10.1177/0146167203261884PERSONALITY AND SOCIAL PSYCHOLOGY BULLETINKray et al. / STEREOTYPE ACTIVATION IN NEGOTIATIONS

Stereotype Reactance at the Bargaining Table:The Effect of Stereotype Activation and Poweron Claiming and Creating Value

Laura J. KrayUniversity of California, Berkeley

Jochen RebUniversity of Arizona

Adam D. GalinskyLeigh ThompsonNorthwestern University

Two experiments explored the hypothesis that the impact of acti-vating gender stereotypes on negotiated agreements in mixed-gender negotiations depends on the manner in which the stereo-type is activated (explicitly vs. implicitly) and the content of thestereotype (linking negotiation performance to stereotypicallymale vs. stereotypically female traits). Specifically, two experi-ments investigated the generality and limits of stereotypereactance. The results of Experiment 1 suggest that negotiatedoutcomes become more one-sided in favor of the high power nego-tiator when masculine traits are explicitly linked to negotiatoreffectiveness. In contrast, the results of Experiment 2 suggestthat negotiated outcomes are more integrative (win-win) whenfeminine traits are explicitly linked to negotiator effectiveness. Intotal, performance in mixed-gender negotiations is stronglyaffected by the cognitions and motivations that negotiators bringto the bargaining table.

Keywords: stereotype; gender; negotiations; reactance; threat; power

The current zeitgeist is one of political correctness,which touts equality between various races, genders,ages, and sexual orientations. Individuals and institu-tions that are insensitive to this cultural movement aresubject to negative consequences, ranging from ostra-cism and scorn to potential litigation. Although it hasbeen well documented that people are affected by cul-tural stereotypes (Greenwald & Banaji, 1995; Greenwaldet al., 2002), the spirit of political correctness mandatesthat discussion of the content of these stereotypesremains muted. Explicit acknowledgements of gender

differences stand out saliently against the background ofsocial correctness. Given this cultural climate, an inter-esting question to consider is how the explicit linkage ofgender to ability affects the thoughts, motivations, andbehaviors of individuals.

In this article, we build on previous research that doc-umented the ironic effects of explicitly and blatantlylinking gender to negotiation ability (Kray, Thompson,& Galinsky, 2001). In general, when individuals perceivea threat to their freedom, they tend to react against it byexerting their freedom more forcefully than they other-wise would (Brehm, 1966). In the negotiation domain,this reactance process has been shown to occur throughthe activation of negative gender stereotypes. Specifi-cally, Kray et al. (2001) found that informing negotiatorsthat stereotypically masculine traits predict perfor-mance at the bargaining table and that these traits differby gender led to a counterintuitive outcome: Femalenegotiators outperformed their male counterparts.Thus, instead of confirming and validating the connec-tion between gender and negotiation ability, women viti-ated the negative stereotype by reacting against it. More

399

Authors’ Note: This research was supported by two grants from the Na-tional Science Foundation (SES-0073682 awarded to Laura Kray andSES-9870892 awarded to Leigh Thompson). The authors would like toacknowledge the helpful comments of David Levine on an earlier ver-sion of this article. Correspondence concerning this article should beaddressed to Laura Kray, Haas School of Business, 545 Student ServicesBuilding, #1900, Berkeley, CA 94720; e-mail: kray@haas.berkeley.edu.

PSPB, Vol. 30 No. 4, April 2004 399-411DOI: 10.1177/0146167203261884© 2004 by the Society for Personality and Social Psychology, Inc.

specifically, women who were told that it is men who havethe upper hand at the bargaining table identified withcounterstereotypic traits and set higher and moreaggressive goals. The ability to marshal empoweringcognitions in the face of what might seem to be the worstperformance conditions is all the more surprising giventhat under most conditions, men do outperform womenat the bargaining table (Kray et al., 2001; Stuhlmacher &Walters, 1999; Walters, Stuhlmacher, & Meyer, 1998).

Kray et al. (2001) suggested that it is the explicitendorsement of negative gender stereotypes and thesubsequent reactance against this proclamation thatleads women to outperform men. Indeed, when theendorsement was implicit, men reaped better outcomesthan did women. An important question that we addressin more detail in this article is how the explicit activationof gender stereotypes affects performance. The findingsfrom Kray et al. (2001) suggest that explicitly activatingthe masculine stereotype in mixed-gender negotiationsprompts negotiators to assess their strengths and weak-nesses and, in so doing, become more focused on theirrelative power within the negotiation context. More gen-erally, we propose that negotiators attend more to thecontent of an activated stereotype when it is explicitlylinked to gender, resulting in a concomitant increase inthe display of the stereotypical traits positively associatedwith performance, compared to when the connectionbetween gender and performance is left implicit. For thenegotiator disadvantaged by the stereotype this involvesacting in a counterstereotypic manner. The mention ofgender serves as a trigger that guides subsequent behav-ior. Therefore, when the masculine stereotype is explic-itly activated, negotiators should become more aggres-sive, contentious, and focused on claiming resources.But when the feminine stereotype is explicitly activated,negotiators should become more cooperative andfocused on creating resources.

To test these hypotheses, we varied whether a genderstereotype was activated implicitly versus explicitlywithin mixed-gender dyads in two experiments. We alsomanipulated the power distribution within the dyads,varying whether the male or the female negotiator hadthe power advantage. If the content of an activated ste-reotype matters, then the extent to which power is exer-cised at the bargaining table should vary after theexplicit connection of a masculine stereotype comparedto a feminine stereotype. In Experiment 1, we varied themanner in which stereotypically masculine traits werelinked to negotiation effectiveness and measured theextent to which outcomes mirrored power asymmetries,a gauge of stereotypically masculine behavior. In Experi-ment 2, we varied the manner in which stereotypicallyfeminine traits were linked to negotiation effectivenessand measured the extent to which negotiators con-

structed integrative agreements, a gauge ofstereotypically feminine behavior. Another advantage ofintroducing a power asymmetry to the bargaining tableis that it allowed us to explore the limits of stereotypereactance, which we describe in greater detail below.

Stereotype Threat Versus Stereotype Reactance

Stereotype threat describes the concern a person feelsabout confirming, as self-characteristic, a negative ste-reotype about one’s group (Steele, 1997). This threatbecomes reality because concern over confirming thestereotype produces anxiety, lowers expectations, andreduces performance; thus, it unwittingly confirms thestereotype (Spencer, Steele, & Quinn, 1999; Steele &Aronson, 1995). Kray and colleagues found that simplydescribing a negotiation as diagnostic of ability wasenough to produce a male advantage at the bargainingtable.

Because many of the traits associated with effectivenegotiators are stereotypically masculine in nature (Krayet al., 2001), subtly activating the masculine stereotypetends to lead to outcomes that confirm the stereotype ofmale dominance at the bargaining table; that is, whenpeople are told that effective negotiators are rational,assertive, unemotional, and have a high regard for theirown interests (stereotypically masculine traits), malenegotiators have an advantage over female negotiators(Kray et al., 2001). Despite the fact that gender was notmentioned in connection with these traits, the resultingdifference in aspirations, opening offers, and profitsacross gender was significant.

The male advantage can be reversed when men andwomen are told that stereotypically feminine traits areimportant determinants of negotiation success. Pro-claiming that effective negotiators “express theirthoughts verbally, have good listening skills, and possessinsight into the other negotiator’s feelings” leads to afemale advantage at the bargaining table. More specifi-cally, Kray, Galinsky, and Thompson (2002) found thatsubtly linking these stereotypically feminine traits witheffective negotiating led women to have higher aspira-tions, make more aggressive opening offers, and negoti-ate more profitable agreements for themselves than didmen.

These findings suggest that stereotype threat pro-cesses apply to both men and women in negotiations;implicitly activating negative stereotypes about one’sgroup leads to decrements in performance for the ste-reotyped group. In general, stereotype threat appearsnot to be limited to traditionally disadvantaged groups;when traits that are stereotypically connected to a tar-get’s social category are linked to negative performanceexpectations in a domain, the target will tend to feelstereotype threat and confirm the stereotype (see also

400 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

Leyens, Desert, Croizet, & Darcis, 2000; Stone, Lynch,Sjomeling, & Darley, 1999). The above pattern is consis-tent with assimilation effects, whereby the implicit andsubtle activation of a stereotype tends to lead to behav-iors that are consistent with the stereotype (Bargh, Chen,& Burrows, 1996; Wheeler & Petty, 2001). These findingssuggest that the linkage between stereotype activationand performance depends on the content of the acti-vated stereotype, that is, whether the activated traits sug-gest a masculine or feminine advantage.

But stereotype activation does not always lead to theassimilation effects characteristic of stereotype threat.Kray et al. (2001) also demonstrated that explicitly andblatantly acknowledging that the association betweenstereotypically masculine traits and effective negotiatingis linked to gender differences can ironically produce afemale advantage at the bargaining table in mixed-gender dyads. Kray et al. argued that the explicit activa-tion of the gender stereotype was perceived as a limit tothe female negotiator’s freedom and ability to perform,thereby invoking stereotype reactance. Kray et al.defined stereotype reactance as the tendency to behavein a manner inconsistent with a stereotype. Explicitlyactivating a negative stereotype resulted in stereotypereactance. This finding is consistent with previousresearch regarding contrast effects: When a perceiver’sattention is blatantly drawn to the link between a stereo-type and a social category, it can produce behaviors thatare inconsistent with the stereotype (for a related discus-sion of blatant activation and contrast effects, see alsoDijksterhuis et al., 1998; Martin, 1986; Strack, Schwarz,Bless, Kuebler, & Waenke, 1993). It appears that themanner in which stereotypes are activated determineswhether they create an assimilation effect (stereotypethreat) versus a contrast effect (stereotype reactance),which suggests that the linkage between stereotype acti-vation and performance depends on the manner inwhich stereotypes are activated.

Building on results from previous research on stereo-type reactance (Kray et al., 2001), we contend that theexplicit activation of a gender stereotype leads negotia-tors to engage in behaviors consistent with the content ofthe activated positive stereotype. In an effort to avoidbeing pigeon-holed, the negotiator disadvantaged bythe stereotype may engage in counterstereotypic behav-iors. By realizing that the stereotypical perception oftheir ability is invalid, negotiators who are the target of anegative stereotype may adjust their goals and strategiesaccordingly. For example, Kray et al. (2001) found thatthe explicit activation of the masculine stereotype ledfemale negotiators to identify with stereotypically mascu-line (counterstereotypic) traits as their key strengths anddisidentify with stereotypically feminine traits. Further-more, explicit masculine stereotype activation led

women to set higher aspirations than implicit masculinestereotype activation (Kray et al., 2001, 2002).1

This finding supports the notion that the explicit acti-vation of the masculine stereotype focuses attentiontoward one’s strengths and power in the negotiation.Reactance processes likely result from stereotypically dis-advantaged negotiators’ awareness that the stereotypicalperception of their ability is invalid in the current con-text, that both negotiators are actually on a level playingfield, or that they themselves hold a power-based advan-tage. If our understanding of the process through whichreactance occurs is accurate, then it suggests that negoti-ators should only react when they possess sufficientpower in the situation to stand their ground at the bar-gaining table. In all of the examinations of stereotypereactance to date, both negotiators had reasonablystrong alternatives to the negotiation and thus couldleverage their power. However, reactance against theexplicit activation of negative stereotypes may not occurwhen negotiators lack power. Manipulating the powerdistribution across the bargaining table, or creating a sit-uation in which the stereotypically disadvantaged negoti-ation is at a power disadvantage, allows us to test thishypothesis.

EXPERIMENT 1

The purpose of the current experiment was twofold.First, we aimed to deepen our understanding of how theexplicit activation of the masculine stereotype affectsnegotiation performance. Second, we sought to identifya boundary condition for stereotype reactance. Toachieve these goals, we created a power asymmetry at thebargaining table by manipulating the attractiveness ofnegotiators’ options apart from the current negotiation.

Although there are various sources of power relevantto negotiations (French & Raven, 1959), the ability towalk away is regarded as the key source of bargainingpower (Raiffa, 1982; Thompson, 2001). On average, thenegotiating party with a better BATNA (best alternativeto a negotiated agreement) reaps more beneficial out-comes from the negotiation; the strength of a negotia-tor’s BATNA positively influences aspirations, and theamount of resources created and claimed at the bargain-ing table (Fisher & Ury, 1981; Pinkley, Neale, & Bennett,1994). Given that we contend that the explicit activationof gender stereotypes prompts negotiators to examinetheir bargaining strengths, it also should lead them toexamine their power by focusing on their BATNA whenit is salient. Therefore, the explicit activation of a mascu-line stereotype should magnify the relative advantage ofthe negotiator with the stronger BATNA compared tothe negotiator with the weaker BATNA. If the femalenegotiator who is disadvantaged by the stereotype is also

Kray et al. / STEREOTYPE ACTIVATION IN NEGOTIATIONS 401

at a power disadvantage, then this should limit her abilityto react against the negative stereotype.

To test the above hypotheses, we examined two fac-tors—the relative power advantage within the dyad andthe manner in which masculine stereotypes were acti-vated (i.e., explicit vs. implicit). We hypothesized thatthe explicit activation of stereotypes would lead negotia-tors to engage in behaviors consistent with the content ofthe activated masculine stereotype. Becausestereotypically masculine behavior at the bargainingtable involves being assertive and focused on one’s owninterests, we predicted that the tone of negotiationswould be more contentious in the explicit activationcondition. For women, this would involve engaging incounterstereotypic behaviors. With both negotiatorsbehaving competitively, we expected the explicit activa-tion of gender stereotypes would lead to more one-sidedoutcomes in favor of the high-power negotiator. Toexplore how the behavior of negotiators was affected bythe experimental manipulations, we videotaped thenegotiations.

Method

Participants and design. Our participant sample wascomposed of 94 undergraduate students in a businessschool at a large, southwestern university who partici-pated in this study in exchange for course credit and thechance to win a monetary prize based on negotiationperformance.2 To test our hypotheses, we used a 2 × 2 × 2mixed-model design, with manner of masculine stereo-type activation (explicit, implicit) and gender poweradvantage (female, male) as between-dyad factors andpower (high, low) as a within-dyad factor. The task con-sisted of a negotiation of an employment contractbetween a recruiter and a job candidate. Role assign-ments (female recruiter/male candidate vs. female can-didate/male recruiter) were counterbalanced. Becauseno differences were observed for any of the dependentvariables across role assignments, we collapsed acrossthis variable. Dependent measures included both nego-tiation outcome and negotiation process measures.

Procedure. Negotiators were randomly assigned to amixed-gender dyad and to a negotiator role. Each dyadnegotiated in a private room. All participants wereinformed of the importance of negotiations to everydaylife and that their performance on the task would bediagnostic of future negotiation success. The experi-menter read the general instructions to participants,which indicated that the goal of each negotiator was toearn as many points as possible in the negotiation. Sev-eral monetary prizes were offered to provide a perfor-mance incentive. Participants were given private roleinformation that indicated their preferences in thenegotiation. After reading their role instructions,

negotiators completed a prenegotiation questionnairethat included measures of the negotiator’s goal and res-ervation price. On 5-point scales, participants also indi-cated their perceived power based on their role andtheir BATNA (ranging from very powerless to very powerful)and the attractiveness of their BATNA (ranging from veryunattractive to very attractive). Upon completion, thevideo camera was turned on and the negotiators weregiven up to 30 min to negotiate.

Negotiation task. The negotiation task was the NewRecruit simulation (Neale, 1997). The task consisted ofan employment negotiation in which a job candidateand a recruiter had to settle on several issues relevant toboth parties (i.e., salary, benefits). The negotiationincluded eight issues in total. Preferences were inducedby assigning points to issues and instructing participantsto achieve as many points as possible. Negotiators couldearn between –8,400 and 13,200 points. Two issues werepurely distributive, meaning that the parties’ prefer-ences were in complete opposition. Two issues werecompatible (i.e., the parties’ preferences were identi-cal). The remaining issues formed two pairs of issueswith integrative potential, meaning that one party caredmore about issue A and the other party more about issueB. If both parties conceded on the issue they cared lessabout, they both could benefit and reach the maximumnumber of points.

Masculine stereotype activation manipulation. As part ofthe general instructions, the experimenter informedparticipants in all conditions,

As researchers, we are interested in examining the vari-ous personal factors that affect people’s ability to per-form in important negotiations. For example, previousresearch has shown that the most effective negotiators innegotiations like the one that you’ll do today are rationaland assertive and demonstrate a regard for their owninterest throughout the negotiation, rather than beingemotional and passive.

Participants in the explicit activation condition were fur-ther told, “Because these personality characteristics tendto vary across gender, male and female students havebeen shown to differ in their performance on this task.”

Power manipulation. The manipulation of poweradvantage occurred through an “urgent message”attached to the negotiators’ private role information.The high-power negotiator had a message that statedthat another party (i.e., employer or candidate) was will-ing to settle on a contract worth 4,500 points and thattherefore the negotiator can “rest assured now that youcan always achieve an agreement worth 4,500 points.”The alternative offer was described as “quite favorable.”The low-power negotiator was told that another party

402 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

was willing to settle on a contract worth 2,200 points,which was described as “not favorable.” Both negotiatorswere informed that the average agreement was worth3,000 points. To manipulate gender power advantage,half of the dyads were composed of women in the high-power role and men in the low-power role and the otherhalf of dyads were composed of women in the low-powerrole and men in the high-power role.

Dependent measures. The primary outcome measurewas the number of points the negotiators achieved (i.e.,their pay-offs). Although we did not have any specifichypotheses about whether our experimental manipula-tions would affect the integrativeness of the negotiation,we also examined joint pay-off, or the sum of the twonegotiators’ pay-offs. Joint pay-off represents the degreeto which negotiators logrolled (Froman & Cohen,1970), or traded off issues to capitalize on differentstrengths of preferences and the degree to which negoti-ators identified compatible issues.

The purpose of the video-coding was to examine howpower and stereotype activation affected the negotiationprocess. Two different judges reported their overallimpressions of the negotiation and the negotiators. Thistype of coding was meant to capture both verbal andnonverbal communication between negotiators (Porter& Geis, 1981). Ratings were made for the first half andthe second half of the negotiation separately. Correla-tions between the ratings for the two halves of the negoti-ation were generally high (average r = .72); we thereforeaveraged across periods. The interrater reliability wasalso high (Cronbach’s α = .85) so analyses were con-ducted on the average scores of the two coders. Judgeswere blind to our hypotheses and manipulations. Toavoid cueing judges to gender differences, they wereinstructed to rate the behavior of the recruiter and can-didate roles, not of male and female negotiators. Theoverall negotiation was rated as to the extent that it wascontentious and cooperative. For example, a coopera-tive negotiation was one in which negotiators madeexplicit references to compromise regarding an issue. Incontrast, a contentious negotiation was one in which thesame issue would be dealt with by threatening to resort toone’s BATNA or accusing the other negotiator of being

uncompromising. Judges also coded whether negotia-tors made reference to their BATNA and, if so, howoften. Negotiators were evaluated in the extent to whichthey exhibited the following stereotypically masculinetraits: assertive, passive (reverse scored), controlling ofthe negotiation, and powerful. All ratings were made ona 7-point scale (ranging from 1 = not at all characteristic to7 = extremely characteristic). Finally, we recorded the lengthof the negotiation as a gauge of its difficulty.

Results

Manipulation checks. Consistent with our expectationthat the manipulation of BATNAs would affect percep-tions of power, high-power negotiators regarded theirBATNA as more attractive (M = 3.52, SD = .96) than didlow-power negotiators (M = 2.24, SD = .93), t(48) = 4.79, p <.001. High-power negotiators also believed they derivedmore power from their BATNA (M = 4.12, SD = .44) thandid low-power negotiators (M = 3.24, SD = 1.05), t(48) =3.86, p < .001. As a result, high-power negotiators sethigher reservation prices (M = 4,930, SD = 1,151) thandid low-power negotiators (M = 3,672, SD = 1,953), t(48) =2.78, p < .01. High-power negotiators perceived a greateradvantage through their role (M = 3.92, SD = .64) thandid low-power negotiators (M = 3.40, SD = .87), t(48) =2.41, p < .05. Role assignment, however, did not affectperceptions of power, F < 1.

To determine the salience of the power asymmetry, wealso coded whether negotiators made reference to theirBATNA during the negotiation. Overall, 84% (n = 50) ofnegotiators mentioned their BATNA at least once dur-ing the negotiation, and this frequency did not differ byexperimental condition. In sum, our manipulation ofpower based on each negotiator’s alternative to the cur-rent negotiation was successful.

Performance. Because negotiation outcomes dependon the two parties forming an agreement, the unit ofanalysis was the dyad. To compare the relative perfor-mance of the two negotiators, we conducted a mixed-model ANOVA of the pay-off score with power as awithin-dyad factor and manner of stereotype activationand gender power advantage as between-dyad factors(see Table 1 for average pay-off scores of male and

Kray et al. / STEREOTYPE ACTIVATION IN NEGOTIATIONS 403

TABLE 1: Negotiator Performance by Gender, Masculine Stereotype Activation, and Gender Power Advantage

Female High Power Male High Power

Explicit (N = 6) Implicit (N = 5) Explicit (N = 9) Implicit (N = 5)

M SD M SD M SD M SD

Male 4,200 1,244 4,640 1,894 7,067 1,850 5,700 878Female 6,017 1,148 5,200 245 3,667 1,083 4,980 950Joint pay-off 10,217 1,898 9,840 2,017 10,733 1,389 10,680 1,073

female negotiators). Consistent with our expectationthat power would be associated with better performance,high-power negotiators (M = 6,168, SD = 1,450) per-formed better than did low-power negotiators (M =4,252, SD = 1,316), F(1, 21) = 14.47, p < .001, η2 = .41.

Consistent with our hypothesis that the extent towhich the high-power negotiator would outperform thelow-power negotiator would depend on the type of ste-reotype activation, the main effect for power was quali-fied by an interaction with manner of stereotype activa-tion, F(1, 21) = 5.31, p < 05, η2= .20. As shown in Figure 1,the high-power negotiator outperformed the low-powernegotiator in the explicit activation condition, F(1, 14) =20.67, p < .001; however, the difference between thehigh- and low-power negotiators’ performances was notstatistically significant in the implicit activation condi-tion, F(1, 9) = 1.70, ns. No other effects emerged as statis-tically significant.

Joint pay-off. Consistent with previous research, jointpay-off was assessed by summing the individual outcomescores of the two negotiators. A two-way ANOVA withmanner of stereotype activation and gender poweradvantage as between-dyad factors showed no significanteffects (all Fs ≤ 1.1).

Negotiation process ratings. To examine the tone of thenegotiation, we conducted separate analyses of variancefor each measure with stereotype activation and genderpower advantage as between-dyad factors. Only oneeffect emerged as statistically significant. The negotia-tion was perceived to be more contentious underexplicit stereotype activation (M = 3.75, SD = 1.54) thanunder implicit stereotype activation (M = 2.38, SD = .85),F(1, 20) = 7.37, p < .05. This finding is consistent with thehypothesis that explicitly activating gender stereotypesled negotiators to engage in behaviors consistent withthe content of the masculine stereotype, thereby raisingthe level of perceived contentiousness of thenegotiation.

To examine the extent to which negotiators engagedin stereotypically masculine behaviors, we conducted amixed-model ANOVA with power as a within-dyad factorand stereotype activation and gender power advantageas between-dyad factors. Before doing so, we formed anindex indicating the extent to which behaviors werestereotypically masculine by averaging responses to thefollowing negotiator assessments: assertiveness, reverse-scored passivity, control of the negotiation, and power.Reliability of the index was high, α = .95. Consistent withthe hypothesis that the explicit activation of masculinestereotypes would lead negotiators to act in a mannerconsistent with the stereotype, a three-way ANOVAshowed only a significant main effect for stereotype acti-vation, F(1, 20) = 8.17, p < .01. Negotiators of both sexeswere perceived as more stereotypically masculine afterexplicit stereotype activation (M = 5.77, SD = .56) thanafter implicit activation (M = 5.14, SD = .60). No other sig-nificant effects emerged.

Duration of negotiation. We expected negotiations totake longer when stereotype activation was explicitrather than implicit, which is consistent with the negotia-tion process ratings that suggest the level of contentious-ness was higher after explicit activation. This expecta-tion was confirmed by a main effect for stereotypeactivation, F(1, 19) = 4.88, p < .05. Negotiations took lon-ger on average after explicit (M = 28 min 28 s, SD = 10min) than after implicit stereotype activation (M = 19min 36 s, SD = 7 min 30 s).

References to BATNA. To assess the degree to which thenegotiation was focused on power, we coded the numberof references made by negotiators to their BATNAs. Asignificant two-way interaction effect between power andgender power advantage emerged for this measure, F(1,21) = 6.01, p < .05, η2 = .23. High-power male negotiatorsmade reference to their BATNA more often (M = 2.64,SD = 2.59) than did low-power male negotiators (M = .64,SD = 1.03), t(23) = 2.42, p < .05, but this difference did notreach significance for female negotiators (M = 1.73, SD =1.80 vs. 1.07, SD = 1.59), t(23) = .97, ns. In the female/low-power condition, men mentioned their BATNAmore frequently than did women (M = 2.64 vs. 1.07),t(13) = 2.24, p < .05. The opposite was the case in thefemale/high-power condition, in which women referredto their BATNA more often than did men (M = 1.73 vs..64), t(10) = 1.83, p < .10.

Discussion

We predicted that the manner in which the masculinestereotype was activated would affect the use of power. Inaddition, we expected that explicitly linking gender toperformance would lead negotiators to engage in behav-iors consistent with the content of the activated positive

404 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

0

1000

2000

3000

4000

5000

6000

7000

Implicit Explicit

Masculine stereotype activation

Low power

High power

Figure 1 Experiment 1: Mean performance by masculine stereotypeactivation and power asymmetry.

masculine stereotype to a greater degree than when theconnection of gender to performance remained at animplicit level. Note that for women, this would lead tobehaviors that are inconsistent with the classic femalestereotype and consistent with the classic male stereo-type. Because the current experiment linkedstereotypically masculine traits to performance, we pre-dicted that negotiators in the explicit condition wouldfocus on and leverage their power to a greater extentthan they would when the masculine stereotype wasimplicitly activated. With both negotiators focused onpower, we expected women to be ineffective at reactingagainst the negative stereotype about their gender whenmen had the power advantage over them.

The results of the experiment support our hypothe-ses. The high-power negotiator obtained more of theresources when the masculine stereotype was explicitlyactivated. The content of the activated stereotypeimplies that having a high regard for one’s own interestsand acting assertively are important for negotiation suc-cess. The pattern of data suggests that both negotiatorsin this mixed-gender context exhibited these traits to alarger degree after explicitly mentioning the relevanceof gender to performance. For women this involved act-ing in a counterstereotypic fashion. This finding is con-sistent with the hypothesis that explicit masculine stereo-type activation leads to more attention to and use ofpower in the negotiation than implicit masculine stereo-type activation.

By using an integrative bargaining task with multipleissues, the current experiment enabled us to examinethe performance of each negotiator individually. Theresults suggest that across the board, both gendersexcelled when they were in a position of power. The pat-tern of data also is consistent with our understanding ofstereotype reactance processes. When masculinity isexplicitly brought to the forefront of negotiators’ minds,female negotiators react against the negative implica-tions of the stereotype by increasing their focus onbehaviors that are consistent with the content of the posi-tive stereotype and counterstereotypic of their gender.However, reactance only occurs for women when theyhave sufficient power through which they can react.

In addition to the evidence garnered by examiningnegotiated agreements, the process measures weobtained also provide support for our hypothesis thatthe negotiations would ultimately become more aggres-sive and contentious after explicit stereotype activation.First, our coding of negotiators’ verbal exchangesrevealed that high-power negotiators were more likely tomake reference to their BATNA than were low-powernegotiators. Second, judgments of the tone of the nego-tiation suggest that it became more contentious after theexplicit linkage of gender to negotiation performance

than when the linkage was made on an implicit level.This finding, in combination with the outcome mea-sures, is consistent with previous research that has shownnegotiation outcomes tend to be more one-sided as theybecome more contentious (Brett, Shapiro, & Lytle,1998). Third, negotiators’ behaviors were perceived tobe more stereotypically masculine after the explicit acti-vation of the masculine stereotype than the implicit acti-vation. Finally, reaching an agreement took longer afterthe explicit activation of the masculine stereotype, sug-gesting it was more difficult under this circumstance.

We expected the manner in which stereotypes wereactivated to moderate the relationship between poweradvantage and performance. More specifically, weexpected the degree to which high-power negotiatorsmade more references to their BATNA compared to low-power negotiators to be magnified after explicit stereo-type activation. Instead, we observed a pattern suggest-ing that across the board, high-power negotiators mademore references to their BATNA than did low-powernegotiators, without any consideration of the manner inwhich stereotypes were activated. Noting the high pro-portion of the negotiators who referenced their BATNAat least once suggests that negotiators may have adjustedtheir behaviors in the explicit activation condition bymaking more extreme demands rather than repeatedlymaking mention of their BATNA.

In sum, the findings of the current experiment sup-port the hypothesis that explicitly mentioning a genderadvantage leads to behaviors that are consistent with thestereotype of the advantaged gender andcounterstereotypic of the disadvantaged gender.Because the current experiment connectedstereotypically masculine traits to performance, theexplicit linkage of gender to performance led to moremasculine behaviors. Although these results support ourhypotheses, the current findings do not allow us toaddress an important alternative explanation; that is, it ispossible that simply mentioning gender in connectionwith performance in a mixed-gender negotiation leadsnegotiators to be more competitive. If this is so, then thecontent of the stereotype should not matter. To rule outthis alternative explanation, it is necessary to connectstereotypically feminine traits to performance and showthat it has a differential impact on performance. We con-ducted Experiment 2 for just this purpose.

EXPERIMENT 2

Experiment 1 showed that when stereotypically mas-culine traits are explicitly linked to performance,women and men who have power use it to their advan-tage to a greater degree than when the connectionbetween gender and performance is implicit. This use ofpower comes at a disadvantage to the party who is low in

Kray et al. / STEREOTYPE ACTIVATION IN NEGOTIATIONS 405

power. Our fundamental hypothesis is that explicitly acti-vating stereotypes increases the display of behaviors thatare consistent with the stereotype. In the case of claimingresources in competitive negotiations, it is masculine ste-reotypes that are most efficacious. In addition to theamount of resources claimed at the bargaining table,another gauge of negotiation success is the amount ofresources created. The results of Experiment 1 showedthat explicitly activating a masculine stereotype, whichseems most relevant to the competitive component ofnegotiations, did not affect the creation of resources.

The fact that the explicit connection of masculinetraits to performance in Experiment 1 did not affectjoint profit is not surprising when one considers that, toreach mutually beneficial, integrative trade-offs in multi-issue negotiations, a certain degree of cooperation isnecessary (Pruitt & Lewis, 1975); that is, a high degree ofconcern for the outcomes of the other party is critical forformulating integrative agreements (Pruitt & Rubin,1986). Because the masculine stereotype speaks solely toa concern for one’s own outcomes and is consistent withthe view that negotiations are a win-lose, competitiveendeavor, the explicit activation of the masculine stereo-type should not improve integrative outcomes. In con-trast, the classic feminine stereotype, characterized bygenerosity, attending to interpersonal relationships, andfluid communication, is certainly consistent with theskills thought to be important for the effective creationof integrative agreements (Pruitt & Lewis, 1975).

What does it mean for an outcome to be integrativeand how does one achieve such integrativeness? Anagreement is considered more integrative than anotherone when the integrative agreement makes at least onenegotiator better off and neither of the negotiatorsworse off (Thompson, 2001). The construction of inte-grative agreements typically involves sharing and solicit-ing information with one’s negotiating partner regard-ing priorities among issues. Both negotiators canimprove their outcome by conceding on a low-priorityissue in exchange for their most preferred outcome on ahigh-priority issue, a technique called logrolling. Bylogrolling instead of splitting all of the issues down themiddle, negotiators can expand the pie (Froman &Cohen, 1970). Because logrolling requires negotiatorsto concede on low-priority issues, a certain degree ofcooperativeness is needed.

As described above, we expect the explicit connectionof gender to performance to increase behaviors that areconsistent with the content of the stereotype. More spe-cifically, we hypothesized that explicitly linkingstereotypically feminine traits to performance shouldlead to more cooperative negotiations, resulting in moreintegrative outcomes. By demonstrating that the explicit

connection of feminine traits to performance increasesjoint pay-offs, we can rule out the alternative hypothesisthat explicitly mentioning gender in mixed-gendernegotiations simply leads to more competitiveness, with-out regard to the content of the activated stereotype.

Method

Participants and design. Our participant sample wascomposed of 68 MBA students in a business school in thewestern United States who participated in this study aspart of a classroom exercise.3 To test our hypotheses, weused a 2 × 2 × 2 mixed-model design, with manner of fem-inine stereotype activation (explicit, implicit) and gen-der power advantage (female, male) as between-dyadfactors and power (high, low) as a within-dyad factor.Role assignments (female buyer/male seller vs. femaleseller/male buyer) were counterbalanced.

Negotiation task. The negotiation task was the Playersimulation (Schroth, 1997). The task consisted of anegotiation between a motion picture director and pro-ducer. The two parties had to agree on 11 issues pertain-ing to the production of a motion picture (i.e., malelead, director’s salary). As in Experiment 1, preferenceswere induced by assigning points to issues and instruct-ing students to strive to achieve as many points as possi-ble. In theory, individual negotiators could earn up to16,200 points and the maximum joint pay-off was 14,000.Three issues were purely distributive, six issues had inte-grative potential, and two issues were compatible. Thesix issues with integrative potential were constructed sothat the negotiators could concede on a low-priorityissue to get their preferred outcome on a high-priorityissue. Therefore, the size of the negotiating pie, theamount of resources created at the bargaining table, is adirect function of how well the negotiators logrolled onthe issues with integrative potential.

Feminine stereotype activation manipulation. Embeddedin the task instructions was the manipulation of stereo-type activation. Following Kray et al. (2002), negotiatorsin all conditions were told,

This negotiation serves as a helpful diagnostic tool forstudents to assess their negotiating skills at the begin-ning of the course. In the past years, this negotiation hasprovided a revealing measure of negotiating abilitybased on personal bargaining style. As we will discussthroughout the course, highly skilled negotiators have akeen ability to express their thoughts verbally, good lis-tening skills, and insight into the other negotiator’s feel-ings. Negotiators who display these skills tend to achievehigher outcomes than those who are unexpressive, poorlisteners, and lack awareness of the other negotiator’semotions.

406 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

In the explicit condition only, the following statementwas added to the end of the paragraph, “Because thesepersonality characteristics tend to vary across gender,male and female students have been shown to differ intheir performance on this task.” All participants were re-minded that negotiating effectively is a difficult skill tomaster and not all students would achieve the highestoutcome possible.

Power manipulation. As in Experiment 1, we manipu-lated power through an “urgent message” attached tothe negotiators’ private role information. The high-power negotiator had a message that stated that anotherparty (i.e., director or producer) was willing to settle on acontract worth 6,000 points and that therefore the nego-tiator can “rest assured now that you can always achievean agreement worth 6,000 points.” The alternative offerwas described as “quite favorable.” The low-power nego-tiator was told that another party was willing to settle on acontract worth 3,000 points, which was described as “notfavorable.” To manipulate the gender power advantage,we varied whether the female or the male negotiatorreceived the high-power role.

Dependent measures. As in Experiment 1, the primaryoutcome measures were the scores for the male andfemale negotiators and their joint pay-off. The amountof joint pay-off is a direct function of how well the negoti-ators logrolled on the issues with integrative potential.Prior to the negotiation, participants also completed aquestionnaire in which they indicated their BATNA, res-ervation price, and target.

Results

Manipulation checks. As a check of the efficacy of themanipulation of power, we asked negotiators to indicatetheir BATNA prior to negotiating. Consistent with ourpower manipulation, high-power negotiators indicateda more attractive BATNA (M = 5,804, SD = 798) than didlow-power negotiators (M = 3,103, SD = 557), F(1, 49) =206.89, p < .001. Likewise, high-power negotiators (M =6,254, SD = 1,523) set higher reservation prices than didlow-power negotiators (M = 3,977, SD = 1,278), F(1, 49) =28.11, p < .001. The degree to which negotiator targets

differed across experimental conditions was not statisti-cally significant.

Performance. To compare the relative performance ofthe two negotiators, we conducted a mixed-modelANOVA of each negotiators’ pay-off scores with power asa within-dyad factor and manner of stereotype activationand gender power advantage as between-dyad factors(see Table 2 for a summary of the means). Consistentwith our expectation that power would be associatedwith better performance, high-power negotiators (M =7,110, SD = 1,008) performed better than did low-powernegotiators (M = 5,076, SD = 1,273), F(1, 27) = 28.14, p <.001, η2 = .51. However, as predicted, the manner inwhich the feminine stereotype was activated did notqualify this main effect for power. Other than a maineffect for stereotype activation, noted below, no otherstatistically significant effects emerged.

Joint pay-off. To analyze joint pay-off, we conducted anANOVA, including stereotype activation and power asbetween-dyad factors. Consistent with the hypothesisthat the explicit connection of stereotypically femininetraits with negotiation performance would lead to moreintegrative outcomes, joint pay-off was higher in theexplicit activation condition (M = 12,661, SD = 907) thanthe implicit activation condition (M = 11,794, SD =1,294), F(1, 27) = 4.03, p = .05, η2 = .13; that is, the explicitactivation of the female stereotype led negotiators to bemore effective at logrolling those issues with integrativepotential. In fact, by breaking down the joint score intothe three types of issues, it is clear that the stereotype acti-vation manipulation only affected joint pay-off throughthe integrative issues, F(1, 29) = 5.30, p < .05, and not fordistributive issues, F(1, 29) < 1, ns, and compatible issues,F(1, 29) < 1, ns.

Discussion

This experiment contributes to our understanding ofhow the content of stereotypes and the manner in whichthey are activated affect agreements. Consistent with ourexpectations, explicitly linking stereotypically femininetraits to negotiation effectiveness in an integrative bar-gaining task affected the degree to which negotiatorscrafted an agreement reflective of the feminine stereo-

Kray et al. / STEREOTYPE ACTIVATION IN NEGOTIATIONS 407

TABLE 2: Negotiator Performance Across Gender, Feminine Stereotype Activation, and Gender Power Advantage

Female High Power Male High Power

Explicit (N = 9) Implicit (N = 9) Explicit (N = 5) Implicit (N = 8)

M SD M SD M SD M SD

Male 5,022 1,469 5,022 927 7,120 719 7,106 1,171Female 7,611 1,316 6,717 598 5,590 1,440 4,750 1,429Joint pay-off 12,633 822 11,739 850 12,710 1,056 11,856 1,731

type; that is, negotiators created more integrative agree-ments after explicit stereotype activation than implicitstereotype activation. Specifically, explicit activation ofthe feminine stereotype led negotiators to be moreeffective logrollers, more capable at conceding on low-priority issues for their preferred outcome on high-priority issues. Explicit activation of the feminine stereo-type created resources and expanded the pie. The find-ings of the current experiment support the hypothesisthat the explicit activation of gender stereotypes leadsnegotiators to engage in behaviors that are consistentwith the content of the positive stereotypical traits linkedto negotiator effectiveness.

The current experiment replicated the effect ofpower from Experiment 1. On average, the negotiatorwith more power was able to claim more resources at thebargaining table than the negotiator with less power.Although Experiment 1 observed an interactionbetween power and manner of stereotype activation, thispattern of data did not emerge in the current experi-ment. Instead, we observed a significant effect of stereo-type activation on joint pay-off. A comparison acrossexperiments suggests that the content of the stereotypedetermines its impact at the bargaining table. Whereasexplicitly linking stereotypically masculine traits to per-formance led to greater competitiveness in Experiment1, explicitly linking stereotypically feminine traits to per-formance led to greater cooperativeness in the currentexperiment.

An interesting question that arises from an examina-tion of the means in Table 2 is whether men and womenwere differentially affected by the manner in which thefeminine stereotype was activated. The pattern of meanssuggests that women’s point totals varied depending onhow the stereotype was activated, whereas men’s pointtotals were relatively stable across conditions. Althoughthe inferential statistics do not support the reliability ofthis proposition, the experiment clearly suffers fromlow-power. As a result, the question of whether thegreater joint gain in the explicit activation conditionresulted solely from the behavior of women as opposedto both women and men is open. With the current data,it is impossible to determine whether women were bothmore proactive in creating and claiming value in theexplicit condition. To address this question, futureresearch will need to involve an analysis of the negotia-tion process, not unlike what we conducted in Experi-ment 1.

GENERAL DISCUSSION

The purpose of the current set of experiments was togain a deeper understanding of how the explicit linkageof gender to performance affects negotiating behavior.The current experiments suggest that performance is

intimately linked to the content of gender stereotypesand the manner in which they are activated. Of impor-tance for understanding gender differences in negotia-tions as a situational phenomenon, this research showsthat both men and women are sensitive to the content ofan activated stereotype and adjust their behaviorsaccordingly.

We hypothesized that the explicit activation of genderstereotypes leads negotiators to engage in behaviors con-sistent with the content of the activated positive stereo-type; for the disadvantaged negotiator, this leads tocounterstereotypic behaviors. Because focusing onpower is stereotypically masculine, we expected agree-ments to reflect a greater power struggle after theexplicit activation of masculine stereotypes. Consistentwith this hypothesis, the explicit activation of masculinestereotypes led to more one-sided outcomes in favor ofthe high-power negotiator (Experiment 1). This was truefor men and women alike. When the traits that werelinked to effective negotiations were stereotypically femi-nine, evidence of a heightened power struggle afterexplicitly mentioning gender was absent. Instead, nego-tiators seemed to act in accordance with the femininestereotype and behaved more cooperatively, with theexplicit activation of feminine stereotypes leading tohigher joint pay-offs (Experiment 2). In total, this set ofresults suggests that the manner in which stereotypes areactivated affects the salience of viable alternatives to thecurrent negotiation and the degree to which those alter-natives determine the division of resources.

This research advances our understanding of the pro-cess through which stereotype reactance works. Krayet al. (2001) defined stereotype reactance as the ten-dency to behave in a manner inconsistent with a stereo-type. This definition suggests that when the stereotype isexplicitly activated, the negotiator subjected to the limit-ing stereotype will engage in counterstereotypic behav-iors. We found evidence of this process across two experi-ments. Because the content of the masculine stereotypeis related to aggressiveness, concern with one’s owninterests, and overall competitiveness, to react againstthe stereotype is to claim and secure a profitable divisionof the pie. In contrast, the feminine stereotype includesattributes related to insight and concern for the otherside, such that reacting against the stereotype involvescreating and expanding the pie.

By manipulating the relative power of the negotiators,we also investigated the limits of stereotype reactance. Inboth experiments, a power asymmetry was made salientand negotiators with more power claimed the lion’sshare of resources for themselves. Power is indeed a cru-cial predictor of negotiated outcomes. In the first experi-ment, however, when the masculine stereotype wasexplicitly activated, low-power women did not reverse

408 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

the gender gap and achieve superior outcomes. Thisfinding suggests that reacting against the masculine ste-reotype and achieving an advantage at the distributivebargaining table only occurs when the stereotypicallydisadvantaged negotiator is not also crippled by tangiblepower deficiencies. Although necessary for overcomingpower deficiencies, engaging in counterstereotypicbehaviors does not appear to be sufficient for reversingthe gender gap.

Thus, the explicit activation of stereotypes does notautomatically lead to a reversal of the gender gap innegotiations (Kray et al., 2001) but does so only in con-texts in which the stereotyped negotiator possesses someinitial power and leverage in the negotiation. Blatantlytelling women that they lack the attributes necessary toprevail at the distributive bargaining table produces ste-reotype threat in the absence of power and stereotypereactance when sufficient power is possessed. On a prac-tical note, these findings suggest that an effective way ofcircumventing the negative consequences of stereotypesis to increase more objective sources of power. If thepower of the stereotyped person is weak, an explicit con-frontation with the stereotype is even more threateningand results in greater performance decrements than theimplicit activation of the stereotype. Female stereotypethreat can only become stereotype reactance when suffi-cient power is possessed.

Our understanding of performance differences asresulting from contextual factors is consistent with previ-ous theorizing regarding stereotype threat. Work on ste-reotype threat suggests that differences between socialgroups in stereotype-relevant performance domains arenot the product of inherent, invariant deficiencies of agroup but that underperformance by stereotypedgroups is a situational phenomenon. When individualsfear that their performance may confirm a negative ste-reotype about their social group, they tend to live downto the negative expectations and show performance dec-rements. These effects can occur for men (Kray et al.,2002; Leyens et al., 2000) and women (Kray et al., 2001,2002; Spencer et al., 1999); for Caucasians (Aronsonet al., 1999; Stone et al., 1999), African Americans(Steele & Aronson, 1995; Stone et al., 1999), Latinos(Gonzales, Blanton, & Williams, 2002), and Asians(Cheryan & Bodenhausen, 2000); and even for socialclass (Croizet & Claire, 1998). These findings demon-strate clearly that stereotype threat is not limited to onestereotyped group but occurs whenever negative perfor-mance expectations exist for a social group that arebased on the possession of stereotypical traits.

This research also enriches our understanding of howpower affects negotiation agreements. How negotiatorsresponded to the salient power asymmetry that existedin both experiments depended on the manner in which

stereotypical traits were linked to performance. We haveargued that the explicit activation of stereotypesincreases behaviors that are consistent with the positivetraits linked to performance. Depending on the contentof the stereotype, power asymmetries were eitheremphasized or deemphasized in the negotiation. Whenmasculine traits were explicitly linked to performance,the negotiation process became more competitive andthe negotiator with the power advantage prevailed to alarger extent than when masculinity was implicitly con-nected to performance. When feminine traits wereexplicitly linked to performance, power asymmetrieswere not emphasized; instead, agreements becamemore integrative in comparison to when femininity wasonly implicitly linked to performance. In total, thisresearch suggests that the degree to which power isattended to and exercised depends on the negotiationcontext.

Limitations and Future Directions

We have shown that when negotiators possess suffi-cient external resources, in the form of attractive alterna-tives to a negotiation, they will be able to react against theconstraints of negative stereotypes. Across both experi-ments, the division of resources was driven by the relativeattractiveness of negotiators’ BATNAs rather than therelative advantage implied by the stereotype. Futureresearch should explore whether there are psychologi-cal resources, such as self-esteem or self-affirmation(Steele, 1988; Steele, Spencer, & Lynch, 1993), that canincrease the likelihood that the blatant activation of ste-reotypes will lead to stereotype reactance. Only when ste-reotypes are explicitly activated and when the stereo-typed individual possesses sufficient resources, physicalor psychological, with which to react is stereotypereactance expected to occur.

Although stereotype reactance is not specific to onesocial group, it has only been demonstrated within thecontext of negotiations. Future research should explorewhether explicitly telling someone that their socialgroup will prevent them from attaining greatness willironically lead them to strive for and achieve excellencein other performance domains and with other stereo-types. Because we define stereotype reactance (see Krayet al., 2001) as the tendency to behave in a mannerinconsistent with a stereotype, it may not always be possi-ble for a person affected by a negative stereotype toengage in counterstereotypic ways. Behavior in negotia-tions is somewhat malleable (deciding to be more asser-tive is something that most people can do) andcounterstereotypic behaviors have real consequences onnegotiated outcomes. In contrast, it is not clear that per-formance in intellectual (Steele & Aronson, 1995),quantitative (Spencer et al., 1999), or athletic (Stone

Kray et al. / STEREOTYPE ACTIVATION IN NEGOTIATIONS 409

et al., 1999) tasks, domains in which stereotype threathas been implicated, is similarly malleable. Thinkingquantitatively or jumping higher may not be so easilyaccomplished. The domain of performance might wellbe another limit to stereotype reactance.

It also should be noted that our manipulations of ste-reotype threat and reactance in the domain of nego-tiation specify the traits (and hence the behaviors) nec-essary to succeed at the bargaining table. Indeed, weselected specific masculine traits (Kray et al., 2001) andfeminine traits (Kray et al., 2002) that are actually con-nected to negotiator effectiveness (Raiffa, 1982). Indoing so, we specified how negotiators could react. Per-haps explicitly specifying counterstereotypic behaviorsin these other performance domains (e.g., related tointellectual and athletic ability) would enable reactanceto take place. Until such effects have been demon-strated, a tentative conclusion to draw is that the stereo-type threat process is more ubiquitous and commonthan stereotype reactance.

Because the content of the masculine stereotype isrelated to distributive tactics and dividing the pie, powermoderated the ability to react against that stereotypewhen it was explicitly activated. However, because thecontent of the feminine stereotype is related tointegrativeness and the ability to expand the pie, powerdoes not appear to moderate the effect of manner of ste-reotype activation (i.e., implicit vs. explicit) on perfor-mance. Future research should explore possible moder-ators of the ability of men to react against their limitingstereotype, to look at what limits men from acting incounterstereotypic ways following explicit stereotypeactivation. For example, accountability to constituentshas been shown to make negotiators more rigid and lessintegrative in their bargaining (Carnevale, Pruitt, &Seilheimer, 1981). It might be the case that men whoare accountable will resist being integrative due to afear of appearing weak to their constituents, therebyconfirming the negative stereotype rather than reactingagainst it.

Another direction for future research is to explorewhether the impact of stereotypes on negotiation perfor-mance differs depending on the degree to which negoti-ators identify with the domain. Indeed, stereotype threatprocesses have been shown to be larger for participantsidentified with the performance domain (for a meta-analysis of this effect, see Walton & Cohen, 2003). Simi-larly, one could expect that reactance would be morelikely to occur in people who strongly identify with thenegotiation domain compared to people who do notregard negotiating ability to be highly self-relevant. Bymeasuring domain-identification prior to the manipula-tion of stereotype activation, this hypothesis regarding apotential boundary condition could be tested.

Conclusions

Across two experiments, we have demonstrated boththe pervasiveness and limits of stereotype reactance. Ste-reotype reactance can occur for both sexes at the bar-gaining table. However, whether stereotype reactanceoccurs will depend on the power and available resourcespossessed by the stereotyped negotiator. Although a gen-der gap has been documented in negotiations(Stuhlmacher & Walters, 1999), our research approachfocuses on properties of the negotiation that are notinherent to the individual negotiator. We show an inter-action between the person, whether it be a male orfemale negotiator, and the situation, as determined bythe relative power and set of cognitions that each negoti-ator brings to the bargaining table. A key message of thisresearch is that by focusing on improving objective alter-natives to the current negotiation and harnessingcognitions that are advantageous to their particular gen-der, negotiators are able to leverage the negotiation totheir advantage.

NOTES

1. A complementary finding was obtained with respect to men, whowere more likely to identify stereotypically feminine traits as their keynegotiating weaknesses after masculine traits had been explicitlylinked to effective negotiating compared to a control condition. Krayet al. reasoned that the positive stereotype identity bestowed on men inthe explicit activation condition might have created a psychologicalburden, a concern about living up to the high standard expected ofthem (see also Brown & Josephs, 1999; Cheryan & Bodenhausen, 2000;Shih, Ambady, Richeson, Fujita, & Gray, 2002).

2. We excluded from the analyses impasses, situations in which anagreement could not be reached that was equal or superior to theBATNAs (best alternatives to a negotiated agreement) of both negotia-tors, which resulted in a final sample of 50 participants. Impasse rateswere not affected by any of our experimental manipulations. To deter-mine whether the dyads that were excluded differed from the dyadsthat were included in the degree to which they exhibitedstereotypically masculine traits, we compared the process ratings fromthe videotapes across groups. The degree to which included andexcluded dyads differed on stereotypically masculine traits was not sta-tistically significant.

3. Once again, we excluded dyads who impassed from the analyses,which resulted in a final sample of 64 participants. Impasse rates werenot affected by any of our experimental manipulations.

REFERENCES

Aronson, J., Lustina, M. J., Good, C., Keough, K., Steele, C. M., &Brown, J. (1999). When White men can’t do math: Necessary andsufficient factors in stereotype threat. Journal of Experimental SocialPsychology, 35, 29-46.

Bargh, J. A., Chen, M., & Burrows, L. (1996). The automaticity ofsocial behavior: Direct effects of trait concept and stereotype acti-vation on action. Journal of Personality and Social Psychology, 71, 230-244.

Brehm, J. W. (1966). A theory of psychological reactance. New York: Aca-demic Press.

Brett, J. M., Shapiro, D. L., & Lytle, A. L. (1998). Breaking the bondsof reciprocity in negotiations. Academy of Management Journal, 41,410-424.

410 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

Brown, R. P., & Josephs, R. A. (1999). A burden of proof: Stereotyperelevance and gender differences in math performance. Journal ofPersonality and Social Psychology, 76, 246-257.

Carnevale, P. J. D., Pruitt, D. G., & Seilheimer, S. D. (1981). Lookingand competing: Accountability and visual access in integrativebargaining. Journal of Personality and Social Psychology, 40, 111-120.

Cheryan, S., & Bodenhausen, G. V. (2000). When positive stereotypesthreaten intellectual performance: The psychological hazards of“model minority” status. Psychological Science, 11, 399-402.

Croizet, J., & Claire, T. (1998). Extending the concept of stereotypeand threat to social class: The intellectual underperformance ofstudents from low socioeconomic backgrounds. Personality andSocial Psychology Bulletin, 24, 588-594.

Dijksterhuis, A., Spears, R., Postmes, T., Stapel, D., Koomen, W., vanKnippenberg, A., et al. (1998). Seeing one thing and doinganother: Contrast effects in automatic behavior. Journal of Personal-ity and Social Psychology, 75, 862-871.

Fisher, R., & Ury, W. (1981). Getting to yes: Negotiating agreement withoutgiving in. Boston: Houghton Mifflin.

French, J. R. P., & Raven, B. (1959). The bases of power. In D. Cart-wright (Ed.), Studies in social power. Ann Arbor, MI: Institute forSocial Research.

Froman, L. A., & Cohen, M. D. (1970). Compromise and logroll:Comparing the efficiency of two bargaining processes. BehavioralScience, 30, 180-183.

Gonzales, P. M., Blanton, H., & Williams, K. J. (2002). The effects ofstereotype threat and double-minority status on the test perfor-mance of Latino women. Personality and Social Psychology Bulletin,28, 659-670.

Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition:Attitudes, self-esteem, and stereotypes. Psychological Review, 102, 4-27.

Greenwald, A. G., Banaji, M. R., Rudman, L. A., Farnham, S. D.,Nosek, B. A., & Mellott, D. S. (2002). A unified theory of implicitattitudes, stereotypes, self-esteem, and self-concept. PsychologicalReview, 109, 3-25.

Kray, L. J., Galinsky, A., & Thompson, L. (2002). Reversing the gendergap in negotiations: An exploration of stereotype regeneration.Organizational Behavior and Human Decision Processes, 87, 386-410.

Kray, L. J., Thompson, L., & Galinsky, A. (2001). Battle of the sexes:Gender stereotype confirmation and reactance in negotiations.Journal of Personality and Social Psychology, 80, 942-958.

Leyens, J. P., Desert, M., Croizet, J. C., & Darcis, C. (2000). Stereotypethreat: Are lower status and history of stigmatization precondi-tions of stereotype threat? Personality and Social Psychology Bulletin,26, 1189-1199.

Martin, L. L. (1986). Set/reset: Use/disuse of concepts in impressionformation. Journal of Personality and Social Psychology, 51, 493-504.

Neale, M. A. (1997). New recruit: Dispute resolution research center exer-cises. Evanston, IL: Northwestern University.

Pinkley, R. A., Neale, M. A., & Bennett, R. J. (1994). The impact ofalternatives to settlement in dyadic negotiation. OrganizationalBehavior and Human Decision Processes, 57, 97-116.

Porter, N., & Geis, F. L. (1981). Women and nonverbal leadershipcues: When seeing is not believing. In C. Mayo & N. Henley (Eds.),

Gender and nonverbal behavior (pp. 39-61). New York: Springer-Verlag.

Pruitt, D. G., & Lewis, S. A. (1975). Development of integrative solu-tions in bilateral negotiations. Journal of Personality and Social Psy-chology, 31, 621-630.

Pruitt, D. G., & Rubin, J. Z. (1986). Social conflict: Escalation, stalemate,and settlement. New York: Random House.

Raiffa, H. (1982). The art and science of negotiation. Cambridge, MA:Belknap.

Schroth, H. (1997). The player: Dispute resolution research center exercises.Evanston, IL: Northwestern University.

Shih, M., Ambady, N., Richeson, J. A., Fujita, K., & Gray, H. M. (2002).Stereotype performance boosts: The impact of self-relevance andthe manner of stereotype activation. Journal of Personality and SocialPsychology, 83, 638-647.

Spencer, S. J., Steele, C. M., & Quinn, D. M. (1999). Stereotype threatand women’s math performance. Journal of Experimental Social Psy-chology, 35, 4-28.

Steele, C. M. (1988). The psychology of self-affirmation: Sustainingthe integrity of the self. In L. Berkowitz (Ed.), Advances in experi-mental social psychology: Vol. 21. Social psychological studies of the self:Perspectives and programs (pp. 261-302). New York: Academic Press.

Steele, C. M. (1997). A threat in the air: How stereotypes shape intel-lectual identity and performance. American Psychologist, 52, 613-629.

Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellec-tual test performance of African Americans. Journal of Personalityand Social Psychology, 69, 797-811.

Steele, C. M., Spencer, S. J., & Lynch, M. (1993). Self-image resilienceand dissonance: The role of affirmational resources. Journal of Per-sonality and Social Psychology, 64, 885-896.

Stone, J., Lynch, C., Sjomeling, M., & Darley, J. M. (1999). Stereotypethreat effects on Black and White athletic performance. Journal ofPersonality and Social Psychology, 77, 1213-1227.

Strack, F., Schwarz, N., Bless, H., Kuebler, A., & Waenke, M. (1993).Awareness of the influence as a determinant of assimilation versuscontrast. European Journal of Social Psychology, 23, 53-62.

Stuhlmacher, A. F., & Walters, A. E. (1999). Gender differences innegotiation outcome: A meta-analysis. Personnel Psychology, 52,653-677.

Thompson, L. (2001). The mind and heart of the negotiator (2nd ed.).Upper Saddle River, NJ: Prentice Hall.

Walters, A. E., Stuhlmacher, A. F., & Meyer, L. L. (1998). Gender andnegotiator competitiveness: A meta-analysis. Organizational Behav-ior and Human Decision Processes, 76, 1-29.

Walton, G. M., & Cohen, G. L. (2003). Stereotype lift. Journal of Experi-mental Social Psychology, 39, 456-467.

Wheeler, S. C., & Petty, R. E. (2001). The effects of stereotype activa-tion on behavior: A review of possible mechanisms. PsychologicalBulletin, 127, 797-826.

Received September 16, 2002Revision accepted May 17, 2003

Kray et al. / STEREOTYPE ACTIVATION IN NEGOTIATIONS 411