helping others) that may operate quite independently of obvious situational con- slderations. One may entertain benevolent and generous impulses toward others as a conditioned response to social rewards in specified situations.6 But one may also achieve a disinterested and universalistic conception of the value of helping others, regardless of the immediate presence or abSence of social rewards, as Kohlberg has argued and demonstrated in his ultimate stage of moral development. It should be noted, however, that Kohlberg's highest state of moral development is a relatively rare achievement (and all the more so at younger ages when £15m and ageriential. levels of 50' 'principled“ - 3y altruis race their situations ( totality)- Thus Lavolv'mg on the “virtue " as some val lacking. Pr cohesive, an farces opera the norm of Ed therefor 1M. ~~-lve behav I v 5‘31“?" ac is+ he respo Div-g. , wmstance 0f c it «6 much m or :‘iiples .F. s ..- I (as I ;1 . ‘* SltUat Consciou. t w HQ 61 .dolescents are still growing cognitively and ').7 Still, altruistic acts flourish at all .al life, in spite of the paucity of truly lorality. The clear inference then, is that .c motivations, particularly for children, rurce to the social reward potential of given e., as in Kohlberg's stage of "good-boy" we can imagine relatively simple situations r Positive-Active moral issues, for instance >f sharing in a peer group when one member :d surplus commodity that another member is riding that the group is relatively stable and there are no seriously competing motivational .ng, one is very likely to be susceptible to :ciprocity that characterizes peer relations experience pressure to engage in the Positive- >r of sharing his surplus. In this situation, .eves the status of a moral stereotype. It se that "everyone knows" is right for the the only action that a "good" person can take. Lrse, life abounds with moral situations that complicated and conflictful than the above >m the individual point of view, many complex >ns can only be ultimately resolved through application of higher level moral reasoning: advanced stages of moral deveIOpment suggest.8 severt‘ieless ‘ real life P" CROlCe' ‘ . “ 4 satetent 04 X By . tions to th expectation {start soci this study: :ters tor Jegree of . 1:15 uences E fluxes wh am' 62 Nevertheless, these simple examples do remind us that in real life prevailing social-situational contexts often have a way of channeling the content, logic and style of moral choice. Statement of the General Hypothesis By expanding our focus from specific social situa- tions to the more or less consistent and enduring normative expectations that broadly characterize the culture of dif- ferent social groups, we arrive at the central premises of this study: (1) Socialization experiences predispose group members toward culturally preferred moral standards; (2) Degree of aperture-closure in a community importantly influences the intensity and efficacy of socialization pro- cedures which attempt to implant these standards. Conse- quently, conformity to culturally preferred standards is lalr£>c>thesized to be a positive function of system closure. Wfication of the seieral Hypothesis Now that I have theoretically identified Negative- Positive and Passive-Active as situationally important moral dimensions, we can ask: What difference do these two I“(’ral dimensions make for the hypothesis of socialized conformity? What is the relationship of these moral dimen- SiOns to the social dimension of system aperture-closure? Ihare a1 held nor: groups. Ziegative-l tion of be easy to se :crality a tionship 1' By Erized by “titties. titrated f. Eject-'3 0f ¢ .63 In developing the notion of system aperture-closure: I have already pointed to the proliferation of commonly held norms and accepted sanctions that characterizes closed groups. Now, when we consider that the combination of Negative—Passive morality embodies ruleeprescribed absten- tion of behavior and the avoidance of punishment, it is easy to see that system closure and Negative-Passive morality are closely connected. One aspect of this rela— tionship is suggested below. By definition, closed social systems are charac- ter‘ized by intensive "boundary" defining and maintainance aetivities. "Undesirable" behaviors and attitudes-“whether generated from outside the system or from within-ware the Objects of system defenses and function to identify for meItlk>ers the unique limits of the system (Cohen, 1966; Erickson, 1966) . This is why "undesirable" behaviors tend always to elicit more explicit attention and efforts to regulate than "desirable" behaviors. These regulatory efforts are more successful in closed systems wherein specific prohibitions are more readily articulated, shared and sanctioned when violated. There exists, in short, a greater range of rules that can be authoritatively invoked to explicitly define a greater I-'al'lge of unacceptable behaviors. It follows, then, that the more prohibitions, the greater the emphasis on a N(Native-Passive morality. This must be true, since, by implication, the opposite of unacceptable behavior is "good," and often “0 :le refrain sociallY def In C‘ fixation of : aghasized' iefined Set ' flexible app fiverse sit“ i5 zanifes‘te 33.3"; actiVEl The within a gr0 system apert saxe--positi is in an ope closed syste Passive stan Spa: systems reduced emph sly have It -‘2:. a1 explor 64 special behavior is required at all——one need from engaging in activities that have been ned as "bad." ntrast to Negative—Passive morality, the com- ositive-Active moral orientations, as I have s not dependent upon a concrete and narrowly f prescriptions and prohibitions. Instead, ication of general moral values to specifically tions is required. The “good," the "moral," by doing good, by engaging in overt behaviors affirm and express deeply rooted principles. :xistence of PositiveeActive moral values p is not necessarily dependent upon degree of re—closure. There may be as many—-and even the e values espoused as ideals in a closed system system.10 However, the social machinery of s is usually geared to emphasize Negative- ards rather than Positive—Active standards. on the other hand, with their relatively sis on Negative-Passive standards, theoreti~ re resources and energies available for actual tion and positive moral concerns. fore, if members of closed systems tend to be y "more moral," it is most likely to be eater emphasis on conformity to Negative- ards. But, at the very least, the "moral of open group members should be reduced or iisappear a1 ml standa Now fine relatioz zoral confo 3.3?fopriate L}! .oL‘ "1:0 meses . :losely tie :3: YEt bee L'flv- L‘,‘ “:rOK-s.eses CEd‘JIes hd‘ 65 disappear altogether when conformity to Positive-Active moral standards is considered.11 Now that the original general hypothesis concerning trie relationship between open-closed social structure and moral conformity has been specified, it would normally be appropriate at this point to develop more specific test h3f£>c>theses. However, because the test hypotheses will be closely tied to sample and measurement features that have not yet been presented, I will delay stating the specific h‘3’E><>theses to be tested until all relevant empirical pro- cedures have been described in the next chapter. n .A A leSC :3:zet' ls tsp: ' n In ‘Y i... a“ fierh; Ihagre C ‘early hi. Riv-Lien . filoody e; the Wests. "he Peder 53*“ the I St’hggle 1 the issue here regal. POple 11nd. hood, there a 'cancer CHAPTER I I'm-FOOTNOTES " 1Empirical evidence to support the usefulness of loose-tight" as a structural variable, along which a Yariety of different social and cultural groups may range, 18 reported in later field studies (Strauss, 1966; Strauss and Strauss, 1968). 2Nisbet's elaboration of Simmel on this point has clear historical relevance for the Mormons-—the group which I ave chosen as the "closed system" for this study. The Early history of Mormonism is replete with violent conflict between the fledgling sect and civil authority (e.g., the loody expulsion of the Mormons from Missouri and Illinois, the westward flight to the mountain "sanctuary" of Utah, e Federal army expedition to Utah soon afterwards to put flow the Mormon "insurrection," the prolonged and bitter s“truggle between the U.S. government and the Mormons over t a issue of polygamy, etc.) . For many years the Mormons Were regarded as a fanatical, traitorous, and perverted Peeple under the absolute dominion of a tyrannical priest- ood, therefore, constituting a dangerous social threat and “cancer on the body politic." For their part, the Mormons regarded their persecutions asmanifestations of a Satanic sclueme to thwart the Kingdom of God and their persecutors as "gentiles" and wicked men. The Mormon ideal was to reIl'lain apart from the ”ways of the world" and avoid all but as Sential contact with outsiders. In their aloofness and Self-sufficiency, the Mormons were often accused of secret plotting and disloyalty to local as well as national government. Indeed, certain para military and para political groups were secretly organized among the Mormons (e.g., the , anites," the "Council of Fiftyf') ,. albeit for avowedly 21f ferent motives than those attributed by Mormon enemies. rnd' of course, the closely guarded Mormon temple rituals Memained a source of "gentile" suspicion, but for the ormons, a source of great spiritual strength. A summary :ccount of these turbulent early years, along with an dxcellent sociological analysis of Mormon history's modern a}? legacy, is found in Thomas O'Dea's (1957), The Mormons. a 66 fi 10 call. Ac ve s is overt. rstly fc social ce {e-gu pE etc.). c had high public co gs t1 itfcminc Pcsitiveig ever an i} hate to a} EQtiOII his for the 0t 415. i‘alues, pe SGCieties ermitiVe 1 24., . I"'eaCt1v: + hat are al religions 5 find teachin ithe cond ‘prrCXinate 67 Negative-Active and Positive—Passive are both logically possible combinations. For example, a Negative- Active situation may be said to exist whenever an individual is overtly engaged in deeds defined as moral but does so mostly for external reasons (e.g., fear of punishment, social censure, etc.) , or for primarily selfish reasons (e - 9., personal hOpes for salvation, advance in status, etc - ). Concrete examples of potential situations of this kind might include participation in a popular "worthy cause," ptlblic contributions to charity, preaching against sin, informing on wrong-doers, etc. On the other hand, a Positive-Passive situation may be defined as existing when- eVer an individual has the potential to harm and/or contri- bute to another's disadvantage but does not act or put into mot ion his advantage because of empathy/sympathy/"mercy" for the other. 4Apart from these abstracted and transformed social Va lues, perhaps the most obvious direct linkage between mo ern moral systems and the positive cult in primitive so(Fieties is through imitative rites. Thus, in the most pr lmitive religious groups, cultists engage in sacred imita- t-‘Lve activity by ceremonially emitting the sounds and ges- tures of the totem animal while garbed in furs or feathers that are also representative of the animal. In modern religious systems, the believer sets before himself the life and teaching of the holy founder(s) as examples to emulate in the conduct of his own activities in order to more closely approximate and approach that which is deemed most sacred. . Comprehensive general reviews of social science literature on moral issues can be found in Wright (1971), Hoffman (1970), and Aronfreed (1968). The suggestion that much altruism is an instrumen— Fally conditioned response carries empirical as well as J‘l'i‘tzuitive weight. For example, Weis et a1. (1973), pro- uCed experimental evidence implying that the impulse to he 1p others in need is importantly reinforced by anticipated {Eduction of guilty feelings, knowledge of adherence to the 0m of social responsibility," and reduction of fear of potential social sanctions. so 7Kohlberg has conducted longitudinal studies of 1:. ,me of his subjects for over a decade. He reports con- tlhuing "stage" advancement in some of these subjects as hey enter into their twenties (1970:73-75) . .. 8It is useful to note, in this regard, that Kohlberg‘s Inoral dilemma" items always involve a conflict between moral erlnciples pg; §_e_, rather than being concerned with influ— enCe source. Often times the conflicting principles may aSily be translated into Negative-Passive vs. Positive-fictive; : ‘_ h: e.g., respect is these his is other. '1 gr'mciples is stats. Inste ccnflict cent face ct Oppog Recs that I gssiticn of c situation. 3 am Item no gins aim“: 1 en Pffiposed by Edith 0n M. 197 h s {orth- ehhlim Cm ‘13 10%]: n 12 fig: 1C Emnfel hm h he1e di “shuns: 3‘03 a i it hit). hf C0115; Elects, Chisels I 68 e - g - , respect for law vs. respect for human rights. Thus, in these dilemmas, one must choose one moral principle over the other. This kind of direct conflict between moral principles is not present in the original Cornell MDT items. Instead, only one moral rule is involved, and the conflict centers on abiding or not abiding that rule in the gace of opposing social pressures. Several of the new MDT items that I have developed do imply a more complex juxta- Position of conflicting moral principles in the same Situation. See especially Item #29, "The Crying Next Door," and Item #30, "The Rules of the Game," Appendix I. 9 In the process of revising a draft of this disser- tation, I encountered an amazingly similar hypothesis pr oposed by Bronfenbrenner in his very recently published a*I‘tnzlcle on Israeli moral socialization (Bronfenbrenner, it al., 1975). On page 485, Bronfenbrenner briefly alludes o a forth-coming publication that will try to further Explain the clustering of high Communist MDT scores compared 0 lower non-Communist countries' MDT scores (see Table l, p r 12 for a reproduction of this distribution). According o Bronfenbrenner, this research will "examine the relation écross 13 countries between the average total score obtained in the dilemmas experiment and an index of sociopolitical pluralism developed by Vincent (1971) . The index, derived fJi‘ora a factor analysis of the political characteristics of 121 nation-states, is based on such features as the presence of constitutional limitations on the executive, competitive elections, freedom for oppositional parties, etc. The 9orrelation between the two measures was a significant «.89; in substantive terms, this means that the less pluralistic 1:118 political structure of the country, the more likely are 1“his school children to subscribe to conventional moral va-lues." By focusing on political organization and political eImpression, this index would appear to most closely corre- S'p<3nd to my general open-closed dimension of "Conformity— DeViance." Bronfenbrenner's earlier mentioned hypothesis c°hcerning "degree of overriding national commitment" is not explicitly tied to discussion of Vincent's political pluralism index. However, I would categorize such a Sational concern under my more general dimension of "Consen- us . n Left unconsidered by Bronfenbrenner and his colleagues are those social arrangements that would fall under my gel'leral dimension of "Linkage." 10This might not hold true for very primitive closed so<=:i.eties wherein ritual and taboo hold sway over virtually e"er-y aspect of group life, leaving "only the narrowest H rgin to action not governed by rules" (Piaget, quoted in offman, 1970:268) . E cafes involvi hypothe {6785.6 :hildre Hus- 1 3"» TEE 69 11Interestingly enough findings from an earlier E31:<311fenbrenner, Devereux and Rogers (1968) study (not j.r1\7<31ving the MDT) provide suggestive support for this trgflp<>thesis. Analysis of interview data in this study r:ee\r£3aled that Soviet children,in contrast to American children, placed higher priority on neatness and obedience karat: less priority on telling the truth. ~—~2mfi l‘. p CHAPTER III METHODS AND PROCEDURES, PART 1: TEST POPULATIONS AND SAMPLES ntroduction In this chapter, I will identify and describe the test populations that will represent points along the Open-closed continuum, and discuss the criteria and procedures used in drawing samples from the test popula- tions. RATIONALE FOR SELECTION OF COMPARATIVE POPULATIONS L311. e Mormons When we think of open vs. closed systems as a continuum, we should be able to think of different social 91-: cups as being located at different positions between the Open and closed poles, and then derive specific hYDotheses that take into account the differential impact that these structures may have upon the socialization of mcral values. Thus, even though we may assume that U.S. s°oiety in general constitutes a relatively "open“ system, 70 _ ‘4-7 1 ast Mus , ? ‘vos \ L -‘ 71 1y exist many less inclusive groupings that ub-systems that are relatively closed. Examples immediately to mind include any one of several nized and self-contained religious sects, such erites, the Amish.and Mennonites or the Black Another such group--much less exclusive and but thereby also more accessible for study-«is . This is the group I have chosen for the dy as representing a comparatively closed sMormons, or members of The Church of Jesus atter Day Saints, are judged to represent, on ounds, an example of a relatively closed social some of the following reasons. Children grow-v~ .in the Mormon system are typically exposed to a :r of highly organized, coordinated and articu— lization contexts. Most of the experiences ,thin these contexts are intended to reinforce .s of the Mormon belief and value system. has often been noted--both by outside observers .es, O'Dea, 1957, 1966; Kluckholn and Strodbeck, :11 as by the Mormon Church itself—-that srovides a literal way of life for its active .hrough a comprehensive range of programs, per- Litments, and lay responsibilities. Youth ; in Church sponsored activities is particularly .ncluding involvement in adult-oriented 72 as well as in programs specifically designed youth needs.2 are are numerous observable features of the munity which lend weight to its being concep- s a closed social system, particularly in Utah main body of the Church is located. Ecologi— 1t 65 percent of the population of Salt Lake :apital and largest Urban center of the state the headquarters of the Church—~is Mormon. increases in most rural areas of the state with sities being practically all Mormon.3 The Church 1 good deal of influence, if not independence, in my social, political, educational, and economic : instance, the Church operates its own far-flung stem which supplies needy members with assis- Lng from foodstuffs, clothing, and household > home rent and loan payments.4 The Church owns as the largest church affiliated university in sponsors several junior colleges, and maintains le system of seminary and institute programs :oordinated with the public schools. The Church 3 a daily newspaper as well as a number of high 1 family magazines and has controlling interest ar-Mountain area's largest radio and television Politically, Church and State cannot help but chat more intimate terms than elsewhere.5 At 1e arena of local politics, Mormons dominate the 73 sscene, and it is a rare occasion when other than Mormon qgovernors, congressmen, and senators are elected to office. Again, the point in listing some of the activities in which the Mormon Church engages is to establish the range and pervasiveness of its contacts among members and hence its stature as a fully functioning social system. Given the fact that Utah Mormons are confronted by the Church at many levels in their lives, there are two addi- tional and interrelated factors that are crucial in giving coherence and closure to the system. One is the ideolo- gical grounding that infuses all levels of Church activi- ties. The second is the highly centralized and hierarchi- cal authority structure which governs the Church. To illustrate: An essential element of Mormon belief is that Spiritual knowledge and doctrinal truth must be obtained through direct revelation. Although each individual may receive partial revelation to resolve personal problems, only the President of the Church (who is also "Prophet, Seer, and Revelator") is entitled to receive revelation pertaining to matters of dogma and Church policy in general. Since this revelation is believed to flow directly from a divine source, there can be little questioning of the truthfulness of principles that are declared to be its products. Accordingly, Church members are taught from an early age to obey Church doctrine with lee 74 doctrines, to honor priesthood authority, and to comply with leadership decisions. The leadership and governance of the Church itself is based on a complex network of priesthood authority. The outermost roots of this network begin with organized, local groups of lay priesthood holders (including all males, 12 years old and over) and extend back up increas— ingly hierarchical levels of organization and authority to a body of men known as "The General Authorities." This body is subdivided, with the top positions occupied by "The Quorum of Twelve" and "The First Presidency" (i.e., the President and his counselors). The decisions these men make, and the programs they establish, are funneled back down the levels of organization. Both formal and informal pressures to support policies emmanating from above are high, and conformity is almost always achieved.6 It must be kept in mind that the aperture-closure of any system is a relative characteristic. Thus, while Mormonism may be described as a relatively closed system, it should be obvious that there are other social systems that are decidedly more closed, including, for instance, the Soviet Union. Perhaps the greatest problem in this regard is the degree of system inclusiveness. That is, the process of Mormon socialization takes place within a §3§~§ystem that ultimately must and does accomodate many of its expectations with those of the larger, secular 75 society. In contrast, the U.S.S.R. and other total societies, such as The Peoples Republic of China, are themselves the secular and sovereign definers and con— trollers of major socialization demands. Because of this basic difference, we can assume that the over—all amount of closure in the Mormon system is, perforce, considerably less than for the Soviet Union or other totalitarian systems. Thus, Mormon children are not as consistently exposed to an ideologically determined socialization experience as we saw to be the case for Soviet children. Mormon children also undoubtedly have contact with a greater number of conflicting socialization agents and potentially contradictory role models through the public schools, non-Mormon friends, mass media, and other institutions of the larger U.S. society. Nevertheless, if Mormon society is not as closed as Soviet society, nor even as closed as such secularly withdrawn American groups as the Amish or Black Muslims, it is probably more closed than many other sub-systems within the U.S. and elsewhere.7 Given this assertion, the logic of comparative inquiry requires the selection of additional groups for analysis that are assumed to be less closed than the Mormons. U.S. Catholics are one of the groups that I have chosen for this purpose. “'1 f::"f:"‘_~"-" R L Lat..c 1 C«Osec Mormon in t N \Q ‘c‘. t k End 76 Catholics Catholicism in the U.S. today rarely defines a comprehensive way of life for its members as it might have done centuries ago, or as Mormonism tries to approach as an ideal in Utah. On the other hand, as an organiza— tion, the Catholic Church is, of course, every bit as hierarchical and based on lines of priesthood authority as the Mormon Church. And, as with the Mormons, Catholic theology is absolutistic. Dogma is believed to be ultimately derived from a divine inspiration that has the capacity to be invoked at any time through ex cathedra pronouncements of the Pope. Where Catholics are strong in numbers and high in orthodoxy, therefore, one might reasonably expect that they would develop a moderately closed system. But the fact is that Catholics do not typically form intensive, full—blown communal enclaves in the U.S. today. There remains, however, an important comparative reason for choosing Catholics as a pOpulation to compare with the Mormons. This is that all religious institu- tions are ostensibly commited to preserving and promul- gating a particular set of moral and ethical standards. Thus, in addition to its social and community features, Mbrmonism is,.after all, an organized religion. There~ fore, the question that might be raised is, does religion i9 general, as an institutional system of moral beliefs, tend to induce greater conformity to adult standards of Ozen 77 proper behavior among children of faithful adherents of anyreligion?8 Or are there predictable variations among religions, depending not so much upon the content and style of religious beliefs, but rather upon the organiza- tion and expression of religious life? If we find, for instance, that Catholics and Mormons—-both Christian religious groups that share broadly similar conceptions of religious authority—-produce children who rank dif— ferently on measures of moral conformity, then we would need to attribute this difference to features of social organization that go beyond belief systems EE£.§S. (ceteris paribus). U.S., CrossfSectional .The final comparison group included for testing in the present study is U.S. public school children. Rather than simply make inferences from the original Bronfen- brenner data obtained in upstate New York, new samples have been drawn in order (1) to see if the original findings for U.S. school children are replicable over time and geographical location; (2) to introduce controls for social class, which was not done previously; and (3) to obtain U.S. children's responses to additional measures beyond those used by Bronfenbrenner. I am assuming that heterogeneous groups of U.S. school children are representative products of a more oPen social system. This is, of course, relative to 78 Mormon children, who are presumed to be more firmly engaged by a distinct sub—system within the larger U.S. system, and even to Catholic children, at least those who attend Parochial schools and are thereby exposed to a more particularistic and consistent set of values and expectations. DESCRIPTION OF SAMPLES Each of the samples drawn from the three major comparison populations (e.g., Mormon, Catholic, Public School) is essentially a purposive, non~random sample. However, efforts were made to draw all of the samples from units which allowed for the possibility of matching subjects on the following characteristics: age, sex, socioeconomic status, and race. .Actual procedures of sample selection varied somewhat for each test population. Situational constraints accounting for this variation are discussed more fully below for each sample. Mormon Children Sampling Units. The usual sampling unit for the Moral Dilemmas studies has been the sixth grade, elemen— tary school class room. But since the Mormon Church no longer Operates primary or secondary educational programs CMbrmon children attend public schools along with non- l‘l<:>rmons), it was necessary to identify some other, compara- kfile sampling units. These were located from among several Y‘ n; (3 Pl (.0 (I? _r—'. 79 youth programs sponsored by the Mormon Church. Probably the most visible and important of the youth programs is the Mutual Improvement Association (MIA), which provides an organized outlet for recreational, athletic, theatrical, musical and speech activities, as well as religious instruction. Most Mormon youth attend MIA meetings which are held in neighborhood church buildings on a weekly basis. It was decided, therefore, that MIA classes would serve well as meaningful, reasonably accessible, sampling units. I Age groupings for separate MIA classes are made at two year intervals beginning at age 12, and usually run through the late teens to early twenties. Sizes of classes vary considerably but fall roughly in the range of 10-20 youths per class. The classes themselves are usually sex segregated, but a good portion of actual MIA activities are carried out in mixed sex groups. Since the Bronfenbrenner studies (including the MDT) have been geared toward children at approximately a sixth grade level in school (ll-l/Z - 12~l/2 years old), the ultimate sample drawn on Mormon children included only boys and girls from the youngest MIA age groups.9 How- ever, this solution alone would still create a sample of somewhat older children compared to Bronfenbrenner's Samples and my own Catholic and Public School children Samples, because no Mormon child would be under twelve 80 years of age, while, on the other hand, a good many chil- dren would be over twelve. Therefore, another Mormon program, established for younger children, was resorted to as an additional sampling unit. This program is known simply as "Primary" and is essentially a junior version of MIA; again, classes are based on two-year intervals up through eleven years of age. When a boy or girl turns twelve, he/she "graduates“ into the youngest MIA group. Sampling Selection. The purpose and design of the research did not allow for random sampling of individual subjects within MIA and Primary classes; administration and interpretation of the MDT assume that any given set of respondents constitutes a natural peer and reference group. This is why I have referred to MIA and Primary classes as sampling units. But the question now becomes, how are these larger units to be selected? Ideally, a sampling frame of all MIA and Primary groups would be obtained, and then a random selection of these units would be carried out. There are two reasons why this procedure was not actually employed: (1) Church reluctance to officially sanction the study or provide a listing of MIA and Primary groups;10 and (2) limited resources (time, money) with which to organize and conduct the study among the many separate Mormon groups which would result from random sampling . , . 0:- oi... 81 An alternate scheme was to rely upon my own familiarity with Mormon Church organization and personal contacts with Church leadership at local levels in certain areas in order to recruit enough MIA and Primary groups to yield a minimum number of subjects for the study. Since I could be reasonably confident that most MIA or Primary groups are fairly well matched on age and sex distribution-- irrespective of sampling procedure-my major recruitment concern was to identify distinct social class differences among the groups to be chosen. Lacking more precise information from official Church sources, the social class distinctions that I even- tually made are at a rather crude level, namely "middle“ and ”lower" classifications based on the residential area from‘which MIA/Primary groups draw their members. These classifications were facilitated by Mormon administrative organization of Church membership into geographic units known as "Stakes" (roughly equivalent to a Catholic Diocese) and "Wards" (a smaller unit within the Stake: roughly equivalent to a Catholic Parish). Population size of these two units varies a good deal; a Stake may include as many as 5,000 members, while a Ward may number as many as 800. Since Mormons constitute a high percen— tage of the over-all population of Salt Lake City, the QGOgraphic Space taken up by individual Wards is rela— ‘tively small-~8-10 square blocks would be about the aVerage. This fact makes location of particular Wards ands 362'" 1 . .OCd. flfln ‘ Vehia. 82 and Stakes within different SES areas a somewhat more precise task. Through personal contacts I was able to secure penmission to conduct my study in individual Wards from local Church authorities of several areas. One Stake, comprising seven wards, and located in the central city area, became my "lower" social class source of MIA/ Primary groups. "Middle“ social class groups were obtained from single Wards in three separate Stakes, each Ward being located in an affluent Salt Lake suburb. Final sample size and distribution of characteristics are shown in Table 2. Table 2 Sex and Social Class Characteristics for Mormon Sample Lower Middle Sex Class % Class % Totals Boy 19 38 45 53 64 Girl .31 62 4o 47 71 Totals 50 85 135 Qatholic Children Sampling Units. Sixth grade parochial school ClaSses constitute the sampling unit for Catholic children. This choice simply mirrors, in a uniquely Catholic set- ting, the previous Bronfenbrenner, gg_gl. studies of Inflolic school children. The advantages of school class ICC-I1! of t1 assu' dict wQuld 83 rooms are, of course, that they facilitate administration of test and questionnaire forms, as well as permit the assumption that children responding in a class make up meaningful peer and reference groups for each other. Further assumptions for the Catholic sample are that most children attending parochial school come from fairly strong Catholic home environments,11 and that the paro— chial school itself constitutes an environment which is permeated with specifically Catholic ideals, symbols, perspectives, expectations, etc. Thus, class room peer groups are commonly confronted with at least a potentially consistent set of general, Catholic, standards. The ques— tion remaining to be answered later in the analysis is the degree to which this presumed set of standards is «diluted by other, "non-system" inputs. Sampling Procedures. In some ways, it might have laeen desirable to sample parochial classes from the Salt Ikake City area. It might be argued, for instance, that this procedure would hold constant all other environmental :Eactors except Mormon and Catholic community and religious (lifferences. However, since Mormon influence is so strong iit most levels of life in Salt Lake City, while Catholics are a distinct and dispersed minority, it is more likely that the Catholic experience there will be more atypical than for Catholics elsewhere. For instance, it w<>uld be difficult to assess the degree to which certain local 84 local Catholic perspectives might be a reaction against the pervasive Mormon influence. These considerations, compounded by lack of Catholic contacts in Salt Lake City as well as time and cost limitations, made local parochial schools in the Lansing, Michigan area seem a more feasible sampling alternative. Therefore, permission was obtained from the Lansing Catholic Board of Education to administer test materials in schools whose principals and faculty were willing to c00perate. Cooperative agreements were reached with four schools, two of which draw students from predominantly lower income and central city areas, while the other two schools' boundaries coincide largely with.middle income, residential areas. As with the Mormon sample, these classifications of "lower“ and "middle" social class are arrived at informally on the basis of familiarity with the areas in question and discussion with school administrators. lFinal sample size and distribution of characteristics for (Zatholic children are shown in Table 3. Eiyblic School Children Sampling Units. Sixth grade classes are again the sampling units employed here. As with the earlier Bronfenbrenner samples of U.S. school children, assumptions are made that the public school does not actively promul- gate an ideologically coherent set of moral standards, nor N do (3] whici Horn: 85 Table 3 Sex and Social Class Characteristics For Catholic Sample Lower Middle Sex Class % Class % Totals Boy 38 59 46 40 84 Girl 26 41 70 60 96 Totals 64 116 180 V do children who attend public schools all come from homes which share relatively consistent parental values and normative expectations. Sampling Procedures, The same arguments obtain for ggt_drawing public school samples from the Salt Lake City area as was the case for the Catholic samples. An .additional argument is that a fairly large proportion of students in Salt Lake public schools would, of course, be Mbrmon. I turned, therefore, to Michigan public school (systems for sampling possibilities. Ultimate selection <3f schools was constrained by the following requirements: (1) willingness of a given school district to participate .in the study; (2) need for participating schools to be llocated in an urban setting in order to match the previous hhormon and Catholic samples; (3) relatively clear "middle“ a~hd "lower" social class distinctions between participat— iJIg schools; and (4) homogeneity of students‘ race, since vi 9; Re if. PC 86 both the Mormon and Catholic samples were almost all white. All of the above criteria were found to be satisfied by schools in Michigan's Battle Creek school district, including the initial requisite of official 12 Fulfillment of the social class and race cooperation. requirements among Battle Creek schools was established on the basis of results from a 1971 State-Wide school survey conducted by the Michigan Department of Educational Research and Assessment. Relevant data from this survey included both a social class index score13 and the pro- portion of white vs. nondwhite students for each school and school district in the state. The particular Battle Creek schools chosen for inclusion in the study all had white student proportions of at least 95 percent.14 At the same time, according to the index, two of these schools were rated low on students' average social class rating, while two other schools that were chosen ranked at relatively high average social class levels. Final sample size and distribution of characteristics for the Public School sample are shown in Table 4. 87 Table 4 Sex and Social Class Characteristics of Public School Sample Lower Middle Sex Class % Class % Totals Boy 29 48 60 54 , 89 Girl 31 52 52 46 83 Totals 60 112 172 CHAPTER III-FOOTNOTES 1Relevant descriptions and discussion of Hutterite social life are found in Peters (1965) and Schludermann & Schludermann (1969, 1971). Howard (1973) gives a succinct account and analysis of the Black Muslim movement. Redekop (1969) presents a rich and detailed analysis of "Old Colony" Mennonite group life in Mexico. This group has Splintered off from the mainstream Mennonite organiza- tion of the U.S. and represents a fascinating modern day example of a religious sect withdrawing from, and strug- gling to remain pure of, the "world.“ I have made some preliminary investigation into the possibility of extend- ing the testing reported in the present study to samples of Old Colony Mennonite children. If and when obtained, such sample results would represent much more clearly the effects of closed system functioning on moral socialization. 2Vernon (1962) found that one consequence of such high involvement appears to be unusually high religious self identification for Mormon youth. Thus, among Boy ;Scouts representing nine different religious denominations, Mormon Scouts were found to have the highest percentage of religious responses to the Twenty Statements Test ("Who Am 1?”). Interestingly enough for this study, the only other group of boys that approached the Mormon level of religious identification were Catholic Scouts. 3See Christiansen (1963) and Anderson (1968) for pOpulation statistics. 4See Welfare Plan of the Church of Jesus Christ of Latter-Da Saints (Salt Lake Eity: Generalzahurch welfare Committee, I952). 5One form that political involvement assumes is through both indirect and direct lobbying of legislative issues by Church leadership. For instance, Richardson and Fox (1972) analyzed voting shifts on abortion bills over a two year period in a Western state legislature and present strong evidence that Mormon legislators changed their stands to coincide with a negative position announced by the Mormon Church during the interim period. Williams 88 (1966) de indulged Congress M fame Congress; and the l the repea at the st into 'Chg briefi .gs Posed cha 3W3 ”rig? Chute. 's and of fic 6] bureaucrE liVeS is modern MC Zigrant c‘ assumes t one who j attend hi In additj the Ward! is, the 5 «Program is 801383;! inactive tional p at the n aSCertai to discc and 0the than to fErenCe ‘9 Par; salvati, more] 89 (1966) describes more overt political tactics occasionally indulged in by the Church: “Communiques to members of Congress are periodically sent by the First Presidency. TWo famous ones were the 1946 admonition to the Utah Congressional Delegation to oppose a peacetime draft . . . and the 1965 letter to all Mormons in Congress to resist the repeal of 'right-to- work' laws. Another technique at the state level is to call Mormons in the legislature into 'Church headquarters during the biennial session for briefings on bills of concern to the Church (e.g., pro- posed changes in liquor laws)." The long history of Utah's own "right-to-work" legislation, in relation to the Mormon Church's concentrated efforts to sway both public support and official decisions, is well detailed by Davies (1966)- 6Illustrative of the Mormon Church's "tight" bureaucratic structure and its ramifications in members‘ lives is Kunz and Brinkerhoff‘s (1970) description of modern Mormon migration procedures: "If the Mormon migrant does not make contact in the new ward, the Church assumes the responsibility of searching for him.‘ Any- one who is rumored to be Mormon and who does not begin to attend his new ward will be visited by the local bishop. In addition, he is visited by representatives from each of the ward's departments or auxiliary organizations, that is, the Sunday School, Priesthood, Relief Society, Mutual Improvement Association, Primary, and so forth . . . . It is somewhat difficult for those who may desire to be inactive in the Mormon Church because of these organiza— tional pressures . . . If a Mormon does not "report in" at the new ward and has not allowed his past ward to ascertain his current residence, the Church then attempts to discover his whereabouts by contacting his relatives and others who know him. Their vast record system enables them to make these contacts. Thus, the method of trans- ference of membership records and the intensive effort on the part of the Mormons to be concerned with the member's salvation and 'activitv' tend to pressure them toward [more] conformity . . . (p. 219)." Vallier's (1962) com- parison of Mormon and "reorganized" Mormon (an early Splinter group from the main Church body) missionary systems arrives at some of the same general conclusions. Namely, that the Mormon Church has evolved (partly as a response to peculiar historical and geographic situations) into a cohesive, bureaucratized social system that seeks to integrate the individual member at all levels of life. 7 For instance, Douglas and Mauss (1968) constructed a "secularization" scale (based mostly on demographic variables such as education, rural/urban origin, region of origin, and age) and administered it to Mormon and non- Mormon samples drawn from a medium size Utah community. They fc "high" the M0: for the samples predomi of reli index w tion, I attitud decided munalit always ; best kn Ship be- their 9. “d All} religio: to be S E 5e" Yout tit “eXt old age disc a“lysis 10 f0! offi StUdY. 90 They found that 50 percent of the non—Mormon sample was "high" in presumed secularization while only 9 percent of the.Mormon sample was rated "high." Even more relevant for the present study is Anderson's (1968) comparison of samples drawn from cities that were designated as being predominantly Mormon, Catholic, and Protestant on an index of religious communality (the four variables making up the index were: friendship choices, religious self-identifica- tion, religious participation, and inter-faith marriage attitudes). Anderson found that "the Mormon group quite decidedly displayed the most accentuated pattern of com- munality." ‘ 8The empirical evidence for this hypothesis has always been shaky and contradictory at best. Two of the best known positive findings include the modest relation— ship between childrens' attendance at Sunday School and their general honesty, found by Hartshorne and May (1930), and Allport‘s (1954) conclusion that "intrinsically" religious (but not "extrinsically" religious) persons tend to be slightly less prejudiced. 9Some of the MIA groups sampled wound up having so few youths in the youngest age group (12-13 years old) that it seemed necessary in several cases to include the next-oldest age group (14-15 years) in the sample. This age discrepancy in some of the Mormon samples resulted in analysis problems that are discussed in Chapter IV. 10The letter reproduced in Appendix III is the final Mormon Church response to several inquiries of mine asking for official permission and cooperation in carrying out my study. 11This assumption is empirically supported by the relatively high percentage of parents and relatives reported to be Catholic by children in my study: Fathers = 93.2 percent, Mothers = 96.6 percent, Father's Relatives = 77.9 percent, Mother's relatives = 79.1 percent. 12Earlier attempts to secure official cooperation from other area school districts were not successful. In general, public schools seem to be coming increasingly resistant to outside researchers Who want to administer questionnaires or test instruments during school time. The Battle Creek School District, however, has been attempting to implement a special program in the area of values and citizenship development, and thus welcomed my study as a potential source of evaluation. 1 Michigan tiormaire family in items 5ch father/mo or apartm your hous run does 3 Has anyon last year for indiv IngEd fr. class) to social cl. as 'lower low fortiu all a‘v’era: informati, tics My ] y‘aiChlgan : ~e 1972 14 91 13The social class index used for children in Michigan schools was made up from responses to 27 ques- tionnaire items that inquired into relevant background and family information. Typical of the questionnaire were items such as: How many years of schooling did your father/mother complete? Do your parents rent the house or apartment you live in? How many grown-ups live in your house or apartment? How many cars and trucks that run does your family have? Does your father have a job? Has anyone in your family traveled in an airplane in the last year? And so on. Mean social class index scores for individual Michigan schools, based on these items, ranged from the low forties (indicative of low social class) to the high fifties (indicative of upper middle social class). The Battle Creek schools that I selected as ”lower class“ for this study all averaged scores in the low forties, while the schools I selected as "middle class all averaged scores in the low fifties. More precise information concerning scale construction and characteris— tics may be found in Technical Report of the 1970—71 Michigan Educational Assessment BatteryifiNinth Report, June 1972. ' l 1480 far, variability in nonwwhite MDT scores has not been systematically tested. However, according to a personal communication from Edward Devereux (see Appendix III), plans for a dissertation project involving Black children in Urban ghetto settings were being formulated at about the same time my own project was being carried out. CHAPTER IV METHODS AND PROCEDURES, PART 2: THE MORAL DILEMMAS TEST Introduction The Moral Dilemmas Test (or MDT) constitutes the first of two major research instruments that were applied to the Mormon, Catholic and Public School samples. Dis— cussion and evaluation of this instrumentu-and the experi- mental procedures employed with it will be broken down into two parts; one dealing with the original, Bronfen- brenner-Devereux MDT items and the second with the ggw_ MDT items that I constructed and added to the original test. THE ORIGINAL MDT General Content and Format Description of Test Items This test has been developed out of the Bronfen- brenner andDevereux studies of the past decade, and presently consists of 36 hypothetical situations designed to reveal the degree to which a child will resist 92 pressure standard. items as: 93 pressure from peers to violate a presumed adult moral standard. The general format that each of these test items assumes is illustrated in the following example: The Lost Test You and your friends accidentally find a sheet of paper which the teacher must have lost. On this sheet are the questions and answers that you are going to have tomorrow. Some of the kids suggest that you not say anything to the teacher about it, so that all of you can get better marks. What would you really do? Suppose your friends decide to go ahead. Would you go along with them or refuse? Refuse to Go Along Go Along With With my Friends My Friends I I Absolutely Fairly Guess Guess Fairly Absolutely Certain Certain So So Certain Certain Other of the original moral dilemmas include such situations as: going to a movie recommended by friends but disapproved by parents; joining friends in pilfering fruit from an orchard with a "no trespassing" sign; wear— ing clothing styles approved by peers but not parents; running away after accidentally breaking a window, and so on. (See Appendix I for complete set of Moral Dilemma Test Items.) Scoring of Test Items The six response categories for each moral dilemma item are assigned scale scores in the following way: Behavic -2.5 I lhsolutel Certain Use of a Ree? 1mg t klthe a: howevel. , control m: ‘1' SGts of tESt fOr expel-”me Thus 0 ti qi‘Jen t' mtal a peer Or ‘- brenner from .1 94 Presumed Adult Preferred Behavior Urged by Peers , Behavior -2.5 -l.5 -0.5 +0.5 +1.5 +2.5 I I Absolutely Fairly Guess Guess Fairly Absolutely Certain Certain So So Certain Certain Use of a negative-positive scoring continuum facilitates keeping track of basic peer vs. adult orientation later on in the analysis. Scale direction on the test form itself, however, is reversed for half of the items in order to control for response set. The 36 Moral Dilemmas are sub-divided into three sets of 12 items each, each set representing a separate test form. These three test forms correspond to three experimental conditions which will be described below. Thus, the maximum score that any subject can receive on a given test form is £30; a score of +30 would represent total adult orientation while -30 would represent total peer orientation. For earlier 10 item forms, Bronfen- brenner has reported split-half reliabilities ranging from .75 to .86 and a total reliability across forms of .94 (Bronfenbrenner, 1967). Test Conditiggs and Experimental Design The test conditions alluded to above, which required the establishment of three separate (but 95 presumably equivalent) Moral Dilemma forms, are as follows (see Appendix I for full procedural instructions): The Base condition: This is always the first condition presented. Under this condition, subjects are administered ggg_of the MDT forms having first been told that no one but the researcher will see their answers, and that individual anonymity will be strictly maintained. This MDT form is retrieved upon completion by subjects. The Adult_condition: This may be either the second or third condition, depending upon design specifi— cations. In other words, its order of presentation is never fixed, rather it varies from subject group to sub- ject group. In either case, a ngw_MDT form is administered to subjects along with the revelation that this time individual responses to each new item will be charted and shown to parents and teachers. Completed forms are then gathered. The 29 £_condition: New MDT forms are distributed, and subjects are asked whether they are curious to know how their class-mates react to situations like the ones that have already been presented. Inevitably, sufficient curiosity is aroused to permit administration of these new moral dilemmas with the understanding that individual results will be charted and shown only to the children themselves while the teacher and other adults are not Present. Again, whether this is presented as the second —-- — —H or third cc group accci Th IESpondent' :anipulati VS. peer 8 int “Aberpret TI". :0: gaugi score is View of 1 his bEha‘. of the e aCIQSS t (ii the C 96 or third condition varies systematically from group to group according to the design being used. The crucial variable introduced by these above manipulations is the degree and direction of change in respondents' answers between the poles of supposed adult vs. peer surveillance. Bronfenbrenner and his colleagues interpret the resulting main scores in the following way. The Base score is taken as the point of departure for gauging the effects of social pressure. The Adgl§~ score is presumed to reflect most directly the child's view of his parents' norms for the appropriateness of his behavior. The §§g£_score is interpreted as a function of the expectations of age-mates . . . . The £2531 score across the three conditions measures the general tendency of the child to subscribe to adult vs. peer-approved alternatives (Bronfenbrenner, gt_§l,, 1975). An important ethical question is also raised for the researcher in carrying out manipulations that are based on deception of subjects. At least one pair of researchers (Beloff and Patten, 1970) has tried to avoid this problem with the MDT procedures by having children imagine that parents or peers would have the opportunity to see various of their answers. However, this imagina- tive procedure proved ineffective in producing sufficient variation in responses compared to the more authentic appearing conditions created by deception. Beloff and Patten re results a insist a. sion of lects an misled their re out with A in Previ Order tC dilemma Effects below : 97 Patten reluctantly concluded, therefore, that reliable results are obtainable only by deception. But they insist all the more strenuously on the need, at the conclu- sion of testing, for cpen and complete debriefing of sub— jects and apprOpriate researcher apologies for having misled the children about the potential consequences of their responses. This debriefing procedure was carried out with all subject groups in the present study. A double Latin square design has been established in previous studies employing the Moral Dilemmas Test in order to control for order (i.e., of test conditions and dilemma forms) and setting (i.e., classroom and school) effects. A representation of this design is presented below: School I School II School III Experimental Class Class Class Class Class Class Condition I II IV V VII VIII 1st Base Set 2 Set 2 Set Y Set Y Set X Set X 2nd Adult Set Y Set X Set 2 Set X Set Y Set 2 3rd Peer Set X Set Y Set X Set Z Set 2 Set Y Class III Class VI Class IX lst Base Set 2 Set Y Set X 2nd Peer Set Y Set X Set 2 3rd Adult Set X Set 2 Set Y orevious idealize grouPs i equivale "SCthl. aVailabl grade C] (MY One schOOls {00mg ' ‘ SchOOl not tak POssibi mags 1 98 As can be seen, this design assumes that either three schools-~each with three classes-~have been sampled for every test population in the study, or, alternately, that three schools with two classes and three schools with one class have been sampled. Sampling realities for this study, as discussed previously, did not permit strict adherence to this idealized design. For instance, each of the ten Mormon groups included in the sample must be considered as the equivalent of a single “classroom" within a single "school." Among the four Parochial schools that were available for inclusion in the study, two had two sixth grade class rooms, while the remaining two schools had only one classroom each. The situation for the Public schools was similar, with two schools having two c1a33+ rooms, one school having one classroom, and the final school having three classrooms. A further consideration not taken fully into account by the original design is the possibility of sampling from schools of different social class levels within a given test population. Given these design incongruencies, the procedure that I fell back upon was simply to randomly select the order presentation of MDT forms while systematically varying the order of adult/peer conditions for each class. This resulted in each MDT form being administered under each test condition at least once in each.of the total Mormon, Catholic and Public School samples. However, order di social c conditio middle c 99 order distributions turned out to be unbalanced between social class subsamples, i.e., some possible form-by— condition combinations were not achieved for both lower and middle class Catholics and Public school children. Underlying Dimensions of the MDT Items Careful reading of all 36 items across the three Moral Dilemma forms reveals a clear intention of the test to represent six different kinds of moral choice situations. These six general situations keep recurring under the guise of different specific details. They include the degree to which a child is willing to: 1. Attempt to dissuade friends from engaging in mischief; 2. Resist personal involvement in mischief which friends are actually carrying out; 3. Inform on friends to adult authority once an act of mischief has been committed; 4. Be obedient to parent wishes to refrain from behavior of which parents disapprove; 5. Express an achievement or self-improvement orientation; 6. Express altruistic behavior toward a peer who is being discriminated against by other friends. 100 Each of these six a priori identifiable situations is given expression by two different dilemma items per form, making a total of six measuring items for each of the six situations across all of the forms. Empirical Support Reported for the A_§ri5rifibimens1onality of the MDT Items ‘” While a number of researchers have used the MDT, and several of these have informally agreed with the six situational dimensions listed above (Beloff and Patten, 1970; Beloff and Temperly, 1972), Mason (1972) notes that very little empirical information concerning the actual validity and reliability of these dimensions has been available. Thus the question arises: are the items that supposedly measure a particular dimension actually related to each other and, further, are the ostensibly same dimen- sions across test forms also actually related to each other? It is possible,for example, that one of the items making up the "dissuasion" dimension of "form X" might refer to only a very minor prank situation, while the second "dissuasion" item might refer to a situation involving a more serious breach of conventional rules. The two items would thus be situation specific rather ‘than constituting separate measures of a general dimen— sion. The same potential problem is perpetuated at the level of comparing dimensions across test forms; i.e., L! of deal wit measurin point, 0 bility o adninisrT 101 22E2.0f the items measuring “dissuasion” in form X might deal with less serious issues than the two items purportedly measuring 'dissuasion" in form Y or 2, etc. This last point, of course, refers to the crucial issue of compar- bility or equivalence of the three MDT forms. As part of a study of Middle Class Canadian chil- dren, Mason attempted to deal with these questions by administering all 36 of the original Moral Dilemma items under a "Base" condition. Responses were analyzed by a principle components factor analysis using an equimax rotation. The resulting factor structure of the 36 items corresponded very closely to the six situational dimen- sions defined previously on an a_priori basis; Mason labeled these factors (1) Reforming, (2) Conforming, (3) Informing, (4) Taste, (5) Achievement, and (6) Sympathy, noting that "Reforming“ and "Conforming" were highly correlated, while several "Taste" items (i.e., obedience and parent wishes) also loaded heavily on Reforming- Conforming. Having essentially verified the intended dimen- sional structure of the Moral Dilemmas Test, Mason also examined separate form equivalence and found the inter- form reliabilities to be consistent and moderately high, i.e., r = .78, r = .72 and ryz = .78. The estimated xy xz reliability for all 36 items pooled across forms was .91. 102. Critique of the Original MDT Face validity and Mason's analysis show that the intended MDT dimensions emerge, and that these dimensions are fairly reliable, both within and between forms. However, the theoretical question raised earlier (see Chapter I, pp. 26-27) is appropriate to reintroduce at this point: How adequate are these dimensions for arriv- ing at a general conception of moral choice and moral behavior? Conformity to conventional rules, obedience to parent wishes, achievement orientation, informing on deviant friends, and sympathy for mistreated peers do not constitute a very impressive Eaggg‘of crucial moral issues. For instance, it is not obvious why "achievement orienta- tion" per §g_ought to be regarded as a basic moral issue, and, while "informing on friends" certainly poses an interesting moral issue, it is not clear that a "pro" response would gain the moral approbation of most adults. Beloff and Patten (1970) felt strongly enough about the inappropriateness of "informing" type items that they elimated them as a variable. in their own analysis. Note, ‘however, that the “informing" type item is very well tailored for the kind of peer "honor" system instituted in.the Soviet Childrens' Collective. That informing is *viewed from opposite poles in different kinds of social systems is itself, of course, an indication of open vs. closed patterns . neglecte represer. dilemma dilemmas rules (6 Propert; I€lative mg a r; the Fred El31311621er tioh to 103 Not only are many other important moral values neglected by the test, but the value dimensions that are represented tend to be measured by relatively trivial dilemma situations. For example, the Bronfenbrenner dilemmas designed to reflect conformity to conventional rules (e.g., presumably of honesty, respect for others' property, etc.) are too often based on mere mischief or relatively minor pranks, such as soaping windows or plac— ing a rubber snake in a teacher's desk drawer. Another conceptual weakness of the MDT items is the predominantly negative way in which they are presented and expressed; many of the dilemmas place far greater emphasis on passive and/or reactive resistance to tempta- tion to do "wrong" rather than on affirmative action to do "right." MODIFICATION OF THE MDT Description of New Items Constructed to Supplement the OriginalMDT In response to the above criticisms, I attempted a more adequate "sampling of the domain“ of moral values 'underlying choice situations by constructing 12 additional items. Each of these twelve new dilemmas was designed to require an active moral choice that is based either on a presumably positive desire to help or benefit someone else, or on an abstract ideal. required his bike ride wit contest heated; a chore other f‘ 104 Among the new items, for instance, the child is required to choose between such alternatives as: lending his bike to someone who needs it vs. continuing on a bike ride with friends; voting for a friend who cheats in a contest vs. voting for a disliked peer who has not cheated; following through on a promise to substitute in a chore for a friend vs. backing out of the promise when other friends urge going to a movie instead; defending a peer against unfair criticisms made by the pOpular leader of the group vs. going along with the rest of the group by not saying anything, and so on in a similar vein through the remaining new items (see Appendix I for a complete set of new items). Besides constructing items that stress positive, active responses, my initial intention for the new items was to introduce four additional value dimensions to the MDT: (1) justice, (2) altruism, (3) sharing, and (4) promise-keeping. The addition of these particular moral dimensions seemed both justifiable and desirable on a_ priori grounds of importance and relevance, and also in light of existing research in the area of moral develop— ment. Construction of dilemmas that actually reflected the desired value dimensions proved a more difficult task than writing dilemmas so that they required Positive-Active choices. Pretesting of new items to establish their dimensional validity and reliability would have been, of course , arranger: bined wi jects to Thus, E being re 105. course, the ideal way to proceed. However, delays and arrangement problems with potential pre—test subjects com- bined with looming time commitments with actual test sub- jects to preclude a pre+test analysis of the new items. Thus, pgigp to their actual administration in the study being reported here, the new moral dilemma dimensions could only be evaluated in terms of their £233 validity. On this basis, I classified three new items under each of the four §_priori moral dimensions. Then, one item representing each a_priori dimension was randomly assigned to each of the three MDT forms (i.e., four items measuring four dimensions for each form). Empirical Evaluation of the Expanded MDT (OriginalfiPIus New ItemSYBased on Results from the Mormon, Catholic and Public School Samples ~ In its final version, the expanded MDT consists of 48 total items. The three forms of the MDT are thus made up of 16 items each, 12 of these items representing the original test with its six moral sub-dimensions, while the remaining four items of each form represent the new moral dimensions. All Mormon, Catholic and Public School groups were administered this expanded version of the MDT according to the design specifications and test conditions described earlier. Prior to examining the dimensionality of the expanded MDT, it is important to develop some confidence in the basic equivalence of the three MDT forms for the samples revealec forts. parabili in this additior alence S We might 535 Cons ”9 could SeriOUSr Of items 106 samples tested. The Mason results cited earlier (p. 101 revealed fairly good equivalence between the original MDT forms. Still, we would like to be sure that the same com— parability holds true for the particular samples tested in this study, especially since the forms now include additional, untested items. If_the assumption of equiv- alence cannot be made with confidence, then even though we might later conclude that the three forms are validly and consistently measuring the same underlying dimensions, we could not be sure that these common dimensions are equally weighted for each form. For instance, the content of items making up a given moral dimension in form X might imply greater seriousness for the average respondent than the content of items measuring the same dimension in forms Y and z. This could then lead to biased mgap scores for X. Since not all subject groups received X under the same experi- mental condition (e.g., Base, Adult, or Peer), the biased score could in turn lead to erroneous group by experi- mental condition comparisons. Eguivalence. A difficulty arises, however, in attempting to determine degree of MDT form equivalence. Unlike the Mason study, no group of subjects in my study received all three MDT forms under a single experimental condition. This fact prevents me from simply intercorre- lating individual separate form scores as Mason did. The alternate procedure that I resorted to was to compare MDT Form 107 mean group scores obtained for the three MDT forms under the Base condition. First, I combined all sample groups that had received form X under the Base condition, then all groups that had form Y under the Base condition, and finally all groups that had form Z under Base. The results of this comparison are as presented in Table 5. Table 5 Mean MDT Scores Under Base Condition Only, All Groups (Mormon, Catholic, Public School) Combined2 MDT Mean Score Mean Score Form N Original MDT Items New MDT Items X 164 44.62 16.37 Y 149 44.48 15.45 Z 174 44.57 16.01 Total N = 487 differences between differences forms, p < .99 between forms, p < .12 Thisglobal analysis shows that the three MDT forms for the original items produce virtually identical overall scores when each form is administered under the Base condition. However, the distribution of 22! item mean scores across MDT forms tends to produce slightly discrepant results. The particular pattern of form by condition combinations that emerged in this study only permits a sub-analysis for Mormon and Public School children; the Catholic Base co: essentia (Table 6 108 Catholic sample did not produce the necessary X, Y, Z by Base combination. The two subanalyses that are available essentially replicate the results of the global analysis (Table 6). Table 6 Mean MDT Under Base Condition Only, Mormon and Public School Subanalyses Mormons Public Schools Mean Mean Mean Mean Score Score Score Score MDT Original New MDT Original New Form N MDT MDT Form N MDT MDT x 52 43.33 15.23 x 45 45.42 16.51 Y 39 42.38 14.81 Y 61 43.34 15.00 2 44 45.09 15.92 2 66 45.09 16.36 Total Total N = 135 p<.53 p<.l3 N = 172 p<.50 p<.ll One consistent difference between forms that does emerge from the above subanalysis is that Form Y always has the lowest mean for both the original and new items. While the absolute difference is not large, its existence does imply a potential degree of bias that could dilute the validity of comparative results in later analysis.3 Overall, however, the degree of form equivalence to be inferred from these results appears to be reasonably good. whether MDT obta like to sions be Populatj a 3&9923 across E this lex argue w Sub-Sam] Combine items t deVelop Unix/erg assure, run thrl max 1'01: factOr to an C “love: by the analyt (s Se (1 109 Dimensionality. We may now proceed to determine whether the expected dimensional structure of the expanded MDT obtains for the samples of this study. Since we would like to assume with some confidence that the moral dimen— sions being measured are more or less the same across test pOpulations, the logical place to begin the analysis is at a global level, i.e., by pooling responses to all items across all samples. If a dimensional structure emerged at this level that is both distinct and expected, it would argue well for the relative generality of the MDT and ease the necessity fixrconducting numerous separate analyses of sub-sample reSponses.4 This global analysis was accomplished by submitting combined sample responses (Total N = 487) to all 48 MDT items to a two-step factor and cluster analysis procedure developed by John Hunter (1969) of the Michigan State University Psychology Department. First, using this pro- cedure, a Pearson r correlation matrix of MDT scores was run through a principle components factor analysis (vari— max rotation, communality in the diagonal). These initial factor analysis results were then automatically subjected to an oblique, multiple groups cluster analysis which allowed further refinement of the initial factors produced by the factor analysis.5 Four major factors finally emerged from this analysis and are presented in Table 7 in summary form (see Appendix IV for complete data presentation). Cluster 3 I II III IV 110 Table 7 Major MDT Item Clusters Emerging From Cluster Analysis. All Sample Scores Combined, Experimental Conditions Ignored Alpha N Items Reliability Cluster A Priori Items in Cluster for Cluster I Inform Authority 6 .86 II Dissuade Friends 12 .84 Resist Temptation III Obey Parents 12 .80 Achievement IV Sympathy 18 .84 New (Positive-Active) Total Item N 48 These findings support the intended dimensionality of the MDT and suggest a more inclusive way to conceptualize the moral dimensions represented. Although the six §_priori_ dimensions are shown to be reduced to four, the new combi- nations are clearly interpretable in terms of the original factors. For instance, "Dissuading Friends" and “Resist- ing Temptation" items (Factor II) are temporally, sub- stantively and cognitively tied together in the MDT, and, of course, this connection was also reflected by the high intercorrelations Mason found between these two types of dilemma items in his study. Mason did not report a high correlation between "Obey Parents" and "Achievement" dimensions, while my analysis brings these two toqether into a single factor (III).6 However, within this single factor, able by of six 1 while t1 items . 7 111 factor, the two subvdimensions are still plainly discern- able by noting the rank ordering of factor loadings; five of six "Obey Parents“ items have the highest loading, while the weakest loadings are obtained for the Achievement items.7 Factor IV shows that my attempt to clearly distin- guish such moral sub—dimensions as justice, promise- keeping, etc., was not successful. However, I was success- ful in creating new dilemmas that all share a strong common element, and also in showing this same common element to underlie the original MDT “Sympathy" items. This factor thus constitutes a very important general category of moral dilemmas, corresponding to my basic intention of supplementing the original MDT "Sympathy“ items with additional positive-active type items that require a "good" choice to be based on (1) an active behavioral component that affirms a moral principle, (2) an altruistic motivational component (i.e., perceived benefit for someone else by action), and (3) non—trivial situan tional content. Since Factor IV empirically supports my general conception of Positive-Active type moral dilemmas, I would like also to find evidence-~beyond face content evaluation—- to justify reduction of the remaining MDT dimensions into a second general factor representing Negative—Passive type dilemmas. NegatiV' mral c‘ ponent convent based c quence situati I€p0rt "Res 112 Relative to PositiveeActive dilemmas, I perceive NegativeePassive dilemmas in the MDT as forcing a "good" moral choice to be based on (1) a passive behavioral com- ponent that stresses resistance against temptation to break conventional rules; (2) a motivational component implicitly based on fear of potential negative sanctions as conse- quence of breaking rules; and (3) relatively trivial situational content. The argument can begin with reference to Mason's report of high intercorrelations found between "Dissuade," "Res .," and "Obey" dimensions (actual factor loadings and/ or r's are not given by Mason). This empirical finding of shared commonality among situational dimensions that are specifically different can also be teased out of my own data through a multiple groups cluster analysis. In this analysis, these original dimensions were arbitrarily designated as the clusters, and then a correlation matrix of these clusters was produced with the results as shown in Table 8. Here we see that "Dissuade," "Resist," "Inform," and "Obey" dimensions are indeed highly interrelated with an overall intercorrelational §'= .76. The "Achievement" dimension is also interrelated, but to a lesser extent (intercorrelational I = .61).8 Finally, all of the above interrelationships within moral subdimensions, as well as the distinction between Negative vs. Positive general dimensions, are succinci two facl analysi: ,4 g4 ‘- 113 succinctly mirrored when we arbitrarily rotate to only a two factor solutigp_in a principle components factor analysis (Table 9). Table 8 Intercorrelations for A Priori Clusters, Original Cornell MDT Items. Samples Combined, Test Conditions Ignored. Cluster Reliability Alphas in Diagonal l 2 3 4 5 1. Dissuade (.78) 2. Resist ~95 (.75) 3. Inform .69 .75 (.87) 4. Obey .69 .85 .65 (.73) S. Achievement -57 ~62 .54 .71 (.63) In this solution, only one, 3 priori "Negative— Passive" item is not clustered under Factor I. Factor I also includes four of the six Achievement items although these rank Myron their factor loadings and also have low communalities. Factor II is entirely made up,of g priori Positive-Active items, with the exception of one "Dissuade" and two "Achievement“ items. As in Factor I, these two Achievement items have low loadings and low communalities. The preceding evidence appears mostly to support the basic conceptual distinction that I wish to draw 114 Table 9 Factor Analysis of Expanded MDT, Samples Combined, Test Conditions Ignored Factor Loadings A_Priori -- DimenSIOn Item Name I II Comm. Inform 37. Ticket #3 70 16 46 Inform 10. Fruit #3 67 12 41 Inform 73. Snake #3 67 13 38 Inform 3. Halloween #3 67 13 39 Inform 44. Test #3 63 12 33 Inform 28. Window #3 60 15 31 Dissuade 35. Ticket #1 59 19 35 Obey 39. Movie 58 26 41 Resist 36. Ticket #2 57 26 37 Obey 7. POpular Friend 54 16 32 Resist 2. Halloween #2 54 23 36 Obey 24. Special Talk 51 24 34 Dissuade 42. Test #1 50 25 30 Dissuade 21. Snake #1 49 30 33 Resist 22. Snake #2 48 20 27 Resist 43. Test #2 48 31 31 Dissuade 26. Window #1 46 43 36 Dissuade 1. Halloween #1 46 29 29 Obey 33. Hat 46 28 29 Resist 27. Window #2 44 37 30 Achieve 40. Game 44 17 23 Obey 18. Sneakers 41 27 25 Resist 9. Fruit #2 36 34 22 Achieve 38. Afternoon 33 14 14 Obey 4. Sweater 31 24 16 Achieve 11. Summer Camp 28 24 15 Achieve 25. Personal Project 27 13 10 New M.D. 32. Amusement 21 57 36 115 Table 9-—Continued. .Factor Loadings A_Priori Dimension Item Name I II Comm. Sympathy 41. New Kid 24 55 36 New MJD. 34. Queer Kid 25 53 34 New.M.D. 48. Bike 16 49 30 Sympathy 6. Sick Friend 19 49 27 Sympathy 12. Foreign Kid 17 49 26 Sympathy 20. Class Picnic 27 48 28 New M.D. 45. Assignment 21 48 29 New M.D. l3. Grocery 17 46 24 New M.D. 47. Snowy Day 24 46 27 New M.D. 14. Ice Cream 0 44 18 New M.D. 46. School Trouble 40 43 30 New M.D. 29. Crying 2 42 16 Dissuade 8. Fruit #1 31 36 18 New M.D. 30. Rules 9 33 10 New M.D. 15. Contest 23 32 13 New M.D. 31. Substitute 9 31 8 Sympathy 17. Baby Sitter 18 31 13 Achieve 5. Class Project 17 25 New M.D. 16. Mistake 9 20 Achieve 19. T.V. Show 18 19 116 between types of moral dilemmas. The distinction is between (1) NegativeePassive morality, defined in the MDT by "Dissuade," "Resist," "Inform" and "Obey" items, and (2) Positive—Active morality, defined in the MDT by “Sym- pathy“ and all new Shepherd items of my own construction. A final distinction must be made for MDT "Achieve ment" items, which appear to constitute a relatively independent factor from the first two. This independence is not only empirical (i.e., the fragmented and low load- ings of "Achievement" items on factors I and II; reduced Achievement cluster intercorrelations with all of the other a priori clusters), but also theoretical since the Achievement items, judged by their content and underlying assumptions, do not appear to fall neatly into either Negative-Passive or Positive-Active moral frameworks. Indeed, it is not clear why Achievement motives and behaviors (i.e., personal success, self advancement) ought to be conceptual- ized, peg g2, as constituting a legitimate moral dimension at all. Later in the analysis, therefore, Achievement items will be treated as a separate dimension. Analysis of Separate MDT Forms. Having examined the dimensional structure of the expanded MDT at a global level (i.e., all 48 items pooled) and finding support for my conceptualization of basic moral dilemma types, I was now concerned to see if this same general dimensionality held true for gagh separate MDT form. My earlier compari— son of mean scores for separate forms under the Base 117 condition revealed a fairly good equivalence in over-all score response. However, that comparison only demonstrated that none of the forms was significantly biased in favor of either adult or peer oriented responses. However, it did not establish that the moral dimensions underlying each form were necessarily the same, either in content or intensity. But, if I can now identify the same dimensional structure in each form, then the theoretical validity of the distinctions that I am trying to make will be enhanced, as will the general comparability of separate MDT forms. My procedure at this point was the same as for the previous global analysis, except this time, of course, responses to each form were analyzed separately. Thus, a factor analysis was performed on the item correlations of each form (see Table 10). This first step produced only a _ single factor for each of the three forms, a puzzling and initially discouraging result, given the distinct factors produced when the forms are combined.9 However, closer consideration of these results does not necessarily lead to the conclusion that moral distinctions are invalid for separate MDT forms. One possible explanation that sug- gests itself is that, in the reduced form, moral distinc- tions are muted by a kind of general "method“ factor imposed by the dilemma format that is built into all of the items, regardless of specific item content. That is, the "good“ or "preferred" choice for each MDT itema—in terms of conventional standards«-is almost always apparent 118 Table 10 Separate Factor Analysis Results for Each MDT Form, Samples Combined (N = 487), Test Conditions Ignored Factor Loading Moral Item Dimension # I Neg—Pass. 3 .65 Form X Neg—Pass. 10 .62 Neg-Pass. l .61 Neg-Pass. 2 .61 Neg-Pass. 9 .58 Neg-Pass. 8 .54 Neg-Pass. 7 .52 Pos-Act. 6 .46 Pos-Act. 2 .45 Pos-Act. 13 .43 Achiev. ll .41 Poe-Act. 15 .41 Neg-Pass. .38 Achieve. .32 Pos-Act. 14 .26 Pos—Act. 16 .06 Neg—Pass. 26 .65 Form Y Neg—Pass. 21 .64 Neg-Pass. 23 .62 Neg-Pass. 27 .62 Neg-Pass. 22 .61 Neg-Pass. 28 .59 Neg-Pass. 24 .56 Neg-Pass. 18 .52 Pos-Act. 32 .52 Pos-Act. 20 .51 Pos-Act. 17 .36 Poe-Act. 30 .33 Pos-Act. 31 .32 119 Table lO—«Continued Factor Loading Moral Item ~ Dimension # I Achieve 25 .31 Poe-Act. 29 .26 Achieve. 19 .24 Neg-Pass. 36 .68 Form Z Neg-Pass. 37 .67 Neg-Pass. 39 .64 Neg—Pass. 43 .64 Neg-Pass. 35 .63 Pos-Act. 46 .61 Neg—Pass. 42 .60 Neg-Pass. 44 .59 Pos-Act. 34 .55 Neg-Pass. 33 .54 Pos-Act. 41 .54 Pos-Act. 45 .47 Achieve 40 .46 Pos-Act. 47 .45 Pos-Act. 48 .41 Achieve 38 .34 120 to the respondent. In other words, there is a general social desirability factor that is inherent in all of the dilemmas. Of course, for tg§E_purposes, the existence of this general factor is essential--without it, comparison of responses between Adult, Peer and Base conditions would be meaningless. But it does obscure dimensional distinc- tions between different kinds of "good" or "moral" choices. This may also explain why, when all_of the MDT items are pooled across forms, the distinct factors that do emerge are nevertheless highly intercorrelated.lo One interesting and suggestive consistency that does emerge from the single factor of each form is that the rank orderingof the item loadings always shows a similar pattern, namely, it is always the case that Negative-Passive items load highest, while Positive-Active items load lowest. The achievement items also have low general factor loadings across forms, although their actual rank ordering within forms is not consistent. In order to get around the analytical obstacle thrown up by the social desirability factor, I resorted again to a multiple groups cluster analysis. In this analysis, three clusters were arbitrarily formed around Negative—Passive items, Positive—Active items, and Achievement items, respec- tively, for each of the three forms. Not only did this pro- cedure allow me to compare the reliability of these §_priori clusters and the loading of each §_priori cluster item (see pp. 125-127, this chapter for more detail on these 121 results and their application in determining strong and weak items), but it also produced a cluster-by-form corre- lation matrix which allowed me to apply the logic of Campbell and Fiske's (1959) "Multitrait-Multimethod Matrix": in assessing the validity of my three moral dimensions. According to Campbell and Fiske, a multitrait— multimethod matrix consists of scores or measurements of one or more purported "traits" or characteristics that have been obtained from several different methods. For instance, one might be interested in validating a parti- cular "personality" profile that includes such character— istics as "optimism," "assertiveness," “self confidence," etc. Scores on these characteristics may be obtained in different ways, including, for example, simple observa- tional ratings by the researchers, by the subject's peers and by the subject himself. In order for the "trait" scores to be considered as representing valid dimensions of personality, the matrix must show that a given trait score is highly related to "same trait“ scores across the different methods of measurement. This provides an indi- cator of convergent validity. 0n the other hand, the matrix should also show that supposedly different traits: do not correlate too highly with each other, either within the same method or across different methods. This provides an indicator of discriminant validity. Extension of this evaluative model to my own data results in the following translation: (1) "traits" 122 become the equivalent of my cluster dimensions, Negative— Positive, Positive-Active and Achievement, and (2) "methOds" become the equivalent of the three MDT forms, x, Y, and Z. The matrix format produced by this transla- tion and the resultant pattern of correlations is presented in Table 11.11 Probably the first pattern to catch our attention in this matrix is the consistently high intercorrelation among all of the values in the matrix. A superficial reaction to these high correlations might be to question the discriminative validity of the dimensions. However, if my speculation concerning the existence of a general social desirability factor underlying all of the MDT items is correct, then we should expect these high correlations. The most telling checks--and the ones that the multitrait multimethod matrix is precisely designed to allow us—-are whether or not (1) different form measures of a given dimension correlate particularly with each other, (2) these "same—dimension" correlations are higher than the correlations between "nonsame—dimensions" (i.e., corre- lations having neither dimension nor form measure in common), (3) a given dimension correlates higher with separate form measures of the same dimension than with different dimensions that happen to be measured by the same form, and (4) a consistent pattern of interrelation- ships emerges within all of the form by dimension compari— sons . 123 m m m N m m N m m u z Epr Amm.v om. up. «a. ms. ms. ms. «4. mm. mnoa Ams.v as. _ mm. om. op. 54. mm. mm. mmom N such ism.v ms. om. mm. ms. mm. mm. mmmz Amm.v we. mm. as. as. mm. ~no< ism.v an. we. mm. mm. Nmom » Sues Amm.v he. mm. mm. Nomz . . . a Ass V mm mm not Aom.v mm. Hmom x snow has.c Hmmz mnod mmom mmoz msod Nmom mmmz H504 Hmom Hmmz msowmsmawn N Euom N Show x Show fiascommwo can: mononucmumm cH mowuflaflnmwdmm mamadv omuosmH mcowuwccou some .Ahme u zv omcwnsoo moamfimm .xwuumz gEHOMfluHSSIGOwmamEAOHDHsZ= HH manna 124 Employment of these criteria produces encouraging evidence for both discriminant and convergent validity of the three moral dimensions across MDT forms. First, we note that the largest coefficients in the matrix belong to ”same-dimension" correlations (i.e., "validity coefficin ents"). Next, we see that all validity coefficients are larger than their row and column counterparts in botthhe "heterodimension-heteroform" triangles, and the "hetero— dimension—monoform" triangles. Finally, we see that the pattern of intercorrelations within all of the heterodimen- sion triangles is quite consistent. Interestingly enough, the lowest correlations (although still substantial) in the matrix are those involving the Achievement dimension. One problematic piece of information provided in the matrix is the low alpha reliabilities (main diagonal, in parentheses) for both the Positive-Active and Achieve- ment dimensions. These weak alphas are partially accounted for by the small number of items that make up the dimen- sions in each form (only two items per form for Achievement, six items per form for Positive-Active).12 A final observation to be gleaned from the matrix concerns a kind of "form bias." Note from the following subset of correlations extracted from the matrix (Table 12) that the relationship between Negative-Passive and Positive- Active dimensions is always greatest within the same form. 125 Table 12 Intercorrelations Between Negative-Passive and Positive-Active Dimension Scores. Within Same Form vs. Different Form Comparisons xNeg-Pass YNeg-Pass zNeg-Pass * * YPos-Act .58 .79 .70 * zPos-Act .56 .60 .72 *Highest r within same form. I think this pattern may be another reflection of the general method or "social desirability" factor discussed above. That is, all items, regardless of dimension, have a discernably "good" or adult preferred solution, hence all items correlate highly. In turn, the comparative strength of these correlations is increased by the direct association imposed on all items within a particular MDT form. This pattern is weaker for achievement items across forms (Table 13), suggesting that achievement (at least as repre- sented by the MDT items) does not pose as clearly a "good" vs. "bad" moral dilemma for children respondents. MDT Item Analysis: Criteria for Retention, Dele— tion and Standardization. In addition to producing a factor correlation matrix and alpha reliabilities for the factors, the multiple groups cluster analysis also generates a detailed item analysis of each factor. This allows me to try to increase the empirical strength of my three dimensions 126 Table 13 Intercorrelations Between Achievement by Negative- Passive and PositivevActive Dimensions Across MDT Forms X Y Ach Ach Ach xNeg_Pass .56* '.52 .65 YNeg-Pass .47 .52 .76 zNeg_Pass .46 .45 .77* xPOS_Act .62* .47 .42 YPOS_ACt .44 .46 .46 zPOS_Act .47 .39 .60* *Indicates highest r within same form. for later analysis by identifying and then deleting "weak" items from the MDT. The criteria that I established for identifying weak items include checking to see if (1) the communality of a given item with other items in the same cluster is low (2 .20), (2) the loading of a given item in its designated cluster is low (2 .30), and (3) the loading of a given item is high or higher on some other cluster than on its own. Criteria (1) and (2) are concerned with "convergent validity." Criterion (3) is concerned with “discriminant validity." Since all of these criteria need to be taken into account in relation to each other, they should be regarded as guidelines rather than taken as separate, absolute rules. 127 Thus, for example, in making an ultimate decision whether to retain or delete, an item may rank "low" on one of the criteria, but satisfactorily on the others, and therefore be retained, or vice versa and be rejected. In border- line cases of this sort (which were relatively rare--a low rating on one criterion usually was associated with equally low ratings on the other criteria), I gave most weight to the discriminant power of an item, reasoning that this would provide the clearest results later in the comparative analysis of Mormon, Catholic and Public School responses. The item criteria were applied to Egg sets of results, namely, the global cluster analysis which ignored forms and pooled items, and also the separate items-within- forms cluster analysis. Decisions to delete items listed in Table 14 were based on the information presented in that table.14 128 .ouo .N snow cw “wumsao podumom on» so msflcmoH hm. m 0cm » Euom cw “mumsHU postmom map so mcflnmoa ev. m can =H* some uwsum= ..m.w ~mEHom so: mumnmmmm muduuu muummumucsoo HmcoflmcmEHo Hagan suw3 mocflomoa cow: wam>wpmamh wm3onm Oman macaw moonat N hm. om. mm. mm. mm. om. cannons Hoonom .mv« M om. Hm. 0H. mm. mm. mo. musuflumnsm .Hm w mm. om. mo. mm. mm. ma. yam sham .AH x mo. «a. mo. ma. do. No. mxwumwz .ma x mm. me. ma. mm. hm. ma. ummucoo .mH mEmuH uodumom x mm. mm. mm. me. mm. om. m: 3occ«3 .nm a Hm. mm. mm. mm. om. mm. H¢ 3ocswz .mm# x mv. mm. om. as. we. mm. m* moss pasum .m t x as. mm. pm. me. me. cm. H* moms uflsum .m t x mm. om. NH. Hm. ov. ma. Housm3m .e mEouH .mmmmlmmz Show .pmoq .pmoq .EEOO .cmoa .omoq .8800 Houmsau “mumsao Hmumsau Hmumsau uodnmom mmmmtmmz postmom mmmmumwz mace Ehom so: cmcoflmmd cwnuflz cmcwnsou meson 992 Had mewuH an: xmmz How muasmmm mammamsfi VH OHQMB CHAPTER IVw-FOOTNOTES 1Use of the Latin Square design helped alleviate this problem in the Cornell studies; i.e., any form-by— condition combination is assured for at least one group in every sample. 2Note that the mean MDT scores being compared here are based on the average sum of the absolute value for all items. Thus, the large absolute difference that appears between "original" and "new" MDT scores merely reflects the number of items that constitutes each set (e.g., "original" MDT item N = 12; "new" MDT item N = 4). The important comparison that is made, of course, is relative degree of consistency within item sets for each form. 3The bias, if it exists, would be most problematic for Catholics, since no Catholic group received Form Y. under the Peer condition. On the other hand, four of the six Catholic groups did receive Y under the Adult condition. In this case, Catholic Adult scores may be slightly depressed, while Peer scores may be slightly elevated. In fact, it does turn out to be true that Catholic Adult scores are surprisingly low (see Chapter IV), but then so also are Adult scores for Public School children. 4Note that analysis at this level ignores the possible effects of Base, Adult and Peer test conditions. Again, no group received each test form under each.oondi— tion. 5The factors produced by the factor analysis then become the clusters for the multiple groups analysis. A correlation matrix of these clusters is generated with communality estimates in the diagonal of the matrix. This matrix allows for checks on (1) within cluster, item corre- lations, (2) between cluster, item correlations, and (3) individual item correlations with each cluster taken as a whole, correcting for attenuation. One is then able to focus on the behavior of any single item, both in relation to its assigned cluster and its potential relation to any 129 130 other cluster. Based on this information, initial itemv to-cluster assignments can be changed until, through successive multiple group analyses, an Optimum item to cluster "fit" is achieved. 61t should be remembered that my factor analysis was performed on data obtained under the three experimental conditions, while the data for Mason's factor analysis was not obtained under any experimental conditions. 7One explanation for the explicit linkage between "achieve" and "obey“ items is that Achievement standards may be viewed by children as extensions of their parents' wishes. This interpretation finds some support in the literature on acquisition of Achievement motivation as well as from the fact that Achievement items were the only MD's that consistently elicited increased scores under the adult condition for all test groups in this study. 8Although not shown here, correlations between Negative-Passive and Positive-Active clusters are also quite high (although about .20 points lower, on the average, than the Negative-Passive intercorrelations). Discussion of the meaning and explanation of these high correlations is carried out in more detail beginning on page 122 where cluster-by—form intercorrelations are examined. ~9Because only one factor was generated, the cluster analysis that normally follows automatically in the PACKAGE routine was not activated. 10Note that a minimum eigen value of 1.0 is the criterion used by the Varimax routine in PACKAGE (see Hunter, 1969) to rotate to a new factor. When this conven- tional cut off criterion was over-ridden (using an oblique factor solution) by specifying three factors, regardless of eigen value, the desired Negative—Passive vs. Positive- Active distinction is fairly well sustained for each of the MDT test forms, i.e., Form x Form Y Factor I = 6 of 8 Neg-Pass items 5 of 8 Neg-Pass items with highest loadings; with highest loadings; Eigen = 3.62 3.93 Factor II= 5 of 6 Pos-Act items 5 of 6 Pos-Act items with highest loadings; with highest loadings; Eigen = 0.66 0.69 131 11The intercluster correlations are based on raw score means for the items in each cluster and are corrected for attenuation. 12Thus, when forms are not analyzed separately and items are combined, the resultant alpha reliabilities are: Positive—Active (with 18 items) = .83; Achievement (6 items) = .63. 13No Achievement items were deleted, because there are only two Achievement items per form to begin with. Note that, among the items deleted, several did have relatively high factor loadings and communalities (e.g., Window #1 and #2, School Trouble, etc.). However, these items also had such high loadings on other factors, both within and between forms, that they offer little discrimi- nant validity. CHAPTER V METHODS AND PROCEDURES, PART 3: THE INDEX OF SYSTEM ENGAGEMENT Introduction The second major research instrument used in this study will apply only to the Mormon and Catholic‘samples.l It consists of a number of questionnaire items that inquire of the respondent's background characteristics and relation- ships with other people (see Appendix I for a complete reproduction of questionnaire items). Out of the responses to these questions, I have attempted to construct an index that measures one dimension of aperture-closure. The intent of this chapter is to explain the rationale and describe the construction of this index. General Description of the Index 9f System Engagement (ISE) My intention in creating this index of System Engagement (ISE) is to measure a theoretically important aspect of system aperture-closure, namely the degree to which individuals are integrated into or 132 133 "engaged" by the system. By implication, system engage— ment (SE) has primary reference to the "consensus" dimen- sion of system aperture-closure. More specifically, system engagement may be defined by the number, intensity, duration, salience and variety of contacts that an individual has with system related activities and with other system members or representatives who share—~and reinforce in each other--a common set of standards. The general focus of the questions, therefore, is to find out whether parents, siblings, relatives, friends and neighbors are Church members and, where possible, whether these "significant othersu are also active members. Other questions inquire of the child's own degree of Church involvement and participation in Church sponsored activities. Data produced from these kinds of questions, as measures of "system engagement," are obviously limited. For one thing, the "background" questions used certainly do not cover all potentially important points of system contact. Mere importantly, responses to these items represent chil- drens' perceptigns or guesses of relevant conditions rather than empirical certitude about those conditions. For instance, a child may only be guessing at how long a parent has been a Church member; he may underestimate, or over- estimate (or even deliberately distort) participation rates of parents in Church activities, etc. Finally, the items identify only extrinsic, not intrinsic, involvement levels. 134 But, however crude, information obtained from the original items should allow me to at least roughly differentiate among respondents according to their apparent over-all level of system related contacts. This over-all level of SE will be expressed by an index score that is arrived at through the rationale and procedures outlined below. Let us begin first with a general overview of steps before going into a more specific explication. As already indicated, there are five a priori dimensions or sources of system influence for which I have collected some potentially useful information. These are: (1) Parents (both Mother and Father); (2) Relatives (both Mother's and Father's); (3) Peers (ranging from best friends to acquaintances); (4) Neighbors; and (5) Church activities.2 In turn, I assume that each of these first four sources may vary along three relevant attributes: (a) Church member- ship; (b) degree of Church activity or participation; and (c) Degree of "closeness" or intimacy with the respondent. Data concerning each of these attributes are pet~ available for all of the sources. But where data are avail- able, from separate questionnaire items, the first step is to merge this data in order to create a sgal§_that repre- sents each of the five source dimensions. After this is accomplished, and since most of the data have been measured at an ordinal level, the next step is to convert each ini- tial scale score into rank order scores.3 The final step 135 is to combine all of the rank order dimensional scores into a single index representing degree of “system engagement." Validating and Scaling the Dimensions ofSyStemEngagement Parents. There are several questions about parents that serve as indicators of parental gimp exposure to the system, including: (1) are father/mother members? (2) Were they ”born members" or are they converts? (3) And if con- verts, for how long (five year interval response categories)?4 Responses to these questions can be simply combined into a single rank order scale (one for each parent) in the follow- ing manner: Score Rank Membership Time 6 = Life (Born Member) 5 = 10 + years convert 4 = 5 to 10 years convert 3 = l to 5 years convert 2 = 0 to 1 year convert 1 = Non-member Thus, extreme scores on this "Time" subdimen- sion are defined by parents who are born members vs. parents who are not members at all. Middle range scores are obtained by ”convert" parents according to the length of their conversion time. The relevant assumption underlying the "Time“ sub- dimension is that the longer parents themselves have had contact with and have been exposed to system standards, the 136 more likely these standards will be reflected, transmitted, stressed, and reinforced when parents rear their children. Another question concerning parents asks how often father/mother attend Church meetings, and thus represents an activity sub-dimension: Score—Rank Activity Rate (Attendance) 5 = More than once/week 4 = Once/week 3 = Once/month 2 = Few times/year l = Does not attend This goes beyond nominal system identification, and may be seen as one crude indicator of the actual degree to which parents are engaged by the system.5 The relevant assumption underlying this "Activity“ sub-dimension is that parents will tend to place greater stress upon system stan- dards in rearing children if they make personal investments in the system and/or have frequent, ongoing contacts with the system. Now, given this information for "Time" and "Activity," I wish to combine (again, separately for each parent) these two sub-dimensions into a summary rank score representing potential degree of parent influence as system agents. The face validity for making this combination receives moderate empirical support from the correlations between the variables in question (Table 15). Thus, given the available data, a maximum amount of potential system influence is claimed for the parent who is 137 Table 15 Correlations Between Parents' Attendance and Membership5 Mormons (N=121) Catholics (N=178) Father Mother Father Mother Attend Attend Attend Attend Member .34 .26 .72 .39 Born Member .36 .31 .38 .29 a lifetime member and, at the same time, attends meetings at a rate of more than once per week. Conversely, a minimum amount of system influence is assigned to the parent who is neither a member nor attends meetings. The remaining intermediary scores are assigned by the same logical combi- nation of "Time" and "Activity" categories as presented in Table 16. Table 16 Time/Activity Combinations Time Activity Combined Rank Score Rank Rank Parent Influence 6 5 ~ 11 138 Relatives. Data on parents' relatives is even less exact. For instance, one question identifies the rough proportion of relatives who are nominally engaged by the system as members. But there is no information telling us how long relatives have been members or how actively they participate, as members, in the system. Thus, in order to use this variable, the more tenuous assumption must be made that system membership pg£h§g_implies a cer- tain exposure to, and acceptance and adoption of, system norms. A second question does allow me to specify somewhat more the potential influence that relatives may exert on” the child. This question concerns the degree of intimacy the child experiences in relation to his relatives, or "how close" he feels the relatives are to him and his family:7 The assumption here is that if relatives have frequent and close associations with the child and his family, their potential influence should increase. This "closeness" variable, in combination with the membership variable, should also provide another potential vehicle for socialization into system standards. Again, as for the "Parent" dimension, the empirical association between the variables in question (Table 17) seems strong enough to justify their logical combination. tr 139 Table 17 Correlation Between Relatives and "Closeness" to Respondent's Family Mormons (N=121) Catholics (N=l78) Father Mother Father Mother Relative Relative Relative Relative Closeness Closeness Closeness Closeness Member .39 .27 .31 .37 The procedure for actually combining the two vari- ables into a summary rank score for “Relatives" is logi— cally very similar to the procedure outlined for the "Parent" dimension. That is, the maximum score for these combined variables is achieved by that set of relatives who are all members and who also have an extremely "close" association with the responding child's own family. But, the minimum score is obtained when pg relatives are members, regardless of how close the relatives may be to the child's family. This must be true, since I am trying to take into account different sources of system engagement. While "close" relations with non—member relatives may certainly constitute a positive normative influence, none of this influence can be seen as flowing from the relatives"gwp engagement with the system. There exists a further complication in arriving at a scoring scheme that appropriately weights the two com- bined elements. Since the questions being responded to do up 140 low for listing and linking of specific relatives, not be sure yhigh_relatives a child is thinking of 3 reports being “close“ to his relatives whether ember“ relatives or his "non-member" relatives? The n seems to be fairly well resolved if we add the ness" variable gply_when all or most relatives are ed as being members, and, further, that “closeness" be eligible for adding only if it is rated at a avel. Weak ties to relatives cannot be expected to tate much influence, even if all relatives are system S. The explicit working out of these criteria for com— variables is as follows. First the separate ranking iables: ve Relative Ship Rank Score Closeness Rank Score 1 Not Close 1 n 2 Somewhat Close 2 3 Quite Close 3 4 Extremely Close 4 the rank combinations of these two variables are in Table 18. EEEEE: Information concerning peers is limited to spondent's perception of his friend's nominal Church fication. As is the case with relatives, no addi— information is available concerning actual degree ends' system involvement and activity. Therefore, 141 H u H H N u H N m u H m v u H v H u N H N n N N m u N n v u N v H n m H N n m N m n m m h u m v H u v H N n v N w n v m m u v v wuoom mmoHU “sham: muoom mmoHo HmnEmz mnoow DonU HDQEDS muoom DmOHU Romeo: xcmm xcmm xsmm xsmm mGOHmsmchnsm mconcmEHpesm msoncmEchsm mGOHmcoEchsm QOHmcmEHQ =D>HDMH0m= Hem mcoHumcHnEOU mCHuoom mumswm0H0\mHnmumnEmz mH OHQMB 142 development of a peer dimension, given my data, must again depend upon the assumption that membership alone implies at least a certain amount of exposure to and sympathy with system standards. Some further assumptions must also be made, including (1) friends have relatively frequent and close associations with one another; (2) friends are important norm carriers and norm modifiers; (3) friends who have been socialized within the same system context tend to reinforce, in each other, basic standards derived from that system; and (4) the "closer" friends are, the more direct influence they exert on each other. The last assumption listed above is given concrete expression by questions that ask for the membership status of three different friendship categories: All Friends, Close Friends, and Three Best Friends. These distinctions now allow us to take into account a broader range of potential peer influence. Suppose that all three of a respondent's pggg friends are pgt_members, even though the respondent lives in a predominantly Mormon or Catholic community. We would be justified in expecting some weaken— ing of system influence for such a child, relative to other children in the community. But, on the other hand, we would also have to take into account the total weight of other peer contacts. If all other friends (except the three best) are members, then we would anticipate this "positive" influence to somewhat counteract the 143 potentially "negative" influence of the three, non-member, best friends. The simplest way to assess these varying peer contributions, given the nature of the data, is again to produce rank scores as functions of the combined responses to the three "Friend" items.8 The logical connection 7 between these categories is borne out empirically by the set of intercorrelations in Table 19. Table 19 _Intercorrelations Between Categories of Friends Rated as Church Members9 Mormons Catholics (N=121) (N=l78) l 2 3 l 2 3 1. All Friends --- --- 2. Close Friends .67 —-— .43 _-- 3- 3 Best Friends .52 .64 ——- .37 .36 ——— As with the "Parent" dimension, a simple additive procedure is used for combining these three variables to yield a single rank score. As can be seen from the variable categories presented below, the maximum rank score for "Friends" is 14, while the minimum rank score is 3. 144 All Friends Close Friends ..3 Best Friends ‘7' 1— W — Rank ' Rank Rank Score % Member Score % Member Score ..% Member 1 None 1 None 1 None A Few 2 A Few 2 One 3 Half 3 Half 3 Two 4 Quite a few 4 Quite a few 4 Three 5 All 5 All Neighbors. Questions yielding data on respondents' neighbors are the same as those previously discussed for respondents' relatives, namely (1) proportion of neighbors who are thought to be Church members, and (2) amount of contact neighbors are perceived as having with respondents' family. “Assumptions and rationale for combining responses to these two questions, therefore, are the same as those used for developing a scale of relatives' influences. This includes assigning the minimum score to any combination that includes "no neighbors members," regardless of how much contact neighbors may have with respondent families, and further, that the "contact“ variable only be counted in cases where "most“ or "quite a few" neighbors are identified as members (see pp. 140-1 for similar response categories and the procedure for designating rank order combinations). The relevant correlations are shown in 10 Table 20. 145 Table 20 Correlations Between Neighbors Membership and Contact with Respondents‘ Family Mormons Catholics Family Contact Family Contact Member .12 .30 Church.Activipy. The final a priori dimension for my SE index is based on a number of questions that inquire of the respondent's ggp_participation in system-related activities. Of course, specific youth activities across Mormon, Catholic, or any other system are not likely to be strictly comparable in content. At a more general level, however, I am making the assumption that participation pg£_ s3, in activities provided by the system, leads to increased interaction with both peers and adults in con- texts that are organized precisely to further reinforce system norms and values. If a particular system places comparatively greater stress on providing and encouraging organized activities, then this fact itself becomes an important variable whose potential socialization conse- quences need to be taken into account. Activities for Mormon children are highly struc- tured. Those reported in this study include: Sacrament Meeting, Sunday School, MIA/Primary, Priesthood Meeting, Seminary, and Family Home Evening.11 Both Priesthood and 146 Seminary participation are age restricted (12 and older), while Priesthood is further restricted to males only. Catholic children's activity is less institution“ alized and more "voluntary," thus making broadly applicable activities more difficult to list. Those finally arrived at for purposes of this study include: Choir, Clubs (i.e., "Children of Mary"), C.Y.O. (Catholic Youth Organization), Lent, Altar Boy (males only) and Confession.12 Participation in all of these activities, except "Family Home Evening" for Mormons and "Confession“ for Catholics, is rated on a simple "Yes-No" format, depending upon the respondent's avowed "usual" involvement. Family Home Evening and Confession items, however, permit breaking down of participation rate into four categories. Therefore, before these last two variables can be appropriately com- bined with the other activity items, it is necessary to reduce these four response categories into the same parti— cipate/non—participate dichotomy that represents the other _activity items. This is done because I have not established ,criteria for differentially weighing activities. Regard— less of intrinsic importance attached to activities by system members, each activity is "counted" only insofar as it provides a potential context for recurring socialization experiences. The rationale for collapsing "Family Home Evening" and "Confession" is based on the rate of participation normally expected of active members by each system 147 respectively. For Mormons, observance of “Family Home Evening" is expected on a regular, once«a~week basis. For Catholics, confession by parochial school children is generally encouraged more on a monthly basis. It also turns out that the most balanced empirical dichotomy is obtained when these expected participation levels are uSed as cut-off points (see Table 21). An over-all ranking of Church activity can now be generated by simply tallying the affirmative activity areas for each respondent. Maximum and minimum participa~ tion score examples are provided in Table 22. Combining Subdimensions Into a CompositelIndex offiSystem Engagement ' A final question remains prior to developing an over—all Index of System Engagement: How well do these §_priori defined and constructed SE subdimensions actually go together? Is it justified to regard their combination into a Summary score as a reasonable indicator of system engagement? A moderately positive answer is found in the intercorrelations among all of the SE subdimensions taken together (see Table 23). Although there are exceptions,13 most of the corre- 1ations are positive and high enough to make us suspect that the dimensions, taken together, do share something in common for both Mormon and Catholic children. ThiS- 148 N m . m m a . 2 a xmmimoco m m . mm H a . S m 552895 H m.mm H N.mH N Hmmw\mmEHB 3mm H m.m H m.mH H Hm>mz ouoom w muoom w muoom ovum xcmm mmsomwmm xsmm mmcommmm xcmm sOHummHoHuumm omosomm cwosomm HmchHHo mOHHonumo mCOEuoz mmHanHm> sonmmmsOU\mchm>m 080m NHHEMN now NumEEsm msHuoom HN OHQMB 149 Table 22 Minimum-Maximum Church Activity Scores Participant Non-Participant Sacrament/Choir 2 1 Sunday School/Club 2 l MIA/CYO 2 l Seminary/Lent 2 1 Priesthood/Altar Boy 2 1 Family Eve/Confess 2 1 Maximum Score 12 Minimum Score 6 150 In- oN. NH. NH. NH. ma. ma. uuu NH. 0H. 49.: «a. mN. NN. NDN>ND6< .N cognac -uu SN. ca. ma. No. NH. nu- 6N. so. mN. 6N. mm. .m muonamNmz nu- mN. NN. om. mN. In: NH.. oa.u mN. ON. .m mumma sun NH. me. no. us: 6N., mN. ma. .6 m>NDmHmm ruoz In. NH. Nm. nu- mN. as. .m m>aumamm gums us- an. un. mm. .N guano: In- nun .H umnumm N m m s m N H N m m s N N N mOHHonumo chEnoz mconsoEHpnsm mm macEd mCOHumHoHuooumusH MN 0.739 151 conclusion is bolstered by the results of a Varimax factor analysis which, as shown in Table 24, produced loadings on only one factor. Table 24 Factor Analysis of SE Subdimensions (Yielding Loadings on Only One Factor) Mormons Catholics Factor Loadings Factor Loadings Variables I . I 1. Father .78 .63 2. Mother .72 .55 3. Father Relatives .52 .57 4. Mother Relatives .24 .46 5. Peers .27 .48 6. Neighbors .46 .28 7. Church Activity .35 .31 Given this reasonably encouraging evidence of commonality among the a priori subdimensions, the conclud- ing step is to combine these in such a way that a single Index score is produced. Such an index is most conveh niently interpreted (and also most conveniently modified when adding new dimensions in future research efforts) if it can be standardized so that the possible range of scores that the Index may take on (regardless of the number of dimensions included) always lies between 0.0 and 1.0. 152 My procedure for achieving this standardization was, first, to examine combined Mormon-Catholic frequency distributions for each of the seven SE subdimensions. Using the distribution of scores as guides, 52! rank order scores were assigned to each respondent for 2222. dimension in terms of the categories as presented in Table 25.14 Table 25 Standardization of Rank Scores for SE Subdimensions Degree of SE New Rank SE Subdimensions Influence Score 1. Father High 5 2. Mother Medium High 4 3. Father Relatives Medium 3 4. Mother Relatives Medium Low 2 5. Peers Low 1 6. Neighbors None 0 7. Church Activity _£Q§;_£§g;g_comparison of original and new rank scores for each of the SE subdimensions reveals the accuracy of "translation" to be very high, as seen in Table 26. 153 oe.N; . oo.N Nue>Nuoa .N Na. Na. uoeemamz .e mm. mm. mumoe .m mm. mm. Hem ago: .4 ea. em. Nam some .N mm. mm. negro: .N Na. Na. serene .N N e m e N N N N e m e N N a mmmmmm Hmesmsuo wGHOUm xqmm Uwu HUHMUGMUW mOHHOSHMU mGOEHOZ mmuoom xamm COHmGOEHD mm Umnwvumwcmum dam meOOm xcmm COHmSOEHQ mm HmcmeHO c0m3uwm mGOHumHOHHOUHOHGH 154 Confident that my secondary ranking decisions had essentially altered the values of the original scores, us now able, to combine all of the dimensions into a 1ary Index of System Engagement. The desired 0.0 to potential scoring range was achieved by the following ula: ISE = l/N (2,2 ) \U .0 '2 ll Number of dimensions Sum of dimension scores 20 V' Standardized maximum value for each dimension xample, if a child were ranked at level 5 (highest) on even of the SE subdimensions, then his total SE Index would be: SE = 1/7 (3%) = 1/7 x 7/1 = 1 the other extreme, a child were ranked 0 on all of odimensions , then l/7 (0/5) 1/7 x 0/0 SIB 0.0 CHAPTER V--FOOTNOTES 1I have not attempted to construct an index of :m Engagement for Public School children. These men came from a variety of backgrounds, including 1 religious identification. Because of this ’round variety, it is simply assumed that the .c school children, as a group, do not experience the fic sub-system engagement expected of Mormons and lics. And, of course, questionnaire items relevant 'etting at Mormon or Catholic engagement would need to dified appropriately for each identifiable public 1 subgroup, assuming that subgroups could be identi- Even the general background questions that I did f public school children met with resistance from teachers. These offending items (e.g., parents' ious identification, degree of religious participa- etc.) were not responded to in one of the public ls. 2Sibs' religious activity was originally intended potential system engagement variable. However, the ionnaire item was poorly structured, and so much sion and obvious error was created by the request timate sibs' activity levels, that I finally cone d to drOp Sib activity as a variable from the analysis. 3An additional reason for developing rank order s is that the original raw scale scores have different ranges (i.e., minimum-maximum scores). These dif- ces need to be standardized so that eventually all e dimensional scores can be combined into a single with a possible range of 0.0 to 1.0. 4If parents were reported as dead, divorced or ome other reason absent from the home, then children instructed to reSpond in terms of the adults who had ost responsibility for the children over the longest d of time. 155 156 5Obviously, one of the reasons why activity alone :rude indicator is that it only documents outward Lance.without measuring the corresponding inward ude that is presumably crucial when we explain trans— on of values that give meaning to particular activities. 6Note that the N's for both Mormon and Catholic es are smaller than in the previous MDT analysis. is due to cases where respondents had personal as for leaving early (during the questionnaire phase) hus did not have time to provide responses to all of ackground items. 7"Close" was interpreted to respondents as meaning well do you get along with your relatives?" 8This procedure does not allow for differentially ting the three "friend" categories according to their ive importance as sources of influence. However, the {ive importance of these categories is itself probably 'iable, depending on individuals and overall situa- . involved. 9The difference in strength of correlations for »n vs. Catholic friends may be revealing the differ- : in population density of Mormons and Catholics in ’ reSpective communities. That is, at all levels of Ldship, Mormon children are simply confronted with a ;er proportion of Mormon peers to associate with. 10The low correlation for Mormon neighbors may also .y reflect the ecological reality of a predominantly >n population. For instance, about.the same proportion >rmon and Catholic children report having at least :e a bit" of contact with neighbors (Mormons = 51%, >lics = 55%). However, Mormon children identify a larger proportion of their neighbors as also being >n than Catholic children report Catholic neighbors Lte a few or most Mormon Neighbors" = 78% vs. "Quite v or most Catholic neighbors" = 42.1%). This means while Mormon and Catholic families are about equally enial in their neighbor relations, Catholics are ably thinking mostly of their Catholic neighbors, e Mormons are less exclusive. This makes sense when emember that in Utah, Mormons are a distinct majority much less need to actively carve out and maintain a rate sense of identity. 11"Family Home Evening" is an activity in which :xxparents and children are supposed to meet together guayer and a spiritual lesson as well as social enter— ment. Children are expected to contribute to all es of the evening's activities. 157 12Several Catholic informants, including a Parish 't, suggested these. I decided to exclude attendance ss as an activity, since such attendance is compul- le‘the Parochial schools included in this study. 13The negative exceptions are mostly found in lations between parent's relatives and peers for n children. One possible explanation of these nega- correlations for Mormons (while the same correlations ositive for Catholics) may again be traced back to‘ ation differences. As argued previously, it is very y that Mormon children will have Mormon friends, dless of their family's church status, simply because »reponderance of potential friendships available are In. Thus, Mormon children whose relatives tend not to :mbers are still likely to have other Mormon children rriends which, in turn, could be accounting for the :ive correlations obtained on the three variables in ;ion. 14The rationale for establishing rank order cut- >oints depended heavily on the frequency distribution :iginal scores for each dimension (i.e., an attempt to 1ce the proportion of cases in each of the five new 3 as closely as possible), but also upon logical eval— >ns as to what numerical score could be reasonably ght of as "high" or "low" or "medium," etc., relative 1e dimension being measured. Judgements were most 1 called for in the intermediary ranks, since the ame ranks were usually defined automatically by Lmum-minimum" scores. CHAPTER_VI FORMULATING AND TESTING RESEARCH HYPOTHE SES iuction In this chapter I will present and discuss the fic research hypotheses that have been derived from my a1, theoretical hypothesis. After each research hypo- s has been discussed, analysis of data to test that .hesis will then be summarized and evaluated. The 'al hypothesis was stated (Ch. II, p. 62) as follows: lore closed the social system, the greater the con— ;ty to conventional standards. HYPOTHESES INVOLVING THE INDEPENDENT VARIABLE The first variable specified by this general hypo- .s implies that social systems differ in their degree of :ture-closure." Aperture-closure refers to linkages and anges with other, contingent systems as well as to rnal functioning of single systems. Selection of system populations from which to draw samples was determined on g priori, grounds; i.e., 158 159 ystems were chosen that, on a common sense level, appeared axmryin the openness of both their external exchanges 1d.huernal Operations (See Chapter III). However, once >mparative systems had been decided upon logically, I then mghtemmirical evidence to test my choices by attempting to asunaa particular dimension of system aperture-closure at e level of internal functioning which,I identified as onsensus." gothesis I: System Engagement For instance, the more closed the system, the eater the amount of significant contact that should exist ;ween system members and system standards. Conversely, as tems become more open, the amount of contact between tem members, in terms of system standards, should become uced and/or diluted because of wider circulation and eptance of competing ideas, values, lifestyles, etc. 3 aspect of aperture-closure, then, concerns degree of n and.value consensus and is purportedly measured by my [ex of System Engagement" (ISE) . The potential importance of this measure lies not ixi'verification of my choice of particular system lations, relative to each other, but also because it ests a way to replace "proper names" (i.e., nominal em categories such as Mormons, Catholics, Russians, ) with a variable (i.e. , degree of system closure). As vorski and Teune (1970) strongly emphasize, such a 160 cunnation is requisite for the advancement of theory on comparative analyses. Specifically, one of the major premises of my study .t Mormon children, and to a lesser degree Catholic tial School children, are implicated in relatively l subsystems, compared to U.S. children in general asented by Public School children in this study). armore, comparative differences in degree of closure an Mormon and Catholic sub-systems should also be strable. For reasons discussed in Chapter V, degree of system ement for Public School children is not empirically red in this study. The assumption that Public School ren represent the most open of the systems sampled is rted only on 3 priori logical grounds. Therefore, the test hypothesis can be stated only to include the »n and Catholic samples. This first hypothesis is: I. Mormon children will score higher on the ISE than Catholic children .ofgypothesis I Comparison of mean scores obtained on the ISE ads that Mormon children did score higher, on the age, than Catholic children (Mormon SE = .674 vs. olic SE = .589; see Table 27, line 8). However, even (flIthis difference is in the predicted direction and is istically very significant, the absolute difference in 161 <3f system engagement, as reflected by these two scores,