“I Do the Dishes; You Mow the Lawn”: Gender Effects in Stereotypically Feminine Negotiation Tasks

Research on gender effects in negotiation has largely relied on stereotypically masculine negotiation paradigms (e.g., selling a car). Globally, though with relatively weak effects sizes, this research shows that women tend to underperform men in negotiations. The present research examines gender effects in negotiations involving typically female-related topics. Specifically, we examine perceptions and performance of mixed negotiating dyads. We manipulate the topic of the negotiation between dyads: typically feminine (i.e., household duties) vs. neutral (i.e., organization of a debate). Results show an interaction between the gender of the negotiator and the negotiation topic. Among men, the topic of negotiation impacts neither perceptions nor negotiation performance. Among women, in contrast, aspirations and performance decrease in the stereotypically feminine negotiation condition compared to the neutral one. Furthermore, this effect of topic of negotiation on performance among women is mediated by (lower) aspirations of women in the feminine topic condition. We discuss results in light of recent research on gender and negotiation, as well as of broader theories such as role congruity and stereotype threat.

Mots clés en Français: Genre; Négociation; Stéréotypes; Performance; Thématique de négociation Does a couple discussing their next holidays negotiate in the same way as when this same couple discusses the division of household duties? Probably not, one would say. If nothing else, the first situation is essentially positive and likely to trigger collaboration between the parties, whereas the second one is more aversive and more prone to elicit competitive behaviors. However, beyond the valence of the negotiation topic, another factor distinguishes these two situations: their level of gender typicality. The "holiday" negotiation is indeed much more neutral in terms of gender than the "household duties" negotiation.
A meta-analysis by Stuhlmacher and Walters (1999) examined gender differences regarding negotiation performance. Their results show that men tend to get better negotiation outcomes than women, although this effect is relatively weak. In addition, gender differences are stronger in distributive negotiation settings (i.e., competitive negotiations in which the parties' interests are mutually exclusive) than in integrative ones (i.e., when interests are not mutually exclusive). The fact that the negotiation structure modulates gender differences suggests that research should take into account the role of contextual factors when examining gender effects. Finally, Stuhlmacher and Walters point out that the large majority of the studies included in their meta-analysis uses stereotypically masculine negotiation tasks (e.g., negotiating to buy or sell a car, a plane, or software products) and that none of these studies has examined gender differences in performance on more "feminine" negotiation topics. Deborah Kolb (2012) recently made a similar argument. According to her, previous research has mainly relied on stereotypically masculine negotiation tasks and this might in part explain why gender effects have systematically emerged.
In the present paper, we take these concerns into account and examine gender differences in negotiations involving stereotypically feminine vs. gender-neutral topics. Addressing this question is important for at least two reasons. First, examining gender differences as a function of the stereotypicality of the negotiation task will allow us to have a better understanding of results obtained by previous research. Specifically, we can get a better insight on the (potentially) confound between the (mainly "masculine") negotiation tasks and the gender of the negotiator in determining negotiation performance (for a similar argument see Kolb, 2012). Second, an increased awareness of the psychological mechanisms at play in mixed-gender negotiations might help us better understand how to deal with the obstacles that women face in this type of situations. The importance of developing knowledge in this field becomes even clearer when one takes into account the pervasiveness of both formal and informal negotiation situations.

Role congruity and performance
To the best of our knowledge, only two studies have directly examined how people react to gendered negotiation topics and, more specifically, to stereotypically feminine ones. In a first study, Julia Bear (2011) analyzed negotiation avoidance as a function of participants' gender and negotiation's topic. Results show that women avoid more strongly stereotypically masculine negotiations (i.e., monetary compensations at work) while men, in contrast, tend to avoid more strongly negotiations that focus on typically feminine topics (i.e., access to a lactation room). In another study, Bear and Babcock (2012) have shown that performance in negotiation follows a similar pattern. That is, men tend to obtain better results than women in a masculine-framed negotiation task whereas this difference disappears in the "feminine" version of the same task.
The authors explain their results in line with the role congruity theory (Eagly, Karau & Makhijani, 1995;Eagly & Karau, 2002). This theory suggests that performance at a given task is a function of the matching between gender roles and contextual features. For example, although globally gender differences in management performance are trivial, men perform better and experience a higher sense of accomplishment when the managerial role is defined in accordance with the masculine stereotype. In sharp contrast, when the managerial role is described as requiring typically feminine skills, women are the ones that benefit the most from the situation (Eagly et al., 1995). Kray, Galinski, and Thompson (2002) showed similar results in negotiations. These authors experimentally manipulated the stereotype of a "competent" negotiator. When the competent negotiator stereotype included typically feminine traits, therefore associating the representation of what is a good negotiator to the stereotype of women, women tend to increase their performance and to obtain better results in the negotiation process.
Based on the research reviewed up to here, one could hypothesize that when negotiations involve stereotypically feminine topics, women should outperform men. This should be due to the increased relevance of feminine attributes for the task at stake and thus to the greater congruency between women's roles and the task at hand. In these contexts, women should feel more confident to act assertively which would then lead them to obtain better results at the end of the negotiation process.

Implicit stereotype activation
In contrast to the role congruity hypothesis, it is also possible to formulate the reverse hypothesis and to argue that a traditionally feminine negotiation topic could trigger worse performances among women (compared to men). This effect would be due to the implicit activation of gender stereotypes by the negotiation topic itself. Indeed, contextual cues are among the first and most important determinants of self-categorization and, by consequence, of (self-)stereotype activation (Yzerbyt & Demoulin, 2010). Indeed, a recent study (Demoulin & Teixeira, in prep) confirmed that gendered negotiation topics increase the activation of gender stereotypes. In this study, women were presented with one of two fictional negotiations and asked to imagine that they would have to negotiate with a romantic partner. In the "stereotypical" condition, the negotiation concerned the division of household duties whereas in the control one it concerned the division of tasks in the process of finding a new house. Results showed that women in the stereotypical condition described themselves more in line with the feminine stereotype than women in the neutral one. Specifically, women that faced a negotiation about household duties perceived themselves as warmer but less competent (Glick & Fiske, 1999) than women in the control condition. In addition, self-stereotyping tended to decrease women's perceptions of relative power in the negotiation as well as their intentions to invest energy and to show firmness in the discussion. We propose that such stereotype activation can, in turn, give rise to the stereotype threat phenomenon.
The stereotype threat phenomenon (Steele & Aronson, 1995) refers to a decrease of performance observable when individuals belong to a group that is negatively stereotyped on a given task (e.g., women in math) and when the situation is presented as highly diagnostic of individuals' abilities on the task (e.g., a math test). This performance decrease does not occur when the task is presented as nondiagnostic or among members of non-negatively stereotyped groups (e.g., men). Stereotype threat effects have been shown to occur among women in negotiations. According to Kray, Thompson, and Galinsky, the majority of the traits associated with the prototype of a good negotiator overlaps with the masculine stereotype (e.g., assertive, rational, constructive, intelligent; Raiffa, 1982). In contrast, traits associated with poor negotiation performances overlap with the feminine stereotype (e.g., weak, emotional, irrational). On this basis, Kray and colleagues suggested the existence of a naïve theory that associates gender to performance in negotiations in such a way that women are perceived, by default, as worse negotiators than men (see also Kray & Thompson, 2005). In line with the stereotype threat phenomenon, these authors show that when gender stereotypes are implicitly activated and when the task is framed as diagnostic of an individual's negotiation skills, women get worse results than men in general, and worse results than women in a non-diagnostic condition. In addition, men benefit from stereotype activation and get better results in the diagnostic condition compared to the non-diagnostic one (Kray et al., 2001).
Interestingly and by contrast, the explicit activation of gender stereotypes triggers the opposite effect. Having the negative stereotype made explicitly salient leads women to behave more aggressively and assertively and this assertiveness allows them to get better outcomes than their male counterparts. This reversed effect is explained by the authors in terms of psychological reactance. Latter research has shed light on two boundary conditions of women's reactance: the existence of power differences that disfavor them (Kray, Reb, Galinski & Thompson, 2004) and the belief in the innate nature of negotiation skills (Kray & Haselhuhn, 2007;Kray, Locke & Haselhuhn, 2010).
According to research on the stereotype threat in negotiations, and in sharp contrast with the role congruity hypothesis, it is thus possible to make the alternative hypothesis that "feminine" negotiation topics would implicitly activate gender stereotypes and, in line with the stereotype threat phenomenon, trigger worse (vs. better) female (vs. male) negotiation performances.

The present research
The present experiment aims to examine the influence of a traditionally feminine negotiation topic (compared to a gender neutral one) on the negotiation performance of women and men in mixed-gender negotiations. As mentioned before, two competitive hypotheses are being tested. First, in line with the role congruity theory (Eagly et al., 1995;Eagly & Karau, 2002) and the results obtained by Bear (2011) and Bear and Babcock (2012), a stereotypically feminine negotiation topic (compared to a gender neutral one) should increase negotiation performance of women compared to men due to a higher congruence between feminine roles and the task at hand. Second, and in sharp contrast, recent research on stereotype activation in typi-cally feminine negotiation tasks (Demoulin & Teixeira, in prep) and on the stereotype threat phenomenon (e.g., Kray et al., 2001) suggests that women should obtain worse results than men in "feminine" negotiations compared to neutral ones because such situations implicitly activate the stereotype of women as incompetent. In addition, men should benefit from the stereotype activation and get better results in the former than in the latter tasks (Kray et al., 2001).
Finally, given the importance of aspiration points (i.e., the pre-negotiation target outcome of a negotiator; Neale & Fragale, 2006) in determining the final outcome of negotiations (e.g., Kray et al., 2002;Stevens, Bavetta & Gist, 1993;Tellhed & Björklund, 2011), we predict that the impact of the interaction between negotiation topic and negotiator's gender on negotiation performance will be mediated by negotiators' aspirations (i.e., a mediated moderation according to Muller, Judd and Yzerbyt, 2005).

Method Participants
Sixty-six participants accepted to participate in the experiment (M age = 29.65, SD = 13.67). Participants were recruited among acquaintances of students who attended a negotiation and conflict management class. The students (whom we will refer to as "experimenters") were instructed to organize a negotiation between a woman and a man. They were further told: 1) that the negotiation dyads could not be formed by people living in the same house or who were having a romantic relationship with one another; 2) that participants could not be informed that their gender served as basis for their recruitment; and 3) that the negotiation should take place in a quiet room without the presence of external observers. Experimenters were unaware of the hypotheses and neither experimenters nor participants received any compensation for participation in the experiment. The majority of the sample was Belgian (93%) and had French as mother tongue (97%).

Material and procedure
Participants were informed that they were going to take part in a fictive negotiation. They were asked to take their negotiator role seriously and were handed written instructions describing the negotiation.
The negotiation consisted in the distribution of four tasks between the two negotiators. Each task had seven possible distribution options and participants were instructed to get to an agreement on all four tasks. Each distribution option was associated to a specific amount of points that the participant would win in case the option was agreed upon by both negotiators. Participants were further told that their matrix of points was different than that of their partner's. In total, across the four tasks, a negotiator's potential gain varied between 0 and 800 points.
The between-subjects manipulation of the negotiation topic was embedded in the description of negotiation task. In the "gendered" condition, participants had to imagine that they lived together with their negotiation partner and had to share household duties. We selected this topic for the stereotypically feminine condition following data from the Organisation for Economic Co-operation and Development (OECD) concerning "Time used for work, care and daily household chores" gender rates (LMF2.5, http://www.oecd.org/social/family/database. htm). Among the 18 founding countries of the OECD, on average women spend 15.8% of their time on household chores compared to 8% for men. In the "gender-neutral" condition, besides living with their negotiation partner, participants had to imagine being part of the same association. This association organized thematic debates about societal issues. The negotiation consisted of sharing the different tasks related to the organization of these debates. The two versions were distributed randomly to the "experimenters" by the negotiation class's professor. Each experimenter organized only one dyad and, as mentioned above, was unaware of the hypotheses.
After making sure that participants had correctly understood the instructions and the negotiation matrix, the experimenter handed out a pre-negotiation questionnaire to participants. This first questionnaire assessed first-offer intentions, aspiration points (i.e., participants' goals), resistance points (i.e., yielding limits), and zero-sum perceptions of the negotiation. The goal of this questionnaire was threefold. First, aspirations were assessed to test for a mediation hypothesis via an anchoring effect of aspirations (see hypothesis above). Secondly, having to think about their first offer, goals and limits allowed participants to fully integrate the information and better prepare themselves for the negotiation. Finally, zero-sum perceptions were measured in order to verify that the two conditions did not vary in terms of perceived competitiveness.
After individually filling in the pre-negotiation questionnaire, participants were given 10 minutes to negotiate with each other. Once the 10 minutes were over, the experimenter stopped the negotiation. Lastly, participants filled a post-negotiation questionnaire designed to assess (1) objective outcomes; (2) satisfaction with both the process and the results; (3) self-and other-stereotypes of warmth and competence; 1 and (4) demographics.

Measures
Scores concerning first-offer intentions, aspirations, resistance points, and final outcomes were calculated by summing up the number of points associated with the distribution option selected for each of the four tasks. Zero-sum perceptions were measured with a four-item seven-point likert scale taken from Demoulin & Teixeira (2010, e.g., "in this negotiation, there will inevitably be a winner and a loser"; α = .73).
Satisfaction with the negotiation process was measured with 3 items (e.g., "To what extent are you satisfied with the negotiation process?"; α = .72) and satisfaction with the result through one item ("To what extent are you satisfied with your negotiation result?"). The inclusion of a post-negotiation satisfaction measure is highly common in this type of study. However, we did not have specific predictions concerning this measure.
Self-and partner-stereotypes were assessed along the two fundamental dimensions of social perception: competence and warmth (the Stereotype Content Model; for a review, see Cuddy, Fiske, and Glick, 2008). Participants were asked to indicate to what extent, in the course of the negotiation, they thought that they, themselves, as well as their negotiation partner were warm (4 items: kind, sociable, nice, and warm) and competent (4 items: capable, efficient, intelligent, and competent; αs between .77 and .93). Self-and other-stereotypes were included for exploratory purposes. We wanted to check for differences in stereotype activation as a function of the experimental condition. In principle, stereotype activation should be higher in the gendered condition compared to the neutral one. This effect was indeed previously observed (Demoulin & Teixeira, in prep). However, the fact that, in the present experiment, stereotypes were measured after the negotiation (and not before) means that the negotiation itself might have contaminated initial levels of stereotype activation. After finishing the post-questionnaire, participants were debriefed and thanked.

Results
Data from two dyads were excluded from the analyses: One dyad whose final outcome was more than 3 SD above the mean and one dyad that did not get to an agreement on two of the four tasks after 10 minutes.

Pre-negotiation measures
We performed multi-level analyses on negotiation performance. This procedure allows us to take into account statistical fluctuations due to the specificity of the dyad in which a given participant is placed (i.e., a random intercept per dyad was estimated allowing to isolate the effects specific to the predictors). Analyses were conducted using SPSS and the Albright and Marinova syntax (2010). Each of the dependent variables was predicted by the topic of negotiation (-1 = gender-neutral; 1 = feminine), the gender of the participant (-1 = men; 1 = women), and their interaction. 2 When an interaction was found, simple effects were examined using "dummy" coding procedures (i.e., assigning the value "zero" to males or females depending on the gender for which the effect of topic of negotiation was being examined).
Zero-sum perceptions. No significant effects were found concerning this variable. This absence of effects is important because it rules out any alternative explanation of the results in terms of differences in perceptions of competitiveness as a function of gender, condition, or their interaction.
First offer intentions and resistance points. No significant results were found for resistance points (ps > .12). With regards to first offer intentions, results show a marginally significant interaction effect, B = 29.52, t = 1.85, p = .07, 95% CI [-62.12, 3.07]. Simple effects analyses show that differences in terms of first offer intentions are not significant among men, p > .63. In contrast, for women, first offer intentions are marginally lower in the household duties condition (M = 489.37, ET = 153.90) than in the neutral context (M = 582, ET = 136.13, b = 46.31, t = -1.72, p = .09, 95% CI [-14.39, 199.64]). Although these analyses were essentially exploratory, they tend to suggest that women make somewhat lower first offers to an interaction partner when they envisage negotiating on household duties than when they believe that they will negotiate on a topic that is gender-neutral.

Negotiation performance
Again, we performed multi-level analyses on negotiation performance. Negotiation performance was predicted by the topic of negotiation (-1 = gender-neutral; 1 = feminine), the gender of the participant (-1 = men; 1 = women), and their interaction. All predictors were entered as fixed factors and the dyad as a random factor.

Mediation analysis
According to our hypothesis, the larger performance gender gap (in women's disfavor) found in the "feminine" condition compared to the neutral one is due to the lower aspirations set by women in the former condition compared to the latter. Aspirations should therefore mediate the impact of the interaction between topic of negotiation and gender on negotiation outcomes. In other words, we should find a mediated moderation (Muller, Judd & Yzerbyt, 2005) in which gender moderates the path between the independent variable (i.e. the topic of negotiation) and the mediator (i.e. aspirations). In addition, (as in simple mediation) when the mediator is entered in the equation, the impact of the Topic x Gender interaction on performances should decrease. Furthermore, the indirect effect should be statistically significant.
As seen above, the interaction between Topic of Negotiation and Gender had the predicted effect on negotiation outcomes: women obtained worse results in the "feminine" condition compared to the neutral one and compared to men. The same Topic x Gender interaction should also impact the mediator (i.e., aspirations). This effect is present as well: women had lower aspirations in the "feminine" condition compared to the neutral one and compared to men. Finally, in order to test our model, we computed gender difference scores for aspirations and performance and conducted bootstrap analysis using the PROCESS macro for SPSS (Hayes, 2013, Model 7, 5000 resampling). This procedure allows not only to verify the decrease in the total effect (i.e., the effect of the independent variables on the dependent one) when taking the mediator into account in the equation (i.e., the direct effect), but also to test the significance of the indirect effect.
Analyses confirmed the significant Topic x Gender interaction on the aspirations difference-scores (B = -58.26, t = -3.29, p = .001, 95% CI [-93.74, -22.78]), showed a significant effect of the mediator on negotiation performances (B = .41,t = 5.78,p < .001,), and a significant index of mediated moderation SE = 18.74,). In other words, bootstrap analyses support our hypothesis that the differential impact of topic of negotiation on performance as a function of negotiator's gender is explained by differences in aspirations.

Post-negotiation measures
Self and other stereotypes. Given that the target of judgment differed for women and men, we conducted separate analyses by gender of the participant. For each of the two groups, we performed a repeated measures ANOVA with Topic of Negotiation (feminine vs. neutral) as between-subjects and Target (self vs. partner) and Dimension (warmth vs. competence) as within-subject factors.
Satisfaction. Concerning satisfaction with the negotiation outcomes, no effects reached significance (all ps >.25). Satisfaction with the negotiation process was only affected by gender of the negotiator (B = .47,p = .001,), with women reporting being less satisfied than men (Ms = 4.51 & 5.45,SDs = 1.18 & 1.05,respectively).

Discussion
The present study puts forward the importance of the negotiation topic in determining negotiation behavior. Specifically, women involved in stereotypically feminine negotiations had lower aspirations and, by consequence, got worse results than both women who negotiated in gender-neutral contexts and men. Indeed, differences in aspirations of women and men mediated the (detrimental) impact of the ("feminine") negotiation topic on women performances.
In line with previous research (Bear & Babcock, 2012), we had formulated a hypothesis according to which a traditionally feminine negotiation topic would trigger better performances among women (compared to men or to other women in neutral contexts). The idea behind this hypothesis was that women would feel more comfortable and have greater expertise when negotiating feminine topics. This (increased) confidence would lead them to have higher aspirations and, in turn, better performances. Our data does not seem to support this hypothesis.
We had also envisaged the reverse effect as an alternative hypothesis: women would perform worse in stereotypically feminine negotiations than in neutral ones and than men in general. According to this alternative hypothesis, this would be due to the fact that the former tasks would more easily activate gender stereotypes which should lead women to perform less well (Kary et al., 2001(Kary et al., , 2002(Kary et al., , 2004. Our results provide support for this latter hypothesis. The inconsistency between our results and those obtained by Bear and Babcock (2012) is noteworthy given that the stereotypicality of the negotiation topic is manipulated in both studies. However, a closer examination of the two experimental manipulations provides some insight into the reasons for this inconsistency. In the present study, the "feminine" negotiation topic was the distribution of household duties, whereas in the Bear and Babcock's study the "feminine" negotiation condition consisted in selling a piece of jewelry. These two negotiation topics are quite different in terms of the social norms they evoke. While household duties are likely to activate gender inequalities, the jewelry topic is probably associated to a situation in which women are more competent than men. It is therefore possible that, in the former situation, women develop lower aspirations (influenced by an unfavorable norm) than women in the latter (boosted by a social norm that puts forward their superiority over men).
Another aspect that we find important to mention is the absence of effects of negotiation topic among men. Indeed, research on stereotype threat effects in negotiation has sometimes found an increase in performance of the positively stereotyped group (Kray et al., 2001;. That being said, gender effects tend to be weaker among men (cf. Kray et al., 2001;Bear, 2011) and the majority of stereotype threat research has focused on negatively stereotyped groups. One plausible explanation for this absence of effect could be that stereotype activation (as a consequence of negotiation topic) is less pronounced among men than among women. This proposition is in line with research showing that powerful people tend to give priority to information related to their own goals (Slabu & Guinote, 2010). One could therefore argue that men would be especially focused on performing well in the negotiation and would give little importance to other kinds of information (i.e., the negotiation topic). However, future research is needed to clarify potential effects of negotiation topic among men.
It is also important to point out the absence of effects on zero-sum perceptions of the negotiating parties. This suggests that differences in women's performance cannot be attributed to differences in perceptions of competitiveness. Furthermore, this result is in sharp contrast with the expectations of the "experimenters". Indeed, informal debriefings with the student experimenters suggested that the students predicted a better performance of women in the "feminine", "household duties" task because they expected women to be especially belligerent in this condition.

Implications and future directions
Our results are in line with research on stereotype threat in negotiations. To the same extent that traditional research on this phenomenon has been generalized to other groups such African-Americans (Stone, Lynch, Sjomeling & Darley, 1999) or psychology students (Croizet, Després, Gauzins, Huguet, Leyens & Méot, 2004), gender effects in negotiation should be applicable to other types of groups. We are probably dealing with intergroup phenomena that go beyond the specificity of the gender categorization and are more about the stereotype content of the groups at stake (Demoulin & Teixeira, 2010).
In addition, despite of the rather artificial environment in which the present experiment was conducted, the practical implications of our study should be pointed out. Our results suggest that women have lower aspirations when negotiating about issues related to pervasive gender inequalities and that these aspirations lead to lower performances. We can therefore envision women as the "weaker link" in negotiations concerning work-life balance (e.g., childcare facilities, flexible working hours, maternity leaves, etc.). Being aware of these mechanisms is of high importance if one takes into account that men not only tend to avoid discussing these topics but also normally occupy managerial, high-power positions (Bear, 2011). In addition, the perception of a family-friendly work environment is positively correlated with employee retention and job satisfaction (Scandura & Lankau, 1997). In practice, one way to minimize the impact of these biases is to frame negotiations as neutrally as possible while institutionalizing discussions on topics that are important for women's well-being at work.
The present research adds to the literature on gender differences in negotiations by showing that stereotypically feminine negotiation topics might act against women. The contrasted results obtained in the present study and