The Proximate Causes of Employment Discrimination Author(s): Barbara F. Reskin Source: Contemporary Sociology, Vol. 29, No. 2 (Mar., 2000), pp. 319-328 Published by: American Sociological Association Stable URL: http://www.jstor.org/stable/2654387 . Accessed: 15/04/2013 17:08 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to Contemporary Sociology.

http://www.jstor.org

This content downloaded from 184.171.106.7 on Mon, 15 Apr 2013 17:08:35 PM All use subject to JSTOR Terms and Conditions

Symposia 319 Giddens,Anthony. 1984.TheConstitution ofSociety: 1996. "Confirming Allusions:Towardan Empirical AccountofAction."American Journal of Outlineof theTheoryof Structuration. Berkeley: University ofCalifornia Press. Sociology 104 (1): 161-216. Goffman, Erving.1959. The Presentation of Selfin Smith,DorothyE. 1987. The Everyday Worldas Everyday Life.GardenCity,NY: Doubleday. Problematic:A FeministSociology.Boston: . 1963.Behavior in PublicPlaces:Noteson the Northeastern University Press. . 1990.The Conceptual Practices ofPower:A SocialOrganization ofGatherings. New York:Free Press. Feminist Sociology of Knowledge.Boston: Heider,Fritz.1958. The Psychology of Interpersonal Northeastern University Press. Relations. NewYork:Wiley. Tavuchis,Nicholas.1991.Mea Culpa:A Sociology of Heimer, Carol,andLisaR. Staffen. 1998.FortheSake Apology andReconciliation. Stanford, CA: Stanford University Press. of the Children:The Social Organizationof Responsibility in the Hospitaland the Home. West, Candace 1996. "Goffmanin Feminist Chicago:University ofChicagoPress. Perspective." Sociological Perspectives 39 (Fall): Hochschild, ArlieRussell.1983.TheManaged Heart 353-69. 1995. Commercialization of Human Feeling.Berkeley: West, Candace and Sarah Fenstermaker. "DoingDifference." Gender andSociety 9: 8-37. University ofCalifornia Press. Latane, Bibb and JohnM. Darley. 1970. The West, Candace and Don H. Zimmerman. 1987. "DoingGender."Gender andSociety 1: 125-51. Unresponsive Bystander. NewYork:Meredith. Rollins,Judith. 1985.Between Women:Domestics and White, HarrisonC. and VilhelmAubert.1959. TheirEmployers. Philadelphia: TempleUniversity "Sleep: A Sociological Interpretation." Acta Press. Sociologica 4, fasc3:1-16. Rossi,AliceS., andPeterH. Rossi.1990.Of Human Zajonc, RobertB. 1980. "Feelingand Thinking: Bonding. Hawthorne, NY: Aldinede Gruyter. PreferencesNeed No Inferences."American Rubin,Zick.1973.Liking andLoving: An Invitation to Psychologist 35 (2): 151-75. SocialPsychology. New York:Holt,Rinehartand Zimmerman, Don H. andCandaceWest.1975."Sex Winston. Roles, Interruptiolls and Silences in Schegloff,Emanuel A. 1968. "Sequencing in Conversation." Pp. 105-29 in Language and Sex, Conversational Openings." American editedbyB. ThorneandN. Henley.Rowley, MA: Anthropologist 70: 1075-95. Newbury House.

The Proximate Causes of Employment Discriminationt BARBARA F. RESKIN

Harvard University

Highontheagenda ofsociology istounderstandknowledge thatcanbeusedtoeradicate jobdisH Ifsociological research is to conH theorigins andconsequences ofinequality. This crimination. tothebattle against injustice, weneedto understanding is potentially oneofourimporHtribute moreattention tohowinequality isproH tantcontributions topublic policy. Examples of direct duced. In the following pages, I suggest that suchsociological research topicsinclude access research findings from our sister discipline, to qualityeducation, welfare "reform" and canhelpusunderstand both poverty, and theamount of job competitionsocialpsychology, the original and the proximate causes of employH between immigrants andnativeHborn lowHwage workers. In thisessay, whichfocuses on gender mentc lScrlmlnatlon. 1 nls sometlmes lnterc 1So approach thatdistinguishes original and race/ethnic discrimination in the workHciplinary) and proximate causes may be useful and even place,I arguethatthe standard sociological necessary in other specialty areas where socioloH approaches toexplaining workplace discriminaH seektocreate a more justsociety. tionhavenotbeenveryfruitful in producinggists In thetwentieth century, mostsociologists concerned withreducing employment discrimiH 8 Theseideasbenefited fromthe comments of nation assumed thatoncewedemonstrated that LowellHargens andWilliam Bielbyandofthe discrimination persisted, our evidence would Contemporary Sociology editors andeditorial board makers whowoulderadiH members. I was also helpedby talkingwith finditswaytopolicy Thus,sociologists and Marilynn Brewer. Anylogical orfactual errors are catethisdiscrimination. othersocialscientists developed a variety of entirely myresponsibility. .

.

.

.

..

This content downloaded from 184.171.106.7 on Mon, 15 Apr 2013 17:08:35 PM All use subject to JSTOR Terms and Conditions

.

.

.

320 Symposia

innovative techniques to assesstheextentof dominant groups to discriminate against menzH employment discrimination. Researchers groups. conH bersofsubordinate I remain convinced ductedsophisticated analyses perspective establishing accurately race thatthistheoretical charH and gender acterizes thebehavior disparities in various ofsomepeople. ButinterH employment outcomes, net of qualifications; confirmedgroupconflictis not the only sourceof oreventhemostimportant through surveys employers' one. aversion to hiring discrimination, discrimination peopleofcolor(Kirschenman as theresult andNeckermanByconceptualizing behavior, 1991;Bobo,Oliver,Valenzuela, we cannotidentify andJohnsonofconflictHbased causesof discrimination 2000);anddesigned that ingenious waystoestimate the proximate from other processes. Insum, I argue that theprevalence ofdiscriminatory treatment (Fix results approach thatnzany sociologists andStruyk 1993;Blumrosen, Bendick, Miller, thetheoretical has not generated and Blumrosen 1998).In terms ofourpolicy embraceintellectually models ofthecausesofemployment impact, however, wemight havespent ourtime explanatory Ifourgoalinstudying discrimiH better in counseling labormarket entrants or discrimination. howtoreduce it,conflict working as humanresource specialists. If we nationistodiscover theories arenotparticularly fruitful inhelping us wantto usesociology to reduce discrimination why discrimination occurs regularH in thetwentyHfirst century, we needto move tounderstand ofwork organizations. beyonddemonstrating thatemployment disH lyintensofthousands In this essay, I argue thatweshould turn our crimination exists, andinvestigate whyit perH to howas wellas whydiscrimination sistsinwork organizations. To dothis,weneed attention occurs, andI propose thatsocialcognition theoH to expandourconceptualization ofdiscriminaH boththesequestions. I maketwo tiontorecognize thatitoccurs asa result ofnonH rycananswer although someemployment disH conscious cognitive processes, as wellas from claims.First, results from peoplepursuing their thedeliberate negative treatment ofpeopleof crimination groupHbased interests or prejudices, muchdisH colorandwhite women. stemsfromnormalcognitive Theprominent sociological explanations for crimination (thesubject ofsocialcognition theory) discrimination atthebeginning ofthenewcenH processes that occur regardless of individuals' motives.2 turyare grounded in conflict theory(e.g., theproximate cause ofmostdiscriminaH Blumer 195c°; Blalock1967,1982;Reskin 1988; Second, andhowpersonnel practices in Martin1992;Jackman1994; TomaskovicHtioniswhether organizations constrain thebiasing effects Devey1993;Tilly1998).According to a conH work cognitive processes. flictHtheory perspective, the beneficiaries of oftheseautomatic In brief, social cognition theory holdsthat systems ofinequality protect their privileges by categorize othersinto usingthe resources theycontrolto exclude peopleautomatically andoutgroups. Thevisibility andculH members ofsubordinate groups. Thus,thesetheH ingroups tural importance of sex and race and their roleas oriesexplaindiscrimination in termsof the core bases of stratification make them almost strategic, selfHinterested actions bymembers of basesofcategorization.3 Having cateH privileged groups whointentionally exclude and automatic others, peopletendto automatically exploit subordinateHgroup members toprotect or gorized think, andbehave toward [particular memH advance their owninterests. However, conflictH"feel, bers of the category] the same way they . . . feel, theoretic approaches todiscrimination aredefiH think, and behave toward members of that social cientin important respects. Mostimportant, moregenerally" (Fiske,Lin, and theydo not identify the specificprocesses category Neuberg 1999). Importantly, categorization is through whichgroup motives giveriseto outH accompanied by stereotyping, attribution bias, comesthatpreserve groupinterests, andthey bias.These,in turn, introduce cannot explain thevariation inemployment disH andevaluation biasesintoourperceptions, crimination acrosscontemporary workplaces. l sex,race,andethnic recollections, andevaluations of As a result, theyhavenotproven fruitful in interpretations, others. These biases are cognitive rather than identifying remedying mechanisms. I should notethatmostofmypastresearch theseprocesses as "normal," I mean assumes thatintergroup competition prompts In describing 2

onlythatnormal mental functioning requires cogj

' In addition,theydo not generateresearch hypotheses thatarefalsifiable.

j

e

1 jr

nltlve slmpllrlcatlon.

3

In thisessayI usetheterm race as shorthand for race,color, ethnicity, andnational origin.

This content downloaded from 184.171.106.7 on Mon, 15 Apr 2013 17:08:35 PM All use subject to JSTOR Terms and Conditions

Symposia 321 motivational; in otherwords, theyoccurindeH fordifferential treatment (Hughes1945).5 pendently ofdecision makers' group interests or However, I propose thatitisprimarily through theirconscious desireto favor orharmothers categorization anditsconcomitants thatsexand (Krieger 1995:1188). racearebasesforunequal treatment. The expected outcomes of thesehabitual Inherent inthecategorization ofpeopleinto inHandoutgroups isthetendency cognitive processes areraceandsexdiscriminaH toexaggerate tion. But discrimination is not inevitable. betweenHgroup differences, whileminimizing Organizational arrangements can activateor withinHgroup differences, especiallyamong oftheoutgroup. suppress socialpsychological (Anexample ofthisis and cognitive members processes (Baronand Pfeffer "I knowanX [outgroup 1994:191).We thephenomenon: cateH gory] saidit,butI can'tremember cannot ridwork organizations ofdiscrimination whichX" untilwerecognize boththatmuchemployment[Fiske1998:372]).Conceptually, socialcategoH rization resembles thesociological conceptof buteachplays a different rolein processes, andthatit occursbecauseofwork differentiation, accounts ofdiscrimination. Whilea organizations' personnel practices. Sociologists'theoretical perspective seescategorizaH knowledge ofsocialandorganizational behavior socialpsychological and notnecessarily groupH qualifies usforthistask.After summarizing the tionas automatic serving, sociologists viewdifferentiation as a cognitive processes thatproduce employment mechanismof stratification discrimination, I propose thatsociology in the fundamental whichdominant groups preserve their twentyHfirst centuryshouldexaminehow through position (e.g.,todivideandconquer employment practices mediate whether these privileged 1979],ortojustify unequal treatment processes glverlsetoclscrlmlnatory outcomes. [Edwards [Reskin 1988] ). SocialCognitionProcessesas the ExogenousCausesofDiscrimination IngroupPreference A largebodyofresearch in cognitive psyH Categorization ismorethana dataHreduction chology suggests thattocopeina complex and devicethatourbrains usetodealwiththebarH demanding environment, peopleare"cognitiverageof stimuli to whichtheyare exposed. misers" whoeconomize Classifying bythrough peopleintoingroups categorizaH andoutgroups tion,ingroup preference, orlessautomatically stereotyping, andattriHleadsmore todistorted perH bution bias(Fiske1998:362).Theseprocesses,ceptions andbiasedevaluation ofingroup and members, andhencetodiscrimination sometimes characterized as cognitive "shortHoutgroup andBrown cuts," 1998).InHversus occur regardless ofpeople's outgroup feelings toward (Brewer defines thepoolofothers othergroupsor theirdesiresto protector membership towhom improve withwhom their they ownstatus seekequal (Fiske1998:364).4If peopleareattracted, and whoserveas theirreference unchecked, theycan produce outcomes that treatment, (BaronandPfeffer 1994).Ingeneral, "perpetrators" peoH neither intend norrecognize. group plearemorecomfortable with, havemoretrust SocialCategorization in,holdmorepositive viewsof,andfeelmore The categorization ofpeopleintoingroupsobligatedto members of theirown group and outgroups is a rapid,automatic, nonconH(Perdue, Dovidio,Gurtman, andTyler1990). scious process.By conservingcognitive As a result, peopletrytoavoidoutgroup memH resources, automatic categorization helpspeople bers, andthey favor ingroup members inevaluaH manage anenormous volume ofincoming stimH uli(Fiske1998:364,375).Inkeeping withcogH 5 Ridgeway (1997)offers a related analysis. While nitive impulses toward efficiency, categorization sheconcurs withthepsychologists whoseworkI intoinH andoutgroups often isbasedonsexand citeontheimportance ofcategorization, sheconracebecauseoftheirwidespread availability as strues categorization as an emergent property of "master statuses" thathavelongbeenthebases interaction. Although shedoesnotaddress the .

c

.

.lscrlmlnatlon

.

.

.

.

.

orlglnates

ln

.

automatlc

..

cognltlve

.

4

Fora demonstration, taketheImplicit Association Tests for racism,sexism,and ageismat www.yale.edu \implicit \ (Greenwald andBanaji 1999)

effect ofsexcategorization except through interactionalprocesses, herconclusions on theconsequences ofcategorization resemble someofthose reviewed here.Shealsoprovides a useful account oftheeffect ofgender categorization ongender statusbeliefs.

This content downloaded from 184.171.106.7 on Mon, 15 Apr 2013 17:08:35 PM All use subject to JSTOR Terms and Conditions

322 Symposia

tionsandrewards (Brewer andBrown 1998:567; retention, andrecallofinformation aboutother Fiske1998:361).Thus,at leastinthelab,the people" (Krieger 1995:1188). unequal treatment associated withgroup memH The cognitive processes involved in stereoH bership results moreoften from ingroup preferHtypingmakestereotypes tenacious.People encethanoutgroup antipathy. unconsciously pursue, prefer, and remember Givenwhitemen'spredominance in many "information" thatsupports theirstereotypes workplaces, minorities' andwhitewomen's remembering staH (including eventsthatdid not tusasoutgroup members probably discount, andforget contributes to occur),andignore, inforH thedevaluation ofjobsthatarepredominantly mation thatchallenges them(Fiske1998).From femaleand predominantly ofsocialcognition minority. This thestandpoint theory, stereoH account ofdevaluation areadaptive: Peopleprocess suggests thatweshould types information totheir observe theovervaluation stereotypes morequickly ofmen'sandwhites' thatconforms information, activities in settings and theyare in whichmenandwhites thaninconsistent to stereotype aretheingroup whentheyareunder asa jobHlevel phenomenon. morelikely timepressure, partly becausestereotyping conH Stereotyping serves mental resources (Fiske1998:366;Fiske Stereotypes are unconscioushabits of etal. 1999:244).Research onpeople's efforts to thought thatlinkpersonal attributes to group suppress stereotypes is relevant.6 In one study, membership. Stereotyping isan inevitable conH subjects instructed toavoidsexist statements in comitant of categorization: As soon as an a sentenceHcompletion taskcouldcomply when observer noticesthata "target" belongs to a they hadenough time, butwhenthey hadtoact stereotyped group(especially an outgroup),quickly thestatements theyconstructed were characteristics thatarestereotypically linked to moresexist thanthoseofsubjects whohadnot thegroup areactivated intheobserver's mind, beentoldto avoidmaking sexiststatements. evenamong people whoconsciously reject thestereo- Andaccording toa comparison ofsubjects who types(Bodenhausen, Macrae, andGarst1998). wereandwerenotinstructed tosuppress stereoH To appreciate theimportance ofstereotyping for types, theformer couldrefrain from expressing discriminatory outcomes, itishelpful todistinHstereotypes, but in a "rebound effect," they guish descriptive andprescriptive stereotypes. expressed stronger stereotypes in subsequent Descriptive stereotypes, whichcharacterizejudgments thandidsubjects whohadnottried howgroup members are,influence howweperH to suppress theirstereotypes in thefirst place ceive othersand interpret theirbehavior. (Bodenhausen etal. 1998:326). Descriptive stereotyping canprecipitate discrimH Bias ination becauseitpredisposes observers toward EvaluationBias and Attribution StereotypeHbased expectations and ingroup interpretations thatconform tostereotypes and actasdistorting lenses through which blinds themtodisconfirming possibilities (Fiske favoritism assess others' performance andaccount 1998:367),especially whenthebehavior that observers successes andfailures (Crocker, Major, observers must makesenseofissubject tomultiHfortheir stereotypes pleinterpretations (e.g.,sheworked latebecause andSteele1998:539). Descriptive observers' expectations and hencethe women arehelpful, rather thansheworked late affect theyconstruct. Whentheactions because shewants a promotion). Thus,descripHexplanations conform toourexpectations, wetend tivestereotypes distort observers' impressions of ofothers theirbehavior to stable,internal thebehavior ofmembers ofstereotyped groups. to attribute propensities (e.g.,ability), whilewe attribute Prescriptive stereotypes are generalizations withourstereoH abouthowmembers ofa group aresupposed tobe, actionsthatare inconsistent expectations tosituational (i.e.,exterH basedusually on descriptive stereotypes ofhow typeHbased nal) or transient factors (e.g., task difficulty, luck, theyare.Thesenormative stereotypes serveas Inthisway, stereotypeHbased expectaH standards against which observers evaluate othH oreffort). tionsgiverisetobiasedattributions. ForexamH ers'behavior. Bothdescriptive andprescriptive thatmenaregoodat stereotypes influence whatweremember about ple,giventhestereotype maletasks, competent performance others andtheinferences wedrawabouttheir customarily behavior (Heilman1995:6). Thus,stereotypes serve as"implicit theories, biasing inpredictable When"attentional resources" are limited, stereowaystheperception, interpretation, encoding, typing increases (Fiskeet al. 1999:237). 6

This content downloaded from 184.171.106.7 on Mon, 15 Apr 2013 17:08:35 PM All use subject to JSTOR Terms and Conditions

Symposia 323 bymendoesn'trequire an explanation; men's lydosothansubordinate glOUp members areto failures do, however, and observers tendto stereotype members ofdominant glOUpS (Fiske attribute theseunexpected outcomes to situaH etal. 1999:241).In addition, under conditions tionalfactors suchas badluckorlackofeffort,ofperceived threat, themorestakeobservers noneofwhich predict future failure. Incontrast,haveinthestatus quo,andhencethemoreto women arestereotypically notexpected to do lose,themorelikely theyaretostereotype outH Goodwin, and Fiske1998: wellatcustomarily maleendeavors, so explainHgroups(Operario, ingtheir failure iseasy:Theylacktherequisite 168).The senseofentitlement thataccompaH ability (aninternal trait) andhencearelikely to niesdominantHgroup status islikely togivedomH failinthefuture. Incontrast, their successes are inantgroupmembers particular confidence in unexpected, sothey must haveresulted from sitH theirstereotypes. Thispropensity is reinforced uational factors thatdo notpredict future sucH bythefactthatpowerful observers actively seek cess(SwimandSanna1996;Brewer andBrown information thatconfirms their stereotypes and 1998:560). disregard disconfirming information. However, Ingroup preference andoutgroup derogationpriming the powerful withegalitarian values lead to similar attribution processes. Because leadsthemto paycloserattention to informaH outgroup stereotypes observers expectingroup members to succeed tion that contradicts and outgroup members to fail,theyattribute(Operario et al. 1998:172-73).Finally, memH ingroup success andoutgroup failure tointernal bersofhighHstatus ingroups showmorebiasin factors, andingroup failure andoutgroup success favor ofingroup members thando members of to situational factors. Observers also tendto lowHstatus ingroups (Brewer and Brown1998: characterize behavior thatis consistent with 570). theirexpectations in abstract terms andunexH pected behavior inconcrete terms. Forexample, The ProximateCausesofDiscrimination to socialcognition theory, bias giventhesameact arriving latefora meetH According and discrimination result from the individualH ing an observer wouldrecallthatan ingroup processes summarized above. member wasdelayed, butthatanoutgroup memH levelcognitive psychologists agree,however, that beris a tardy person. Oncea behavior hasbeen Cognitive biasescanbecontrolled (Fiske1998:375). interpreted andencoded intomemory, itisthe these causes ofdiscrimination are interpretation, nottheinitial behavior, towhich Thus,theproximate factors thatpermit orcounter the peoplehaveready access(Krieger 1995:1203). thecontextual ofthesehabits ofthebrain. Thecourse I Thus,observers would predict thattheoutgroupeffects inthetwentyHfirst century is member, butnottheingroup member, would be urgeforsociology to investigate howorganizational practices can tardy inthefuture. checkthesefactors. Experimental research on Powerand Cognitive Biases contextual factors thatappearto minimize the Up tothispoint, I havetreated discriminaH likelihood ofstereotyping anditsbiasing effects a starting pointforthisenterprise. tionmotivated bystatus politics or antipathyprovides and discrimination thatresults automatically Thesefactors includeconstructing heterogeH from unconscious cognitive processes as ifthey neousgroups, creating interdependence among weremutually exclusive. Although cognition ingroup andoutgroup members, minimizing the researchers havegivenrelatively little attentionsalience ofascribed status dimensions inpersonH totheir relationship, a handful ofexperimental nel decisions, replacing subjective datawith decisionmakers studies indicate thatpowerdifferentials condiH objectivedata,and making tionthesecognitive processes. Thesestudies accountable fortheir decisions. haveshown thatalthough thepropensity tocatH Ofcourse, organizations' ability toapplythe egorlze lsunlversa ,occupylng a posltlon otpowH findings fromexperimental research to the ermayprompt peopletoinvest extra effort into exogenous causesofdiscrimination depends on categorizing others(Goodwin, Operario, and theexternal validity oftheexperimental results Fiske1998). In addition, poweraffects the described above.Workorganizations arevastly degree towhich peopleactonthepropensity to morecomplex thanlaboratory experiments. In work organizations arehothouses that stereotype. Peoplecan'tafford tostereotype othH particular, power andstatus differences. Thus,the ersonwhom they depend because theyneedto nurture assessthemaccurately, buttheycan afford to first taskforsociologists perhaps incollaboraH is todetermine stereotype subordinate groups andaremore likeH tionwithsocialpsychologists .

.

.

r

This content downloaded from 184.171.106.7 on Mon, 15 Apr 2013 17:08:35 PM All use subject to JSTOR Terms and Conditions

324 Symposia

whether thecognitive processes thatI have gorization areteams, divisions orbranches, job reviewed operate thesamewayinwork organiHgroups, andtheorganization itself.8 Withrespect zations asthey dointhelab.7 Ifthey do,thenext to the lastof these,the moreorganizations stepis to investigate theproximate causesof emphasize organizational culture, theeasierit employment discrimination: thesocial,contexHshould be toexpandtheingroup toencompass tual,andorganizational mechanisms thatsupH all employees. Organizations canmaximize the pressor exacerbate theseexogenous causes. impact ofheterogeneous groupings byreinforcH BelowI summarize theexperimental research. I ingingroup identification through taskinterdeH hopereaders viewthissummary as a set of pendence, job rotation, and othercollective propositions thatspecify howorganizations can activities. Sociology shouldplacehighon its prevent nonconsclous cognltlve processes trom agendaforthetwentyHfirst century a study of culminating inemployment discrimination. organizations' ability to minimize ascriptively basedcategorization byemphasizing other cateH Heterogeneity ofWorkGroups gories and the impact of such reHengineered Categorization is toofundamental to cogniH onstereotypes andattribution bias. tiveprocessing to be prevented, and ingroup groups favoritism isremarkably hardtoeradicate, even Interdependence forpeoplewitha material stakein endingit Intergroup contactthatexposespeopleto (Brewer andBrown1998:566). ButorganizaHindividuating information aboutoutgroup memH tions candiscourage thecategorization ofpeople berschallenges outgroup stereotypes, and hence basedontheir sex,race,andethnicity, andthus reduce bias.Butforintergroup contact to reducesex and racediscrimination. Creating should change ingroup members' perceptions, the latter workgroups anddecisionHmaking groups that aboutoutgroup areheterogeneous withrespect to theseascripHmustattendto information members (Goodwin et al. 1998: 681).TheconH tivecharacteristics should suppress ingroup prefH ditions that should foster such attention areenuH erenceandoutgroup derogation, stereotyping, merated in the contact hypothesis. This andconcomitant biaspersonnel decisions. (In argues thatintergroup contact alters addition, ifneither ingroup noroutgroup memH hypothesis ingroup members' attitudes onlyifthegroups bersnumerically dominatedecisionHmaking come together with a common goal,haveinstiH groups, personnel outcomes arelesslikely tobe tutional support for their joint enterprise, and linked togroup membership.) contact inequalHstatus Ofcourse, organizations whoseworkgroups havecloseandsustained (Brewer andBrown 1998:576-78). arewellintegrated bysexandracearenotthe positions Thelogicofthecontact hypothesis assumes oneslnWh1Ch C lSCrlmlNatlOn 1S a serlous pro members' interdependence with lem.Organizations in whichworkgroups are thatingroup outgroup members encourages the former to segregated maybe ableto createsuperordinate counterstereotypic information aboutthe identities (i.e.,more inclusive ingroups) thatare notice andthusto form moreindividuated and independent ofsexandrace(Brewer andBrown latter 1998: 583). In laboratory By the same logic, experiments,accurateimpressions.9 researchers can artificially members' createcategories dependence to ingroup on outgroup whichsubjects becomeattached, evenon the members shouldmotivate theformer to seek most trivial basis, soworkers should bereceptiveaccurate information aboutthelatter. Basedon torecategorization basedoncharacteristics that thisexpected association, Goodwin andhiscolH are contextually salient(Fiske1998: 361). leagues (1998:694)contended thatsupervisors Organizations maybe abletocreatesuchcateH whoknowthattheirsalaries dependon their goriesby usingexisting functional categories forthese categories tosupplant thatarerelevant sex-and andhencecognitively availH ofcourse, categorization, category membership ableto workers, ortheymaybe ableto create race-based beassociated withsexorrace. newcategories thatsupplant ascriptively defined cannot mayinitially resistthesecollective categories as thebasisforthecognitive processH9 Workers arrangements, however. In addition, whenthe esdiscussed here.Among possible basesofcateH context .

.

.

.

.

r

.

8

7

changes, theformer groupings arelikely to re-emerge (Brewer andBrown 1998:582-83).In Bielby( 2000) believestheyshouldbe even other words, intergroup contact isnota quickfix; stronger inwork organizations thanin laboratory itmakes a difference onlywhenitoccurs through experiments. a permanent transformation oftheworkplace.

This content downloaded from 184.171.106.7 on Mon, 15 Apr 2013 17:08:35 PM All use subject to JSTOR Terms and Conditions

Symposia 325 themore subject itistomisinterpreH subordinates' productivity or evaluations will responding, judgetheir subordinates moreaccurately. tation. In worksettings, thismeansthatrecolH Intergroup competition basedonstatus charH lectionsand evaluations thatare basedon acteristics is counterproductive becauseit unstructured observations areparticularly vulH encourages eachgroup tostereotype theother. nerableto race or sex bias (Fiske,Bersoff, Fiskeandhercolleagues (1999:241A2)specuH Borgida, Deaux, and Heilman 1991). latedthatthishappens because group membersOrganizations shouldbe ableto minimize race devote most of theiravailablecognitive and sex biasin personnel decisions byusing resources to obtaining accurateinformation objective, reliable, andtimely information that abouttheirteammates, rather thanabouttheir isdirectly relevant tojobperformance (Heilman opponents.l° Thus,cooperative interdependence 1995).Forobjective measures tominimize interH can reducestereotyping, whilecompetitivegroup bias,organizations must provide a detailed interdependence increases it. specification ofall performance criteria along withprecise information foreachcandidate for Salience eachcriterion (Krieger 1995:1246).Employers Anything thatfocuses observers' attentionshould further reduce attribution errors byrouH ona stereotyped category "primes" stereotyping, tinely collecting concrete performance dataand anditdoesso without theobserver's awarenessimplementing evaluation procedures in which (Heilman 1995;Fiske1998:366).Forexample, evaluators relyexclusively onthesedatawithout menwhowereprimed withstereotypic stateH attributions explaining them. ments aboutwomen weremorelikely to aska femalejob applicant"sexist"questions and Accountability exhibit sexualized behavior (andit tookthem The biasing effects ofstereotypes andother longer thannonprimed mento recognize nonH cognitive distortions on evaluative judgments sexistwords; Fiskeet al. 1999).Thus,a comH arereduced whendecision makers knowthat mentaboutpregnancy, a sex discrimination theywillbe heldaccountable forthecriteria lawsuit, ordiversity immediately before a comH theyusetomakedecisions andfortheaccuracy mittee evaluates a female jobcandidate islikely oftheinformation uponwhichtheybasetheir toexacerbate sexstereotyping intheevaluation decisions (SalancikandPfeffer 1978;Tetlock (Heilman1995).The process ofpriming may 1992;Tetlock andLerner 1999).l2 Ifevaluators meanthatinjunctions toa search committee to knowthattheywillbe heldaccountable for lookclosely atfemale orminority candidates can theirjudgments before beingexposedto the backfire, tainting theevaluations ofwomen and information onwhich theywillbasetheir judgH minorities. Similarly, whenwomen andmenare ment,accountability not onlyreducesthe interacting andgender isrelevant topurpose of expression ofbiases, italsoreduces biasinnonH the interaction, culturalgenderstereotypesconsciouscognitive processes, such as the become"effectively salient"(Ridgeway 1997: encoding ofinformation (Tetlock1992).The 221).llOrganizational contexts can alsomake benefits of accountability vanishundertime category membership salient. A highly skewed pressure, however (TetlockandLerner 1999). sexorracecomposition ina work group islikely Indeed,timepressure, mental"busyness," and to activatestereotypes (Bodenhausen et al. information overloadall common incontemH 1998:317). poraryworkorganizations exacerbatethe effects of stereotypes on judgment andmemory FormalizedEvaluationSystems (Bodenhausen et al. 1998: 319). Stereotyping and its concomitants distort Theprocesses underlying theimportance of howweinterpret thebehavior ofoutgroups, and accountability no doubt help explain how thevaguertheinformation to whichwe are antidiscrimination andaffirmative actionlaws andregulations increase jobaccessforpeopleof women. Goals,timetables, and '° Interpersonal (i.e., one-on-one)competitioncolorandwhite reduces stereotyping, because competitors' success monitoring all effective affirmative action depends on havingaccurate information about mechanisms-hold organizations responsible for their opponent. " As Ridgeway (1997:221) observed, thediffuse nature ofsexstereotypes makes themrelevant in many situations.

12

See DiTomaso (1993)fora description ofXerox's successful useofaccountability.

This content downloaded from 184.171.106.7 on Mon, 15 Apr 2013 17:08:35 PM All use subject to JSTOR Terms and Conditions

326 Symposia

sex-andrace-balanced hiring andthesexand as on structures thatpreserve a discriminatory race composition of theirjob assignmentsstatus quo.l3 Thisapproach assumes thatdomi(Reskin1998). Hypothetically, organizations nantgroup members intentionally creatework can achievesimilar results through programsstructures and organizational arrangemellts thatmakedecision makers atalllevelsresponsi-whosepurpose is to preserve orenhancetheir ble forensuring thattheirdecisions are not position. Among many examples I couldoffer is tainted byingroup preference andfortheout- the widespread and deliberate exclusionof comesofthosedecisions. minorities and womenfrompoliceand fire departments (Chetkovich 1997;alsoseeReskin Conclusions 2000).Whenpeople's group position motivates Allcommon socialscientific theories ofdis- them todiscriminate, exclusionary organizationcrimination, as well as the dominant legal al practices aresuperficial causesofdiscriminaapproachto discrimination (Krieger1995, tion,andtheyrequire different interventions.l4 1998),locateitssource inintrapsychic process- Organizations, I haveargued, can reducedises suchas prejudice, ignorance, thesenseof crlmlnatlon lssulng tromnonconsclous COglllthreat, andthedesireto maintain orimprove tiveprocesses. Remedying discrimination that one's position.They differ, however,in results from dominant group members' deliberwhethertheyview the consequencesof ateconstruction ofexclusionary personnel pracintrapsychic processes as motivated or auto- ticeswillrequire race-and gender-conscious matic.Theories thatassumethatdiscrimina-interventions, including formal charges ofsextionismotivated byantipathy toward orfearof /race-based discrimination, collectiveaction another group viewdiscrimination as an aber- organized onthebasisofstatus groups,ls orinterrationwithina generally fairreward system vention byregulatory agencies, including sex(Black1989).According to socialcognition andrace-conscious remedial affirmative action. theory, incontrast, thebasiccognitive process- Thesamecharacteristics sex,race, ethnicity es throughwhicheveryone'sbrain sorts aretheprimary basesofbothautomatic cogthrough datadistort allourperceptions, biasall nitivecategorization and socialstratification; ourattributions, andinduceall ofus to favor indeed,theircentrality in each process reiningroup members. Laissez-faire decision making forces themin the otherprocess. Moreover, in workorganizations and otherdomains, automatic cognitive categorization andrace-or including schools, voluntary organizations, and sex-based socialstratification have the same thefamilytransforms thesebiasesintodis- result: privileging ingroup members whoareusucrimination against outgroup members. Ifthe ally white males of Europeanancestry. cognitive processes thatleadtodiscrimination Moreover, bothcognitive-based and conflictare universal, as experimental evidencesug- group-based processes comprise "countless small gests,thentheycause a hugeamountof acts by a changing cast of characters, . . . that employment discrimination thatis neither incrementally and consistently limit the intended normotivated byconscious negative employment prospects of one group of workers feelings toward outgroups. Andtheorganizawiththoseofanother" (Nelsonand tionalpractices thatdetermine howtheinput compared Bridges 1999: 243). Individually, either process of individuals contribute to personnel decileads to the accumulation of advantages and dissions, andhenceprecipitate, permit, orprevent theactivation ofcognitive biases, aretheproximatecausesofmostemployment discrimina- Atleastonesocialpsychological theory, "realistic tion. group conflict" theory, attributes discrimination to Although group I and otherssuspectthatmost conflict (Brewer andBrown 1998:565). Fordiscussion ofsuperficial causes, seeLieberson employment discrimination originates in the (1985)orReskin (1988). cognitive processes I have summarized, we in challenging intentional, biasS should notlosesight ofthefactthatdiscrimina-t5 Forexample, based racial discrimination by Shoney's tionalsoresults from conscious actions thatare Restaurants, theNAACP LegalDefenseFund motivated byignorance, prejudice, orthedelibpublicized an 800number forcomplaints against erateefforts bydominant group members topreShoney's, generating boththebasisfora class servetheir privileged status. Twentieth-centuryactionlawsuit andsupporting evidence from white sociologyhas focusedon theseconscious supervisory employees whosupported thelawsuit processes ofexploitation andexclusion, as well (Watkins 1993). ,^

13

14

This content downloaded from 184.171.106.7 on Mon, 15 Apr 2013 17:08:35 PM All use subject to JSTOR Terms and Conditions

Symposia 327 in thelast Mostofourprogress biases pline'scredit. bothcognitive Sometimes advantages. has century severaldecadesof thetwentieth discrimination orconflict-based andprejudiceextent discrimination's Ridgeway'sbeenin documenting effects. withreinforcing areatwork, in Wehavemadelessheadway thiswithrespect andpersistence. illustrates (1997:227)analysis andverylittlein itspersistence be so understanding willgender "Onlyoccasionally togender: itbecausewehave outhowto reduce thatmenwillactself- figuring salientin thesituation whydiscriminahowo1* theorized theirinterests[,notcorrectly as mento preserve consciously ofitresults thatmuch I haveargued ofgen- tionoccurs. activation background but]therepeated If I'm processes. cognitive nonconscious from interactions, derstatusovermanyworkplace causeis theorganizathenitsproximate in subtleormoresubstantialright, behavior biasing thatpermitor preventit. ofmenactingin tionalpractices theeffect produces degrees, the contain howandwhenorganizations menfeel Exactly evenwhenmany interest, gender their on the be high should biases cognitive of effects sex." totheir loyalty nospecial century. forthetwenty-first agenda ofthedomi- discipline's As I saidabove,somemembers people against discriminate actively nantgroup as well References origin, race,sex,national basedontheir 1994."TheSocial Pfeffer. N. andJeffrey Baron,James orien- Psychologyof Organizationsand Inequality." suchasage,sexual characteristics asother HereI havequesandreligion. weight, tation, 57: 190-209. Quarterly SocialPsychology thatI andmanyother Bielby, William T. 2000. "How to Minimize tionedtheassumption ofworkplace Workplace Gender and Racial Bias." to thestudy brought sociologists 29: 120-29. Sociology Contemporary thatmost century: in thetwentieth inequality New York: Justice. actions Black,Donald. 1989.Sociological thepurposive from results discrimination Press. University Oxford whoseekto premembers group bydominant Whiledomi- Blalock, Hubert M. 1967. Towarda Theoryof privileges. their serveandexpand NewYork:Wiley. Relations. Minority-Group nant groupmembersbenefitfromsuch Englewood . 1982.RaceandEthnicRelations. ofraceandsexin thesalience discrimination, Hall. N]: Prentice Cliffs, processesBlumer, andincognitive society contemporary as a Senseof 1958."RacePrejudice Herbert. allows andstereotyping suchas categorization 1: 3-7. Review Sociological Pacific GroupPosition." with- Blumrosen, tobenefit members group mostdominant John]. Miller, W., MarcBendick, Alfred Byassuming and Ruth Blumrosen.1998. ''Employment having totakeanyaction. outtheir againstWomen in Washington ofpurpo- Discrimination theresult islargely thatdiscrimination Project Discrimination track forreduc- State,1997."Employment weareonthewrong siveaction, University Rutgers N]: 3. Newark, Number Report lose routinely Plaintiffs ing discrimination. SchoolofLaw. becausetheycannot Bobo, lawsuits discrimination and D., MelvinL. Oliver,A. Valenzuela, Larry todiscrimi- ]. H. Johnson.2000. Prismatic intended employer provethattheir Race, Metropolis: see Krieger Segregation them(forexamples, nateagainst NewYork: inLosAngeles. andInequality whoshare Andemployers, 2000).16 1995;Reskin RussellSageFoundation. deliberateBodenhausen, involves GalenV.,C. NeilMacrae,andJennifer ourviewthatdiscrimination in Thoughtand Deed: status, Garst.1998."Stereotypes oftheir peoplebecause toharm attempts SocialSCognitive Origins of Intergroup and chargesimplausible finddiscrimination " Pp. 311-35 in Intergroup Discrimination. that rejectthemoutofhand.Therecognition Behavior,edited by Intergroup and Cognition cogniuniversal from stems often discrimination ConstantineSedikides,John Schopler, and lessresis- ChesterA. Insko.Mahwah,N]: LawrenceErlbaum maymakeorganizations tiveprocesses and more of discrimination tantto charges Associates. prac- Brewer,MarilynB. and Rupert] . Brown.1998. theiremployment tomodifying receptive biases ofcognitive theeffect ticesto remove of Pp.554-94inHandbook Relations." "Intergroup editedby D. T. Gilbert,S. T. SocialPsychology, women. peopleofcolorandwhite against NewYork:McCrawSHill. under- Fiske,andG. Lindzey. toexpose, oftrying history Sociology's C. 1997.RealHeat:GenderandRaceirv istoourdisci- Chetkovich, discrimination andreduce stand, 16

N]: Rutgers NewBrunswick, theUrbanFireService. Press. University BrendaMajor,and ClaudeSteele. Jennifer, associated Crocker, ofthelegallimitations Fora discussion of Pp. 504-53inHandbook 1998."SocialStigma." as ofdiscrimination conception withthestandard editedby D. T. Gilbert,S. T. SocialPsychology, to harmpeoplebasedon their actionsintended NewYork:McCrawSHill. Fiske,andG. Lindzey. (1995). seeKrieger sex,race,orcolor,

This content downloaded from 184.171.106.7 on Mon, 15 Apr 2013 17:08:35 PM All use subject to JSTOR Terms and Conditions

328 Symposia DiTomaso,Nancy. 1993. Notes on Xerox Case: Nelson,RobertL. and WilliamP. Bridges.1999. BalancedWorkForceat Xerox.Unpublished. Legalizing GenderInequality: Courts,Markets, and Edwards, Richard.1979.Contested UnequalPayforWomenin America.Cambridge: Terrain. NewYork: BasicBooks. Cambridge University Press. Fiske,Susan T. 1998. "Stereotyping, Prejudiceand Operario, Don,StephanieA. Goodwin,andSusanT. Discrimination." Pp. 357-411 in Handbookof Fiske.1998."PowerIs Everywhere: SocialControl SocialPsychology, editedby D. T. Gilbert,S. T. andPersonalControlBothOperateat Stereotype Fiske,andG. Lindzey. NewYork:McCrawSHill. Activation,Interpretation, and Inhibition." Pp. Fiske,SusanT., DonaldN. Bersoff, EugeneBorgida, 163-75inStereotype Activation andInhibition, editS Kay Deaux, and MadelineE. Heilman. 1991. ed by RobertS. Wyer.Mahway,N]: Lawrence "Social Science Researchon Trial. Use of Sex Erlbaum Associates. Stereotyping Researchin Price Waterhousev. Perdue,C. W.,]. F. Dovidio,M. B. Gurtman, andR. Hopkins." American Psychologist 46: 1049-60. B. Tyler. 1990. " 'Us' and 'Them': Social Fiske,SusanT., MonicaLin,andStevenL. Neuberg. Categorization andtheProcess ofIntergroup Bias." 1999."The Continuum Model:Ten YearsLater." Journal of Personality and Social Psychology 59: Pp. 231-54 in Dual ProcessTheoriesin Social 475-86. Psychology, editedbyShellyChaikenand Yaacov Reskin,BarbaraF. 1988."Bringing theMen BackIn: Trope.NewYork:Guilford Press. Sex Differentiation and the Devaluation of Fix,MichaelandRaymond ]. Struyk, eds.1993.Clear Women'sWork."Gender S Society 2: 58-81. and ConvincingEvidence. Measurementof . 1998. The Realities of Affirmative Action. Discrimination inAmerica. Washington, DC: Urban Washington, DC: American Sociological Institute. Association. Goodwin, StephanieA., Don Operario, . 2000. "Employment andSusanT. Discrimination and Its Fiske.1998."Situational Powerand Interpersonal Remedies.''Forthcoming in Handbookon Labor Dominance Facilitate BiasandInequality." Journal MarketResearch, editedby Ivar Bergand Arne ofSocialIssues54: 677-98. Kalleberg. NewYork:Plenum. Greenwald, Anthonyand MahzarinBanaji. 1999. Ridgeway,Cecilia. 1997. "Interactionand the "Implicit Association Test." www.yale.edu \ Conservation of GenderInequality." American implicit \ . Sociological Review 62: 218-35. Heilman,M. E. 1995. "Sex Stereotypes and Their Salancik, Gerald R. and Jeffrey Pfeffer.1978. Effects in the Workplace:What We Knowand "Uncertainty, Secrecy, andtheChoiceofSimilar WhatWe Don'tKnow."Journal ofSocialIssues10: Others." SocialPsychology 41: 246-55. 3-26. Swim,JanetK. and Lawrence ]. Sanna. 1996."He's Hughes, Everett C. 1945. ''Dilemmas and Skilled, She's Lucky: A MetaSAnalysisof Contradictions of Status."American Observers' Attributions forWomen'sand Men's Journalof Sociology 50: 353-59. Successesand Failures."Personality and Social Jackman, MaryR. 1994.TheVelvetGlove.Berkeley: Psychology Bulletin 22: 507-19. University ofCalifornia Tetlock, Phillip E. 1992. "The Impact of Press. Kirschenman, Joleenand KathrynM. Neckerman. Accountability on Judgment andChoice:Toward 1991."'We'd Love to HireThembut....': The a Social ContingencyModel." Advancesin Meaningof Race forEmployers.'' Pp. 203-34 in Experimental SocialPsychology 25: 331-76. The Urban Underclass, edited by Christopher Tetlock,PhillipE. andJennifer S. Lerner.1999."The Jencksand Paul Peterson.Washington,DC: Social Contingency Model:Identifying Empirical Brookings Institution. andNormative Boundary Conditions on theErrorS Krieger, LindaHamilton.1995."TheContents ofOur andSBiasPortrait ofHumanNature."Pp. 571-85 Categories:A Cognitive Bias Approach to inDualProcess Theories inSocialPsychology, edited Discrimination and Equal Employment byShellyChaikenandYaacovTrope.NewYork: Opportunity." Stanford LawReview Guilford 47: 1161-248. Press. . 1998. "CivilRightsPerestroika: Intergroup Tilly, Charles. 1998. DurableInequality. Berkeley: RelationsafterAffirmative University Action."California ofCalifornia Press. LawReview 86: 1253-1333. TomaskovicSDevey, Donald.1993.Gender andRacial Lieberson, Stanley.1985.Making It Count.Berkeley: Inequality atWork:TheSources andConsequences of University ofCalifornia Press. JobSegregation. Ithaca,NY: ILR Press. Martin,PatriciaYancey.1992."GenderInteraction Watkins,Steve.1993."Racismdu jourat Shoney's. and Inequality TheNation, in Organizations." October18. Pp. 208-31 in Gender Interaction andInequality, editedbyCecilia Ridgeway. NewYork:SpringerSVerlag.

This content downloaded from 184.171.106.7 on Mon, 15 Apr 2013 17:08:35 PM All use subject to JSTOR Terms and Conditions