Journal of Personality Assessment

ISSN: 0022-3891 (Print) 1532-7752 (Online) Journal homepage: http://www.tandfonline.com/loi/hjpa20

Assessment of Shame and Guilt and Their Relationships to Psychopathology David W. Harder , Lisa Cutler & Liesl Rockart To cite this article: David W. Harder , Lisa Cutler & Liesl Rockart (1992) Assessment of Shame and Guilt and Their Relationships to Psychopathology, Journal of Personality Assessment, 59:3, 584-604, DOI: 10.1207/s15327752jpa5903_12 To link to this article: https://doi.org/10.1207/s15327752jpa5903_12

Published online: 10 Jun 2010.

Submit your article to this journal

Article views: 117

Citing articles: 118 View citing articles

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=hjpa20

IOURNAL OF PER5CrNRLF3 ASSESShEhTT 1992, 5951, 584-604 Copyright a 2992 Lawrence Erlb2um Assoc~ates,Inc.

Assessment of Shame and Guilt and Their Relations

-wo srudies are reported.

i Fim, we tested the previously validated Persona! Feelings Questionr~ire-2(PFQZ; t i a r d c S: Zeixa: I ' E Q shame and guilt proneKobiitzelle, 1982) ness measure and the Adapted Shame an2 CruiIt Scale (ASGS; Shame subscale against the newly introduced Self-Conscious Afiec: and Attribution T,nven:ory (SCAAI; Tangney. i9Wj for shame and pi!:dispositions. Fifty~ i n co'liege e nndergraduates cornpieted randomly ordered personai~tyinventories reflecting constructs theoreticaiiy relevant t c the presence of shame and guilt proneness. Zorre!ations between ihe affect measxec and personaky variables shawed evidence oi valkiity for a2 shame scales. The FFQ2 Guilt snbscaie aisc demonstrated construct validity when partieled for shame, bur the X A A I did not. Second. iw tested hypotheses regarding the relative importance of shane and guilt to various symptom cupes (Syr.p:orn Checkhst-91"!-Revised; Derogatis, 1983) using 71 college undergraduates. Both emotions were approxms:ely e q d g related to ell nzjor sympta?~, C L U S ~ ~ T S : bur there was some evidence for differer,tiz': patterns o! reietive importance for sharne and guil: t3 Sifferenr symptom.

Gu:k and, more recerarly, shame have been rhe :wo domnant clvsphoric emotlons :hat have beer. lizLed ciinica'liv with the formatior; and m a n t e n a x e of psychopat.thoiogv. h r m g the Ersr t w c th:rcis of t h s cenrurr, F r e d \i909/:955.)and hls hllowers (e.g., Femchei, 1945, ernphas:zed the rmporcance of rrnconsclous g d t . and the anxi~ry aroused by z t , in the omgins oi svrnptornarologv, particularly for neurotic ronditmnr. C o n ~ e r ~ p o r a rcimicai v theonsa (Broucek, l991; Harder, in press; K m h a n , 1989; Lewis, 19211, rn contrast. have emphasrzed the mporcarace of shame across a wide variety of disorders, :ncluding the neurotic states

Hypotheses have been icrmulareci Eriksm, i959; Lewis, 1971: Piers Gr Singer, i 39 1! regardirig differenr aeveiopmentai pathways for = ) e r s ~ d i r imere e~ characteristicaii:;prone t c guilt or :D shame. Different pe:sonaiit)- patterns have been associated w:b shame proneness and guilt Froneness (Harder S; Lea.%, 1 ?Y 7: Lev&! 19: 1; kki-man, 19'741,And different ~reierreciclinicai treatmm: ~ ~ 1ei.i~ ap~roacheshave been recom:nended itiarder, 'LP9&: K a u f r - , ~13%: 1971;Le&, 1991; Nathanson, 1957)for il-L&\ri&ziseXhibitiRgthe :u.c. dispositions. -i hss! the aDi!i:y to assess individual d:$erences in shame and guilt prcneness airh paper-and-~encil measures would aid both researchers and ~1inic:ans. personah Hypotheses about the re1a:ive inpisrzince of the two &ect t;a:ts i z a c d o n i ~ gin~ symptom famation, 2n2 ir, response ce. :herape;:ti~ interienAkc?cIir,icai screening for these :n7c:Lmporra:lt tions coujd be tested eifi~ient!~. treatment-relevmt emotionai tendencies would beccme possikk 1:-i this articie, we describe five studies :n t h s fieid of inquiry. The first s t . ~ d ~ '7 1 ' centime2 construct validaticn research: kgcc jr Harder ar.2 Lewis i i.33 i 1, S y " ccmpar:cg me :nost proxxsing rneasures of shame and g G t prcneness. Spec:$,-. call!-, we in\-estigared the revsed version of the SeK-Consciou: Azect an2 Atrriburion hventory- (SCAA1; Tangney, 1992 meassre d shame oi g d t proneness and compared it :c previouziy validated shsme subscale: of t h e Aciapred Shame an2 Guilt Scale (ASGS: Harder & Zairria, 1990;P;,obiit.de, 1952) and the BersonaI Feelings Questionnaire-3 (PFQZ: Earder: Rockart, SI Cutler, 198;; Harder Sr Zalma, 13??'; znd to rhe simiiarir estabiishec!:"P Gaik subscaie (E!larcler. 139Cia; Harder & Lewis. i 9 K ; Fiarjp: e t ai., 1%:: Pisrder & Zalxa? 1992). The second stildy was comprised of twc parts. Ic the Brsr part, we examined the possibie vz1idit.y of 3 second revisicr? of die SCAAI. ir, the secmd parrywe direcrii. rested hypotheses regarding :he role cishame and g.dt in different ti7pesof psychoiogicai svnptomatolog):. -AX

:I-.

Initial reiiabilty and construct vaiidirv research (Tangney, i99C) with the recentiv introduced S C A N has snggested that it may \veil be ar, exce!Ient way co assess shame an6 guilt, The SGAAl consists of a :vierc of sitwtions :o which rhe subject might experience seiwai emotional reactions. i he subject races the IikeIihood of thiriking or f e e h g in ways consistent wi:h sheme, guiir, anii.;or other responses. The scale's ad\:anrages include the rating oi herb s h a m 2nd p i i t feelings for each situation, :hc ase of ar? cash comprehended socabdsr-c, and the phenomenologca! natnre of the possible reactions iisced. Hcweier, rhe data reported previonsiy must- be regarded 2 s preliminary f ~ several r reasons. The Shame s&czie ofthe SCAA'i has showc piidence of vaii&v iTangney,

-

586

HARDER CUTLER, ROCKAIR'

199Cjby exhibiting 2. significant relationship with one previous shame measure ithe Revised Shame and Gui'it Scale [RSGS; Hoblitzele, i982!, which itself has shcum only one correiation -with depression -to support its vaiidity (Mobiitzeile, 1963). h addition, the RSGS was used by Tangney (1998; in a revised form that c:oihpsed its separate Shame and Guile subscaies into a reconceptualized shame measure. That procedure may hare been problematic, because the RSGS G d t subscale had not previously demonstrated construct validity as a measure of either ~ i l ort shame (HoDiitzeIie, 1962). Subsequent , with a measure ithe PZSGS;Wobiitzelle, 19821 research (Harder & Z a l ~ a 1993) similar to the RSGS prcvides additionai validity for the SCAAI Shame subscaie, rhough its final sratus as a useful assessment tool still reqsires addit~onalverificarion. Sixilarly, the S C A N guilr instrument has demmstrateci an encouraging, posirive reiationship with Mosher's (1966,1965) Guiit Scde il-. two saxpies of mostly female sul-jec:s (Tangney, 199C.).Yet another icvestigatior; (Harder iSr Lewis, 1987) of promising shame and guilt measures reported a notably invalid (shame-prone rather than guift-prone) patterr. of exterr~aimiidicy correlations 3ecause Mosher dic': no: for Mosherts G i l t Scale among ferriaie sub@s separate shame from g ~ i i ritems in the construction of his scale, this i a t r e ~ finding is perhaps not ton surprising, Nevertheless, the result has left che S C A N Guik subscaie with only equivocai validation to this point. We wanted to examine further the concurrent validity of the SCA.4IR ithe revised, psychomrtrical?y ixproved version of the SCAAI, now called the Test of Self-Ccnscious Affect; Tangney, Wagner, & Gramzow, 1993) in conjunction with additionai i~vestigationof the more established ASGS and PFQZ subscales. Ir: light of current evidence, suggesting that the validity of guiit measures is more difficult to demonstrate convincingly thar: thac for shame scales (Harder, 1992a; Harder et ak., 1991; Harder & Zairma, i990>,we hoped l d impressive!y. that the SCAAlR guilt measure u ~ ~ perform Foilowing the design of previous va!idaticn studies (Harder 6r Lew:s. 1957: Earder & Zalma, 1990), nine personality dimensions were used as external criteria against. which tc. evaluate the validity of the shame an6 guiir measures. These constructs were depression, self-derogaticn, social anxiety, shyness, purc?Ekself-consciousness,private seltconsciomness, narcissism, social desirabiiky, and Ioclas of control, Clicicai an$ emplricar literature (Crouppen, 191';; Harder & Lewis, 1987; Lewis, 19?i, 1985; Smith, 1972; Wright? O'Learyl 5r Baiiiln, 1989) has repeatedly linked depressior. to guilt and, even more strongly, to shame, Hence, positi.i,e reiationships berween valic scales for both of the ciysphoric affects under study and a depression variable were expec:ed. bzt the magnitrmde of assoclarion with shame was theorized to be larger. Self-demgaticn, representing a relativeiy stable tendency to devalue the self, was expected to correlate with shame and g d t measures in a pattern similar to that seen for depression.

Because shame experiences praauce a more giobai seii-cie~dua:icr: than guiit (Lexvis, 1971: iynci, 1958: T a ~ g n e v ,i?P9), shame was expe:ted to show the xronger relacsnship. B e c a s e vdrrerabiiit:; t o shame has been described cknicaiiy an$ theorecicaiiy as ail importam dynaxic underjying narcissism (Broucek, 1Q91;KercSerg, !?7?; Rlorrison, i959 and because the t i y e of narciseisrr! that successr'uiiy defends .. high sek-es:ee= ir: an cngning rrianner seems tc be the variant assessee r y self-report perso~alitymeasures (Earder, 13%; Harder & Lewis, I%;!, trait :~arciss:sm vias Fredic~edts carrelate i~verseh:with shame rroneness an8 nor co e. rniate sign~!~can:l;t t c g d t . Buss

p < .331'i- urcrner! had mu& higher scores- arid r h k m a y preser.: ~ r o b i e m sir;

the fctnre. Aichough m n e rheorisrs (Leffis, i975: Wright et a]., i?%i ha;.e ., argued ma: w70me:: rruiy experience more shame! neithe: the A S S m r :he PFQ2 shewed sdcb a diserence in this ST in prei-i0.5~sru$ies i'Harder et a:.. : i 94, ; Ehrae: & Z a h a , 19'33; Wright et al., 1959. P a r d csrrelzticns, : e x w i n g tk,e ., . , ,influence of thls sex diiterence. were caicuia:ed rnr rne Lypoth~si;edSC,:\.liIR vali,::+ .c , , y correlations (Table 3). No appreciabie cixnges in re:r:jonsiiy ncCUvor. ..LL Eence, the vaildiry c?i the new SCAAIR Shame scaie is supporced, although use of the scale in research would r e a s r e separate interpreratior~sfc7r nren anb women. I he correixims ir: Table 4 presenc a more nixed conciusion regaromg m e . , SC.4.43 Gnh scale. This measnre showed presicteci ncnsgxfiiant reiaticnships wiih narcissism? social anxiety, shyness, and sociai desirabilit;. and 2 prediccec! significant assoc:ation with private self-censciousness ( 7 = .i6,p < .35L Hcwever, expected relacionships wi:h the Seck Depression 1 ~ i ~ e ~ t i = .r~ ,, .. -.Cil and df-derogatioc i: = '06)did net mamiaiize. Aisc. ~ u b i : cse:icnsciousness showed 2 surprisingly high correiaticn (r = .26, 9 < . 3 3 in rhe dxectiori costrary to predicrinr;. The magni::de for slyness i,r = - . 18;was iisc , higher char: expected-in a negative, ratfier trim the more 'likelv scmwha: poskive, direction. Amocg chese results, the m a r :ro:;Siesome for CBnStX1ci , , . ? , -

-?

-

7

.

+.

3

(7

-

validity are the near-zero relationships with depression and self-derogation. i c find no sdch relationships goes strmgiy against the weight of theoretical and clinical evidence linking characteristic g d t dispositions with these two conditions. Further, these datz revealed n c significant association between the SCAAIR Guiit scaie and the more established i"FQ2 Guiir scale j: = .0 s3eme t h a guilt. ~ Traditiona l ! ~ however. , gxiit hrs beec s t r o n g l ~linked with the presumes core c x f l i c t . ~ mderlving hysterical persona!it\; constei1a:ior.s. (Fenichei: 143%. Thus, the stronger relatonship expected for shame shonid be small. -, kegarding ohsessi\le-compzisi\~e synptorns, Lews i,197921 made z reasaned czse? b a k e r e d by cxnicai e x a n ~ l e s(Freud, 1909:1955; teil;i~, i 9; i ivf e y :he:r ( ,

'

7 .

g x z i c conae-tion with g.uilt emorions rhrn wii-h shame. The arg2rner.c "

7

.

centers

7

an the r~rninztiveaspects of cognkior: around gdt-lad,.Pq behavifirs tilat are so '

prominent ir: obsessive states. l'e: Lewis felt that sometimes fears of (shamefdr * i o ~ sof seif-esteerr: in these same indlv:duais a x obscurea fir. more i o n s c j o a . . rcrn:narions whose apparec: conren: invoives g d t . Also niiitatrng againsr Lewis's hyporhesis of a stronger connection between gsiir and ebsess~venessars . , , . Shapiro's (1965': oSservations that o'Ssessi\~e-icxp~is!\-e characrers are iceeniv aware of pcrentiai criticisms from others anit a wideiy held theoreticai posiricc ie.g., Erikson: i959; Piers Sr Singer, IS7I: thac skeme is :he more ~ r i ~ . i t i \ ~ c emotion, occzrring earlier in deve:iopmenr. Eriksoc ;i959: ciained rhat 5eaiinp , , adaptivelv with :ctense feeiinps ai shame and doubt is the cen:raI c;.,i:lenge af rhe anai stage of de\:elopment, which is tied c!:r.ical!ly with obsessive-co-.puis:ve problems (Feniihel,. 1945; Sha~iro,I365). The inrerpersanai autoncxv main:ained by such compulsive characters can be see2 as a defense aga:nst a stror,g shame propensity. Clinicai ohservatior! suggests that such a dispos:tior: ?robably remains largely unconscious for most obsessives. Thus, in o x srudy. whereiri self-reports are expected primariiy to re3ec: i o n s c i o : ~feelings, ~ we made 2 pdic:ior: rhat agrees with Lewis$ expectations: Cbsessive sy';ipta:ns :vex rn be more strongly associated with gnllt.

.

596

HARDER. CUTLER, ROCKART

Symptoms reflecting pathological interpersonal sensitivity, akin to social anxiety and extreme shyness, were expected to show more association with shame than p d t . -historically, both guiit and shame have been linked with pathological anxiety srates (Fenichei, 1945: Freind, 1985/;95?; Gottschzk 5r. Gleser, 1969; Lewis, 1971; Mayman, 1994). As with most earlier theorizing on symptom fornation prior to 1910, the emphasis was on guik This emphasis has shiked toward shame, and changes in psychodycamicaliy inspired cl/nicsi rheory (e.g., Greenberg 5r Mitchell, 1983; Kohut, 1971) have shifted toward interpersonal aspects of parholog-; causa:iors. Lewis (1985) indicated thar shame shod$ be the more important affect in anxiety states, because shame proneness in\&es threats from amoral and moral sources: as weir as fiorr, more immediately anticipated losses of icterpersonal connection, Lewis il919b) also linked shame with hystericai conditions, which (in Fsychoanz!ytic classification\ indude the most. obvious examples of overtly exhibited anxiery symptoms. From the foregoing, we predicted that both gui!: and shame proneness should be sdxterz.r tially reiated to xiacnest anxiety probiems, with 2 mmwhat. grexer association for shame, We held this expectation tentatively, however, because a previous exipirical examisatlor? (PIarier & Lewis, 1981) of this qxstion indicated that shame and gxilt were equally implicaceci. Similarly, problems with excessive Rosdity h a w also been relace2 to both shame and guik /Eenlche%,1945; Freud, 1933; Lewis, i97lj. But much a: this theory has dealt with ~lnconscioushostility, defensively avoided on the persona l i ~surface by depression, reactictn f~rmation,or other rr,aneuvers. The extreme intensity zf :he hurjliliated fury, or shame-rage spirzi described by Lewis (1971; Retzinger, 19571, and the hostile defensiveness against shameful exposure sonetirnes observed in therapy (Levin, 1371) a r s e for a greater contribution o i shame to excessive hostility; yet ic is also apparent that many such individuals suffer mostly fron the depressim created by their blockage of the hosrihy they experience (Lewis, 197l , 1985). From other repcrts (Biines. 1967; Fiscker, 1985: Katz, 1988; Lansky, :%'1?; Wright, 2989) excessively uncontroled hostility would, indeed! be mere linked with shame. Because the focai parhokogies of chese reports-antisocial behavior, borderline personality, conduct disorders; and violence toward ochers- are not cilr,icaly connected with strong or overwhelming guilt feelings ar with the capacity to transfora rage intc guiit-ridden depression, the plausibiliry for a stronger associatior. with sharne is strengrhened. Therefore, we predicted thar shame wwid be more highly a significam correiated with symptoms of'hostBity, though s i l t vr,ouXd alsd s h ~ w h k - .di pre\+ous study (Harder 6r Lewis, 1987) with college students, however, did not find such a &if erence. SaseZ on theoretical assertions (e.g,, Erikson, 1959; Kernberg, !9?0) about the less developmenrally advanced nature of psychotic aspects o i personalit.)!, v r ~ predicted that shame woulci be more reiated than grriir to a clustering of

F~syc:clotic symptoms. Shame reac:ions have been postulated t c o c c ~ as r early as the 8:h month of Life, Iong before a capaciry to experience guilt iNatKansan, < ~ 9 2 , ; and, as sash, u7ouId Ee more ex-iden: in pathoiogies :ha: p r e s ~ r n a c r! ~ ed: horn those eariier periods. A body of research (Eooiey, I%::, indica:ing :ha: . . oaen cri:icisrr; by coresident f a d v menbers i?! schizophrenicc and ps\.&ot:: . depressives increases the probabi!it): of reiapse, suggests that - lr muc? D: the ** c;iticism is of a glohal, shaming nature-shame would be a p t i o g e c i c arkc: fc: individuaiz with pzychoric tendencies. . ?. Phobias can occur :n many speclrrc forms, ~ h i c hhave beer, grouped by p s ~ d x a n a l v t i ctheory (Fenichel, 2945') inro rhose linked rnore with 'Cyster:a (svmSo!~cailvrelared to s e x d conflicts! and rhose Iinlceiri rnore n-:h obsessivecompxlsive concerns {e.g.,dirt and diseasej. Guilt has beer? oDser\d c!irricaily in * , . response t c an mconscious wish represented srmbo!ical!y PY the phobii, v;hiche~;er type k may be. Shame has been considered a control mechaxis:~ Leuis! grrzyed against exhibitionistic andior vqeuristic wishes ~ F e r k h e i1945; , 197;) i ~ .the one type of phobia, bur. aiso as a defense againsr :he hsstile and sadistic rr.puises associated with :he presexua1 jpreoedipai) conflicts !.Erikscr., 195P: S h a ~ i r o ,1945) of another type d phobia. These consideraticns predict roughly equal contributions of both shame and ,-;At to the presence of phobic symptoms. For paranoid svmpromatoiogv, Lew:s ( I 9 7 l i used an argument siniiar tc. the one dread;. discussed far obsessive-compulsive states tc clairr; a greater role for gsilt. Howe\7er, some clinicai ohservers iColby, 1976;Kaufzan, 13%: Skap:rc, r

> .

O.?.

!?

I

/

:

?

1965)heve noted proncxanced shame proneness in cases olpareno:s. Ssppcrriix

ofthis positioc is the phenomenoiogy of p r a n c i d preoccuprions, which after: include beiiefs a b o x being watched criticail-y or hcstiiely, for exposure of one's shortcomings ia shame experience). Hence, we predicted :hat shame and gaiit shouid be equally associated with paranoid svmptcmatolop!..

51

-

!, undergadsate students (42 fernaie an;! 29 maie! from the same xr:iversitr that provided the sarr;.~ie for S d l ; 1. i hese subjects were zisc from an introductory psyc'hoiogy course. a d thei: other characteristics were quite s i d a r to :hose in Study I,

Sl;bjecfs. Snbjects were

Proceiiu~e~As before, research assistants distributed packers oi measures to the subjects. Code r ~ m b e r were s used co keep respsnses anonymous. The scales incirrded in the packets were :he ssme as in Study i ? wkh rwo changes. i he first change was the inclusion of a revised SCAPLR. We created :h:s SCA.413 measure for gui!t proneness by adding eight new items, u h c h . we hoped, ud:! elicit rnore intense guilt responses from ssi'bjects and ccnsequenciy prcvide a more valid weaslrre o i guilt tendencies. A possibie limitatlox s t :he

-

?

.

598

HARDER, CUTLER, ROCKART i

'

eariier version was the reiativeiy miid level of distress that seemea to be srimulated b y rhe siraations depicted in most items. The new items inciuded the foIlowing situations: knocking down ar, eideriy woman while rushing to a class, picking up a $20 bill that has dropped from someone else's pocker, intelectually embarrassing a person of cleariy Iirnired indigence, cheating on an exam, sieeping with a new acqsaintance while one's lover is out of town, sneaking a look at the job evaluation of a competitor, cheating on a dass composition, and choosizg rc?go away ior h n with friends wher. one's widowed and lonely mother has specificaliy reqaested a visit during that same time, Informal pilot resting of t h e s e n e w items suggesteci strongly that the:r use would enhance rhe chances far stccess :i guilt assessment. -i he second charge in the composition of rhe research packet was the addition of the Synptom Checkiist 99-Revised (SCE-90-3; Derogatis, i983; Derogatis, i r p m a n &.CoVi, i973) to assess general psychopathoiogy severir); and a varietv of specific symptom types. Subjects required a bit more ehm i hr KO complete ali research scales. Means and standard deviations of the shame; guilt, and constxcr. validity variables in the srudy are presented in Table 2. 7

.

Results: Vdjlchity Evidence for the SCAAI5 Guilt Scale

-i he new

SC.4AI3 Guiit instrument was correlated with each of the nine construct vaiidity variables used in Study 1 and witk the PFQ2 Guilt scale. hesuits were coxpared :0 the hypothesized vaiid::y pattern for a gcilt measure (see Table 1;. These correlations are presented in Table 4.The resuits were not encourag:ng, which is perhaps an unsurprising outcome afier noring that the SCAAU and rke SCAAIR produced a correlation of .?2. As with the SCAAIR Guiit scale, the new, revised SCAMS showed no significant relationships with depression or selkderagation (7 = .12 anid r = .I;, respectively), though the correiations were in the predicted (lirectioil' Zn addition?unexpected significant or near-significant associations appeared with narcrssisrn ir = - 2 2 , $ < .10) arad sociai desi~abiiit~ (7 = .25, p < 45). Becarese :he new Guiit scale shcwed a strong relationship with gender :r = 34,p < .OT), the sex variable was partialled our of the zero-order correlations. This analysis left no significant relationships wirh conmuct vaiidity variabIez at all and essent:aily n= reiationships with the cruclai depressior, and seif-deroga~ionvariables, Further, the SCAAl-?Guik scaie did not correlate slgrrificantly wich the PFQ2 Gnilt scaie (r = .20), yet i~ d ~ d relate verp strongly with the SCAXIR Shame scale (r = .69: 2 < .001! and significantly with the ASGS Shame scaie (7 29,fi < .G5). We aiso examined the eight new items as a ciuster, with the thought that they might form a separate: valid Guiit scaie. But the pattern of concurrent validity coeificienes was almost ide~ticai to that just described for the SCAM3 7

-

Discxssion: Relationships With Symptom Variables Fearson correiatmns were caic-Iated Set.weer: the 4SGS shame xeasure, deter-

600

HARDER, CUTLER, ROCKART

moderately to strongly, and about eqilally, associared with the entire range of syrrrpr~rnatolog~ shown by undergraduate students. Hostiiity and paranoia were the least associated with shame and guilt, perhaps reflecting, in part, the generaliy well-control'led and nonparanoid nature of a coiiege sample. ;o examine the separate and distinguishably independent associations between shane and gsii: and the symptom ciusters assessed, correlations were calculated again, this time partialiing shame for g d t and guilt for shame iTabIe 5). The ASGS Shame scale now related significa~ti~ to depressior, ir = 24, < 3 1 , obsessive-compu1siveness Ir. = .38, p < .01), interpersonai sensiti.vity ir .29, p < .Eland , phobias (r = 2 4 , 9 < .P),and it showed a :rend with psychoticism (.r = .2@!, Shame also remained significsmtl.y associated with bo:h summary meassres: :he ,331 (r = 2 6 , p < .'35) and the PST (r = 24,P < .05). The PFQZ Guilt scale, with the overiap of shame now removed, correlated significantly wich somatic complaints (r =: .3C, p < .E),interpersonai sensitivity (r = .3E! p < .01), hostility (r = 2 7 , p < +C15),and psychotic manifestations I:= 2 6 , p < 35): and ir showed a trend with anxiety (r = 21). Guiit also remained significantly associated with the GSI (r = .38, g < .35) and the PST (r = .30, P < -05:. Examinatim of the partial correlariorx provided the clearest test of the hypotheses set farward esrlier regarding the relative importance of shame and guilt to various symptom types. Shame char is "pure" of guiit's infiuence appears to be more highly related to depression, obsessive-compuisi\reness, and phobic anxlery, whereas shame-free guilt appears to be more related to s~maticmtmn. psvchotkism, hosAty-anger, overt anxiety, and poss:bly parano:a. These ismail) differences. however, shoujd be accepced tectativeiv. pending rephcaand host:lity were the coefscient tron. Only far obsessi\~e-cornplsi\~e~~es~ differences for shame and guilt as muck as 20. Table 5 indicates that, at this polnt, the sole differentiating hypothesis gaining empirical support was that stated for shame's greacer. reiationship with depressive symptoms. Expected eqzivalent relationships for shame and guilt with Eorh phobic anxiety and paranoid ideation seem, fro= these data, more plausible for paranoia than for phobias. And an hypothesized, somewhat larger magnitude of association between shame and overall psychopathology (GSI and FST variables) was not r~pportedby these findings. In fact! if an.)-difference does exist for the affect variables, it would appear that guiit is slighdy more gloSaliy associated with pathology. Note that, although these partial correiations can indicate the portion of variance of each d e c t not attri'butabie to the other kind of emotion, thev do remove any variance that vdidly reflects the co-occurrence of the two dysphoric emotiona:1tendencies. As such, the partial correiations mzy be underestimations of the true associations between shame and guilt and the syrnp:oa scales. Also note that ail nonsignificant partiai correiations in this anaiysis! except for shame w:th hostility, showed positive magnitudes (effect sizes) o f . l l or greater. Wkh rp

-

large saxples, a'ii such reiationships c o d prove significanr and nonni\.ial i R o a l o ~& Rosenthal, 1959). These ~esnirsare at variance with findings of minimal impoxance for g d t reporred biTangnev et al. (1990),whose m e a s a e of gui!: w2s the promising 5x1 uitimareiy invalid SCAAIR. Presenriy, we x x s t c o ~ ~ c i u dthat e guilt is, indeed, associated with the eztire range of svrn~toma:oiogv obseri~ecia x o n g coilege undergraduares, and that its influence on sach slrmptoms is at least e q d rc that shcwc bv shame. T'ne Ierge number of unsupported hypotheses regarding different relationships for shame and guilt with varions synptom ciuscers soggests :hat more caref:d cli~icalan2 empirical ic:restiga:ions or :>e role phi-ed by :hese two affects need r c be conducted. Ideaily, such research shonld aiso :nrdve the \,erifica:ion of symptoms in ways ocher :bar, seif-repor:. -i he bypassed shame coccept of Lewis (1973, in n.hich s patienr obsessidy ..jminates abou: the roie of :he seif in a shaming experience, z a v accoun: h r some of the unexpectedlv strong association herweer: shame and oSsessi~,~eness chseri%d here, but there are n o readily available esp1anations of :he other .,--.. ,,iarSrici~areddifferences. Especia3y for inrerpersc~alsensi:i\*~ry,:he essent-iaijy .. " , equal partmi cxreiations f x guilt anti shame a x jurprising. Esaminaclzn ot che / . indii:idsai items on the SCL-90-2 scaies offers an e x ~ i n a r i c no: on!,; one addr~ionalunanticipated difference observed here-that berweeri shame and phobic anxiety. Almost- all of the phobic items pertain rc fears of being in ~ u b l i c or in social settings (t-ypicaHy,shame concerns?, and none d e d wirh :he dirt an2 disease concerns Lews (!?7i! ikked more strongly with gui!:. li rhe o.,,eraI1 pccure provided by these data is ar, 2ccara:e one: then perhaps g d r is r ~ x nho r e i n i d v e d with symptoma:oiogv than the current ciinical and theorerical emphasis o n shame wou'id inply. Perhaps, f a example, theore:;ca! and research atrention shouid once again be directed toward some of :he eaiiier grriit hypotheses of the psychodvnarnicists. An imperrant Iini:ation o n the genera1i:ak:iity 3ftt:s study shodci be noted. It has :o d3 with the use of an exciusiveiy undergraduate sampie. Some symptom ,. frequencies, and the subsequent variance in scores, wouia PC milch iarger in . other sanpies, paxicdarly those selecre3 frorr;. c i r x a l or hnspkal popuiations. perhaps ~rcvidiraga better estimate o i true affect-symptom reiationships. The pmminence of shame and gu:it mag also be different in the symptc,ix configurations of generally more diswrbed groups. 5efore generaiizi2g iniau::ousi~ ro these n o r e troubied g r o u p , some validationai work o n the scales s h o d d be conducted with those popuia:ions. P

,

v

7~

A ~ r e v i o u sversior, of this article was presenred

af rhe annuaj neecing of :he

Sociery for PersonaIiry Assess;nent in Washingtor;,

DC,March i992.

602

HARDER, CUTLER, ROCKART

REFERENCES tibramson, L. Y.. Seligman, M. E. T.: & Teasdaie, j. (IoiEl. Learned he!piessness in humans: Critique and reformiation. lounral oj AbnomaI .FsychoioP; 87, 49-74. 9 e A . A. T. (1967). Depression. Causes and treutmenr Philacieiphia: L'niversiq of Pennsylvanic Press. Biimes, M.(i96?j. Shame and deiinquenq. Contempora? Prjchounc!y:s, 3, 113-133. Binder. I. (1970). The ;cia:we proneness to .&me or p i k o: c dimension ofci~ciactersr:vIe. Unpublished doctoral dissertation, University of Michigan. Ann Arbor. Broacek, F. 1. (19911. Skume and the seij. New York: Guilford. h s s , .4. il. (1932:. Seli-consriousness and social axiety. San Francisco: Freeman. Carver. C.. di Giass, D. C. i19?6}. The Self-Consciotasness Scale: A discriminan: va!idirr study. .Joumaiof Pee.sonatiry Assessment, 40,169-172. Coihy, K. N. (I'J76). Clinica! implications oi a simuia:ion mcdei o:' paranoid processes. Arciiiurs nf Genera: Psyciiiar~,33, E54-859. Croupper., G. A. (i979::. Fieid dependence-independence in depressed and normai males, as an indicator of relative proneness t o shame o; guilr and ego kncacning (Doctoral disserrarion, Caiiiornia School oi Professionai Psvchoiogy: 1'375). Dissertatioc Abstracts Intpmational, 3;. 46693-46305. Zrowne, D., & h.lzriowe, 3.:I%?!. A new scale of sociai des~rajilityindependent 31 psychopathology. !ou~nalof Cnwulcing clnd CE:n~culP&b!ogy, 24: 343-354. Crowne, i""., & ?\).;arlowe,E.(i964). T n e a ~ p w motiuc. d hiew \-ork: W'iley Derogatis, L. ;I983:. SCL-91:-i7 manual. St. Petersburg, FL: Ciinicai Psychorne:rics. Derogatis, L.R., Llprr.a.;1, R.S., & Govi, L. (i973j. SCL-?O: .Ax o q ~ t i e n :psychiatric racings scale- A preiirinarv report. P s y c h o p h a m ~ oBuiietiz: '3, 2-28. Eriksor., E. Pi. (1959). identity and the life cycle. Psyhoi'~eiaiiIssues Monograph, !(Whole No.1). Fenichei. 0. (1945i. The pYchoanui~trcthe09 o? neuross. New York: In:erna:ionai Universities Press. Fenigste~n,A, Scheier, M. F., 6i Buss. A. H. (19'753. P>h!ic and private self-ccnsciousness. Assessmerc and :heor?. Jotlmul oj Cor;suir;ng and Clinml R?~iWiog)..43. 52-52; Ferguson. T. J.? Stegge, H.,6r Damhuis, I (Iwi). Chiidreg's ~ndersrandingof guilt and shame. C.hiia Det~eio.m~ent. 62, 827-839. Fischer; S. F. (19%;. ideaticy of two: The phenomerroiogy oisharne in borderiine cieveioprnent and treatment. ?syc;bthera@, 22. 131-i09. Freild, S. (1917). Mmrning and melanchoiia. Ir; Coliected >ape.-. p o i . 4). London: Hogarth. Freud, S. (1933). New ir:troductor) lecturer to psychoanaivsis. New Yo&: Norton. Freud, S. (1953). Three esseys an the theory o:sexuaiiry. in J'. Srrachey

Assessment of shame and guilt and their relationships to psychopathology.

Two studies are reported. First, we tested the previously validated Personal Feelings Questionnaire-2 (PFQ2; Harder & Zalma, 1990) shame and guilt mea...
2MB Sizes 0 Downloads 0 Views