:umuu.‘ FY".- :i‘L‘jfl . 1:111141111 AMAS’AL §E§?©~£§E§ T3 TW‘éSéfiE’ESgEGK L flunk»: 5% We Dogma a? PE. 9. WCEEGRN SMTE ENEVERSKTY Bruce Sandor Vegei “£967 L I B R A R Y i‘vlichigan State University Thisns This is to certify that the .‘ thesis entitled {o in; prjre rr/ GA: ,«7 ./ I [77/7 0ra/ /{€5/>op PI presented by Bruce 56‘ NC/Ur [/6381 has been accepted towards fulfillment of the requirements for _l:)L'_D_ degree in My/ /%Mm Major professor DmWé 7 0-169 ‘J x1 .. e. g. ABSTRACT AMORAL RESPONSES TO TRANSGRESSION by Bruce Sandor Vogel Middle- and working-class delinquent and non—delinquents were required to turn a dial on an apparatus in order to compute arithmetic problems presented to them by the experi- menter. They were unaware that the apparatus had been wired to malfunction. The 96 male, adolescents who participated in the study were randomly assigned to either a punitive or maximal fear condition or to a minimal fear or "guilt" condition. Subject reactions to the simulated breakdown of the apparatus ("transgression") were analyzed by analyses of variance and chi square techniques. Results were expected to support nine hypotheses which were based on an amoral, anxiety— reduction model. Six of these hypotheses concerned the function of transgression responses and contrasted assumptions based on morality theory with those of the anxiety-reduction model. Morality theory equates internalized responses with guilt or conscience development; the anxiety—reduction model views internalized and externalized responses as amoral, instrumental attempts at reducing anxiety. Bruce Sandor Vogel The results were in accord with expectations for five of these six hypotheses and provided suggestive evidence for the sixth hypothesis as well. The amoral, anxiety-reduction perspective adopted in the present investigation was unequiv- ocally supported. The results indicated that internalized and externalized responses to transgression are both elicited by a maximal fear condition, and apparently serve and are designed for the identical function of reducing anxiety. Internalized responses are not specifically elicited by a minimal fear or "guilt" condition, nor do non-delinquents evidence a greater number of such responses than delinquents. Internalized responses therefore, apparently do not represent the behavioral manifestation of guilt. It was concluded that the practice of dichotomizing and categorizing internalized and externalized responses as guilt—indicative and instrumental- avoidance responses respectively, is not justified. Further- more, responses to transgression are inaccurate and inefficient as predictive and as explanatory variables for delinquency. These results carry the implication that psychologists abandon the attempt at establishing a phenomenological state of guilt as measured by internalized transgression responses. While the desire to make such an internal state public so as to pre- dict and explain overt behavior may be worthwhile, the obtained results indicate that psychologists are studying indicators of anxiety rather than of guilt when they study responses to transgression. AV“: V“..- . {BSD me an: ‘ a trans:' r 3 .1. anr: AA“ ..1': \‘ I65 8 34 Ca 9 #- » iSa r E e h t .l V: r «a C Pu . q We. N\V . l A ‘i 4‘ 3n. S . I v. .3 n! .d on st .1 w...“ 3.. C C Bruce Sandor Vogel Three hypotheses dealt with the origin of transgression responses and attempted to support the assumption based on the anxiety-reduction model that response orientation to transgression results from Specific patterns of social rein- forcement directly related to parental disciplinary techniques and indirectly associated with socioeconomic status. The results failed to support expectations. However, this study did not measure parental disciplinary techniques directly and so did not provide a crucial test of these three hypotheses. Faults in the design relative to these three hypotheses suggest further research. Disciplinary techniques used by parents should be directly assessed through subject interviews, before an evaluation of the relationship between parental disciplinary techniques and transgression response orientation is attempted. Furthermore, if socioeconomic variables are examined, investigators may need to rely on more refined criteria of socioeconomic status, since inter—class differences may be diminishing. Finally, it was suggested that the high cognitive struc- ture present in the experimental design may have had a dif- ferential effect in eliciting internalized responses from middle- and working—class subjects. This differential effect may have obfuscated differences between the two socioeconomic classes. Further research which contrasts a high cognitive condition with a minimally structured condition might yet obtain 5: crientati. elicited i learned in Status. “~.|. Donald Charle Bill L DCZier Bruce Sandor Vogel obtain socioeconomic class differences in transgression response orientations. Such a finding would suggest that responses elicited by transgression are originally reinforced and thereby learned in association with factors related to socioeconomic App 0 énAfixadg;;é:étibontmvak\_ Commiatee Chairman Date: EM 1737 /f‘[ status. Thesis Committee: Donald L. Grummon, Chairman Charles Hanley Bill L. Kell Dozier W. Thornton AMORAL RESPONSES TO TRANSGRESSION BY Bruce Sandor Vogel A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1967 —--.- , Q Li (721,73 ‘3'" ibfl’og to Darrell Scott Vogel, future scholar ii .—.___—— ‘_ A —- w—v —_'--— I we: the thesis the course c:mit.ee and upper segree. and bCYS health, I “R rade r idle-, MaSS Ma ae;inqu 'W'al?s ACKNOWLEDGMENTS I would like to thank Dr. Donald L. Grummon, chairman of the thesis committee for his invaluable assistance throughout the course of this research. As chairman of the guidance committee Dr. Grummon consistently offered the encouragement and support which helped me complete the requirements for the degree. Appreciation is also extended to Dr. Charles Hanley and Dr. Bill L. Kell for their very worthwhile suggestions. Special thanks go to Dr. Dozier W. Thornton, committee member, colleague and friend whose casual remark one day when the dis- sertation was in the idea stage, set the tone for the entire research project. I would also like to extend my appreciation to the staff and boys of the Lansing Boys' Training School, Starr Common- wealth, Eastern High School and Pattengill Junior High School who made this investigation possible. Many individuals, too numerous to mention, gave unstintingly of their time and energies to make this project complete. Some however, deserve special recognition: Mr. Tom Reif, a graduate psychology student at Michigan State University, was highly convincing as the distraught experimenter whose scientific apparatus "broke" while being manipulated by each of the 96 subjects. Mr. Reif willingly contributed his talents and energies. He received little in return, other than unneeded training as a thespian. Mrs. Thelma Harwood's willingness to submit to the task of typing and re-typing original drafts entitles her in the least to a piece of the Ph.D. Mr. Al Lilly of Starr Commonwealth went out of his way to make arrangements for the testing of the middle-class delinquent boys who participated in this study. Miss Maxene Goodman helped in the testing of the working—class delinquents and in many less tangible, but equally important ways. Finally, I need to thank Dr. Roberta B. Vogel, colleague, friend and devoted wife who helped me in coding responses, reduced my anxiety when it became unbearable, and managed to continue loving me even in my worst moments. iii ACEGMEDGEEX'LI LIST OF TABLES eon ,, I LOT OF APPER- N'— “3338! I. INTR OD 1' fl rr\ ‘ TABLE OF CONTENTS ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . LIST OF TABLES . . . . . . . . . . . . . . . . . LIST OF APPENDICES . . . . . . . . . . . . . . . Chapter I. II. III. IV. V. INTRODUCTION . . . . . . . . . . . . . . . The Anonymous-Punitive Issue. . . . . . The Socioeconomic Class Issue . . . . . The Delinquency versus Non-delinquency Issue. . . . . . . . . . . . . . . Morality Theory versus the Amorality Position . . . . . . . . . . . . . Summary and Hypotheses. . . . . . . . . METHOD . . . . . . . . . . . . . . . . . . Subjects. . . . . . . . . . . . . . . . Procedure . . . . . . . . . . . . . . . Treatment of the Data . . . . . . . . . RESULTS. . . . . . . . . . . . . . . . . . DISCUSSION . . . . . . . . . . . . . . . . The Anonymous-punitive Issue . . . . . The Socioeconomic Class Issue . . . . . The Delinquency versus Non-delinquency Issue. . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . SUMMARY. . . . . . . . . . . . . . . . . . APPENDICES.................... REFERENCES. . . . . . . . . . . . . . . . . . . . iv Page iii V]. 11 14 16 28 38 41 44 49 52 65 65 68 75 82 84 87 96 near “‘3" 1.0ccupati: Rxbers : A: Onnuh~ m byub‘u O l\') 5. . :rrelat; . P \‘u..r Y . b thou ‘_ 0" 0 Analysis 3. . “"-0-L( ( Mean Trans; for the Ma: 95 Externai "AnalYSis C: SCCres b},] T a " H § usgreSS V' . categcries ‘ . n“‘ o,‘ ' ‘ldtx‘ct ‘v, 2" l; .V‘ u. 36l§nq_e:c. ‘ c . 1% ‘J. MC LIST OF TABLES TABLE 1. 1CL Occupational Classificatory Categories and Code Numbers for Subjects, Taken from the Dictionary of Occupational Titles, 1965. . . . . . . . . . . Correlations for Mean Transgression Scores and Number of Transgression Statements with IQ Scores for the Main Experimental Conditions and for all Groups Combined . . . . . . . . . . . . . . . . . Mean Number of Responses and Standard Deviations for the Main Effect Variables . . . . . . . . . . Analysis of Variance for Number of Responses by Experimental and Control Groups . . . . . . . . . Mean Transgression Scores and Standard Deviations for the Main Effect Variables on a 7-Point Scale of Externalization-Internalization. . . . . . . . Analysis of Variance for Mean Transgression Scores by Experimental and Control Groups . . . . Chi Square Values for Frequency Comparisons of Transgression Responses in Seven Transgression Categories Under Conditions of High and Low Anxiety, Middle- and Working-class Status, and Delinquency and Non-delinquency Status. . . . . . Correlations for Mean Transfression Scores and Number of Transgression Statements with HSPQ In- telligence Scores for the Main Experimental Con- ditions and for;allGroups Combined. . . . . . . . Number of Scorable Responses Given by Subjects 1 through 12 in Each of 8 Groups (N = 96) . . . . Mean Transgression Scores Obtained by Subjects 1 through 12 in Each of 8 Groups (N = 96) . . . . Page 42 45 55 55 55 56 58 64 89 9O APPENDIX A. LIST OF APPENDICES APPENDIX A. Number of Responses Given and Mean Transgres- sion Scores Obtained by Bach Subject. . . . . B. Illustrative Examples of Scoring of Subject Responses to Examiner and Experimenter Probes vi Page 88 91 ‘ 6 Hva + ‘:Shri‘d|.tn » a Fujfi'flse ,9 s ‘ ~u.y‘v & . . ~ ~ ‘I arses are a, . ' 'R s v a! .191 - I‘M :Csl " 5 .I -~_ ‘.\ .f‘h ~V ~ t'tcces ‘flVL ‘t. '5 s::“-\ CHAPTER I INTRODUCTION This investigation attempts to demonstrate that so-called "guilt" responses to transgression are essentially learned instrumental behaviors designed to reduce anxiety. A secondary purpose is to show that Specific types of transgression re- sponses are associated with social class variables, but are unrelated to delinquency. Without attempting to deny that a phenomonological state of guilt may exist in a person after he commits a socially unacceptable act, the present investigation examines the more accessible and definable affect of anxiety as reflected in responses to transgression. Psychoanalytically based writers equate "internalized moral standards" with guilt, superego, or conscience, and externalized morality with fear of externally imposed punishment. Externalized responses are assumed to be unrelated to the experience of guilt (Freud, 1956; Whiting & Child, 1955; Miller, Swanson et al., 1960; Sears et al., 1957; JKohlberg, 1965; Ausubel, 1955). All of these investigators tagree that externalized responses such as denial are designed txa avoid detection and punishment and are therefore instrumental responses which follow transgression. However, internalized responses such as self-criticism are viewed differently. sizce inte ‘ Av A quu» U‘ ‘ i m‘njsr'v-env- Hund- 3..., es» . ' a true ‘” ‘ bun Gui-L d .. «c I. , P'A. U‘Stlr‘nCu V 2365 323: e . ANVA‘ Mix 5 av “4"” ‘e .‘P‘ V - “'1 hJoEJ .‘ V‘es Ga ' ‘ vyar b ’V V 5‘ easy:- a They are not regarded as instrumental to reducing anxiety, since internalized responses are presumed to occur even in the absence of detection or the threat of detection and consequent punishment. Self-criticism therefore is presumed to reflect true guilt or conscience. The investigator has labeled this distinction the Punitive versus Anonymous issue. Whether or not guilt as a phenomonological state does or does not exist, the present investigator contends that psychol— ogists are studying instrumental responses designed to reduce anxiety and not the behavioral manifestations of guilt when tfliey investigate both externalized and internalized responses tc: transgression. He therefore disagrees with those investi— gators who would dichotomize internalized and externalized triinsgression responses into moral and amoral categories resPectively. Transgression in the present investigation is defined as true commission of an act which the subject thinks of as socially iruappropriate and which has received negative or punitive Seuactions by external socializing agents in the past. The major ENerose of the present investigation is to demonstrate that 1Iransgression elicits both externalized and internalized re— sPonses and that both types of response are instrumental to 3reducing anxiety. If any particular type of response to transgression is not effective in reducing anxiety it presumably would extinguish. 'The effectiveness of a response for reducing anxiety in a given individua 1e :n'rzg his pare: respcnse. tween par iisciplin 5:32 exte .‘_ 5., ‘ ' A“ :-Bda¥ Vs. A‘Il‘; ‘3: Vno‘u' -v‘ '3;\‘ WW). A \ lean ‘l I ‘Y O: .n, :‘i: 4‘ individual should depend on his prior conditions of social learning and on the specific disciplinary techniques used by his parents which differentially reinforce specific types of response. Many studies have noted a positive association be- tween parents' use of praise, reasoning and other highly verbal disciplinary techniques and children's apparent independence from external fear and control in reacting to transgression situations (Burton et al., 1961; Ausubel, 1955; Davis, 1947; Sears et al., 1957; Allinsmith, 1960; Heinecke, 1953: Whiting & Child, 1953; Mackinnon, 1938; Kohlberg, 1962). This association rmay be an effect of differential reinforcement of responses tc> transgression under specific types of discipline. This tuition receives support from several studies (Aronfreed, 1961; MacRae, 1954; Peel, 1959). Many surveys have indicated gross differences in dis- Ci{>1inary techniques by our two major socioeconomic classes. ‘Wcufl‘ 5 en .1 .3 L “VII“ -.;.:l ‘. L-auuie-c,L 3‘: 35:53 : 5"”V" Au ‘ “v-4.vu . as "C I ~ H “-5 ‘I‘AA .-y.alltvv H r511 present investigator makes no such assumption. Indeed, if the findings of Aronfreed (1961), MacRae (1954), and Peel (1959) are reliable an amoral orientation is necessary. To assume that working-class individuals are in some way less moral than middle-class individuals simply because they tend to reSpond to transgression with externalized responses can only reflect a Spurious interpretation. Conversely, the finding that middle- class individuals evidence self-criticism and other internalized responses to transgression cannot be justifiably interpreted as meaning that such individuals reflect a higher level of nuarality or conscience development. Failure to investigate the possible association between scxzioeconomic level and type of transgression response may :resnalt in erroneous interpretations of correlational findings. Fcu: example, the positive association between self-criticism responses and avoidance of delinquent acts (Bandura 80 Walters, lEHSS; McCord & McCord, 1956) may be interpreted as meaning that expressions of self-blame following transgression prevent de- lianuency on future occasions. Such an interpretation would kNB consistent with the psychoanalytic view espoused by Sears, bfiiccoby and Levin (1957) that a fully deve10ped superego or Clonscience (as manifested by self-criticism re3ponses) is r1eeessary for the expression of socially acceptable behavior. HO‘wever, if middle-class delinquents are not included in a IAUIportedly representative delinquent sample, as is the case in most investigations, then any obtained relationship between nu Y. « . an Av S .n.» HM; ~Ua .P~ e K nu; .nu an .3 2‘ r. w. a“ .1 ‘7“ CL C» a a: a... n t. «1 V. .o . ~u. .. . L» r. .3 a: .D at . . sfid lo" .1: v . :- F. a ANU Pa 3 S a: .» ~ hi. 3‘ a v t “Us a C. ~\~ o u .- u A: s. . 9 . v . n“ a. E e. v 1 .a. -HH ._. S . n a. .2" «y v-x .. .ns‘x.‘ to "2 - 5‘5 ’~~ nu av .2 .... ~ .. «C .. a Cu v. C» a v u s .5 ~ IV. delinquent behavior and expressions of self—blame following transgression may be explained with reference to social class variables such as disciplinary techniques, rather than with reference to delinquency and conscience development. Kohlberg (1963) wants to establish conscience or morality as a psychological variable which has explanatory value with regard to overt behavior. He interprets the reported associ- ation between avoidance of delinquent acts and self-blame or self-judgment reSponses following transgression as indicating that such internalized responses reflect true guilt or morality. flfiie absence of relationship between fear of punishment or hiding responses and delinquency, resistance to temptation, or moral jtuigment (Bandura & Walters, 1959; Burton, 1959; Kohlberg, 1965) is :interpreted as reflecting the subordinate, amoral status of exd:ernalized responses. A dichotomy between internalized and exiuernalized responses is thus established, the former allegedly iruiicative of true guilt and the latter viewed as instrumental for: avoiding punishment. Externalized responses are therefore PIKasumed to be qualitatively different than internalized re- SENDnses and are viewed as unrelated to morality considerations. As indicated above, the present investigator assumes an Eimoral orientation with regard to both internalized and ex- tiBrnalized responses to transgression. The validity of this iapproach is examined by investigating the types of transgression responses of middle and working-class delinquents and non— delinquents in a transgression Situation. If delinquents ezimlt «A f‘ 35 1v v on; war: .nv “NV“ ‘f‘fl‘fl A Oo‘l‘ j . 'nn tn 0...: 5V “ I l ‘v- n "nV n. W”; . , . T;"_:Iv. AV 'v.n‘u‘v- 1:? 5"” ‘4‘ ‘...e I: .‘L‘flali .. .W “Av ~ \’\ uteV" ‘eb. . ‘v- ‘a s“ aA _ 0.. ‘.‘V ‘ ‘ a‘fi vA I ,- exhibit internalized or so-called "guilt" responses as often as do non-delinquents when social class variables are constant, the morality model which differentiates amoral and moral transgression responses would become less credible. Such a finding might also raise doubts as to the sagacity of attempt— ing to establish conscience or guilt as a psychological vari- able which has explanatory value for overt socially acceptable behavior. The morality and amorality positions are examined in the present investigation under the delinquency versus non-delinquency issue. The Anonymous-Punitive Issue The widely accepted belief that internalized responses are indicative of guilt or conscience is based upon the equally accepted assumption that such responses often occur in the absence of the threat of detection and consequent punishment. Sears, Maccoby and Levin (1957) regard resistance to temptation in the absence of detection or the expression of confession and self-criticism following failure to resist temptation as the behavioral correlates of conscience. Excluding self- criticism, these investigators regard flight and denial as instrumental responses, since they are viewed as attempts at avoiding external punishment. Kohlberg (1963) states that cultural rules of social action may be said to be internalized when situational sanctions or incentives do not determine con- formity, that is, when conformity is intrinsically motivated. . Aususel 31C, a 5 OA 3 fl". uv bu- ‘ terna. C a... h~5 . A. b»\ . 'A y a». ’4 . 1) l. u.€ S A ‘u.~ ‘ x. «:1: .‘. . A ~"~‘ ‘ .- -..~n ‘ Ausubel (1955) defines guilt as an internalized characteris- tic, a sense of moral obligation and responsibility to conform to a culture's social norms that operates regardless of ex— ternal conditions. He regards fear alone as an inadequate deterrant to anti-social behavior. All of these views are based on the psychoanalytic account of superego development which maintains that internalized reSponses originate in reactions to external authority, but with maturation eventually become self—administered even in the absence of supervision. However, defining conscience develop- ment as a declining concern with external sanctions and a growth to behavior motivated by moral standards and values that are independent of situational sanctions or incentives may not be verifiable. How can internally dictated adherence to social norms be discriminated from conventional behavior which is motivated by fear of detection and punishment in the absence of surveillance? Responses which appear to conform to the definition of internalized behavior as stated by psychoana- lytically based writers may in fact be generated by externalized concerns. One investigator states, for example, that confession responses represent a tendency to be upset by imagined disap- proval from others, even when fear of detection is minimal (Rebelsky, 1963). In a similar vein, Whiting (1959) attributes internalized control over transgressions in many different cul— tures to fear of external punishment. The question, "To what extent are internalized standards vulnerable to or caused by w 911633; E is mterna bflvf‘“ ' ’ 12.5...dlll "3‘s“: w‘r‘ ...NC;A Lndn a .u u..\-Al \ vwn' :k-yvpno Hi u . 9-4 r‘?’ 15 ‘ P ”wait A- “we '4- "" \ufiillnuu a ”2:- “firfisrq ~~ “How-I ‘ an A ".. u. _‘ L on“ ~ A n ~““nl . :R A ".__\L‘ \ ~~ c (1‘ 5. . \:\"A I .‘. .“5 ‘ ‘ ~..q"~- ‘A'qb- d‘ ... ‘- ‘ h ‘l‘:~c¥,.: I A s‘.,“§_‘h ‘C \ h\‘~ "9 31‘ n H H .4 1 .b ., h ‘ ~‘ ‘ 1 0‘3; h external pressures and reinforcement?" or "How internalized is internalized morality?", is left unanswered by defining internalization as the ability to behave in a socially pres- cribed manner even in the absence of surveillance and possible punishment. It is probably impossible to completely eliminate the threat of potential detection and consequent punishment from any transgression Situation. Nevertheless, such an aim can be approximated. The anonymous-punitive issue is investigated in the present study by creating a relatively non-punitive, non-detection condition in which transgression occurs and con- trasting this with transgression in a maximally punitive, detection condition. Responses in these different situations can then be examined. If the investigator's assumption is tenable that all transgression reSponses are instrumental in reducing anxiety, the punitive condition Should generate a greater number of both internalized and externalized responses. If on the other hand, psychoanalytic theory is to be supported, the punitive condition should elicit more externalized re— sponses, since psychoanalytic theory also regards such re- sponses as influenced by situational sanctions; but internalized responses, such as self-criticism, Should not be affected by the change in conditions, since such responses are purportedly independent of situational incentives or sanctions. The anony- 'mous or "guilt" condition Should therefore generate at least an equal number of internalized responses. If the anonymous ,. gag: s ‘ 1 “(Nut ~-5 , , viz? *" isn‘t! bl- . ROOAV‘ A — v‘-uvn V .. 'v-u «N6 sun b3. .. ”2:” 5“ hilly: . 'r*-~ ..:~\l“ a "' v ' . ‘ H..g~ “rss. A; b..._."" ‘ v“. 4 Fl ..‘ C v - Hwy J 10 condition elicits significantly fewer internalized responses than the punitive condition the assumption that such responses often occur in the absence of the threat of detection and conse- quent punishment will not be supported. Such a finding would cast doubt on the belief that internalized responses are indicative of guilt or conscience, since this belief would then be based on a questionable assumption. The investigator defines internalized responses to trans- gression as those which usually remain covert or nonverbalized. of those responses investigated in the present study, self- criticism is most indicative of an internalized response to transgression. Confession, apology and reparation must be verbalized. As such they are not viewed as internalized re- sponses. Furthermore, such responses may represent sensitivity to external punitive agents and may therefore indicate a desire to receive or regain another's favor. Consequently, such responses obviously could be influenced by external considera- tions - However, a concern with self-involvement in transgres- Sion which may or may not be associated with confession would tyPically remain covert and unexpressed, giving confession an internalized aspect as well. Self—involvement in confession as contrasted with externalized confession is, therefore, also categorized as an internalized response. Denial of transgres- sion requires an external agent and has no known Specific emotional correlate; it may, therefore, be categorized as a totally externalized response to transgression. In summary .aa'r- .- ...v.. on:- v: 'r v“ ,_. o.. .y- .un .- — ‘; ...u u‘ _ ..._ (I! s.. >- i. ’v I ll' ! '- 11 then, any response to transgression which is usually covert and often unexpressed and which is therefore presumed to occur in response to internal stimuli, is here defined as an internal- ized response to transgression. This definition results in the same categorization of internal and external responses to transgression as would be the case if the categorization were based on psychoanalytic theory. Its advantage however, is in avoiding the morality connotations associated with the more usual definitions of internalized and externalized transgres- sion responses. The Socioeconomic Class Issue Aronfreed (1961) has referred to verbal disciplinary methods as "induction" techniques, since they induce reactions to transgression which tend to become independent of their original source. Thus, a mother who expresses feelings of dis— appointment or hurt in association with her child's transgres- sion, communicates disapproval of his actions and perhaps rejection of the child. This arouses unpleasant feelings in the child that would persist even in her absence. Likewise, termination of this negative affect is less well defined by her presence than would be the case were she to use physical punishment. The use of verbal explanation and reasoning facili- tates resources within the child for evaluating and changing his own behavior. The child's regaining of parental approval after examining his behavior, assuming responsibility for it, . «an ‘ . cud - ov‘rt ..‘“I A, ‘ we. v. 1" fl. 4: \r I ‘l (I) [U 9" u... . ”A. ."“u -.‘ ‘v— .~‘ 12 and making restitution, etc., reinforces these responses to transgression on subsequent occasions. Induction techniques may thus facilitate internalized responses to transgression. Sensitization techniques, on the contrary, place emphasis on the presence of the punitive agent and greater importance on complying with external demands, rather than internal con- straints. They sensitize the child to the anticipation of physical punishment if detection follows violation of external demands or expectations. As indicated earlier, several investigators have found that internalized responses to transgression are associated with induction methods of discipline (Burton et al., 1961; Ausubel, 1955: Davis, 1947; Sears et al., 1957; Allinsmith, 1960; Heinecke, 1955; Whiting & Child, 1955: Mackinnon, 1938; Kohlberg, 1962). However, the tendency to equate internalized responses with moral learning, guilt or deveIOped conscience, has led to the interpretation that parents who use induction methods of discipline produce children who function on a higher level of morality or conscience development, while parents who use sensitization techniques produce children who are "guilt- deficient" or "conscienceless" (Sears et al., 1957). Theorists who use an anxiety-reduction model place greater emphasis on the reported association between internalized responses and middle-class status, and externalized responses and working-class status (Aronfreed, 1961; MacRae, 1954; Peel, 1959). This association is viewed in amoral terms as indicating V'VO'H - ul MI; “C. “.4- . .5. P‘.- u a». or}. w. 4 t . . st Va . vi. 1‘. . d N.» A s .~h \~‘ . . Ce 6. ‘a u,. . . u .I. uu. ,. . . . A: .. Aye a . . A .. Ta. H... u . \ .. ...o.. .nu we 33 a? MC. in a: N: .: .. . ..\. I. l. C» 2. C» 3. a: .... .mu .. v. .t ,r,. A: v. n: v . 2‘ s . wu. v . 5.. . 15 that the various types of response to transgression represent the invariant final result of specific patterns of social reinforcement. The fact that internalized responses have been found to be associated with both parental use of induction techniques and with middle-class status, suggests that a relationship exists between induction techniques and middle-class status. This has been found to be the case (Bronfenbrenner, 1958; Maccoby, Gibbs et al., 1954; Littman, Moore, & Pierce-Jones, 1957; Miller, Swanson, et al., 1960). Apparently, belonging to the middle socioeconomic class prediSposes parents to use induction techniques which in turn results in children's use of internalized responses to trans- gression. Anxiety-reduction theorists interpret use of such responses by middle-class children as learned instrumental responses for reducing anxiety. Thus, from a functional, anxiety-reduction perSpective, internalized responses are viewed as a separate type of transgression response only in the sense that they are effective in reducing anxiety under middle-class disciplinary techniques. They apparently are ineffective in reducing anxiety under sensitization methods. If transgression responses are produced only for the purpose of reducing anxiety, working-class children would not be expected to evidence internalized responses to transgression, since their parents apparently do not reinforce such responses. vv“ _.y I.» .: q.- ~n :5 .: .. .., A\~ 2‘ ._~ -. z. 14 Psychoanalytic theory and the anxiety-reduction model offer different explanations as to how induction techniques might facilitate internalization. Psychoanalytic theory empha- sizes identification with socializing agents and the development of superego. The anxiety-reduction model stresses the role of conditioned anxiety. Before reviewing these theoretical posi- tions and examining relevant empirical studies, one further issue inherent in the morality-amorality controversy needs to be explored. The Delinquency versus Non-delinquency Issue The resistance to temptation criterion which has been viewed by some investigators as a behavioral correlate of conscience (Sears, Maccoby & Levin, 1957) is examined in the present investi- gation by comparing the transgression responses of delinquents and non-delinquents. Sears and associates, assume that conscience development or superego (as indicated by internalized responses to transgression) prevents nonconventional, negatively sanc- tioned behavior. It follows from this assumption that individu- als who transgress and ignore society's prohibitions must have inadequately developed superegos or must be "conscienceless" (Sears et al., 1957). This position is eschewed in the present investigation, since the investigator views internalized re— sponses as attempts at coping with anxiety and as unrelated to conscience development. The morality model would receive support if delinquents demonstrated Specific and characteristic responses to ov:" ..‘a- .A.n_' \. .‘Fyn .— '-~.» a“... A. h I— ~uu‘y' . ‘1 o~__‘~’ :"~.~ "'\ ‘l“~ ., e. . A. h. Av- .‘ .1“ . .. j ‘- I‘v- P, "‘~~: \ _ fi “‘..:_‘ - .‘- ‘ ‘0‘ ~ ‘ h u. - ‘ - . ‘1'. ¥.. . .. E) §.-“‘ .“-I - . “, “n. -‘s CV, L. 15 transgression. The amorality perspective assumed in the present investigation, on the contrary, predicts that delinquents and non-delinquents do not differ in type of response to trans- gression. If internalized responses do reflect superego develop- ment, and if Sears and associates are correct when they state that a fully developed superego is necessary for the expression of socially acceptable behavior, then delinquents should evidence a paucity of internalized responses relative to non-delinquents. The resistance to temptation criterion as a behavioral correlate of conscience has recognized limitations even to psychoanalytically oriented investigators. Kohlberg's (1965) review of research on moral development indicates inconsistent findings for resistance to temptation in and out of the home. Hartshorne and May's (1928-1950) early investigation and their finding'cm’situational specificity for resistance to temptation SUggests that learning to resist temptation under specific types Cf parerwal discipline does not generalize in diverse situations. Kohnmarg (1965) therefore suggests examining guilt reactions to transgression rather than behavioral conformity criteria as indicators of morality. He argues that doing so would circum- vent the situational specificity involved in the resistance to temptation criterion. The present investigation examines both Criteria of conscience development by studying the reactions of delinquents and non-delinquents to a transgression experience, in order to demonstrate that neither criteria is adequate for SuPPOrting the morality model. .. 'r‘!"‘ on“ ‘ .c...a §. '0" IA; A usuav u: 5 In.“ 5 .A A b’u“. “v“... - a r" 1"; uvb‘.n "‘ M... .- ‘ H o 0 n e \ “‘- ‘ h I § ‘~ C .a ‘ Q -‘ - 16 Although situational variables may have a differential effect upon tendencies to resist temptation, delinquents are individuals who fail to resist temptation in a variety of situations. They usually manifest impulsive behavior in the school setting, are often described as "incorrigible" at home and frequently violate the legal and social code of acceptable behavior outside of the home. Delinquents demonstrate a generalized inability to delay gratification. Furthermore, delinquents are typically not categorized as such by society nor are they confined to an institutional setting for rehabili— tation until multiple infractions of legal and social norms have been detected (Matza, 1964) . Thus, if the morality perspective is valid, and if it is worth while to maintain the resistance to temptation criterion as a behavioral correlate of conscience then non-delinquents should give greater evidence of conscience development than known delinquents in a transgression situation. In a sword, the morality model would be supported if non- delinquents evidence a predominance of internalized responses and delinquents manifest externalized, punishment-avoidance responses in a transgression Situation. Evidence inconsistent with the morality position would be provided if the transgres- sion responses of delinquents and non-delinquents do not differ. W31 TheorL versus the Amorality Position Psychoanalytic theory and the anxiety-reduction model Offer differing explanations as to the origin and maintenance Of internally mediated responses to transgression. They also .. .4- l1: 4.. ‘I -v A'4 vs. 0“ (I) s. (ll 1:; «A --_ 17 differ in their respective explanations of the reported associ- ation between parental use of induction techniques and children's use of internalized responses to transgression. While morality theorists regard internalized responses as indicative of conscience, from an amorality perSpective internalized responses are simply viewed as instrumental responses learned under highly verbal and cognitive forms of parental discipline (Aronfreed, 1961; 1965; 1964; 1965; 1966). Thus, children apply parental criticism to themselves in order to reduce the negative affect associated with parental censure, rather than to experience the positive affect associated with nurturant parental characteris- tics as many morality theorists suggest. Aronfreed's contention is therefore Similar to Anna Freud's "identification with the aggressor" (1946) and Mowrer's "defensive identification" (1950), concepts which also reflect behavior designed to reduce anxiety. In short, Aronfreed's several investigations have led him to the conclusion that internalized responses to transgression are anxiety-reducing instrumental responses rather than pain-inducing expressions of psychoanalytically conceived "guilt". He main- tains that an invariant component of internalization is con- ditioned anxiety. In naturalistic socialization, anxiety per- taining to fear of punishment following transgression eventually becomes conditioned to the act of transgression itself. When this happens the presence of external socializing agents who administer rewards and punishment is no longer necessary for the occurrence of the originally anxiety-reducing responses. - v- 1': ,_» awn-0 ‘- :¢I. — -¢.vv --.....J ‘3 A~ ”'1 .~ u; o -. A‘J "O 5rd. l'~ . mm g C I'C‘v- guy _‘ p... ., x A — «s-» 18 Transgression responses may thus become independent of external sanctions. Conditioned anxiety possesses the same motivating properties and therefore facilitates the same instrumental behavior as was originally elicited under the influence of external controls (Aronfreed, 1961). Aronfreed (1966) parsimonously explains the use of internal- ized and externalized transgression reSponseS by an aversive learning model which includes classical and instrumental condi- tioning. Parental discipline induces the attachment of anxiety to external and intrinsic stimuli in accord with the principles of classical conditioning. Application of parental criticism to the self and other responses to transgression acquire reinforc- ing consequences, that is, reduce anxiety, in accord with instrumental learning principles. Reinforcement occurs initial— ly when responses are effective in avoiding parental punishment, and subsequently when responses to the transgression reduce the anxiety which is conditioned to the transgression (Aronfreed, 1966). Amorality theory would thus stress the role of conditioned anxiety in accounting for both externalized and internalized responses to transgression. The morality model differs primarily with regard to the origin of internalized responses. It empha- sizes children's identification with parental standards and values in accounting for internalized responses generally and reproduction of such specific parental social responses to transgression as self-criticism. ivvu. - en!" 4- cu‘ .v"; ._....- a-" U) av- —. . . . .v-‘V —..-‘ A” dun Av-QQ _\ -. “.2 Q :- 'V‘h‘ - b- . sh.“ s; 0 1 (D 19 Freud thought that the experience of guilt, which is governed by the superego, serves as an inhibitor of socially unacceptable aggressive and sexual impulses. Those of us who transgress freely in these areas (i.e., who give free expres- sion to such impulses) have underdevelOped or primitive super- egos. Superego or conscience is martialled as a unitary dimension which deve10ps through identification with parents and in the normal individual gradually comes to serve the func- tion of controlling behavior, a function that was previously served by the parents (Freud, 1956). Freud (1955; 1956), Fenichel (1945), Anna Freud (1946) and Sanford (1955) all ex- press the psychoanalytic view which maintains that children have a dual motivation for identification with parents. A child's primary motivation for identification is based upon a desire to reproduce the parent's characteristics, Since they represent sources of nurturance and affection, and secondly, to reduce the anxiety aroused by the threat of their punishment or re- jection. Ausubel (1955), Bronfenbenner (1958), Miller & Swanson and associates (1960), Sears and associates (1957), and Whiting & Child (1955) all place maximum weight on the first of these two reasons for a child's motivation for identifying with his parents. Thus, a child is believed to adopt his parent's standards and values primarily because he wishes to reproduce and maintain their love and affection. Parents who are warm and nurturant and who use verbal disciplinary techniques and withdrawal of affection as forms of punishment are purportedly .. a. V 0 (li .. Io: ‘- an ‘at- u: «D ll) .- I: e ‘~ I. ‘\ 20 easiest to identify with. Hence, parents who use such tech- niques produce children who are capable of internalizing parental values and standards. Conversely, children of hostile, rejecting parents who use physical forms of punishment as disciplinary techniques cannot identify as readily with their parents, therefore do not develop internalized standards of morality and remain "guilt-deficient" or "conscienceless" (Sears et al., 1957). In the same vein, R. R. Sears (1957) and Whiting & Child (1955) suggest that children apply self— critical standards to themselves, since such behavior is secondarily reinforced. Thus, children imitate or practice parental roles because parental nurturance secondarily rein— forces other characteristics. According to these writers, a child's predisposition to reproduce parental evaluative re— sponses is a function of the extent to which his dependency upon them has been reinforced by their affectionate and nurtur- ant interactions with him. Motivation to reproduce a model's nurturant characteristics is thus presumed to include the internalization of social punishment (Sears et al., 1957). In sum, parental nurturance and affection create the motivation for identification. Identification results in the internali- zation of parental standards and values and in the reproduction of parental social responses, e.g., self-criticism. The assumption that children who do manifest self-criticism in their behavioral repertoires, do so as one aspect of identi— fication with their generally nurturant parents needs more 21 direct experimental verification. Recent investigations (Aronfreed, 1964; Aronfreed, Cutick, & Fagen, 1965) cast doubt that a child's tendency to apply parental criticism to himself is based upon this principle. These investigators state that the most severe limitation to the view that self-criticism is based on a general tendency to reproduce parental roles, lies in the nature of the self-critical response itself. In such behavior, a child adopts parental behavior which is associated with the painful experience of punishment. Kagan (1958), Maccoby (1959) and Whiting (1960) have also reasoned that if secondary reinforcement is in fact responsible for a child's tendency to reproduce parental behavior, it would be more plausible to assume that he would express behavior imitative of parental rewarding, rather than of critical behavior. Thus, these investigators while not anxiety-reduction theorists, nevertheless find fault with the psychoanalytic account of the origin of the self-critical response to transgression. Aronfreed and associates (1965) and Hill (1960) criticise psychoanalytic identification theory on the grounds that the Specific mechanism of reinforcement through which a child learns to apply parental standards to his own behavior has not been made clear. However, Hill and Aronfreed disagree as to the source of reinforcement for the self-critical reSponse. Hill (1960) attempts to demonstrate that the self-criticism response is gradually enhanced through direct external reinforcement. Aronfreed (1965) argues that the response must have 'wy. ll bug‘s cub.» AA 'vay ouugv . .. o I: - ..~. o.‘ " AI. ah I .....:. ’ a A ,j p" a. 'V‘.~‘ .;A m: 'l .i 'v h 3“. “ .. ‘-,; e.“- . . h‘nA - I Q ~.. ‘ . .I ‘ i h \ I- ‘ 'q 5 I‘D I. ‘ ’l 5‘. ‘l 22 self-reinforcing characteristics, since the initial occurrence of self-criticism by a child is usually quite vigorous. Furthermore, it is difficult to conceive of a response that typically remains covert or unexpressed as being reinforced incrementally through external means. The disagreement between these two learning theorists arises because the imposition of negative affect upon the self as in self-criticism is itself a presumably painful, anxiety- filled experience which is difficult to explain with reference to an anxiety-reduction model. It is easier, for example, to conceive of confession, apology, etc., as serving an anxiety reducing function, since such responses are often followed by the removal or palliation of punishment and the regaining of approval from parents; that is, such responses do receive direct external reinforcement as Hill (1960) has demonstrated. Although self-criticism may not receive direct external reinforcement, Aronfreed nevertheless regards it as an instru— mental response. It's origin and maintenance as a reducer of anxiety appears plausible when viewed as occurring in response to induction socializing techniques. Under such disciplinary methods punishment becomes associated with critical labels and verbal explanations. Punishment following transgression results in the experience of anxiety on subsequent occasions when transgressions are committed. Anxiety is terminated when punishment is administered. Punishment itself therefore becomes anxiety-reducing as do the critical labels and verbal explanations ,4-.. (I1 25 associated with it. Thus, on subsequent occasions when trans- gressions occur, the critical labels previously employed by the parents are employed by the child to himself as an object so as to reduce the anxiety associated with his fear of subsequent punishment. Apparently, self-depreciation is less painful or anxiety-provoking than is fear of externally imposed punishment. For such a self-critical process to occur however, parents must use a cognitive and verbal medium in their disciplinary tech- niques. Verbal mediation allows the child to apply critical labels to himself early in the transgression-anxiety-punishment sequence, although they originally occurred, through external sources, at the completion of the sequence (Aronfreed, 1964). Both Aronfreed and Hill maintain that there is no unitary dimension such as superego underlying the internalized suppres— sion of socially unacceptable behavior or the self-criticism that may occur following transgression. Rather, Specific trans— gression responses (such as confession, reparation, self— criticism, apology, denial, externalization of responsibility, etc.) are each learned in accord with specific socialization practices that reinforce or punish such behaviors. All of these responses are regarded as instrumental behaviors which persist in the behavioral repertoires of individuals, because they re— duce the anxiety (rather than the "guilt") present during and following transgression. The surveys which have found that highly verbal, induction, disciplinary techniques are associated with greater evidence of coh .n . (I) . an. 51; -\-. _ 1,, ~,. ‘5‘ ~.., (I! (I! 24 internalization in children subjected to them (Aronfreed, 1961; Sears et al., 1957; Whiting & Child 1955) may be interpreted as indicating that the verbal behaviors employed by parents rather than withdrawal of affection Specifically, or love-oriented discipline generally, are the primary factors responsible for the greater evidence of internalization. Aronfreed (1964) pro— vides evidence which is contrary to the notion (Sears et al., m 1957) that motivation to reproduce a model's nurturant character— istics is reflected in childrens internalization of social ~u punishment. To test his assumption that criticism by parental models is a component of punishment that eventually operates as a signal for the reduction of anticipatory anxiety and is there- fore reproduced in transgression situations, Aronfreed (1964) designed an experiment to assess the relevance of the timing of criticism for the acquisition of the self-critical response. This factor was contrasted to the relevance of the reward characteristics of a model in producing self-criticism. If Sears and associates (1957) are correct, children should be more likely to adopt the criticism of a nurturant, than of a neutral model in a punishment learning situation, regardless of the timing of the criticism in relation to the punishment. Criticism occurred concurrently with either the onset or the termination of anxiety. Aronfreed found that whether the adult model was nurturant and physically affectionate or displayed a neutral demeanor had little effect on the child's adoption of the model's criticism. 714' -. ("I ~ ‘ k. I" IV’ 25 However, the factor of timing of criticism relative to presumed onset and termination of anxiety did appear to be crucial to the child's adopting the model's use of criticism and applying it to himself. The results indicated that self-criticism is acquired as a specific reSponse in aversive learning situations. As Hill (1960) had predicted in an earlier study, the critical label was used by subjects on test trials significantly more often if the label had been learned under conditions designed to give it maximum anxiety—reducing value by signaling termination of punishment, than when learned under conditions of nurturance— affection. The fact that self-criticism occurred only when timing was such as to signal termination of punishment, suggested that the factor responsible for its adoption was anxiety—reduction. Love oriented discipline per se apparently has little effect on the self-application of critical labels previously used by socializ- ing agents (Aronfreed, 1964). In another investigation, Aronfreed reasoned that the tim- ing of criticism relative to onset or termination of punishment may not be as crucial in naturalistic socialization, as in laboratory situations, Since verbal mediation under highly cognitive forms of discipline might spread anxiety uniformly over the transgression-anxiety-punishment sequence (Aronfreed, 1965). Thus, children in the laboratory situation might not apply critical labels to themselves, if these labels are in- serted by the experimenter in the training paradigm prior to if LII .Q ya -u (I) ‘u ‘> n . ‘— U) a. ‘o 26 punishment, since in this case they will merely be associated with the onset of punishment and therefore increase, rather than reduce anxiety. Under naturalistic socialization, however, on subsequent transgression occasions, verbal mediation might facilitate the association of these labels with termination of punishment thereby reducing anxiety. Reduction of anxiety via self—application of criticism on these occasions should rein- force self-criticism responses. Thus, if a child is to apply parental criticism to himself in naturalistic situations, it would seem essential that a highly verbal and cognitively structured context be provided. Aronfreed, Cutick and Fagen tested this assumption in a second investigation, in this case comparing the effects of cognitive structure and nurturance on the adoption of self-critical re— sponses following transgression (1965). AThe findings, as expected, demonstrated that self-criticism occurs Significantly more often when punishment is accompanied by sufficient verbal or cognitive labelling. Application of self-criticism is unlikely to occur when cognitive structure is minimal. Furthermore, the effect of cognitive structure was not dependent on the degree of nurturance provided the child in socialization paradigms. Nurturance per se did not appear to be necessary nor sufficient to account for a child's adoption of a socializing agent's evaluative behavior, nor was this de- pendent on the degree to which punishment indicated a withdrawal of nurturance. ..4 A.» ~u~ r ~ SD .I Iw 27 Morality and amorality theorists thus differ as to whether the self-criticiam response initially represents a guilt-induced attempt at identification or whether it is an instrumental, learned response designed for reducing anxiety. Although the two perspectives differ as to the origin and function of self- criticism, they both regard verbalized self-criticism in trans— gression situations as most clearly indicative of an internally motivated response. Other reSponses to transgression, such as confession, apology, reparation, externalization of responsibility and denial are viewed as progressively less internalized. The seven categories of transgression reSponses examined in the present investigation have been ordered in a 7-point scale with low and high scale values indicative of externalized and internalized responses to transgression, respectively. Each transgression response is weighted from 1 to 7 depending upon which category it exemplifies. Thus a response categorized as denial receives a weight of 1, while a self-criticism reSponse receives a weight of 7. A mean transgression score is derived for each subject by summating the weighted values of his state— ments and dividing this summation by his total number of verbal- ized responses. Each subject's mean transgression score thus indicates his position on the 7-point internalization scale. The seven categories of transgression reSponses are described in greater detail in the next chapter. ‘v w. 5-» V. up. '.~ "— -‘ v. .5 v ~ ~\~ 28 Summary and Hypotheses The Anonymous-punitive Issue The basic theoretical point of view assumed in the present investigation is that both externalized and internalized re- sponses to transgression are instrumental responses designed to reduce anxiety. The investigator thus eschews the dichotomy between externalized or instrumental responses and internalized or guilt responses to transgression assuming an amoral anxiety- reduction perspective with regard to both types of response. The present investigation attempts to demonstrate that a transgression experience elicits both externalized and internal- ized responses and that both types of reSponse are instrumental in reducing anxiety. It does this by creating a punitive or maximal fear condition and an anonymous, non-detection condition which is designed to elicit minimal anxiety. The assumption, based on the morality model that internalized, "guilt-indicative" responses occur even in the absence of the threat of detection and potential punishment, and the further assumption that such responses are independent of situational sanctions or incentives suggest that internalized responses should occur as readily in the anonymous as in the punitive condition. On the other hand, the anxiety-reduction, amoral perspective, adopted in the present investigation suggests that transgression occurring under a maximal threat of punishment should elicit more anxiety, hence a greater need for reSponses to reduce that anxiety than u-F u'. I!) ll) .1. (I) rn.‘ \ «.4 :vu ‘Q - ‘ 29 would transgression occurring under an anonymous, non-detection condition. If both types of response are instrumental to reducing anxiety, the punitive condition should elicit more responses of both types than the anonymous condition. The amorality model maintains that orientation in respond— ing to transgression serves the single purpose of reducing anxiety. An internalized orientation is in no way generally superior to an externalized orientation in realizing this pur— pose. Both orientations are simply viewed as the products of learning under differential social reinforcement. Thus, a maximal fear condition should generate a greater need to reduce anxiety and elicit a greater number of anxiety—reducing re- sponses than a relatively benign condition. The number of re- sponses produced in each condition is the primary dependent variable for evaluating the anxiety-reducing function of transgression responses. Although the quantity of responses should be differentially reflected by the experimental conditions, there is no reason to expect a particular type of response or type of orientation to predominate in one of the two conditions. Rather, both types of response should increase in frequency in the maximal fear as compared with the minimal fear condition. A general increase in the number of responses produced however, should 1mm affect the mean transgression score, since this is derived fnmmthe weighted average of internalized and externalized responses. If the number of low—weighted, externalized responses 30 and heavily-weighted internalized responses were both to in— crease in the punitive condition the amorality model would be supported. Random differences in mean transgression scores between the two experimental conditions would then be expected. Morality theorists, on the other hand, view externalized responses alone as instrumental flight or avoidance responses to fear situations. They would therefore predict the maximal fear condition to elicit externalized responses, but would not maintain this expectation with regard to the anonymous condition, since the necessity for flight or avoidance is minimized in this non-detection situation. However, guilt would be equally elicited in both conditions according to this view, since the actual transgression experience is identical in the two con- ditions. If guilt or internalized responses are averaged with externalized responses in the punitive condition the mean trans- gression score would be lower than in the anonymous condition where the more heavily-weighted internalized responses alone would be averaged. Thus, results indicating higher mean trans— gression scores for Ss in the anonymous condition would be con- sistent with morality theory, while a finding of no difference between the two conditions would be consistent with the anxiety- reduction, amorality, model. These considerations lead to the following hypotheses: Hypothesis I - A punitive or maximal fear condition elicits a greater number of transgression responses than does an anonymous condition which is designed to elicit minimal anxiety. 51 Hypothesis II - On a 7-point scale with low and high scores indicative of externalized and internalized responses respec- tively, Ss in a minimal fear condition do not differ in their mean transgression scores from $8 in a maximal fear condition. Hypothesis III* - (A) - When externalized responses are defined as including denial, externalization of responsibility and externalized confession, Ss who transgress in a punitive or maximal fear condition tend to respond with a greater number of externalized responses than do Ss who transgress in a benign, non—detection condition. (B) - When internalized responses are defined as including self—involvement in confession and self- criticism, 53 who transgress in a punitive or maximal fear condition tend to respond with a greater number of internalized responses than do 85 who transgress in a benign, non—detection condition. The Socioeconomic Class Issue A tendency to manifest a specific internalized or external- ized orientation in responding to transgression apparently results from specific patterns of social reinforcement associated with socioeconomic status. Reported associations suggest that belonging to the middle socioeconomic class predisposes parents to use induction techniques of discipline which results in the *- Reparation and apology have been found to be intermediate between externalized and internalized responses to transgression (Kohlberg, 1964). They are therefore not included in either re- sponse type. 32 learning and use of internalized responses to transgression by middle-class children. The association between use of sensi- tization techniques and working-class socioeconomic status sug— gests that working-class children tend to learn and use externalized responses in reacting to transgression experiences. This investigator doubts the assertion of morality theorists that children apply parental criticism to themselves as one aspect of identification with generally nurturant parents. The association between induction methods of discipline and use of self-criticism responses is more parsimoniously explained by an anxiety-reduction model. Although induction techniques are characterized by love-oriented forms of discipline, such discipline typically occurs within a highly cognitively struc~ tured and verbal medium. Cognitive structuring and verbal mediation rather than nurturance or love-oriented discipline may be the primary factors responsible for internalization in middle-class children. If self-criticism and other internal— ized responses to transgression are learned only under highly verbal and cognitively structured forms of discipline, and if such conditions predominate in middle-class homes, then middle- class adolescents should evidence high internalization and working-class adolescents low internalization or an externalized orientation as reflected in responses to transgression. These considerations prompt the following hypotheses: Hypothesis IV - On a 7-point scale with low and high scores indicative of externalized and internalized responses reSpectively, 53 middle-class Ss obtain higher mean transgression scores than working-class Ss. Hypothesis V - (A) — When internalized responses are defined as including self-criticism and self—involvement in confession, middle-class Ss tend to respond to transgression with a greater number of internalized responses than do working-class 85. (B) — When externalized responses are defined as including denial, externalization of responsibility, and externalized confession, working-class Ss tend to respond to transgression with a greater number of externalized responses than do middle-class Ss. Hypothesis VI - Verbal ability as reflected in scores on an intelligence scale which is heavily weighted for verbal ability is positively associated with mean transgression scores, higher scores indicative of internalized reSponses and lower scores indicative of externalized responses respectively. The Delinquency versus Non—delinquency_Issue Morality theorists recognize two criteria of conscience development or superego. These are resistance to temptation in the absence of surveillance, and expressions of "guilt" or self-criticism following failure to resist temptation, that is, following an act of transgression. Both criteria are examined in the present investigation. The resistance to temptation criterion and "guilt" are evaluated by comparing delinquents with non-delinquents specifically with regard to internalized responses elicited by an act of transgression. 54 Morality theorists assert that the superego indirectly and the experience of "guilt" directly acts as an inhibitor of socially unacceptable behavior. Thus, individuals such as de- linquents, who transgress and ignore society's prohibitions must have inadequately developed superegos or consciences. The amorality perSpective adopted in the present investi- gation defines internalized responses simply as covert usually non-verbalized responses to transgression which are instrumental in reducing anxiety, but are unrelated to such overt behavior as delinquency. The investigator questions the sagacity of attempt— ing to establish conscience or guilt (as indicated by internalized responses to transgression) as a psychological variable which has explanatory value for overt behavior. He asserts that psychologists in fact study anxiety rather than guilt and observe attempts to reduce anxiety when they investigate both internalized and externalized responses to transgression. He therefore eschews the alleged dichotomy between internalized or guilt responses and externalized or instrumental responses, maintain— ing that all transgression responses are instrumental in reduc- ing anxiety. From such an amoral perspective it is no less reasonable to expect delinquents to evidence an internalized orientation in a transgression situation than to expect non- delinquents to do so. If there is substance to the dichotomy in transgression responses alleged by morality theorists, delinquents should evidence a paucity of internalized responses and a preponderance . -"""“ M'" "L 55 of externalized responses relative to non-delinquents. Morality theory would receive further support if non-delinquents obtained higher mean transgression scores than delinquents, since high transgression scores as derived from the 7—point scale used in the present study reflect internalized orientations. Such find— ings would suggest that indicators of conscience development may be used to predict and explain socially acceptable or unaccept— able behavior. However, the present investigator predicts that delinquents manifest as many (or more) "guilt" or internalized responses as do non—delinquents in a transgression situation; furthermore, obtained transgression scores of delinquents and non-delinquents are not expected to differ. Such findings would suggest that equating internalized responses with "guilt", "superego" or "conscience" and then viewing such constructs as relevant predictive and/or explanatory variables of overt be- havior is at best a questionable practice. A relationship between number of transgression responses and delinquency is anticipated, although it is assumed that delinquency is not associated with typ§_of response orientation to transgression. However, rather than expecting that delin— quents give fewer internalized responses than non-delinquents as morality theory would assume, or that delinquents and non- delinquents do not differ in the number of internalized responses elicited by a transgression experience as the amorality per- spective would anticipate, it is predicted that the delinquents in the present investigation will produce a greater number of 36 internalized as well as externalized responses to transgression. This prediction is made despite the investigators' belief that delinquents and non-delinquents in a similar environment do not differ qualitatively or quantitatively in their responses to transgression. However, the environments of the delinquents and non-delinquents in the present investigation are markedly different. The fact that the delinquents are incarcerated in a potentially punitive environment, whereas the non-delinquents are non-incarcerated and are therefore in a more benign environ- ment should create two situations which differ from each other much as do the punitive and anonymous experimental conditions. While this situation is not expected to affect the type of trans- gression response elicited, it should increase anxiety among the delinquent subjects necessitating use of a greater number of instrumental, anxiety-reducing responses to transgression. These considerations lead to the following hypotheses: Hypothesis VII - On a 7-point scale with low and high scores indi- cative of externalized and internalized responses respectively, the mean transgression score obtained by delinquent 83 does not differ from the mean transgression score obtained by non- delinquent control Ss. Hypothesis VIII - Transgression elicits a greater number of trans- gression responses from incarcerated delinquent 85 than from non-incarcerated, non-delinquent Ss. .A' ,- a». ma. 57 Hypothesis IX - (A) - When internalized responses are defined as including self-criticism and self-involvement in confession, incarcerated delinquent Ss tend to respond to transgression with a greater number of internalized responses than do non- incarcerated, non-delinquent Ss. (B) - When externalized responses are defined as including denial, externalization of responsibility and externalized confession, incarcerated— delinquent Ss tend to respond to transgression with a greater number of externalized responses than do non-incarcerated, non- delinquent Ss. CHAPTER II METHOD "Transgression" in the present investigation consists of the simulated breakdown of an apparatus being used by the sub- ject (S)which the experimenter (E) had previously described as "very expensive". Following transgression, all subject re- sponses are recorded with the aid of a hidden tape recorder. S is then taken to a second room where the examiner conducts a standardized interview. Responses to the examiner's four probes are also recorded with the aid of a hidden tape recorder. The transgression response categories were coded by two judges on a 7-point scale in accord with the following defini- tions: 1 - Denial is a negative statement by S to the effect that he is free of blame. Examples are, "I did not turn it too fast", "I know I_didn't break it". Denial also includes deliberate fabri- cations to the examiner's probes. For example, Examiner - "How do you think you are doing in the experiment?" Subject — "Fine", or "Very well", or "OK". 2 - Externalization of responsibility is a verbalization by S to the effect that the machine or some part of it was defective, that it broke because of excessive use, or any response that 58 39 omits the possibility of S's responsibility in the machine's "breakdown". 3 - Externalized confession is an admission that the machine broke or a description of the machine's malfunctioning without explicit reference by S to himself. For example, "The machine broke". ""1 4 - Reparation is a statement by S in which he offers to pay for or fix the machine, suggests that he knows where the appara- ~+— tus might be taken to be fixed, or any response indicating use of S's own resources for corrective action. 5 - Apology is a verbalization by S which expresses regret or remorse. Examples are "I am sorry", or "I didn't mean to break it" C 6 - Self-involvement in confession is an admission that the machine broke with explicit reference to the self. For example, "I broke the man's machine". 7 - Self-criticism is a self—depreciatory verbalization descrip- tive of S's behavior in breaking the machine. Examples are, "I turned the dial too fast", ”. . . too hard", ". . . too far". Self-criticism is also scored for a self-depreciatory generali- zation. Examples are, "I am always goofing things up", or "I am clumsy". Each transgression response was then weighted from 1 to 7 depending upon which category it exemplified. As may be noted .v. (I! FA. u.‘ o ~. “ 40 from the order of the transgression categories presented above, the higher the mean transgression score obtained by any S the closer he is to that end of the continuum representative of internalization. A low mean transgression score reflects an externalized orientation. In order to establish reliability between the two judges who weighted the transgression reSponses, every twentieth state— ment was rated by both individuals. For the twenty-five state- ments thus selected, 100 per cent agreement was obtained. The 590 statements obtained from all subjects were then divided be- tween the judges and each statement was assigned a score on the 7-point scale. The intelligence scale (form A) from Cattell's Junior- Senior High School Personality Questionnaire (HSPQ) was adminis- tered to each S prior to his participation in the experiment proper. The items of this scale consist principally of vocabu- lary definitions, and also include analogies, syllogistic reasoning and concept formation. The scale is used as a measure of verbal ability. Recent data indicates that this intelligence scale has an immediate retest reliability of .85, and a retest reliability of .60 after six months (HSPQ Manual, 1965). Scores on the HSPQ are correlated with mean transgression scores so as to derive a measure of association between verbal ability and internalized orientation. 41 Subjects The non-delinquent controls are tenth grade students at a high school and ninth grade students at a junior high school in the Lansing area. The working—class delinquent subjects have been court committed to the Lansing Boys Training School (BTS). A sufficient number of middle-class delinquents were not available at BTS, since middle—class delinquents comprise only 1 per cent of this population. Middle-class delinquents (also court committed) were therefore obtained from Starr Commonwealth in Albion, Michigan. The two delinquent popula- tions differ in several respects, one being that Starr has a higher percentage of middle—class juvenile offenders and a smaller percentage of Negro boys. All subjects ranged in age from 13 to 17 years with a mean age of 15 years, 5 months. Only Caucasian male subjects participated in the experiment, since it would not have been possible to select an equal number of Negro boys from each of the populations. The occupational Classificatory categories given in the Dictionary of Occupational Titles were used to designate socio— economic class. Subjects are defined as middle-class if the head of the household is employed in a professional, managerial, fines, clerical or service occupation. Working-class families anathose in which the head of the household is employed at a “flue-collar" skilled, semi-skilled, or unskilled occupation (see Table 1) . 42 H.00a wmnz w.¢a mum ea mCOHummsooO ooHHflmeD m.mm film mm mcoaummsooo Umaaflxmlflfimm m.© mum m mCOHpmmDOUO Umaaflxm H.N m N mCOADMQSUoO moa>umm fi.hm N mm mmamm Ucm HMUflumHU m.om filo om Hafiuwmmcme pom .HMUHCgomu .HMCOHmmomoum 2 mo HmQEDZ muumflflsm whomoumu HmCOHummsuoo mmmucmoumm mpoo mo quEDZ mama .mmaufla HMCOHummsooo mo kHMCOHuUHQ on» Eoum coxMB .mpomhnzm How mHmQESZ wUOU Ucm mmfluommumo huODMOHMHmmmHU HMCOHDMQDUOO .fi OHQME 43 Random selection of subjects was not possible in three of the four schools due to the limited supply of subjects who met the criteria of the various conditions. Thus, only 52 boys from the entire Starr population had fathers whose occupations could be categorized as "white-collar". Some of these 52 were not available for participation in the study. The 24 middle- class delinquents included in the investigation thus comprise the entire, available, caucasian, middle—class population at Starr Commonwealth. In View of Kohlberg's (1964) finding that moral development (as reflected by several of the same transgression response categories as are used in the present study) increases with IQ, it was necessary to limit the IQ range of subjects included in the present investigation. A strictly random selection of con- trol subjects was therefore not feasible. Obtaining an equal number of control subjects who fulfilled the middle- and working- class criterion was a further limitation on random selection. It was also necessary to establish the fact that none of the control subjects had engaged in delinquent behavior. Information was solicited from school personnel, case records and from the subjects regarding any contacts that they may have had with legal authorities. Any occurrence of arrest, juvenile court contact, or report of delinquent behavior such as theft, made an S in- eligible for participation in the control group. Several Ss who had been selected to participate in the study had to be elimi- nated, because of admitted or recorded delinquencies. 44 An abundant supply of Caucasian, working-class delinquents was availabletn the investigator from the BTS population. Every fourth boy was selected from an alphabetically ordered list to participate in the investigation. Four groups of 24 subjects each were thus selected, middle- and working-class delinquent groups, and their middle- and working-class non-delinquent controls. The subjects in each group were then randomly assigned to one of two experimental conditions, the anonymous-benign or the punitive-detection condition. Eight groups of 12 subjects each comprising a total §_of 96 were thus included in the investigation. In order to demonstrate that IQ is not associated with ob- tained differences between groups on transgression scores, those IQ scores which could be obtained from case records (61 out of 96 were available) were correlated with transgression scores in each condition. The absence of a significant relation- ship between IQ and transgression scores was thus demonstrated as indicated in Table 2. Each of the obtained correlations exceeds the .05 level. Therefore, with regard to the corre- lation between IQ and the main experimental conditions an hypothesis of random sampling from a zero correlated population is regarded as tenable. Procedure General Instructions In both the anonymous—benign and in the punitive-detection conditions E describes the experimental apparatus as an expensive 45 Table 2. Correlations for Mean Transgression Scores and Number of Transgression Statements with IQ Scores for the Main Experimental Conditions and for all Groups Combined* Mean Transgression Number of Verbalized Condition Score Transgression Statements r N r N Middle-class —.25 50 -.54 50 Working-class +.24 51 -.04 51 Delinquency -.08 48 -.24 48 Non-delinquency +.11 15 -.26 15 Anonymous -.16 50 -.15 50 Punitive +.01 51 —.08 51 All conditions -.05 61 -.20 61 *Note: None of these correlations reach significance at the .05 level. 46 machine used for measuring thought processes in computing arithmetic problems. Each S is informed that part 2 of the experiment takes place in a second room where an examiner discusses the experi- ment with the subject. E stresses the point that the machine is an expensive apparatus which has required a great amount of time and effort to construct. S is asked to carefully turn the dial on the face of the machine, to the positions indicat- ing the solution to the arithmetic problem presented by E. The dial on the face of the apparatus may be set by S at 0, 1, . . . 9. It is rigged so that dialing 8 breaks the electrical circuit. S is asked to dial any two numbers which add up to a particular total. Each problem is assigned in a prearranged order from a list which E keeps before him. Thus, S is asked to dial two numbers which add to 7, 5, 10, 5, . . . 15. In order to dial two numbers which summate to 15, S must dial the 8 and the apparatus will "break". This is dramatized by a buzzer sounding, a bulb theretofore dark lighting and making the words "TPG- not in trail" clearly visible, other bulbs already lit going off, and a voltmeter and ammeter on the face of the apparatus dropping from readings midpoint on their scales to 0. E immediately comments, "What happened?" and frantically turns a non-functional switch in a feigned attempt at correct— ing the damage. This having no effect, E pulls the cord from the electrical socket finally terminating the buzzer. E then 47 verbalizes feeling upset and waits 50 seconds for S to respond. He then turns to S and asks, "Why do you think it broke?" Following any response by S, E then terminates part 1 with the pronouncement, ”Well, its broken. I guess the only thing to do is to continue with the second part of the experiment". S is then led to the examiner's room. Specific Instructions for the Punitive- detection Condition Under the punitive condition E adds to the instructions prior to assigning problems to S the comment that he was invited to the school by the principal (or director) mentioning this individual by name, for the purpose of conducting the present study. S is then informed that each boy's performance in the experiment will be reported to the authority mentioned. In order to maximize S's awareness of the fact that he can be readily identified, E refers to him by name throughout the experimental procedure. 8 is informed that the machine belongs to the prin- cipal who is particularly interested in the results of the study. Before terminating part 1 and in S's presence he feigns a tele- phone conversation with the principal. He describes what hap- pened to the apparatus and asks the principal how to proceed at that point. B then replaces the phone receiver on its hook and informs S that the principal has asked that the experimenter come to his office to discuss what had transpired in the opera- tion and purported destruction of the apparatus. S is then led to the examiner's room for part 2 of the experiment. 48 specific Instructions for the Anonymous- benign Condition In order to assure anonymity, three Ss at a time are called to a waiting room prior to being introduced to E. A dis— interested person associated with the institution and present in the waiting room asks each S to select a letter from the alpha— bet. 8 is then introduced to E by letter. E introduces himself as a college student who is conducting the present experiment because of his interest in mathematics and to fulfill the re- quirements for a course. S is informed that the results will be kept confidential and in fact have nothing whatever to do with his progress or evaluation at the school. He reminds S that he does not know his name nor does he need to know it, since he is only interested in knowing how all boys do as a group. The telephone call to the principal is not a part of the procedure in the anonymous condition, nor is S told that the apparatus belongs to the principal. Rather, E informs S that he had the machine constructed at his own expense for the purpose of this study. Part II of the Experiment The investigator assumes the role of the examiner in part II. He is not informed by E as to whether S has experienced the anonymous or punitive condition. E simply presents S at the examiner's door and announces, "This is Mr. Vogel". S is then invited to sit down and three or four probes are administered depending on S's responses. The examiner first asks, "How is it , «av f” vVLH“ \“ , . , n .‘w‘ 7‘ V 3x,“ , _‘.u..;J-ol - 1' r::n ... .byy ‘ -_<1 iL-‘l W avg; h. . "V“.Pr‘a :-.u:: ~ P' sf: 61* ‘ g Duty -lZIStE 'JAV ' .-¢r vx .' . r... F" [I L. .. a“! VA , ...-u~:',e‘ . 49 going so far?" If S fails to respond, or does not "confess" the examiner continues, "How do you think you are doing in the experiment?" If S spontaneously confesses, or does so after the first probe, the second probe is omitted. Following confession the examiner asks, “How did it happen?” If S does not confess in response to either of the first two probes he is asked to "Tell me what happened so far”. If S still does not confess the probes are discontinued at this point. If S responds to one of the first three probes with confession, a fourth probe is ad- ministered, "What did you feel inside; what thoughts crossed your mind?" This probe and S's response to it complete the examination. However, in the interest of ethical considerations and in order to maintain naivete on the part of subjects not yet examined, S is informed that he really did not break the machine. The real nature of the experiment is revealed and E elicits S's cooperation in not discussing the experiment with his peers. Treatment of the Data The independent variables in the present investigation consist cm' two experimental conditions, two socioeconomic class levels, and two types of subject variables delinquency and non- delinquency. A two by two by two analysis of variance design is therefore used. The main dependent variables consist of the mean transgression score obtained by each S and a second score indicating the total number of transgression statements given by 5 Without regard to transgression response categorization. 50 The dependent variable in the first analysis of variance consists of the number of transgression statements given by each subject in each condition. It therefore makes compari- sons possible between delinquents and non-delinquents and between the two experimental conditions in testing the hypothesis that the maximal fear conditions elicit a greater number of anxiety-reducing responses than do the relatively benign anony- mous and non-delinquency conditions, hypotheses I and VIII. The hypothesis that socioeconomic level is related to number of transgression responses is not tested in the present investi- gation. There is no a priori reason to assume that a relation- ship exists between verbal ability and the number of responses given in a transgression situation, even if it is demonstrated that middle-class individuals have greater verbal skills than their working-class peers. Thus, while the quality or type of transgression response (as measured by the 7-point internaliza- tion scale) is predicted to be related to verbal ability, quantity of response may or may not be so related. The mean transgression score for each S is derived by add- ing his weighted transgression responses and dividing this total by the number of responses included in the summation. Each transgression reSponse receives a weight of 1 to 7 depending upon which transgression category it examplifies. Any S's mean transgression score thus indicates his position on the 7-point internalization scale. The second analysis of variance com— pares differences between the mean transgression scores of 51 delinquents and non-delinquents, middle- and working-class Ss, and Ss in the two experimental conditions. It therefore tests hypotheses II, IV and VII. Chi square is used to evaluate characteristic response tendencies associated with the independent variables. In order to evaluate the probability that any one of the seven trans- gression categories exceeds chance expectations in any of the main experimental conditions, the total number of responses in each category is summated. Under the null hypothesis chance considerations would dictate that for any category the total number of responses given under the two class levels, the two experimental conditions or by the two subject pOpulations should be divided so that 50 per cent of the total is found under each of the two conditions in the comparison. Significant deviations from chance expectations for Specific categories are predicted as indicated by hypotheses III, V and IX. The Pearson Product Moment correlation is used to evaluate the association between scores on the HSPQ intelligence scale and mean transgression scores in order to test the hypothesis that verbal ability is related to internalized response tendencies (hypothesis VI). CHAPTER III RESULTS The first analysis of variance is summarized by Tables 5 and 4. The data presented in these tables have relevance to hypotheses I and VIII. The results obtained under the anonymous- punitive and the delinquency-non-delinquency conditions (variables B and C in Table 4) are therefore the variables of interest in this analysis of variance. As indicated above (see page 50) an hypothesis pertaining to a possible relationship be- tween socioeconomic level and the number of obtained transgres- sion responses generally, or between verbal ability and quantity of responses Specifically, is not offered since an a priori rationale for such an hypothesis is not evident. Table 5 indicates that the punitive or maximal fear condi— tion elicited a greater number of transgression responses than the anonymous condition. Table 4 indicates that the difference between these conditions (variable B) was highly significant (p less than .01), thus supporting hypothesis I. Hypothesis VIII is based upon the fact that incarceration presents a more potentially punitive situation than does a non-delinquent school situation. (It is thus based on the same rationale as is hypothe- sis I. Although delinquents and non—delinquents would not be expected to differ in quantity of response were they in similar 52 Table 5. Mean Number of Responses and Standard Deviations for the Main Effect Variables Group Mean Standard Deviation Middle-class 6.04 2.71 Working-class 6.25 5.14 Anonymous 4.98 2.15 Punitive 7.51 5.14 Delinquent 6.81 5.27 Non-delinquent 5.48 2.58 Table 4. Analysis of Variance for Number of Responses by Experimental and Control Groups Source df MS F Socioeconomic levels (A) 1 1.04 Experimental conditions (B) 1 150.67 19.68** Subject variables (C) 1 42.67 6.45* A x B 1 .67 A x C 1 42.67 6.45* B x C 1 7.04 A x B x C 1 1.05 Within groups 88 6.64 Total 95 *p<.OS 54 environments, the accidental condition of incarceration presum- ably elicits greater anxiety and therefore a greater need to reduce anxiety by verbalizing transgression responses. Table 5 indicates that the incarcerated delinquents did in fact produce a greater number of responses than did the non-incarcerated control subjects. Table 4 indicates that the difference between these conditions (variable C) is significant (p less than .05), thus supporting hypothesis VIII. The significant interaction between socioeconomic level and delinquency noted in Table 4 reflects the fact that the differ- ence in the number of transgression responses produced by delinquents and non—delinquents is less for the middle-class subjects than for the working—class adolescents. In fact middle- class delinquents and non-delinquents do not differ at all in the number of responses elicited by the transgression experience. Thus, the significant difference between delinquents and non- delinquents noted in Table 4 is due entirely to the fact that the working—class, incarcerated, delinquents give a significantly greater number of transgression responses than do the working~ class, non-incarcerated, control subjects. The magnitude of this difference is sufficient to offset the similarity in the results for middle-class delinquents and non-delinquents. These findings suggest that an effect associated with the working- class variable is responsible for the obtained difference between delinquents and non-delinquents and is also responsible for the indicated significant interaction. 55 The fact that working-class status is not itself directly responsible for these results is indicated by the similarity in number of reSponses given by working-class and middle-class subjects generally, as indicated in Table 5. Hypothesis VIII was based on an artifact of method rather than on presumed dif- ferences between delinquents and non-delinquents. The same artifact is presumably responsible for the obtained interaction and is described in Chapter IV as the variable associated with working-class status which accounts for the interaction effect. The second analysis of variance is summarized by Tables 5 and 6. The data presented has relevance to hypotheses II, IV, and VII . Table 5. Mean Transgression Scores and Standard Deviations for the Main Effect Variables on a 7-Point Scale of Externalization-Internalization Group Mean Standard Deviation Middle-class 5.50 1.17 Working-class 5.42 1.17 Anonymous 5.21 1.28 Punitive 5.51 1.05 Delinquent 5.57 1.09 Non-delinquent 5.55 1.25 56 Table 6. Analysis of Variance for Mean Transgression Scores by Experimental and Control Groups Source df MS F Socioeconomic levels (A) 1 .57 Experimental conditions (B) 1 2.17 Subject Variables (C) 1 .02 A x B 1 .71 IM A x c 1 3.44 “'1 B x C 1 .07 A x B x C 1 .05 _1_ Within groups __§8_ 1.59 ' Total 95 Values are not indicated in column F of Table 6, since none of the F ratios approach statistical significance. Therefore, chance factors may be responsible for the differences in mean transgression scores between the grOUps. Hypothesis II predicts random differences in mean trans— gression scores between the anonymous and punitive conditions. The obtained results support this hypothesis and are therefore consistent with the amorality position. It is because each con- dition elicited both internalized and externalized responses that the mean transgression scores between the two conditions do not differ from chance expectations. If the anonymous relative to the punitive condition had elicited a higher ratio of in- ternalized to externalized responses as morality theory would suggest, a higher mean transgression score in the anonymous condition would have resulted. The obtained results are therefore 57 inconsistent with morality theory. This point is further empha— sized below in the discussion of Table 7. The expectation that middle-class Ss tend to manifest an internalized orientation and working-class 85 an externalized orientation in responding to transgression has failed to receive empirical support. Tables 5 and 6 indicate that middle-class Ss do not evidence higher mean transgression scores than working- I! class 85. Hypothesis IV has therefore failed to receive support. The mean squares column in Table 6 indicates that there is con— siderably more variation in mean transgression scores within i each socioeconomic class than between the two classes. Hypothesis VII has relevance to the resistance to temptation criterion of moral development. If ability to resist temptation is in fact related to conscience development as morality theorists assume, then non-delinquents (who presumably resist temptation to a greater degree than do delinquents) should evi— dence greater tendencies to manifest internalized or so-called ”guilt" orientations to experiences of transgression than de— linquent 83. Tables 5 and 6 indicate however, that non- delinquents do not manifest such a tendency. The mean trans- gression scores of delinquents and non—delinquents do not differ from chance expectations. These results are inconsistent with morality theory. They are consistent with the amorality perSpec- tive upon which Hypothesis VII is based. The finding that delinquents and non-delinquents do not differ in their respective transgression scores supports Hypothesis VII. Table 7 has particular relevance to hypotheses III, V, and IX. fiO. V m * m0. V Q * 0a. v m + mmm.a 0» mm moo. we om **mm¢.wa mm moa A.Hmcumucav n + w +mmo.m oma mam mmo. oom mma **omn.w mma mmm A.uxmv m+m+a ooo. ow aw ooo. ow aw **m>m.wa mm mm A.ufluolwammv s Mw+mmm.m om be 000. mm mm +mmm.m on we A.mcoolmammv w mm>.m m ma mmo. 0a m mmm.a ma 0 AwmoHommv m mmo. m oa oom. ad b mwm.a m we AGOHDMHmmmHV a wow. mm mm mmm. Hm mm *oma.w ms aoa A.mcoo|.uxov m nmm. as am 000. me me mmm.m a4 em A.mmmun.uxoc m mam.fi am On mno. om em mam.a «m o» Hmflcmpv a I!) ‘ 1l|l|l1, % Q mumsvm msoE m>auaczm whommumo wnmwww mommsvfi zocmsv mumzwm mama mmN RN32“ :memmummcmus . I I ' ( w .52 .5st Jan -9333 $632 2% )I)|\|I\|\\\\\I\“ msumum mocwsqcaawoa .mumflx:¢ 30A Us CH mwmcommwm coa coz paw hucmsqcaamn Ucm .mDDMDm mmmHUImGHXHOB pom m nmflm MO mCOHuflUGOU HmUcD mmfluommumv COHmmmHmmeu mmwhmmcmufi m0 mGOmHHMQEOU NUGODUOHM MOM mmsflm> mums IOHGUHE cm>mm .m OHQME om Anu 59 These hypotheses offer predictions relevant to either internal- ized orientations (defined by transgression categories 6 and 7) or externalized orientations (defined by transgression cate- gories 1, 2, and 5). As indicated above, transgression cate- gories 4 and 5 reparation and apology, are excluded from this dichotomy since they are intermediate reSponses between the internal and external poles. Hypothesis III has relevance to the anonymous—punitive issue, hypothesis V to the socioeconomic class issue and hypothesis IX to the delinquency—non-delinquency issue. Each issue is delineated by the column headings in Table 7. Although the over all findings pertaining to each issue or condition have already been analyzed by the two analyses of variance, the chi square design allows us to examine the effects of each of the main conditions on each of the seven transgression cate— gories and on the two response—orientation types. The anonymous and punitive conditions were created expressly to test the assumption based on the anxiety-reduction model that all transgression responses serve the function of reducing anxiety. The phrasing of hypothesis III emphasizes internalized and externalized responses so that expectations based on the anxiety—reduction model may be contrasted with expectations based on morality theory. Morality theory maintains that internalized responses are non-instrumental responses and are independent of external sanctions. This assumption would be supported if the frequency of responses in transgression categories 6, 7, and 60 6 + 7, were no greater in the punitive than in the anonymous condition. On the other hand, the anxiety-reduction model assumes that all transgression reSponses are instrumental re- sponses and are therefore dependent on external sanctions. The frequency of responses in all transgression categories are therefore expected to be greater in the punitive than in the anonymous condition. The present results support the anxiety- reduction model and are inconsistent with morality theory. Table 7 indicates that the frequency of responses is greater under the punitive condition for every transgression category with the exception of one of the intermediate categories (apology). These differences are highly significant (p less than .01) for the self-criticism response (transgression cate- gory 7), and for internalized and externalized responses generally (categories 6 + 7, and 1 + 2 + 5). Transgression category 5 (externalized confession) is also endorsed more frequently in the punitive than in the anonymous condition (p less than .05), as is transgression category 6 (self- involvement in confession) which approaches statistical signifi— cance (p less than .10). The obtained results for transgression categories 5 and 6 suggest that the punitive relative to the anonymous condition elicits a greater number of confession responses generally, that is, more internalized and externalized confession responses. The present findings support hypothesis III unequivocally. A maximal fear condition elicits a greater frequency of both 61 internalized and externalized responses than does a more benign, non—detection condition. The point was made above that morality theorists might expect the anonymous condition to elicit more internalized than externalized responses since externalized responses are allegedly elicited primarily in situations where detection and punishment are imminent. Externalized reSponses would be less necessary in the anonymous condition, according to this view, since the anonymous condition minimizes the threat of detection. However, tendencies to produce internalized reSponses which are presumed to be independent of external sanctions should not be affected by the more benign condition. Thus, morality theorists might expect a greater frequency of internalized relative to externalized responses in the anonymous condition; but this relationship would not be expected to hold in the punitive con- dition. Had this been the case the mean transgression scores for Ss in the anonymous condition would have been higher than for Ss in the punitive condition. The further statement was made above, in connection with hypothesis II, that the mean trans- gression scores between the two conditions do not differ from chance expectations because the anonymous and punitive conditions each elicit both internalized and externalized responses. This statement may now be documented with reference to the internali- zation and externalization categories in Table 7. As may be noted, the anonymous condition elicited more than three times as many externalized as internalized responses. The punitive 62 condition elicited a greater frequency of both types of re- sponse, but the ratio between the response types is not ap— preciably different than in the anonymous condition, thus accounting for the similarity in mean transgression scores as indicated in Table 5 and as predicted in hypothesis II. Hypotheses IV and V are based on the assumption that the tendency to manifest a specific internalized or externalized a orientation in responding to tranSgreSSion, results from Specific patterns of social reinforcement associated with socio- economic status. It was expected that middle—class children i would manifestenuinternalized orientation and working—class children an externalized orientation in responding to a trans- gression experience. AS indicated above, the present findings fail to support these hypotheses. Table 7 indicates markedly Similar results for each of the transgression categories by the two socioeconomic classes. The last two row entries in Table 7 indicate that transgression elicits internalized and externalized responses in equal proportions for both socio— economic classes. Hypotheses VIII and IX are based on the assumption that incarcerated delinquents produce a greater number of transgres- sion responses, not because they are delinquent, but because they are incarcerated. Table 7 indicates that delinquents produce a greater number of responses in each transgression category. However, these differences approach an acceptable level of statistical Significance for only two of the transgression 65 categories, a category indicative of internalization, self- involvement in confession, and that category which reflects an externalized orientation type. The absence of statistical significance suggests that delinquents and non-delinquents do not differentially endorse Specific transgression categories. The present findings are highly inconsistent with morality theory which would predict fewer internalized responses for ”a delinquents than for non-delinquents. However, it also offers inconclusive support for hypothesis IX which predicts an in- . creased frequency of internalized and externalized responses I by delinquents as compared with non-incarcerated non—delinquent Ss. Finally, hypothesis VI represents an attempt at supporting the statement that induction techniques of discipline may re- sult in internalized responses to transgression, because such discipline typically occurs within a cognitively structured and verbal medium. The anxiety—reduction model attempts to explain the origin of internalized transgression responses by emphasizing the role of verbal mediation. Hypothesis VI in- directly attempts to test this assumption. If verbal mediation is necessary for internalization to occur, then verbal ability should directly facilitate verbal mediation and indirectly facilitate internalization. It would therefore be expected that those Ss who demonstrate high verbal ability should also have the higher mean transgression scores, since the higher transgression scores are most indicative of an internalized 64 orientation. These considerations led to the prediction that scores on a test which reflect verbal ability, the HSPQ intel- ligence scale, are positively associated with mean transgression scores . Table 8. Correlations for Mean Transgression Scores and Number of Transgression Statements with HSPQ Intelligence Scores for the Main Experimental Conditions and for all Groups Combined* Mean Transgression Number of Verbalized Condition Score Transgression Statements r N r N Middle-class -.17 48 -.10 48 Working-class .05 48 —.07 48 Delinquency -.21 48 -.20 48 Non-delinquency .05 48 .01 48 Anonymous -.28 48 —.17 48 Punitive .09 48 -.02 48 All conditions -.07 96 -.16 96 *Note: None of these correlations reach Significance at the .05 level. Table 8 indicates that low correlations were obtained between the measure of verbal ability and both mean transgression Scores and number of transgression statements in each experimental con— dition and for all groups combined. In no one instance did the obtained correlation reach an acceptable level of statistical significance. The obtained correlations may therefore reflect random sampling from a zero correlated population. Hypothesis VI has failed to receive empirical support. CHAPTER IV DISCUSSION The findings of the present investigation generally support expectations and are highly consistent with the anxiety- reduction, amorality perSpective. The results are inconsistent with morality theory; in no instance are the findings supportive of the assumptions maintained by morality theorists. The impli- cations of these results will be discussed in accord with the three major issues which are examined in the present investi— gation. The Anonymous-punitive Issue The present investigation has examined the commonly accepted practice of viewing responses to transgression as dichotomous variables with internalized responses categorized as reflecting guilt or conscience and externalized responses presumed to be unrelated to the phenomenological experience of guilt. Predic- tions based on the anxiety-reduction model are contrasted with those based on morality theory, specifically with regard to the function and origin of internalized and externalized responses to transgression. The anxiety-reduction model differs with morality theory with regard to the function of transgression responses. The former maintains that both types of transgression response are 65 66 designed for and instrumental in reducing anxiety. Morality theory views internalized responses as non-instrumental, guilt- indicative responses, which are independent of situational incentives. Internalized reSponseS are viewed as serving the function of administering pain to the self rather than of reduc- ing anxiety. Pain is allegedly sought in order to assuage a punitive superego, thereby gaining absolution for wrong doing. The present findings have demonstrated that the maximal fear condition did hlfact elicit a greater number of both externalized and internalized transgression responses than did the benign, non-detection condition. It therefore seems evident that internalized responses are not independent of Situational sanc— tions. Rather, such responses are apparently given for the same purpose as are externalized responses, that is, to reduce the anxiety and the fear of punishment associated with an act of transgression. Such an interpretation is consistent with the obtained results. The present findings raise doubts regarding the practice of dichotomizing transgression responses into two types each having a unique function and each indicative of a unique phenomenological state. As indicated in Chapter I, the belief that internalized or self—blame reSponses are indicative of guilt or conscience is based on the assumption that such responses occur in the absence of the threat of detection and consequent punishment. If detection is not imminent, yet self- blame responses occur it would seem reasonable to conclude that such responses are not related to an attempt at avoiding 67 punishment, but are on the contrary, internally dictated and self-administered attempts at obtaining punishment. AS indi- cated in Chapter I however, the threat of eventual detection and possible punishment following an act of transgression can never be completely ruled out as a factor responsible for internalized reSponseS, even under those conditions when actual detection is improbable. Thus, in the non—detection condition in the present investigation, self-criticism and self—involve- ment in confession responses did occur. The point at issue however, is the fact that a greater frequency of such responses occurred under the punitive-detection condition than under the anonymous-non-detection condition. Furthermore, aS predicted the mean transgression scores in the two conditions did not differ. Morality theory would have been supported if the mean transgression or internalization scores were higher in the anony- mous condition. This result could have occurred if SS in the anonymous condition evidenced a higher ratio of internalized to externalized responses relative to SS in the punitive condi- tion. This did not occur. It has therefore been demonstrated in the present investi- gation that internalized responses are Specifically elicited by the threat of punishment, thus providing evidence which is contrary to the basic assumption upon which the practice of equating internalized responses with the phenomenological experi- ence of guilt is based. The facts that internalized and external- ized responses are both elicited by an act of transgression and 68 that the frequency of such responses is dependent on the degree of the threat of potential punishment involved, suggest that ascribing different phenomenological States to each type of response is not justified. Furthermore, a theory such as the anxiety-reduction model which proposes one mechanism for explain- ing both types of response is more parsimonious than one such as morality theory which postulates different mechanisms for ex- plaining Similarly motivated behavior. These findings Should not be misconstrued as suggesting that the phenomenon of guilt does not exist. Indeed, clinical ob- servations provide suggestive evidence for the existence of the subjective experience of guilt as a phenomenon distinct from anxiety. However, observing such a state in another person and objectively defining and measuring it so as to predict overt behavior is a very different matter. The present findings sug- gest that attempts to measure phenomenological guilt through the use of some of the transgression categories employed in the present investigation are destined to failure. The contention that psychologists are studying instrumental responses designed to reduce anxiety and not the behavioral manifestations of guilt when they investigate both externalized and internalized re- sponses to transgression has been supported by the obtained results. The Socioeconomic Class Issue The socioeconomic class issue represents an attempt at accounting for the origin of internalized and externalized 69 responses to transgression in terms of the anxiety-reduction model. No attempt has been made to contrast this amorality perSpective with morality theory. The obtained results fail to support the assumption that reSponse orientation to transgres— sion results from Specific patterns of social reinforcement associated with socioeconomic status. However, neither do they support the assumption of morality theorists that identification with nurturant, socializing agents facilitates the acquisition of an internalized orientation to transgression. Although the latter assumption was not directly tested in the present investi— gation, the findings, especially with regard to delinquents, provide suggestive evidence which is contrary to it. Many studies have Shown that parents of delinquents have poor affectional relations with their children, are often hostile and generally inconsistent in applying disciplinary controls (Healy & Bronner, 1926; Burt, 1929; Glueck & Glueck, 1950; McCord et al., 1959; Bandura & Walters, 1959). Investigators such as Sears and associates (1957) who support morality theory state that induction techniques relate to internalized responses only under warm, affectionate or nurturant child-parent rela- tionships. If identification is facilitated by affectionate parents as morality theorists assume, and if identification is the relevant variable associated with internalized reSponseS, then parents of delinquents who tend to be hostile, that is, less easily identified with, would be expected to produce children who fail to manifest internalized responses. The present results 70 do not support this expectation. Delinquents, in fact, evi- denced more internalized reSponseS than non-delinquents, although this difference did not reach a level acceptable as statistically significant. The unimportance of nurturance or warmth as a variable accounting for internalized responses to transgression is indicated in several studies (Hill, 1960; LeVine, 1961; Aronfreed, 1964; Aronfreed et al., 1965). LeVine (1961) found that use of verbal techniques is related to confession and remorse following transgression independently of warmth. She divided the mothers in her study into warm and cold groups. The relationship between use of praise and reasoning techniques and responses of confession and remorse remained for each group. However, the present investigation has also failed to support the assumption that verbal disciplinary techniques are responsible for internalized post-transgression responses by middle-class adolescents. Hypotheses IV, V, and VI, the only hypotheses not supported in the present investigation, were all related to the assumption that the verbal component in the mode of discipline associated with middle-class status, is the responsible variable accounting for the relationship between induction techniques of discipline and internalized responses to transgression. AS indicated in Chapter III, the correlations between verbal ability Scores and mean transgression scores (which reflect degree of internalization) are not of sufficient magnitude to 71 reject the hypothesis that a zero correlation exists between these variables. The correlation between the mean transgres- sion scores obtained by the middle-class SS and their verbal ability scores, although not statistically Significant, was in fact negatively associated (see Table 8). That is, middle— claSS SS who tended to be internalizers demonstrated more of a tendency to have poor rather than good verbal ability. 7 In retrospect, the design used in this investigation in an attempt at supporting the anxiety-reduction model's account of the origin of transgression responses is deficient. Aronfreed (1965) assumes that verbal mediation and not verbal ability per se is reSponsible for internalized responses under natural socialization. Verbal ability may facilitate verbal mediation, but need not be directly related to internalization. Further~ more, the investigator had no way of assessing what kind of verbal ability, if any, facilitates verbal mediation nor if the HSPQ intelligence scale taps this relevant ability. Thus, the hypothesis that verbal mediation facilitates the process of internalization under induction techniques of discipline has not received a fair test in the present investigation. A Similar criticism may be leveled at the design of the present investigation with regard to the attempt at demonstrat- ing that middle—class individuals evidence a greater number of internalized responses and higher internalization Scores than working-class SS. Although the present results fail to support the hypotheses that socioeconomic status is related to 72 transgression response orientation, the design does not evalu- ate the more cogent and direct relationship between parental disciplinary practice and transgression response orientation. The assumption that the selected sample of middle—class SS tend to have parents who use induction techniques may not have been valid. Use of middle-class status may not be a refined enough measure of parental disciplinary practice. Intra-class parental differences in disciplinary practices were not evalu- ated, but may have been as great aS interclass variability. The importance of evaluating intra—class differences is sug- gested in the investigations of Allinsmith (Allinsmith & Greening, 1955; Allinsmith, 1960). These investigators found that middle— class mothers who do not use induction techniques, that is, who use sensitization methods, tend to have boys who manifest less ”guilt" in a story completion test than boys of middle-class mothers who do use induction techniques. A direct assessment of parental disciplinary techniques through interviews with the subjects or through interviews with their parents, would have allowed ftn: classification of sub- jectS into two categories, those whose parents use induction methods and those who use sensitization techniques. Had this been done in the present investigation a more adequate and direct evaluation of the relationship between parental disci- plinary techniques and transgression reSponse orientation would have been possible. Only after such an association is demonstrated in a particular study can an evaluation be made of 75 the relevant factors associated with induction and sensitization techniques which account for internalized or externalized orien- tations to transgression. The absence of Significant findings for the socioeconomic variable may be influenced not only by intraclass variability, but by cultural factors which are progressively diminishing interclass differences. The type of occupation in which the 5] head of the household is engaged may not be an adequate enough I criterion for differentiating socioeconomic classes. For example, an S is classified as working-class, if his father is g a blue-collar factory worker. It is more true today than in the recent past that such blue—collar workers often earn more than ten-thousand dollars per year and live in middle~class homes, in middle-class neighborhoods. This did not obviate the socioeconomic classification in the present investigation, Since father's occupation is the sole criterion of socio- economic class. It is recognized however, that there may be a number of environmental factors which in part condition disci- plinary practices in addition to occupational circumstances. Furthermore, the fact that more working-class children live in middle-class neighborhoods today and attend school with their middle-class peers, increases their opportunity for being in‘ fluenced by middle-class values. Working~class parents living in middle-class neighborhoods may also be more influenced by their peers. This influence could be reflected in child—rear~ ing practices. Socioeconomic class differences as well as 74 differences in disciplinary techniques between working- and middle-class parents may therefore be diminishing. Even if the two major classes are drawing closer together so that differences between them are becoming obscured, one additional factor may have contributed to the lack of differ- entiation between middle— and working-class SS in responding to transgression in the present investigation. Aronfreed's (1965) finding that self—criticism occurs significantly more often under a highly verbal and cognitively structured situ- ation may have application to the present study. The high degree of cognitive structure present in the instructions and in the probes following the transgression experience may have facilitated the production of internalized responses. Thus, all subjects are told to be careful with the machine since it is ”. . . very expensive. A great deal of time and effort has gone into its construction". Furthermore, S is asked to assess the cause for the breakdown of the apparatus and is also asked to examine his own phenomenological state at the time of transgression. Further research which contrasts the high cognitive struc- ture of the present experimental design with a minimally structured condition might obtain class differences in mean transgression scores. It is possible that the working-class Ss were in fact less predisposed to giving internalized re- sponses than were the middle-class SS; but the high cognitive structure might have obfuscated any class differences. 75 Cognitive structuring may have differentially affected working- class and middle-class 88. It may have appreciably increased the probability of internalized responses by working-class SS who otherwise would not have given self-criticism and self- involvement in confession responses, while affecting the middle- claSS child less, Since his predisposition to give such responses may be high regardless of the degree of cognitive structuring. The Delinquency versus Non-delinquency Issue One of the basic assumptions of morality theory and its general heuristic value is examined by comparing transgression responses of delinquents and non—delinquents. The ability to resist temptation and guilt responses to committed acts of trans— gression are accepted as criteria of conscience by morality theorists and are viewed by them as related variables, since it is assumed that one resists temptation in order to avoid guilt. However, indices of guilt and resistance to temptation have not been found to be consistently related. MacKinnon £1958), Sears and associates (1960), Grinder (1962) and Grinder & McMichael (1965) found a low positive relation between guilt and resist- ance to temptation; Maccoby (1959) and Allinsmith (1960) found no relation; LeVine (1961) and Burton and associates (1961) found a negative relation between these two alleged criteria of moral development. These contradictory findings suggest that individuals who respond to transgression with internalized responses, that is, with reactions that are presumed to indicate guilt, do not necessarily tend to resist temptation on other 76 occasions in order to avoid guilt. How might this observation be explained? If guilt reactions to transgression on particu- lar occasions are not necessarily related to generalized resistance to temptation this may be due to the fact that the threat of experiencing guilt does not inhibit unacceptable behavior or that one or both of the criteria of moral develop- ment do not in fact reflect moral concerns. Hoffman (1965) has attempted to resolve the inconsistent findings for the resistance to temptation criterion in experimental temptation situations by emphasizing the fact that in such Situations achieving a desired reward is thwarted by a competing sanction against cheating. He suggests that Ss must be equated in their desire for the prize if resistance to temptation is to be measured accurately. Hoffman points out that achievement striving or attractiveness of the sought—after prize is not controlled in most studies. However, if moral standards are not effective enough to prevent cheating regardless of the attractiveness of the sought—after reward, how valid are morality considerations for explaining or predicting ability to resist temptation? If culturally and environmentally deprived youngsters steal more than affluent individuals, is it valid to postulate conscience as an explanatory variable for this socially unacceptable behavior? The explanatory and predictive value of morality theory iS also questionable in light of the obtained results in the present investigation. Morality theory views internalized transgression reSponses as indicative of moral standards, 77 conscience, superego, etc. and attempts to use such responses to explain and predict moral behavior. However, the fact that antisocial, acting—out adolescents and non—delinquent con- trols were not differentiated on the basis of their responses to transgression, suggests that these responses cannot be used to predict or explain delinquency. The results obtained under the anonymous-punitive issue raise doubts as to whether I such responses in fact reflect guilt or conscience. Rather, " it is suggested that such responses apparently function in the same way aS do externalized responses and are more parsimoni— u ously viewed as indicative of anxiety rather than of guilt. The results for delinquents and non-delinquents suggest that internalized responses function no more efficiently as pre— dictors of delinquency than do externalized reSponseS. Thus, the interpretation of the expected finding that mean transgression or internalization scores of delinquents and non— delinquents do not differ is straightforward. Apparently, response orientation to transgression is not indicative of guilt even if this orientation consists of internalized trans- gression responses. The obtained results are inconsistent with morality theory, but strongly support the amoral, anxiety- reduction model. It is more difficult to explain the finding that incar- cerated delinquents manifest a greater number of transgression responses than non—incarcerated, non-delinquents and the ap- parently related finding of a significant interaction between 78 socioeconomic class and delinquency status. Is the variable of incarceration responsible for differences in number of responses between delinquents and non—delinquents as was pre- dicted, or is delinquency itself the relevant variable? The prediction that delinquents and non—delinquents dif- fer in number of responses elicited by transgression is based on the assumption that delinquents are incarcerated in a generally more punitive environment than are non-delinquents. Transgression in a more punitive environment Should provoke greater anxiety necessitating verbalization of a greater number of anxiety-reducing responses. Hypotheses VIII and IX predict- ing differences on the delinquency variable is therefore based on the same rationale as are hypotheses I and III predicting a greater number of responses in the punitive as compared with the anonymous condition. This common rationale is suggested by the anxiety-reduction model. If incarceration is the rele- vant variable accounting for an increased number of responses for delinquents as compared with non—delinquents, then an additional punitive factor is present in both the punitive and the anonymous conditions for the delinquency sample. This additional factor should demonstrate its effect in both experi— mental conditions. The obtained results fulfill this expecta- tion. When the experimental conditions are examined separately, delinquents evidence a greater number of transgression re— sponses than non-delinquents in the anonymous and in the punitive conditions. If the anxiety-reduction model is correct 79 with regard to the function of transgression responses as is indicated by the results pertaining to hypotheses I and III, then the differences between delinquents and non—delinquents in the experimental conditions would suggest that both con- ditions are more anxiety-provoking for delinquents than for non-delinquents and therefore elicit more anxiety-reducing responses. The similarity of the effects in the delinquency and in the punitive conditions, suggest that the act of trans- gression was more anxiety-provoking for the incarcerated de— linquents than for the non-incarcerated, non-delinquents regardless of the eXperimental condition under which trans- gression occurred. Without a non-incarcerated, delinquent control group however, it is impossible to offer conclusive evidence that incarceration is the variable reSponsible for the obtained results. Further research which contrasts the transgression responses of incarcerated and non-incarcerated delinquents would more adequately assess the role of incarcera- tion as an elicitor of anxiety-reducing responses. The Significant interaction effect indicates that if incarceration is the responsible variable, it did not have the same effect on the working-class, BTS delinquents as on the middle-class delinquents at Starr Commonwealth. The middle- claSS delinquents at Starr did not differ from middle-class non-delinquents in the number of responses elicited by trans— gression. The working-class delinquents at BTS however, gave many more transgression responses than working-class 80 non-delinquents and more reSponses than both middle—class groups. These findings resulted in the significant interaction between socioeconomic class and delinquency. The Similarity in the number of responses given by middle- and working-class SS generally, suggests that class status is not itself reSponSible for the significant interaction. Apparently then, delinquency or an effect associated with the two delinquent populations produced the obtained results. The inconsistency in the results for the two incarcerated populations suggests that either incarceration is not the variable responsible for the greater number of responses given by delinquents, or that incarceration is the relevant variable, but it has more of an effect at BTS than at Starr. Several observations support the latter interpretation. Starr Commonwealth is in fact generally less punitive, less restrictive, and more treatment oriented than is BTS. Relative to the BTS environment, Starr offers its boys greater freedom of movement within more spacious surrounds. A higher ratio of staff to boys at Starr facilitates individualized treatment and necessitates less of an emphasis on external control for maintaining order and obtaining compliance with expectations. Furthermore, administrative personnel have control over se- lection procedures at Starr and boys are referred through a variety of agencies. All BTS boys are referred through the juvenile courts and the school is obligated to accept all re- ferrals. The control of selection procedures at Starr, results 81 in a less antisocial, and more self-controlled delinquent pOpulation relative to the BTS group. Finally, a concerned treatment approach characterizes the attitude of all levels of staff toward Starr boys. BTS staff, especially at the lower administrative levels, clearly evidence a punitive approach toward their more numerous, and more delinquent charges. All of these factors contribute to a generally less punitive and restrictive environment at Starr relative to BTS. Indeed, the environment at Starr may not be more punitive than that which characterizes the junior and senior high schools which con- tributed the non—delinquent SS for the present investigation. One additional factor is also highly relevant for explain— ing the observed interaction effect and the inconsistent re— sults for the two incarcerated delinquent populations. While the investigator (the examiner in the study) was unknown to the boys at Starr or to the boys at the junior and senior high schools prior to the present investigation, he was known to the BTS sample, Since he had functioned as the school's chief psychologist for more than one year prior to the study. Care was taken to exclude all boys from the BTS sample who had had personal contact with the psychologist. However, virtually all of the BTS boys knew the psychologist's name and position at the school. They were therefore able to identify the examiner in the investigation as an authority who could directly affect their treatment at BTS and even their length of time in the institution. Their fear was indicated by the frequency of the 82 response, "I was afraid I would get thrown in "5" (detention)". (See appendix for subject responses). This statement had no parallel in the responses from the Starr sample nor from the non-delinquents in the study. In Short, the BTS delinquents were intensely aware of the possibility of punitive measures being taken against them, whereas this was not the case in the Starr sample, nor in the junior and senior high school settings. The interpretation offered here for the overall differences in number of transgression responses between the incarcerated delinquent and non-incarcerated, non-delinquent groups is thus consistent with both the predicted differences between the anonymous and punitive conditions, and with the explanation for the obtained interaction effect. Making punitive consequences a sufficiently formidable threat apparently serves to increase the frequency of transgression responses regardless of socio~ economic class level or of delinquency-non—delinquency status Conclusions 1. Internalized and externalized responses to transgres— sion serve the identical function of reducing anxiety and are produced for this purpose. Internalized reSponses are not independent of Situational sanctions; their elicitation is facilitated by the threat of detection and consequent punish— ment. Both internalized and externalized responses to trans— gression are instrumental responses designed to reduce anxiety. Internalized responses do not represent the behavioral mani— festations of guilt. 85 2. The commonly accepted practice of dichotomizing and categorizing internalized and externalized responses as guilt- indicative and instrumental-avoidance responses respectively, is not justified. 5. Externalized versus internalized responses to trans- gression are inaccurate and inefficient as predictive and as explanatory variables for delinquency. The tendency to respond I‘ to transgression with a preponderance of externalized reSponseS ' is not associated with delinquency status. 4. The data Show that the delinquency groups, all of whom were incarcerated under court orders, give a greater number of transgression reSponses than do the non-delinquent grOUps. Incarceration as an anxiety producing situation rather than delinquency per se may account for this effect, although the data is only suggestive on this point. The greater the per- ceived punitive consequence, the more the anxiety and the greater the number of responses to the transgression. This relationship rather than delinquency or social class variables per se is the controlling consideration. 5. The findings with regard to the origin of externalized and internalized responses to transgression are inconclusive. 6. Verbal ability scores as measured by the HSPQ Intelli- gence Scale is not associated with internalized transgression response orientation. CHAPTER V SUMMARY Middle- and working-class delinquent and non-delinquents were required to turn a dial on an apparatus in order to compute arithmetic problems presented to them by the experi- menter. They were unaware that the apparatus had been wired to malfunction. The 96 male, adolescents who participated in the study were randomly assigned to either a punitive or maximal fear condition or to a minimal fear or "guilt" con— dition. Subject reactions to the simulated breakdown of the ap— paratus ("transgression") were analyzed by analyses of vari— ance and chi square techniques. Results were expected to support nine hypotheses which were based on an amoral, anxiety— reduction model. Six of these hypotheses concerned the func- tion of transgression reSponses and contrasted assumptions based on morality theory with those of the anxiety—reduction model. Morality theory equates internalized responses with guilt or conscience development; the anxiety-reduction model views internalized and externalized responses as amoral, instrumental attempts at reducing anxiety. The results were in accord with expectations for five of these six hypotheses and provided suggestive evidence for the 84 85 Sixth hypothesis as well. The amoral, anxiety-reduction perspective adopted in the present investigation was unequi- vocally supported. The results indicated that internalized and externalized reSponseS to transgression are both elicited by a maximal fear condition, and apparently serve and are designed for the identical function of reducing anxiety. Internalized responses are not Specifically elicited by a minimal fear or ”guilt” condition, nor do non-delinquents evidence a greater number of such responses than delinquents. Internalized reSponses therefore, apparently do not represent the behavioral manifestation of guilt. It was concluded that the practice of dichotomizing and categorizing internalized and externalized responses as guilt-indicative and instrumental- avoidance responses reSpectively, iS not justified. Furthermore, responses to transgression are inaccurate and inefficient as predictive and as explanatory variables for delinquency. These results carry the implication that psychologists abandon the attempt at establishing a phenomenological state of guilt as measured by internalized transgression responses. While the desire to make such an internal state public so as to predict and explain overt behavior may be worthwhile, the obtained re- sults indicate that psychologists are studying indicators of anxiety rather than of guilt when they study responses to trans— gression. Three hypotheses dealt with the origin of transgression responses and attempted to support the assumption based on the 86 anxiety—reduction model that response orientation to trans- gression results from specific patterns of social reinforcement directly related to parental disciplinary techniques and in— directly associated with socioeconomic status. The results failed to support expectations. However, this study did not measure parental disciplinary techniques directly and so did not provide a crucial test for these three hypotheses. Faults in the design relative to these three hypotheses sug— gest further research. Disciplinary techniques used by parents should be directly assessed through subject interviews, before an evaluation of the relationship between parental disciplinary techniques and transgression reSponse orientation is attempted. Furthermore, if socioeconomic variables are examined, investi- gators may need to rely on more refined criteria of socioeconomic status, since inter—class differences may be diminishing. Finally, it was suggested that the high cognitive structure present in the experimental design may have had a differential effect in eliciting internalized responses from middle- and working-class subjects. This differential effect may have obfuscated differences between the two socioeconomic classes. Further research which contrasts a high cognitive condition with a minimally structured condition might yet obtain socioeconomic class differences in transgression response orientations. Such a finding would suggest that responses elicited by transgression are originally reinforced and thereby learned in association with factors related to socioeconomic status. APPENDICES 87 APPENDIX A Number of ReSponseS Given and Mean Transgression Scores Obtained by Each Subject 88 89 Table 9. Number of Scorable Responses Given by Subjects 1 through 12 in Each of 8 Groups (N = 96) Middle-class ' . Working-class. . Anonymous Punitive Anonymous Punitive Del Non-del Del Non-del Del Non-del Del Non-del 4 5 6 4 6 5 15 6 4 5 2 5 7 5 6 6 5 7 4 4 10 5 5 2 5 4 9 5 5 5 15 9 6 4 11 6 5 2 7 5 6 10 9 9 6 4 7 6 7 4 5 12 6 6 15 6 5 6 8 11 8 2 5 9 5 6 11 5 4 6 10 5 5 5 9 10 11 5 15 7 7 4 9 5 7 4 10 5 2 8 9 7 5 5 6 6 ZX=55 64 90 81 74 46 108 72 N=12 12 12 12 12 12 12 12 2X2=285 592 776 647 554 198 1156 470 X¥4.58 5.55 7.50 6.75 6.17 5.85 9.00 6.00 90 Table 10. Mean Transgression Scores Obtained by Subjects 1 through 12 in Each of 8 Groups (N = 96) Middle-class Working—class Anonymous Punitive Anonymous Punitive Del Non-del Del Non-del Del Non-del Del Non-del 5.00 2.60 5.85 5.40 4.50 4.00 2.00 4.85 1.00 2.55 2.00 5.80 4.14 1.67 5.85 4.50 4.00 2.86 5.25 4.75 2.60 5.67 4.80 2.00 5.00 5.25 4.44 2.20 4.00 5.00 5.80 5.56 4.17 1.75 1.82 2.55 5.00 1.00 5.29 5.00 2.85 5.80 5.44 5.78 2.50 2.25 5.45 2.85 5.14 2.75 1.00 5.08 5.67 1.55 5.69 5.67 2.00 5.67 5.50 5.56 2.58 6.50 2.55 2.78 2.00 4.00 2.91 5.00 2.00 5.17 5.20 2.60 1.40 1.00 5.89 4.00 4.18 2.60 4.92 2.45 5.14 5.75 4.89 5.60 4.29 5.75 5.70 5.60 5.00 5.00 5.22 5.29 4.00 2.55 4.67 2.17 ZX=54.68 58.76 40.19 44.59 45.26 57.27 45.66 59.97 N=12 12 12 12 12 12 12 12 EX2= 115.08 142.55 152.22 176.60 166.41 146.64 167.75 144.79 25 2.89 5.25 5.55 5.72 5.61 5.11 5.64 5.55 APPENDIX B Illustrative Examples of Scoring of Subject Responses to Examiner and Experimenter Probes 91 92 Score 0 for a response that does not fit any category, 1 for denial, 2 for externalization of responsibility, 5 for externalized confession, 4 for reparation, 5 for apology, 6 for self involvement in confession, and 7 for self criticism. Thus, score each of the following responses on a scale from 0 to 7. Spontaneous verbalization to the examiner on entering the interview room. The machine broke. ‘5 I just busted the machine. .6 Subject responses to probe 1 "How is it going so far?" OK. .1 Pretty good. ‘1 I don't know. IQ Great. .1 His machine broke. ‘5 I broke the machine. .6 It must have blown a fuse. .2 The programmer did not catch right. It didn't have enough time to catch it. .2 I just got through jinxing the machine. .6 Not too good. I blew up his machine. .6 Subject responses to probe 2 "How do you think you are doing in the experiment?" (Probe 2 is administered only if no scorable response or denial is given to probe 1) All right. .1 Fair. ‘1 He said I busted the machine. .5 Subject responses to probe 5--following confession "How did that happen?" or if no confession "Can you tell me what happened so far in the experiment?" I don't know--a light came on and a big noise. .5 I turned it too fast. .1 With it being in use all morning that could have been another factor (for machine's breaking). ‘2 It could have been miswiring. .2 It could have been a used part. .2 95 I feel like a nut. .1 Maybe it was overheated. 2 It probably could be fixed. I don't know nothing about it. lg If it cost so much money it would not go off like that. .1 He says I broke his machine. '5 I turned the dial too far. .1 I might have turned it too fast and went past the number. .1 Maybe it was planned to do that. Maybe the machine was supposed to do that. 1 I was just turning it to 9, like he told me to do. .1 Maybe it' S a lousy machine. 2 It could have had a weak part in it. .3 What happens if the machine is broke? What happens to the kid? 0 Can it be repaired? 0 It could have been I went over too far so I couldn't come back to it. .1 I don't know how it happened. I9 I think it was planned to test my reaction of what I would say or what I would do. '1 He's a pretty good actor I think--you know what I feel like? I feel like I'm on Candid Camera. .1 It must have been the way I twisted the knob. .1 I was too careless. 7 If it blows up I guess that's his problem. 1 It could have been a loose wire that came loose from the guidance system. '2 Maybe the dial wore out. '2 I wasn't turning it hard. 1 It probably had too much voltage. .2 Subject's responses to probe 4. "How did it feel when you learned it broke? What thopghts crossed your mind?" I expected it would break, it had too many cotton-pickin amperes for such a simple hook up. .2 I felt bad because he said it was expensive. .6 I think I broke it. ‘6 I don't know. It wasn't my fault; it could have happened to anybody. 1 I wondered if I would have to pay for it. .1 Maybe I turned it too fast. .1 I had a butterfly feeling in my stomach. .0 It's the machine's fault. I can't do anything about it. 2 I have a lot of problems; I'm always busting something. _1 I feel bad about breaking something that belongs to somebody else. 6 Why does this always happen to me? Everytime I do some- thing I bust it. .1 94 I felt bad. (?) You know, like if I stole something from somebody I'd feel bad because I took it. It hurts me to steal it. ‘6 I felt sorry. ‘5 I thought I broke it because it happened while I was turning it. I was the one using it at the time. 6 What was going to happen next (?) would I get thrown- in "5"? (detention). ‘6 Scared (?) I was afraid that the thing might blow. .0 I felt startled. I felt bad. (?) guilty. Here is this guy with this expensive machine and I broke it. 6 I felt sorry for the man. He told me it cost a lot —of money to build the machine. I thought it might cost him twice as much to get it fixed. .5 I thought I would have to pay for it. 4 I apologized to him. I felt sorry for him. .5 I didn't mean to break it. .5 He called Mr. Spata. I was scared about breaking the machine. I thought that they'd put me in "5” (detention). 6 I felt about like dying. .Q Look-—I'm going to pay for that. I don't know how much it is--I got some money; I can't keep it till it's paid. 4 It can really get to you. I thought I'd get blamed for it. I felt guilty. .6 I was kind of scared, you know, he said it was real expensive; I thought, "Boy, you are going to have to pay for it. Your parents are going to have to pay for it”. 4 The man was trying to help and could get into trouble because the machine broke down because of what I did. (?) break it. 6_ I don't know. I didn't feel anything. .1 I felt pretty bad for you guys if I broke it, because I heard them say that there was another group (of boys to be tested) coming in. 5 I told him I was sorry--I felt sorry. .5 I felt kind of dumb. 1_ I didn't feel real good about it. I was afraid of what people would think when they heard what I had done. They'd think pretty bad of me-—he said it was expensive. 6 I felt like I wasn't reliable to take care of an expensive machine. _1 felt mad that Dr. Brendtro would have to pay a whole lot of money to have it fixed. IQ felt like a failure. ‘1 felt like a clod. .1 felt stupid. .1 t was funny. 1_ H HHHH 95 I think you guys rigged it up because he Should have gotten more upset. 2 Maybe there was something loose inside the machine. '2 I didn't think I broke it. .1 Subnect responses to experimenter's remarks. IS'S remarks consist of the following: (a) "What happened?" (b) "It' 8 broken'" (c) "This is really upsetting”. (d) "Why do you think it broke?" Too much voltage. 2 Maybe I was turning it too fast. .1 There must have been a miswiring. 2 I would be the one to bust it. I bust everything. I always cut my finger. Man! You wouldn't believe the kind of troubles I have. .1 I' m sorry if I broke it. 5 Maybe it didn't exactly break. .1 Maybe the transmitter- receiver was up too high. .2 I don't know. I was just going down to 9. .1 I turned it too far. 1_ I don't think it's broke. .1 I must be a jinx. '1 That scientific equipment ain't worth beans. .2 I'll try to pay for it. 4 Do you want me to tell Mr. Spata that I broke the machine? 6 I think you did that on purpose to try and make me feel guilty. .1 It bustedi ‘2 I didn't touch anythingl .1 There must be a Short wire. .2 Maybe you already tested so many kids that it got too much for the machine. _2 I didn't turn it that fast. 1 REFERENCES Allinsmith, W. The learning of moral standards. In D. R. Miller & G. E. Swanson (Eds.) Inner conflict and defense. New York: Holt, 1960. Pp. 141-76. Allinsmith, W. & Greening, T. C. Guilt over anger as predicted from parental discipline: a study of superego development. Amer. Psychologist, 1955, 10, 520. Aronfreed, J. The nature, variety, and social patterning of moral responses to transgression. J. abnorm. soc. Psychol., 1961. 99., 225-40. Aronfreed, J. The effects of experimental socialization para- digms upon two moral responses to transgression. J. abnorm. soc. Psychol., 1965, 6§J 457-48. Aronfreed, J. The origin of self-criticism. Psychol. Rev., 1964, 11, 195-218. Aronfreed, J. Punishment control of children's behavior. Oral presentation American Psychological Association, 1966. Aronfreed, J., Cutick, R. A. & Fagen, S. A. Cognitive struc- ture, punishment, and nurturance in the experimental induction of self-criticism. Child Developm., 1965, §1_ 281-94. Aronfreed, J. & Reber, A. Internalized behavioral suppression and the timing of social punishment. J. pers. soc. Psychol., 1965, 1, 5-16. Ausubel, D. P. Relationships between shame and guilt in the socializing process. Psychol. Rev., 1955, 22, 578-90. Bandura, A. & Walters, R. H. Adolescent aggression, New York: Ronald Press, 1959. Bronfenbrenner, U. Socialization and social class through time and Space. In E. E. Maccoby, T. M. Newcomb & E. L. Hartley (Eds.) Readinggin social psychology (5rd ed.) New York: Holt, 1958. Pp. 400-25. Burt, C. The young delinguent, New York: Appleton, 1929. 96 97 Burton, R. V. Temptation influences on Fantasy. Unpublished paper, 1958. Burton, R. V., Maccoby, E. E. & Allinsmith, W. Antecedents of resistance to temptation in four-year—old children. Child Developm., 1961, 22, 689-710. Cattell, R. B. Handbook for the Junior—Senior High School Personalitpruestionnaire, Institute for Personality and Ability Testing, Champaign, Illinois, 1965. Cattell, R. B. Manual for the Junior-Senior High School Per- sonalitygguestionnaire, Institute for Personality and Ability Testing, Champaign, Illinois, 1965. Davis, A. & Havighurst, R. J. Social class and color differ— ences in child-rearing. Amer. social. Rev., 1946, 11, 698-710. Fenichel, O. The psychoanalytic theory of neurosis. New York: Norton, 1945. Freud, Anna. The ego and the mechanisms of defense. Inter- national University Press, 1946. Freud, S. New introductory lectures on psychoana1ysis, New York: Norton, 1955. Freud, S. The problem of anxiety, New York: Norton, 1956. Glueck, S. & Glueck, E. T. Unraveling juvenile delinquency. Cambridge, Mass.: Harvard Univ. Press, 1950. Grinder, R. E. Parental child rearing practices, conscience, and resistance to temptation of Sixth-grade children. Child Developm., 1962, 55, 805-20. Grinder, R. & McMichael, R. Cultural influences on conscience development. J. abnorm. soc. Psychol., 1965, 22, 505-507. Hartshorne, H. & May, M. A. Studies in the nature of character: Vol. I--Studies in deceit; Vol. II--Studies in self- control; Vol. III--StudieS in the organization of char- acter. New York: Macmillan, 1928-50. Healy, W. & Bronner, A. F. Delingpents and criminals: their making and unmaking. New York: Macmillan, 1926. Heinecke, C. M. Some antecedents and correlates of guilt and fear in young boys. Unpublished doctoral dissertation, Harvard Univer., 1955. 98 Hill, W. F. Learning theory and the acquisition of values. Psychol. Rev., 1960, 21, 517-51. Hoffman, M. L. Child-rearing practices and moral development: generalizations from empirical research. Child Developm., 1965, fig, 295-518. Kagan, J. The concept of identification. Psychol. Rev., 1958, §_5_l 296-305. Kohlberg, L. Sex differences in morality. In Sex role development. E. E. Maccoby (Ed.). New York: Social Science Research Council, 1964. Kohlberg, L. Moral development and identification. In H. S. Stevenson (Ed.) Yearb. Nat. Soc. Stud. Educ. Part I. Child Psychology. Chicago: Univ. of Chicago Press, 1965. Pp. 277-552. Kohlberg, L. Development of moral character and moral ideology In Hoffman, M. L. & Hoffman, L. W. (Eds.). Child develop- ment research, Vol. I, New York: Russell Sage Foundation, 1964. Pp. 585-451. LeVine, B. B. Punishment techniques and the development of conscience. Unpublished doctoral dissertation, North- western Univ., 1961. Littman, R. A., Moore, R. A. & Pierce-Jones, J. Social class differences in child rearing. A third community for comparison with Chicago and Newton, Massachusetts. Amer. social. Rev., 1957, 22, 694-704. Maccoby, E. E. Role-taking in childhood and its consequences for social learning. Child Developm., 1959, 29, 259-252. Maccoby, E. E., Gibbs, P. K. et al. Methods of child-rearing in two social classes. In W. E. Martin & C. B. Stendler (eds.). Readings in child development. New York: Harcourt, Brace, 1954, Pp. 580-96. Mackinnon, D. W. Violation of prohibitions. In H. A. Murray et al., Exploration in personality. New York: Oxford Univ. Press, 1958. Pp. 491-501. MacRae, D., Jr. A test of Piaget's theories of moral develop— ment. J. abnorm. soc. Psychol., 1954, 12, 14-18. Matza, D. leinguency and drift. New York: John Wiley and Sons, Inc., 1964. 99 McCord, W. & McCord, J. PsychOpathy and delinquency. New York: Grune & Stratton, 1956. McCord, W., McCord, J. & Zola, I. K. origins of crime. New York: Columbia Univ. Press, 1959. Miller, D. R. & Swanson, G. E. The changing American_parent. New York: Wiley, 1958. Miller, D. R. & Swanson, G. E. et al. Inner conflict and defense. New York: Holt, 1960. Mowrer, O. H. Learning theory and personality dynamics. Ronald, 1950. Peel, E. A. Experimental examination of some of Piaget's schemata concerning children's perception and thinking, and a discussion of their educational significance. Brit. J. educ. Psychol., 1959, 22, 89—105. Rebelsky, F. G., Allinsmith, W. A. & Grinder, R. Sex differ- ences in children's use of fantasy confession and their relation to confession. Child Developm., 1965. Sanford, N. The dynamics of identification. Psychol. Rev., 1955, 62, 106-18. Sears, R. R. Identification as a form of behavioral develop— ment. In D. B. Harris (Ed.) The concept of development. Univer. of Minnesota Press, 1957. Pp. 149-61. Sears, R. R. Maccoby, E. E. & Levin, H. Patterns of child rearing. Row, Peterson, 1957. Sears, R. R., Rau, L. & Alpert, R. Identification and child training: the development of conscience. Paper read at the Amer. Psychol. Assn., Chicago, September, 1960. U. S. Department of Labor, Bureau of Employment Security, Division of Placement Methods. Dictionary of occupational titles: Suppl. 1. (2nd ed.) Wash., D. C.: U. S. Gov't. Print. Office., 1955. Whiting, J. W. M. Sorcery, sin, and the superego. In M. R. Jones (Ed.) Nebraska symposium motivation. Univer. of Nebraska Press, 1959. Pp. 124, 195. Whiting, J. W. M. Resource mediation and learning by identifi— cation. In I. Iscoe & H. W. Stevenson (Eds.) Personality development in children. Univer. of Texas Press, 1960. Pp. 112-26. Whiting, J. W. M. & Child, I. L. Child training and personality. Yale Univ. Press, 1955.