Sex Roles

, Volume 57, Issue 5, pp 329–339

Gender Differences in Virtual Negotiation: Theory and Research


    • Department of PsychologyDePaul University
  • Maryalice Citera
    • SUNY—New Paltz
  • Toni Willis
    • SUNY—New Paltz
Original Article

DOI: 10.1007/s11199-007-9252-y

Cite this article as:
Stuhlmacher, A.F., Citera, M. & Willis, T. Sex Roles (2007) 57: 329. doi:10.1007/s11199-007-9252-y


Social roles create conflicting behavioral expectations for female negotiators; however, virtual negotiations reduce social pressures. This paper reviews theoretical explanations on why men and women might differ in negotiations that occur through email, telephone, or video. Forty-three negotiation studies comparing face-to-face and virtual negotiations were examined for gender differences. All studies were reported in English but not limited to US participants. While many reports omitted gender information, meta-analytic findings supported the prediction that women would be more hostile in virtual compared to face-to-face negotiations, as well as finding no hostility difference for men between virtual and face-to-face negotiations. While negotiators overall were more successful face-to-face than virtually, results separated by gender did not find this effect.


E-mailGenderNegotiationSocial RolesVirtual Negotiation


Research suggests that women, on the average, experience negotiation differently than men (e.g., Gerhart and Rynes 1991; Stevens et al. 1993; Stuhlmacher and Walters 1999; Walters et al. 1998). On the other hand, negotiations that are not face-to-face but through email, telephone, or video (or what we call virtual negotiations), may present a situation where gender differences may be reduced or even reversed. This paper examines theory and research on gender and virtual negotiation, offers preliminary tests through meta-analyses of the resulting predictions, and suggests avenues for further research. This research is particularly timely given the prevalence of virtual communications and the fact that conflict across many diverse areas (e.g., employment, pay, status, household chores, and work–life balance) are most commonly resolved through formal or informal negotiations.

Women in negotiation often reach less favorable agreements than men (e.g., Dalton and Todor 1985; Dalton et al. 1987; Gerhart and Rynes 1991; Stevens et al. 1993; Stuhlmacher and Walters 1999). In a study of MBA graduates, Gerhart and Rynes (1991) found that women negotiated smaller starting salaries than men did. Men and women received comparable initial offers and attempted negotiations at about the same rate. The difference was in the amount of salary increase men and women secured during their negotiations. On average, men negotiated adjustments that were $743 more than those negotiated by women.

These gender differences in negotiation appear consistent throughout research. Stuhlmacher and Walters (1999) conducted a meta-analysis examining studies that reported gender and negotiation profits. They found that, overall, women negotiated significantly lower outcomes than men. Although the overall effect was small, it was consistent across studies. The gender difference for agreement value was homogeneous across a variety of tasks and contexts.

Some have suggested that negotiation outcomes vary by gender because women approach negotiations differently than men. There is some evidence that women negotiate differently than men. Smeltzer and Watson (1986) found that women’s communications patterns in negotiation differed from men’s. Less assertive (more submissive) communications patterns were found with women exchanging more disclaimers and interruptions than men during collective bargaining sessions. Walters et al. (1998) found, across 62 research reports, women exhibited more cooperative behaviors than male negotiators. The effect was very small, but significant and consistent. Moderators influenced the extent of these gender differences. One intriguing finding showed that the effect was reversed (i.e., women were more competitive than men) in negotiations that involved prisoner dilemma or matrix games rather than traditional face-to-face negotiations. Prisoner dilemma/matrix games involve less communication, less relationship building, and less possible face-to-face contact than explicit negotiations. In many ways, these games have similar (but not identical) characteristics to virtual negotiation. Unfortunately, the Walters et al. meta-analysis was unable to include virtual negotiations in their analysis; too few virtual studies were located that reported the necessary gender information for comparison.

Likewise, others suggest that negotiation outcomes vary by gender in part because negotiation partners are influenced by expectations about the opponent (Kray and Thompson 2005). In a controlled set of field studies (Ayres 1991, 1995, 2001; Ayres and Siegleman 1995), female car buyers were given higher initial offers and ended with higher car prices than male car buyers. These experiments carefully controlled for factors such as negotiation strategy, occupation, attire, attractiveness, economic status, and script. Even the addresses of the car buyers were identical. It appears that the partner’s gender influenced the negotiation opponents (i.e., car salespeople). Given these results, it is relevant to consider how virtual negotiation contexts might minimize partner expectations about gender.

Now is an opportune time to examine virtual negotiation in light of gender theories. No longer do people talk about the potential for e-negotiation and its rapid growth and popularity; the focus now concerns when e-negotiation is appropriate, inappropriate, and what conditions might influence the outcomes. Thus, this paper integrates two areas of literature (gender roles and virtual communication) to offer theory and predictions concerning the impact of gender in virtual negotiation. We turn first to social role theory (Eagly 1987) as an organizing theoretical framework.

Social Roles

Social role theory suggests that people develop expectations for their own and others’ behavior based on their beliefs about what is suitable for members in that role. People may hold multiple social roles simultaneously (i.e., professor, mother, woman, volunteer). A social role has both descriptive aspects (expectations about what role members actually do) and injunctive aspects (what role members ought or ideally should do; Eagly and Karau 2002). Looking specifically at the impact of gender in negotiation, we suggest that multiple and perhaps conflicting social roles are present for women. And, in the case of virtual negotiation, there are specific characteristics about virtual communication that influence the expectations about negotiators and gender roles. It is relevant to consider role congruency (Eagly and Karau 2002) in negotiation, that is, the extent to which negotiators behave similarly to role expectations.

Before discussing the specifics of negotiator roles and gender roles, we first consider how the virtual context affects roles and role expectations.

Virtual Communication and Negotiation

Virtual communication uses communication modes other than face-to-face (FTF) such as telephone, email, or written notes. Several theories can be brought to bear on the roles (specifically the role expectations of actual and ideal behavior) present across communication modes (Daft and Lengel 1984; Latané 1984; Wellens 1986, 1989; Williams 1977). Across all of these theories, the critical factor is the number and type of social cues filtered out. Social cues are the verbal and nonverbal information in the communication process. Verbal cues range from voice hesitance, tone, and immediacy of response while non-verbal cues include eye contact, body language, facial expression, and physical appearance (Tanis and Postmes 2003).

These social cues influence impressions about communicators and the meaning of their message (Daft and Lengel 1984; Latané 1984; Nowak 2003; Sproull and Kiesler 1986; Tanis and Postmes 2003; Wellens 1986, 1989). Tanis and Postmes (2003) found that social cues are significant factors in impression formation. Participants were given either a descriptive biography, a picture, a biography plus a picture, or nothing about another person in a computer mediated discussion. Individuals with more information about their partners expressed less ambiguity in their impressions of the partners. The results also showed that increased social cues promoted the development of more positive impressions (Tanis and Postmes 2003).

Social cues are part of social roles and communicator status. Traditional status levels, and thus roles, are not as salient within virtual communication. By removing or changing the types of social cues, virtual communications equalize social interaction. Virtual communication, in essence, promotes a sense of equality for users. For example, Sproull and Kiesler (1986) examined interpersonal communication through e-mail and found that there was no distinction between messages written by subordinates and supervisors. Interestingly, more uninhibited behavior has also been found in virtual communication than FTF (Kiesler et al. 1984). When communicating over the computer, individuals displayed more “flaming” behaviors than when communicating FTF (Kiesler et al. 1984). Flaming behaviors included name-calling, cursing, and resorting to put downs. This equality as well as reduced inhibition would be consistent with the idea that roles are less prescribed in the virtual environment than FTF interactions.

One explanation for cues being less salient is that virtual interactions represent what personality researchers call a weak situation (Mischel 1977; Weiss and Adler 1984). Weak situations are those in which a person feels more latitude in acting in accordance with personal beliefs, dispositions, and emotions. Strong situations involve clear and direct expectations regarding appropriate behavior. Or in the context of social role theory, weak situations lessen the influence of social roles. Thus, both the expectations about what a virtual negotiator may or may not actually do, as well as the prescriptive elements of what behaviors should occur, are weaker than FTF. The gender roles, and their associated norms, are weaker as well.

We expect that virtual negotiation has less salient cues and less prescribed roles than FTF interactions. Stuhlmacher and Citera (2005) define virtual negotiations as those negotiations that “occur using media other than face-to-face communication (e.g., telephone, e-negotiations, video-conferencing)” (2005, p. 70). Their meta-analytic review of virtual negotiations found that virtual negotiations were more hostile than face-to-face negotiations. Consistent with social role theory, more hostility was present when virtual negotiators were anonymous as opposed to when they were identified. In addition to hostility, anonymity moderated the effect of virtual negotiation on profit. Profit between FTF and virtual negotiators was more similar when virtual negotiators were identified than when they were anonymous. Anonymous virtual negotiators were less profitable than identified opponents.

Another example of e-negotiators being more hostile than FTF negotiators is demonstrated in research on credibility. In Citera et al. (2005), e-negotiators reported being less credible than FTF negotiators. Lying behaviors were significantly correlated with credibility. As further evidence of weaker situational strength in negotiation, Citera and Beauregard (1997) found that personality and negotiation behaviors correlated higher for e-negotiators than for FTF negotiators. In particular, individuals advocating deceit as a useful negotiation tactic (scoring high on the deceit scale of the Machiavellian personality scale) were less credible in e-negotiations than those scoring low. No differences were found for FTF negotiations. The correlation between deceit and self-reported credibility was considerably different in e-negotiations (r = −.42, p < .01) than FTF negotiations (r = −.24, n.s.). Likewise the correlation between deceit and lying behavior was stronger in e-negotiations (r = .46, p < .01) than FTF negotiations (r = .15, n.s.). Negotiators were more likely to act in accordance with their dispositions in e-negotiations, the weaker situation, and less likely to fall back to dispositions in FTF negotiations, the stronger situation. This is also consistent with social role theory, in that the weaker situation (e-negotiation) had less injunctive norms on what negotiators ought to do.

Thus, we suggest that FTF negotiation is a strong situation because there are clear norms and expectations for negotiation and social interaction. In contrast, virtual negotiation represents a weak situation because its restrictive bandwidth and reduced social cues provide ambiguity in how to act. The weak situation reduces social attention to both descriptive and injunctive social norms. In virtual negotiations, then, social roles hold less influence on negotiation behaviors for all parties than their FTF counterparts.

Gender Roles and Negotiator Roles

The process of virtual negotiation becomes more complex when we consider the place of gender roles. In most societies, men and women have different social roles, and then are expected to have the skills to fulfill these roles. Relevant to negotiation are societal expectations for women to communicate in a friendly, warm, supportive, and selfless manner (Eagly and Carli 2003). Women are expected to have and exhibit more communal characteristics than men. In contrast, men are expected to adopt more agentic behaviors. Men would be expected to interact with more assertiveness, confidence, self-promotion, dominance, and one-up-manship than women.

It is important to note that in addition to being expected to behave in certain ways and perceived as having the skills to fulfill various roles, individuals may accommodate to the roles. For example, Eagly and Wood (1999) discuss that women may acquire relationship or domestic skills from being in a caretaker role, while men may accommodate to employment roles and incorporate qualities like assertiveness or dominance. Gender-typical roles would then drive expectations and gender norms. Social role theory suggests that certain situations make gender roles more or less salient. In particular, the saliency of gender roles may be influenced by such things as the negotiation group’s composition, communication form, communication style, and type of task (i.e., leadership vs. nurturing).

Several researchers (e.g., Kray and Thompson 2005; Putnam and Kolb 2000) suggest that negotiation in general has been viewed as a masculine role. Stereotypically masculine traits (strong, dominant, assertive, rational) are seen as more important for negotiation success than stereotypical feminine traits (weak, submissive, intuitive, emotional). It is also the case, however, that negotiation research has focused almost exclusively on masculine gendered activities. (For an exception in negotiation research, see Miles and LaSalle 2006 which manipulated the gender context of the task). Much research has portrayed negotiation as a competition or game, with winners and losers, and limited social interaction. In Walters et al. (1998) meta-analysis of cooperation in negotiation, all the tasks were masculine gendered (e.g., negotiating turbo engine parts) and not a single task involved a feminine gendered context (e.g., caretaking).

Putnam and Kolb (2000) discuss that there are many ways parties may come to an agreement, but experimental negotiation tasks have focused on “trade and transaction” rather than more complex situations. Frequently, the experimental situation minimizes relationships and produces very strong situations. Feminine social roles may play an important part in negotiations, building long-term relationships and problem solving. Kolb (2000) contends that collaboration is indeed a desirable negotiator behavior if negotiation is an ongoing part of the relationship, but that negotiation research has been limited to masculine gendered contexts.

While we maintain that effective negotiators (both men and women) likely must draw on a wide collection of behaviors (masculine, feminine, and neutral), depending on the situation and context, there is convincing evidence that negotiating or being a negotiator, is a masculine social role. Role incongruity theory (Eagly and Karau 2002) would then suggest that behavior congruent with a role would be most successful, hence masculine behavior would be seen as most appropriate and successful.

Gender and knowledge of gender stereotypes has been found to affect negotiation performance (Kray et al. 2001, 2002). The implicit activation of the stereotype that men are more effective negotiators than women had a detrimental effect on women’s negotiation outcomes and a positive effect on men’s (Kray et al. 2001). Men’s negotiation outcomes increased and women’s decreased when the negotiation was labeled diagnostic of ability.

When stereotypically female characteristics associated with successful negotiator tactics were activated (e.g., verbal expressiveness, good listening skills), female negotiators’ performance was enhanced while male negotiators’ performance was hindered (Kray et al. 2002). Interestingly, when gender-neutral characteristics associated with negotiator success were activated, men’s performance increased and women’s performance decreased. This result suggests that the default expectation in a negotiation is that men will do better than women.

The roles in negotiation create a dilemma for female negotiators. Although a masculine style is seen as more effective, when female negotiators use this agentic/instrumental communication style they will be perceived negatively. Thus, there is pressure in negotiations for women to use an affiliative/communal style of communication. If women, however, act in accordance with prevailing female roles, they may be less successful in negotiations or may be likely to be perceived as less successful in negotiation than men. If women use the more agentic style, they may be concerned with being sanctioned for their deviation and expect to be judged harshly. The agentic/instrumental style may result in their being labeled as domineering or pushy. In some respects, women have to strike an uneasy balance between using a strategy that is perceived as successful and the negative labels that might result from using it. Women negotiating in a standard face-to-face situation would be expected to interact with cooperation and low hostility or suffer social consequences.

The labeling of the task as negotiation influences reactions to it, and thus creates expectations of the social role. In their recent book, Babcock and Laschever (2003) report substantial differences when the same task is called “asking” vs. “negotiating.” Women were much more comfortable “asking” than negotiating. This would be consistent with the female social role of being submissive and dependent. Babcock and Laschever suggest the reluctance of women to negotiate creates inequities in outcomes that are perpetuated in society. We suggest that the existing social roles influence not only how women act, but how they feel obligated to act, and how others interpret their actions.

Not only is the feminine gender role inconsistent with perceptions of negotiation success, related evidence also suggests that there are gender differences in conforming to the expectations of others. Women are predicted to be more sensitive to others’ expectations than men. Cross and Madson (1997) found that men maintain an independent self-construal while women maintain an interdependent self-construal. That is, men’s cognitions, emotions, and behaviors are less influenced by relationships than women’s cognitions, emotions, and behavior (Cross and Madson 1997). Thus, men in negotiation may focus more on their own outcomes and strive to attain the highest amount of profit while being minimally influenced by the other party. Women negotiators, on the other hand, may be more concerned with preserving the relationship and exhibit more cooperative behavior than men exhibit. Because women are more interdependent in their self-construals, they are more influenced by strong norms for how they are expected to interact than men.

Some have suggested that interactions of gender and status confound interpretations of gender differences. Watson (1994), for example, holds that gender differences in negotiation result from power differences. Social inequalities are often encountered by women; women unconsciously self-categorize themselves into lower status positions when interacting with others. This would be especially influential in competitive situations such as negotiations. Compared to negotiations with less personal information, the social environment in FTF negotiations makes gender a salient category and increases the connection between women and low status (Palomares 2004). According to Kray and Thompson (2005), “Women appear to adjust their behavior—becoming more acquiescing when lacking power and more exploitative when possessing power—to a greater extent than men do” (p. 159). This can also be explained by the competing social roles present for power and gender.

We suggest that social role theory goes beyond gender stereotypes to explain how situations or contexts can influence behaviors, attitudes, and outcomes. According to Eagly and Karau (2002), stereotypes connote a fixed attitude toward a social group; these attitudes about a group are brought to bear on behaviors and attitudes related to an individual member of that group. Thus, a gender stereotype explanation about women’s performance in negotiation would be attributed to attitudes about women. Beyond a fixed stereotype of women, however, social role theory offers predictions on how and when context may influence the behavior and perceptions of the negotiator and the partner. For instance, social role theory would maintain that gender roles are more salient in certain situations (see Eagly and Karau 2002, particularly pp. 576–578). Gender roles are more salient in gendered tasks (negotiating for turbo engine parts vs. childcare), gendered personal appearance (e.g., negotiators in feminine dress; pregnant), gender uniqueness (e.g., face-to-face; one gender as a minority or token), or under situations with limited cognitive resources (e.g., time pressure or information overload) that may trigger reliance on accessible norms. Likewise, negotiator roles would be more pronounced in situations where negotiation is expected (e.g., car purchases), when formally labeled as a negotiation (vs. asking or problem solving), or are FTF versus virtual.

Perceived incongruity between negotiator role and gender would then depend on the definition of the negotiator role, traditional beliefs in gender roles, and the importance or salience of gender. Thus, women would be particularly disadvantaged in tasks with a masculine context that involves interacting with others who hold traditional gender views, and where women are in the minority. Likewise, role incongruity aspects of social role theory would predict that men would be disadvantaged in tasks with a feminine context that involves interactions with others with traditional gender views, and where men are a minority. This pattern of results was found in a meta-analysis of gender and leader effectiveness. Eagly et al. (1995) found that while women were somewhat less effective than men in leader contexts with masculine definitions (e.g., military), men had somewhat poorer outcomes than women in leader roles with feminine definitions (e.g., elementary education, nursing).

In short, our review suggests these points. First, social roles provide frameworks on how to act and how to evaluate the behaviors of other. Negotiation, as it has traditionally been conceptualized, has a masculine social role. In most virtual negotiations, the negotiation task remains a masculine social role with its focus on competition, winning, and limited personal contact. Virtual negotiation, however, makes the gender of the negotiator less salient (compared to FTF negotiation) because there are fewer status and social cues.

Gender role salience then has different implications for men than women negotiators. Instead of focusing on the status of the other individual, women in virtual negotiation may be more likely to concentrate on the content of the interaction (Nowak 2003; Ridgeway 2001). Recall also that psychological distance theory predicts that virtual negotiations lessen the social influence of the other communicator (Latane 1984; Wellens 1986, 1989). The reduced social cues in virtual negotiation may equalize social interaction, making women’s behavior less consistent with a female gender role and more consistent with a negotiation role.

If this is the case, women may be more hostile virtually than FTF, while men’s behavior would not differ. We would predict that in virtual negotiations, women negotiators will be less guided by female gender roles, and there will be less stereotypical affiliative behaviors than in a FTF negotiation. This would result in women being less cooperative and more hostile in virtual than in FTF negotiations.

Specifically, we predict:
  1. Hypothesis 1

    Female negotiators will display more hostile behaviors in virtual negotiations compared to FTF negotiations.

    We also would expect no differences for men based on the mode of communication, in part because being a negotiator is congruent with a masculine role.

  2. Hypothesis 2

    No difference will exist in displays of hostile behavior for male negotiators between virtual and FTF negotiations.

    If being a negotiator is indeed a masculine role, we should see women’s performance increasing when there is less role incongruity. That is, not only would women display role consistent behaviors (less stereotypically feminine behaviors), but this would increase their outcomes in terms of profit or objective performance. Given that the current negotiation paradigms have been almost exclusively masculine or at least gender neutral, we predict:

  3. Hypothesis 3

    Female negotiators will have better objective outcomes, or profit, in virtual compared to FTF negotiations.


To test these predictions, we meta-analyzed the effects of gender and virtual negotiation on two different dependent variables: hostility and profit in the negotiation. In terms of virtual media type, studies were considered that had video negotiation (with seeing and hearing), audio only (telephone, barrier), or text only (notes, email, computer/electronic). By combining effect sizes across existing studies, we hoped to offer some evidence for these predictions concerning gender in virtual negotiation.


Computer databases and recent publications were searched to locate studies to include in the meta-analysis. Rather than search for gender as a keyword, all possible studies comparing virtual negotiation to FTF negotiation were collected and screened for gender information. This was done to retrieve gender data even if gender was not a main focus of the report. The computer databases of PsycINFO (1887 to September, 2006), Educational Resources Information Center (ERIC 1966 to September, 2006), Social Science Research Network (October, 2006) and Digital Dissertations (1861–August 2006), were searched for negotiation terms (bargain-, negotiati-) combined with virtual negotiation terms (face to face, audio, visual access, communication, computer mediated, medium).

The reference list of the studies retrieved and recent books were also searched for relevant research. Several attempts were made to find unpublished research to minimize publication bias. If a presentation was referenced, copies of the presentation and other unpublished works were requested from the author.

Criteria for Inclusion

The sample was limited to studies that manipulated the communication form of the negotiation along with a comparable face-to-face negotiation. In terms of virtual media type, studies were considered that had video negotiation (with seeing and hearing), audio only (telephone, barrier), or text only (notes, email, computer/electronic). A typical study involved a face-to-face negotiation condition as well as at least one condition where parties negotiated by email, telephone, or video. Each party typically relied on a payoff table indicating the profit associated with certain outcomes. Common negotiation tasks involve buying and selling or negotiating a work contract.


The studies were coded for characteristics of reports. Coders recorded the gender of participants (males, females, or not reported/mixed). Most studies included both male and female participants and did not break results out by gender. Only two studies reported separate findings for both men and women. The remaining studies used either only male or only female participants.

Two individuals independently coded the majority of reports. Agreement rate between the two coders was over 95% with discrepancies resolved by discussion. Two dependent variables were examined: negotiators’ hostile behaviors as well as the profit received. Both of these variables were predicted to be related to gender and whether the negotiation took place face-to-face or virtually.

Hostile Behavior in Negotiation

The first dependent variable of interest was competitive or hostile behavior. These behaviors were operationalized as either (a) linguistic measures (e.g., lies, threats, putdowns), or (b) offer patterns (e.g., amount of concessions, number of alternative offers). An example of a hostile offer pattern would be a negotiator who never concedes any values or makes counter offers. The difficulty interpreting this behavior is that while it may indicate hostility, a slow concession rate may also indicate a competitive strategy. However, both language use and offer pattern have been used in previous research as indicators of hostility, or if not outright hostility, then competitiveness. Given the use in the literature of various operations, and the difficulty interpreting motives from behaviors, we felt it appropriate to include both types of measures. It is reasonable that some behaviors might be seen as more hostile than others. Follow-up analyses were conducted to determine if different operationalizations would lead to different conclusions.

If information on both of these types of measures was reported in a single study, we used the linguistic measures rather than offer pattern. In studies that did not report linguistic communication information, offer pattern was used as a measure of hostile behavior. Self-reports of one’s own behavior, or ratings of the opponent’s behavior, were not included as measures of actual behavior within the negotiation. Twenty-five effect sizes from 22 different articles representing 1878 research participants were included.

Negotiator Profit

The second dependent variable was the objective profit of the negotiations across conditions in the studies. The negotiation literature uses the term “profit” to represent the extrinsic value gained from a negotiation. The profit may be calculated from a payoff table of agreement values in money units (e.g., dollars), points, or in the amount of an outcome (e.g., actual starting salary). The value or worth of the agreement is the “profit.” While not familiar terminology outside of negotiation research, we will use the term profit (rather than other terms like agreement value) to indicate the objective, rather than subjective, value of the negotiated outcome.

Profit was operationalized as the points or “money” earned by the negotiators. Studies may have reported an individual or joint (all party) profit. In several studies, both an individual and joint profit was reported. In these cases, individual profit was used. Further, the focus was on objective profit and did not consider the subjective or self-reported quality or satisfaction of the outcome. Thirty-eight effect sizes from 35 different articles representing a total of 3,168 research participants were included.


Hostile Behavior

Results found gender differences in the effect of virtual communication on hostility (See Table 1). As predicted (Hypothesis 1), women were significantly more hostile in virtual than face-to-face negotiations (d = .64, 95% confidence interval (CI)=.32/.97). For men, no significant difference was found between their FTF behavior and their behavior in virtual negotiations (Hypothesis 2; d = .06, 95% CI = −.16/.28). Studies combining men and women, or not reporting gender breakdowns, were also significantly more hostile in virtual than FTF conditions (d = .47, 95% CI = .35/.57). While the effect size when gender composition is unknown is not directly interpretable in terms of gender differences, it does provide a baseline to compare the known effect sizes for men and women.
Table 1

Comparison of face-to-face and virtual negotiation for hostile behavior and profit across gender.


Number of effect sizes

Adjusted effect sizes

95% CI


Within class (Qw)

Between class (QB)







Split by negotiator gender













Mixed/Not reported












Split by negotiator gender













Mixed/not reported






Positive effect sizes indicate more hostile behavior in virtual negotiation or more profit in virtual negotiation than FTF negotiation. Significant Qw indicates rejection of homogeneity within class. Significant QB indicates rejection of homogeneity between classes. *p < .05, **p < .01.

The gender of the negotiators significantly moderated the effect of media on hostility; the between class effect was significantly different between female, male, and mixed gender studies (Q(24)=12.71, p < .01). It is important to note that significant variation was present within all groups and other factors might be important to explain additional variation across studies.


Results found gender differences in the profit between virtual and FTF negotiations (see Table 1). Although not significant within gender categories, the effects were significantly different between classes of female, male, and mixed gender negotiations (Q(37)=26.01, p < .01). While not significant, women did better in virtual negotiations than FTF (d = .25, 95% CI = −.08/.58), as did men (d = .12, 95% CI = .04/−.29). The small number of studies reporting profit by gender requires caution in interpretation. Thus, we have only partial support for Hypothesis 3 in that results differed when gender information was available from when gender was not broken out. It is important to note, however that separating by gender provides somewhat different results than when gender is not specified (c.f., Stuhlmacher and Citera 2005). Curiously, across the set of studies that did not report gender or used a mix of male and female participants, higher profits were found in FTF negotiations than virtual negotiations. Again, it is important to note that significant variation was present within groups. Further exploration of possible moderators of these effects is clearly warranted to determine if characteristics of the studies or the context influenced the results. Ideally, we would want to know if these patterns would hold in larger samples. Unfortunately, the sample size limited the meaningfulness of further breakdowns within gender.

Exploratory Analyses

To explore if the type of measure of hostility influenced the results, three separate additional analyses were run. As mentioned previously, several types of measures were considered hostile based on the language used or the offer pattern. In the first analysis, hostility measures were separated into language or offer pattern measures. Thus, we compared if linguistic or offer pattern measures had different effects in the comparison of virtual to FTF negotiation. While language measures had a bit larger effect than offer pattern measures (d = .45 vs. d = .27) in comparing virtual and FTF, there was not a significant difference between conditions (Q(25)=2.71, p = .10).

In the second exploratory analysis, the hostility measures were coded as “covert” or “overt”. Covert behaviors are ones that may not be obvious to the opponent; overt behaviors would be obvious to the opponent. Covertly hostile behaviors included lies and deception as well as offer pattern measures that were not visible such as large differences between secret aspirations and initial offers. Overtly hostile behaviors included language measures like flaming, threats, and persuasion tactics as well as offer pattern measures like small consensus change and restricted bidding strategies. The overtness of the behavior was not a significant moderator (covert behavior d = .48 and overt behavior d = .36) in virtual vs. FTF negotiation (Q(25)=1.40, p = .24).

Finally, measures were separated into the ones that were more clearly hostile and ones that might better reflect competitive strategies, but perhaps not strong hostility. Lies, flaming, criticism, putdowns were coded as clearly hostile while less hostile and perhaps simply competitive behaviors included low concession rate, amount of persuasion tactics, change in consensus, or low information exchange. In this case, there was a significant moderating effect between measures in the two categories. While both extreme hostile behaviors and competitiveness were more prevalent in virtual conditions compared to FTF, there was a larger difference in hostile behavior (d = .51) than the competitive behavior (d = .29; Q(24)=5.77, p < .05) between virtual and FTF. It is important to note that this difference does not necessarily change our interpretation that negotiators become more difficult in virtual negotiation, but it does suggest that more deviant behavior is seen in virtual situations. Sadly, the results are unable to shed light on gender differences. It would be extremely interesting to run theses analyses by gender; sample size issues make this inappropriate here. It appears that what behaviors are measured is another important aspect in understanding the effects of virtual negotiation.


The last decade has seen a re-emergence in the study of gender differences in negotiation situations (e.g., Babcock and Laschever 2003; Kolb 2000; Kolb and Williams 2000; Kray et al. 2002, 2001; Stuhlmacher and Walters 1999; Walters et al. 1998). These authors and others suggest that there are differences between men and women in responding to conflict and negotiations. How people deal with negotiation contributes to their mental health, employment opportunities, pay, status, and a multitude of other tangible and intangible outcomes. In addition, multiple negotiations are transacted throughout a single day and the cumulative differences between men and women in outcomes may become particularly striking. Despite the fact that effect sizes for gender difference may be small, their collective impact over time can be substantial. Even small, one-time differences can be of great consequence. A car purchase scenario offers a concrete example of the meaning of small effect sizes. Assume a specific car model sells on the average for $25,000 with a standard deviation of $2000. If women on average paid $700 more than men, the effect size (r = .17) would not be significant and would account for less than 3% of the variance in the car price. This would be considered a small effect. While objectively 3% or $700 is labeled a “small” difference, many would consider this 3% difference between group outcomes as unfair and hardly trivial, particularly if this cost difference was consistently found between groups.

Meta-analytic results have found significant, consistent, but small effects. It remains important to understand the circumstances that maximize, minimize or reverse the differences. With the expansion of e-business and the increased likelihood that individuals will engage in e-negotiations, exploration of gender and e-negotiation is particularly timely.

As predicted, our findings suggest that the hostility in virtual negotiation was related to the gender of the negotiator. Women were significantly more hostile in virtual than face-to-face negotiations. For men, no significant difference in hostility was found between FTF and virtual negotiations. Women may tend to be less hostile face-to-face due to social roles that prescribe women to be affiliative or relationship oriented. With the reduced female social role, women tended to be more hostile; this suggests that an agentic/instrumental communication style is the default style on “how to negotiate.” Virtual negotiation allows women to ignore status cues, reduces pressures to be affiliative, and allows women to adapt to the expectations of the task of the negotiation rather than the relationship.

In terms of outcomes of the negotiation, the gender of the negotiators significantly moderated the effect of media on profit. While our results suggested that virtual negotiation led to better negotiation outcomes for women than FTF, it is premature to suggest that virtual negotiation will always lead to better outcomes than FTF for women. In fact, the majority of the studies (where gender statistics were not reported) found the opposite – better outcomes transpired FTF than virtually. It is very important to explore the other factors of the negotiation that could interact with or supersede the influence of gender. It is clear that many variables may magnify or shrink the gender influence and need further exploration.

Future Directions

This paper offers explanations on why we might expect gender to play a role in negotiations and offers avenues for further research. More empirical work is needed to explore dyad composition, gender salience, forms of hostility and the relationship to outcomes. One promising direction for future research could address how the interface and context of an e-negotiation may influence gender effects. For example, situations may differ in terms of social cues, status cues, and anonymity across different forms and means of e-negotiation. Different implementations may influence personalization and relationship building (e.g., “schmoozing,” Morris et al. 2002) which have implications for activating gender role stereotypes and decreasing psychological distance. The gender composition of the negotiation (male–male, male–female, or female–female) may influence the saliency of social roles.

Additionally, the cultural stereotypes about men and women may differ by geographical country or region, by the type of negotiating task (childcare vs. sales), or requirements of the job (self-interest vs. negotiating on behalf of a client). Also useful would be to separate perceptions about male and female negotiators from their actual behavior and outcomes. Active gender stereotyping may be fostering perceptual differences that are even greater than the objective differences that exist between men and women. The literature reviewed earlier suggests that these subtle aspects of the negotiation may very well contribute to gender differences in negotiation, and may be explanations as to why sometimes gender does and does not have an effect.

While changes in social roles are expected as society and the workforce change, these expectations may be slow to change. Indeed, examining studies over time has not found evidence of reduced gender differences in behaviors or outcomes of negotiation (Stuhlmacher and Walters 1999; Walters et al. 1998). With further study, we may be better able to provide information for the training and the practice of negotiation, and the design of e-negotiations. Negotiators should be aware if, and how, behaviors and outcomes might be tied to gender and gender roles. Extending findings from gender and leadership, Eagly and Karau (2002) suggest that women “receive more disapproving and uncooperative reactions than men do when they proceed in an assertive and directive manner. However, these unfavorable reactions may dissipate at least partially when women complement their agentic repertoire with communal behaviors that are consistent with the female gender role, as long as these behaviors do not violate the relevant leadership role” (p. 590). In the case of negotiation, negotiators could then be trained to work effectively with these perceptions, to consider power differences amongst disputants during the process, as well as on specific techniques for changing the situation or handling the ways disputants relate to one another.

Virtual negotiation offers unique challenges and opportunities for men and women. Across virtual negotiation contexts it is possible that gender differences may be substantial under certain circumstances, while minimal in others. Our results provide an important step to clarify this relationship and open further discussion on the topic. This understanding is particularly important given the roles of men and women and the prevalence of virtual and e-negotiation in our workplaces and society.

Copyright information

© Springer Science+Business Media, LLC 2007