Introduction

Researchers who study gender development often distinguish communal and agentic orientations (cf. Bakan 1966). These orientations correspond to socially-prescribed gender roles that are found across many, but not all, cultures (cf. Mead 1935). Stereotypically feminine roles correspond to a communal orientation that focuses on meeting others’ needs. Stereotypically masculine roles correspond to an agentic orientation focused on meeting the needs of the self. In the present study, we examined gender differences and similarities in communal and agentic conflict-management strategies as a function of the context—whether the conflict involved a same-gender friend or a heterosexual romantic partner. Hyde’s (2005) meta-analytic work on the gender similarity hypothesis indicates that many gender differences are small such that men and women are more similar than they are different—differences within groups of men and groups of women are greater than the differences between men and women. We sought to identify conditions under which gender differences and similarities in the endorsement of gender-typed conflict-management strategies emerged using a U.S. sample of undergraduate college students. Further, when gender differences in strategies were found within specific relationship contexts, we sought to investigate whether these differences reflected gender-typed personality traits or the goals women and men were striving to accomplish. Our study was informed by prior research which has typically investigated American, Canadian, and Western European samples.

According to Deaux and Major’s (1987) social-constructionist model of gender, gender-typed behaviors are created and maintained by the transaction of the person and the immediate context. For college students in contemporary Western cultures, two contexts in which gender-typed behaviors may manifest are same-gender friendships and heterosexual romantic relationships (cf., Monsour 2002). Indeed, these relationships may be of particular importance during this time of the life span in the early 21st century because “emerging adult” (Arnett 2000) college students in industrialized cultures develop quasi-independence from family, but delay marriage (relative to earlier birth cohorts within these cultures).

In line with a social constructionist perspective, empirical research on conflict management indicates that strategies for managing conflict depend on the social context—namely, the type of relationship within which the conflict occurs (Creasey et al. 1999; Jensen-Campbell et al. 1996). However, in these studies, the source of the strategy difference is unclear because relationship is confounded with the type of conflict. That is, strategy differences could be due to the relationship type or to aspects of the conflict situation. For example, Jensen-Campbell et al. (1996) used different hypothetical situations (i.e., vignettes) for conflict involving a sibling vs. a friend. One of the conflicts involving siblings described a situation where two siblings wanted to watch different television programs being shown at the same time. One of the conflicts involving friends described a situation where a person did not want their friend to associate with a third friend. Results showed that power assertion strategies were used more with siblings than with friends. However, because both the relationship (sibling vs. friend) and the conflict situation (which television show to watch vs. whom to associate with) changed, it is unclear whether the difference was due to the relationship type or to aspects of the conflict. In the present study, we examined associations between conflict-management strategies, gender, and relationship type while holding the conflict constant across different relationships.

In the present study, using self-report and vignettes to portray hypothetical conflicts, we investigated communal and agentic conflict-management strategies as a function of gender, relationship type, gender-typed goals, and gender-typed personality traits. We first examined gender differences in strategies for managing conflict with same-gender friends and heterosexual romantic partners. Gender differences were predicted based on Maccoby’s (1998, 2000) work theorizing that gender-typed patterns of conflict management develop via socialization in gender-segregated peer groups in childhood. According to Maccoby, gender-typed patterns of conflict management persist across the life span and affect how men and women manage conflict with same-gender friends and heterosexual romantic partners. Second, we aimed to move beyond documenting the existence of gender differences to understanding why gender differences and similarities emerge. Gender differences in conflict-management strategies are often attributed to individual differences in women’s and men’s communal and agentic goals or to their expressive and instrumental personal traits. Thus, we investigated the person’s perspective of the situation (as reflected in their goals) and their personality characteristics (as reflected in gender-typed personality traits) to better understand the association between “gender” (identification as a man or a woman) and the endorsement of communal and agentic conflict-management strategies in specific relationships.

Contextual Specificity of Gender Differences in Communal and Agentic Strategies

A social constructionist perspective of gender guided the present study. From this perspective, in social contexts where people interact with others, social expectations and perceptions based on gender may be activated. These expectations and perceptions of the social situation, in combination with characteristics of the self, influence interpretations of the situation and ultimately affect behavior (Deaux and Major 1987). Different perceptions and expectations might be cued depending on whether the social context involves a same-gender friend or a romantic partner. Thus, gender differences may not necessarily be stable across different social contexts, such as relationships with same-gender friends or romantic partners.

In some respects, same-gender friendships and heterosexual romantic relationships are similar (Furman et al. 2002). Both are voluntary and relatively egalitarian relationships. However, these relationships also differ in important ways. First, early adult’s same-gender friendships are likely to be longer in duration compared to romantic relationships (see Carver et al. 2003). By early adulthood, most people have spent a considerable portion of their lives interacting with same-gender friends. Indeed, gender segregation in children’s social relationships has been documented across a number of cultures (Whiting and Edwards 1988) and persists across the life span (see Mehta and Strough 2009 for a review). In contrast to the pervasiveness of same-gender friendships, the introduction of romantic relationships begins at adolescence (Carver et al. 2003). With this introduction to romantic relationships, also comes the beginning of mature sexuality (Carver et al. 2003; Furman et al. 2002). Given these differences, romantic relationships may be more volatile (exciting, but fragile) than same-gender friendships. Thus, young adults might manage conflict differently with romantic partners than same-gender friends.

Developmental theory and research on the socialization of children within gender-segregated peer groups suggest how gender differences and similarities may emerge in conflict-management strategies depending on the specific relationship context (same-gender friends vs. romantic partners). During childhood, via socialization by same-gender peers, girls learn that to effectively manage conflict, they must use communal strategies with same-gender friends, whereas boys learn that agentic strategies are most effective with their same-gender friends (Maccoby 1998). According to Rose and Rudolph (2006, p. 117), “exposure to same-sex peers elicits and strengthens sex-linked relationship processes.” Maccoby (1998, 2000) theorizes that gender-typed patterns of conflict management that develop via socialization in gender-segregated peer groups in childhood (i.e., boys are assertive or agentic and girls are cooperative or communal) persist across the life span when the conflict involves a same-gender friend (cf. Maccoby 1998, 2000). From this perspective, women would be hypothesized to use communal strategies more and agentic strategies less than men when managing conflict with same-gender friends.

Empirical research is consistent with Maccoby’s (1998) theoretical supposition that within same-gender friendships, women use agentic conflict-management strategies less than men. Specifically, Suh et al. (2004) used event-contingent recording methodology where young adults completed a behavioral check list describing every social interaction (not just conflictual interactions) for 20 days and found that when the interpersonal interaction involved same-gender friends, women reported using agentic strategies such as making a demand on the other less often than did men. Consistent with Suh et al.’s findings, Lindeman et al.’s (1997) study examining adolescents’ responses to hypothetical conflicts involving a same-gender friend found that girls were less likely than boys to endorse agentic strategies that were assertive or aggressive. Thus, we hypothesized that when managing conflict with a same-gender friend, women would be less likely to endorse agentic strategies than would men.

Empirical research is consistent with Maccoby’s (1998) theoretical supposition that within same-gender friendships, women use communal conflict-management strategies more than men. For example, Suh et al. (2004) found that when the interaction involved a same-gender friend, women reported using communal strategies (e.g., expressing reassurance) more often than did men. Consistent with Suh et al.’s findings, studies of conflict management (Black 2000; Lindeman et al. 1997) and interpersonal interactions (Strough and Berg 2000) in adolescence indicate girls are more likely than boys to use or endorse communal strategies when interacting with a same-gender peer (i.e., friend or classmate).

Within the heterosexual romantic relationship context, a different pattern of conflict management occurs. Extant research on conflict management within satisfied/happy and unsatisfied/distressed romantic relationships including dating couples, cohabiting couples, and married couples suggests that women are more likely than men to be demanding or to use agentic conflict-management strategies, whereas men are likely to withdraw from the conflict (cf. Eldridge and Christensen 2002). According to Maccoby (1998), this demand-withdraw pattern of conflict negotiation stems from childhood socialization by peers within gender-segregated contexts. In married couples, husbands are less likely to use the agentic strategies that they learned to use with same-gender friends in childhood to manage conflict with their wives. As suggested by Maccoby, in childhood boys learn that using agentic strategies does not work to manage conflict with girls—girls do not respond favorably to boys’ use of agentic strategies. Due to lack of experience using communal strategies to manage conflict, combined with the knowledge that agentic strategies are ineffective, husbands become frustrated or distressed. To decrease their distress, they withdraw from the conflict. In response to their husband’s withdrawal, wives use agentic/demanding strategies in an effort to engage the husband. Building on this work leads to the hypothesis that when managing conflict with romantic partners, women would endorse agentic strategies more than men.

The use of communal strategies to manage conflict within heterosexual romantic relationships is not well understood. Researchers either have not included communal strategies in their studies (i.e., Gottman et al. 1998) or existing studies yield inconsistent results. For example, Feldman and Gowen (1998) asked adolescents to rate a list of strategies indicating how often they used each strategy during disagreements with heterosexual romantic partners. Girls rated communal strategies such as compromising greater than did boys. However, Suh et al.’s (2004) event-contingent recording study on young adults found that in interactions (including, but not limited to conflictual interactions) with heterosexual dating partners, women reported using communal strategies such as listening attentively to the other less often than did men. Perhaps the inconsistent findings can be attributed to the different ages of participants studied. Thus, we based our hypothesis on Suh et al.’s work because, similar to the present study, their work focused on young adults. We hypothesized that when managing conflict with a romantic partner, women would be less likely to endorse communal strategies than would men.

Explaining Gender Differences: Goals and Conflict-Management Strategies

Berg and Strough’s (2010) contextual model of interpersonal everyday problem solving provides a framework for understanding the process whereby conflict-management strategies may vary as a function of gender and relationship context (see also Strough and Keener in press). Their model emphasizes that strategy differences may reflect aspects of the context that are salient from the perspective of the person—as reflected in the goals she or he is striving to accomplish. Thus, in the current study, when gender differences in strategies were found within specific relationship contexts, we investigated whether these differences reflected the goals women and men were striving to accomplish.

Goals differ in their level of specificity (Strough and Keener in press). Higher-level goals are broad goals reflecting life tasks and goals for the future (e.g., Zirkel and Cantor 1990), such as getting married or starting a family. Other goals are more narrow or context specific, such as control goals for wanting to be in charge of a specific situation (e.g., Pickard and Strough 2003). In the present study, we focused on goals related specifically to the conflict context. When solving everyday interpersonal problems or conflicts, goals often focus on the relationship (Rose and Asher 1999; Strough et al. 1996). Interpersonal goals may focus on the needs of both the self and another person (e.g., mutual participation when working together on a task) or may focus solely on ensuring that one’s own needs are met (e.g., control goals for getting one’s way, see Strough and Berg 2000; Strough et al. 1996). In the current study, we use the term “communal goals” to refer to mutual participation goals that entail a focus on the needs of both the self and another person. We use the term “agentic goals” to refer to goals that entail a focus on one’s own needs. Thus, communal goals reflect the motive to collaboratively manage the conflict, whereas agentic goals reflect the motive to exert authority or to be the person who gets his or her own way.

Consistent with Berg and Strough’s (2010) model of interpersonal problem solving, research (e.g., Rose and Asher 1999; Sorkin and Rook 2006) has shown that goals (what a person wants to happen or accomplish in the conflict situation), are systematically related to strategies (the steps or actions a person takes to resolve the conflict). Rose and Asher (1999) found that children who responded to hypothetical peer conflicts with accommodation-compromise or communal strategies (i.e., “I would say that he could pick the game now if I could pick the game after that.”) were more likely to also endorse relationship maintenance or communal goals (i.e., “I would be trying to stay friends.”). There is also evidence to suggest goals are useful in understanding the association between strategies and individual difference characteristics such as gender (cf., Berg and Strough 2010). For example, Strough and Berg (2000) found that when working with a same-gender classmate, preadolescent girls were more likely than boys to use communal, affiliative conversation strategies. However, when boys and girls reported goals that focused on communal concerns (mutual participation) their use of communal strategies was similar. Building on this research and theory, we hypothesized that when gender differences are found and after controlling for gender, gender-typed goals would predict unique variation in conflict-management strategies.

Gender-Typed Personality Traits: Association with Conflict-Management Strategies

Theorists increasingly emphasize that gender typing is multidimensional—there are many dimensions related to one’s gender. Gender typing includes not only one’s categorization as a male or female, but also an array of other dimensions, including, but not limited to, personality traits, attitudes, and behaviors (cf. Huston 1983). Yet, when investigating conflict-management strategies, researchers often solely focus on self-reported gender—operationalized as the person’s report of whether they are male or female. In the present study, in addition to considering the person’s report of their gender as male or female, we also considered gender-typed personality traits as a dimension of gender.

At a conceptual level, communal orientations are concordant with gender-stereotypical “feminine” (Bem 1974) and “expressive” (Spence and Helmrich 1981) personality traits such as being compassionate or sensitive to the needs of others. Agentic orientations are concordant with gender-stereotypical “masculine” (Bem 1974) and “instrumental” personality traits (Spence and Helmrich 1981) such as being competitive and assertive. Based on the degree to which a person endorses traits stereotypically ascribed to men or women, a person’s gender-typed personality is considered to be more or less instrumental (masculine) or expressive (feminine). Research shows that compared to boys and men, girls and women are more likely to endorse expressive traits (Eagly 1987; Feingold 1994; Leszczynski and Strough 2008; Spence 1993). Gender differences in the endorsement of instrumental traits are less apparent in younger birth cohorts and have decreased across historical time (Strough et al. 2007; Twenge 1999).

Although gender differences in the endorsement of gender-typed traits have been documented (to varying degrees depending on the particular trait), it is clear that both boys and men and girls and women endorse expressive and instrumental traits (Bem 1974; Constantinople 1973; Spence and Helmreich 1978; Suh, et al. 2004; Twenge 1997; Unger 1979). Thus, it could be that men and women who endorse expressive traits may also endorse communal conflict-management strategies and men and women who endorse instrumental traits may also endorse agentic conflict-management strategies. Empirical research supports this supposition. Specifically, Suh et al.’s (2004) findings showed that gender-typed traits were associated with strategies for interpersonal interactions involving same-gender friends and with heterosexual romantic partners. The present study aimed to extend Suh et al.’s findings which are based on peer interactions in general to more specific types of interactions by examining conflictual interactions with same-gender friends and romantic partners. Based on prior research, we hypothesized that when gender differences were found and after controlling for gender, gender-typed traits would predict unique variation in conflict-management strategies.

Summary of Hypotheses

Drawing from a social constructionist perspective, we hypothesized that for communal strategies, (1) there would be an interaction between gender and relationship type. We based specific predictions on developmental research and theory on the socialization of boys and girls within same-gender contexts and hypothesized: (1a) when managing conflict with a same-gender friend, women would be more likely to endorse communal strategies than would men (see Lindeman et al. 1997; Suh et al. 2004). Also, we hypothesized (1b), when managing conflict with a romantic partner, women would endorse communal strategies less than would men (see Suh et al. 2004). For agentic strategies, drawing from a social constructionist perspective we hypothesized that (2) that there would be an interaction between gender and relationship type. Based on developmental research and theory we hypothesized: (2a) when managing conflict with a same-gender friend, women would endorse agentic strategies less than would men (see Lindeman et al. 1997; Suh et al. 2004), and (2b), when managing conflict with a romantic partner, women would be more likely to endorse agentic strategies than would men (see Gottman et al. 1998).

To better understand differences between men and women in strategies for managing conflict, we investigated the extent to which goals and personality traits uniquely predicted strategy endorsement after controlling for the person’s self-reported gender. To this end, when we found gender differences in strategies, we then examined associations among strategies, goals, and traits. Thus, Hypothesis 3 and 4 were only tested when Hypotheses 1a, 1b, 2a, or 2b were supported. Drawing from Berg and Strough’s (2010) model of interpersonal problem solving and Huston’s (1983) multidimensional conceptualization of gender, when endorsement of communal strategies varied by gender, (3) we hypothesized communal goals and expressive traits would explain unique variance in communal strategies after controlling for gender. When endorsement of agentic strategies varied by gender (4) we hypothesized agentic goals and instrumental traits would explain unique variance in agentic strategies after controlling for gender.

Method

Participants

Participants (N = 334; 131 men, 203 women) were college students between the ages of 19 and 25 years. See Table 1 for sample characteristics. Participants were students enrolled in a psychology class at a large public university in the Mid-Atlantic United States. This class, introduction to human development, fills a general education requirement as part of the university curriculum and is a required class for students majoring in exercise physiology and nursing. Psychology majors choose between this class, social psychology, and abnormal psychology to fill a major requirement. Students completed the measures used in this study to fulfill a research experience requirement. Answering questions was voluntary and completing the measures used in the present study was one of many ways students could meet their course requirement.

Table 1 Sample characteristics
Full size table

Procedure

Participants completed measures using an online data management system (Sona Ltd.). Participants reported demographic information and completed the Personal Attributes Questionnaire (PAQ; Spence and Helmreich 1978) and the Peer Conflict Questionnaire, which assessed conflict-management goals and strategies. The Peer Conflict Questionnaire consisted of vignettes describing hypothetical conflicts with same-gender friends and romantic partners that were created by the authors for this study. Following each vignette, participants rated communal and agentic conflict-management strategies or goals. The order of presentation of vignettes corresponding to same-gender friends and romantic partners was counterbalanced—order effects were assessed, but not found. Participants completed the Peer Conflict Questionnaire at two different times: once in the beginning of the semester to measure strategies and again later in the semester to measure goals. The first time participants completed the questionnaire, each vignette describing a conflict (e.g., about which concert to attend) was read and a list of six strategies (randomly ordered for each vignette) was presented. Participants then rated the degree to which they were likely to use each strategy. The second time that participants completed the questionnaire each vignette was read and a list of six goals was presented. Participants then rated the degree to which they were likely to use each goal. A distractor task (i.e., a reaction time task) was completed between each set of hypothetical conflicts (i.e., between the sets involving different relationship types).

Measures

Construction of the Peer Conflict Questionnaire

The vignettes on the Peer Conflict Questionnaire that portrayed hypothetical conflicts and the items that assessed strategies and goals were developed for the present study (see Appendix). The items used to assess goals were adapted from existing goal measures (Pickard and Strough 2003; Strough and Berg 2000). The items used to assess strategies were adapted from people’s own descriptions of their strategies for solving everyday problems (Strough et al. 2008). In adapting the goal and strategy items, we drew from research that examines links between goals and strategies (Rose and Asher 1999; Sorkin and Rook 2006).

Conflict vignettes were developed from prior research on interpersonal problem solving (i.e., Rose and Asher 1999; Strough et al. 2008), a pilot study with college students, and a focus group. Initial versions of five vignettes were presented to undergraduate students (N = 88) who rated the degree to which each vignette was easy to understand on a 1 (very untrue) to 5 (very true) numerical scale. Mean scores for the five vignettes ranged from 4.21 to 4.84; SD = .64 to .93, indicating each vignette was easy to understand. To ensure participants were not biased in their responses to conflict based on one person in the hypothetical conflict having a more convincing case than the other person, we followed Jensen-Campbell et al.’s (1996) approach. Specifically, participants rated the degree to which each person involved in the hypothetical conflict had a convincing case to get their own way on a 1 (very untrue) to 5 (very true) numerical scale. Mean scores on the five vignettes ranged from 3.21 to 4.23; SD = 1.02 to 1.32. Based on these results, some vignettes were edited and the vignette with the lowest score was replaced. A small focus group (N = 3) of undergraduates was then consulted. Based on feedback from this focus group, a new vignette was created that followed the structure of the other vignettes; the other vignettes were slightly modified to increase the degree to which each person described in the vignette had a convincing case to get their own way.

The Peer Conflict Questionnaire

Participants completed the Peer Conflict Questionnaire, which consisted of a total of ten vignettes that portrayed five hypothetical conflicts in two types of relationships (i.e., same-gender friend or romantic partner), (see Appendix). For each type of relationship, the conflict was held constant such that the only difference in the vignettes for same-gender friends vs. romantic partners was the description of the type of relationship (i.e., same-gender best friend vs. boyfriend/girlfriend). Based on reliability analyses of strategy and goal ratings (i.e., Cronbach alphas), one of the five vignettes was dropped for both types of relationships.

Strategy Scores

Participants were instructed that strategies refer to actions he or she would take to solve a problem or resolve an issue. Participants read statements that corresponded to specific strategies that could be employed to solve the problem and indicated their response on a 1 (strongly disagree) to 4 (strongly agree) numerical scale. For each vignette, there were two communal strategies (i.e., “I would talk with my boyfriend/girlfriend [best friend] to decide which concert to attend”; “I would work with my boyfriend/girlfriend [best friend] to find a way to decide which concert to attend”) and two agentic strategies (i.e., “I would stand my ground and explain why I should get to decide which concert we attend”; “I would step aside and let my boyfriend/girlfriend [best friend] decide which concert we attend”). Some agentic strategies were reverse scored. Strategies were randomly ordered for each hypothetical vignette.

A participant’s score for each type of strategy was the mean of his or her ratings of a given type of strategy (communal or agentic), across the four hypothetical vignettes. Thus, each participant had four strategy scores, each of which was based on eight items (e.g., 2 communal strategies for each of the 4 vignettes): (1) communal same-gender friend strategy score (α = .80; Range = 1.88–4.00); (2) agentic same-gender friend strategy score (α = .62; Range = 1.00–3.88); (3) communal romantic partner strategy score (α = .82; Range = 2.00–4.00); and (4) agentic romantic partner strategy score (α = .70; Range = 1.00–3.88).

Goal Scores

Participants were instructed that goals refer to what he or she would want to happen. Participants read statements that corresponded to specific goals for resolving the conflict and indicated their response on a 1 (strongly disagree) to 4 (strongly agree) scale. For each vignette, participants rated two communal goals (i.e., “I would want both of us to decide which concert to attend in a way that satisfies both of us;” “I would want both us to have a voice in deciding which concert to attend”) and two agentic goals (i.e., “I would want to do what is best for me and decide which concert we attend,” “I would want to convince my boyfriend/girlfriend [best friend] to go to the concert that I want”) on a 1 (strongly disagree) to 4 (strongly agree) scale. Goals were randomly ordered for each hypothetical vignette.

A participant’s score for each type of goal was the mean of his or her ratings of that type of goal (communal or agentic), across the four hypothetical vignettes. Thus, each participant had four goal scores each of which was based on eight items (e.g., 2 communal goals for each of the 4 vignettes): (1) communal same-gender friend goal score (α = .77; Range = 2.00–4.00); (2) agentic same-gender friend goal score (α = .80; Range = 1.00–4.00); (3) communal romantic partner goal score (α = .80; Range = 1.13–4.00); and (4) agentic romantic partner goal score (α = .80; Range = 1.00–4.00).

Gender-Typed Personality Traits

Gender-typed personality traits were assessed via the 24-item Personal Attributes Questionnaire (PAQ; Spence and Helmreich 1978). The PAQ was administered with the goal assessment. Participants rated eight instrumental traits (e.g., 1 = not at all aggressive vs. 5 = very aggressive) and eight expressive or feminine traits (e.g., 1 = not at all emotional vs. 5 = very emotional) on a 5-point scale. Responses for all traits within the two subscales were summed and averaged such that higher scores indicated greater endorsement of the trait. Therefore, there were two scores for each participant: an average instrumental and an average expressive score. Participants also rated eight items that assessed androgyny; however these items were not used in the present study. In the present study, the PAQ instrumental and expressive scales had coefficient alphas of .74 (Range = 2.00–5.00) and .77 (Range = 2.50–5.00), respectively.

Results

Prior to conducting analyses that addressed hypotheses, a Multivariate Analysis of Variance (MANOVA) was conducted to examine potential gender differences in all of the variables of interest. There was a significant effect for gender, F(6,326) = 17.24, p < .001, η 2p = .29 (see Table 2). Specifically, women were significantly more likely than men to endorse communal strategies for conflict involving a same-gender friend, agentic strategies for conflict involving a romantic partner, communal goals for conflict involving a romantic partner, communal goals for conflict involving a same-gender friend, agentic goals for conflict with a romantic partner, and expressive traits. Women were significantly less likely than men to endorse instrumental traits. Women and men were equally likely to endorse communal strategies for managing conflict with a romantic partner, agentic strategies for managing conflict with a same-gender friend, agentic goals for managing conflict with a same-gender friend.

Table 2 Gender differences and similarities in study variables
Full size table

Prior to conducting analyses that addressed hypotheses, we also conducted correlational analyses to examine relations among goals and strategies and to explore the extent to which these correlations varied by gender. Correlations for men and women are presented in Table 3. We used Fisher’s r-to-z test to examine whether correlations varied significantly by gender. Unless otherwise stated, the correlations discussed below did not vary by gender.

Table 3 Intercorrelations among variables of interest by gender
Full size table

Associations Among Strategies and Goals

For both same-gender friends and romantic partners, the correlation between communal strategies and communal goals was .45. Agentic strategies were significantly correlated with agentic goals for conflict involving a same-gender friend (r = .42) and for conflict involving a romantic partner (r = .51). The association between agentic strategies and agentic goals for managing conflict with a same-gender friend was significantly stronger for women (r = .48) than for men (r = .29), p = .05, albeit only marginally so.

Associations Among Strategies and Traits

Expressive traits were significantly correlated with communal strategies for conflict involving a same-gender friend (r = .28) and for conflict involving a romantic partner (r = .27). Instrumental traits were significantly correlated with agentic strategies for managing conflict with a same-gender friend (r = .13), but were not significantly correlated with agentic strategies for managing conflict with a romantic partner.

Associations Among Strategies

Endorsement of communal strategies for managing conflict with a same-gender friend was significantly correlated with communal strategies for managing conflict with a romantic partner (r = .66). This association was significantly stronger for women (r = .73) than for men (r = .54), p < .001. Endorsement of agentic strategies for managing conflict with a same-gender friend was significantly correlated with agentic strategies for managing conflict with a romantic partner (r = .49) and this association was significantly stronger for women (r = .63) than for men (r = .32), p < .001.

Conflict-Management Strategies as a Function of Gender and Type of Relationship

Communal Strategies

A mixed-model ANOVA was conducted to test Hypothesis 1—that an interaction between gender and relationship type for the endorsement of communal strategies would be found. The within-subjects factor was relationship type (same-gender friendship, romantic relationship) and the between-subjects factor was the person’s self-reported gender (male, female). The main effect of relationship type on endorsement of communal strategies was significant. Both men and women were more likely to endorse communal strategies when conflicts occurred within romantic relationships than within same-gender friendships (see Table 4), F(1,332) = 10.92, p = .001, η 2p = .03. The main effect of gender on endorsement of communal strategies was significant. Women were more likely than men to endorse communal strategies, F(1,332) = 5.59, p = .02, η 2p = .02 (see Table 5). As hypothesized (Hypothesis 1), these main effects were modified by an interaction between gender and relationship, F(1,332) = 5.17, p = .024, η 2p = .02. To further investigate this interaction, simple effects were examined. Consistent with Hypothesis 1a, when the conflict involved a same-gender friend, women were more likely than men to endorse communal strategies, F(1,332) = 9.40, p = .002, η 2p = .03, (see Table 4). In contrast to our prediction (Hypothesis 1b), a simple effect of gender was not found in romantic relationships – women and men were equally likely to endorse communal strategies when the conflict involved a romantic partner, F(1,332) = 1.50, ns, η 2p = .004 (see Table 4).

Table 4 Communal and agentic strategies for managing conflict
Full size table
Table 5 Predictors of agentic & communal strategies in friendships & romantic relationships
Full size table

Agentic Strategies

A mixed-model ANOVA was conducted to test Hypothesis 2—that an interaction between gender and relationship type for the endorsement of agentic strategies would be found. The within-subjects factor was relationship type (same-gender friendship, romantic relationship) and the between-subjects factor was gender (male, female). The main effect of relationship on endorsement of agentic strategies was significant. Both men and women were more likely to endorse agentic strategies when conflicts occurred within same-gender friendships than romantic relationships, F(1,332) = 46.63, p < .001, η 2p = .12 (see Table 4). The main effect of gender on endorsement of agentic strategies was significant. Women were more likely than men to endorse agentic strategies, F(1,332) = 4.78, p = .03, η 2p = .01 (see Table 4). As hypothesized (Hypothesis 2), these main effects were modified by an interaction between gender and relationship, F(1,332) = 31.35, p < .001, η 2p = .09. To further investigate this interaction, simple effects were examined. In contrast to our hypothesis (Hypothesis 2a), when the hypothetical conflict involved a same-gender friend, women were not less likely than men to endorse agentic strategies F(1,332) = 0.82, ns, η 2p = .002, (see Table 4). In accord with our hypothesis (Hypothesis 2b), in romantic relationships, women were more likely than men to endorse agentic strategies, F(1,332) = 19.75, p < .001, η 2p = .06 (see Table 4).

Predicting Strategies from Goals and Gender-Typed Personality Traits

To better understand the gender differences in strategy endorsement reported above, we examined whether expressive and instrumental personality traits and communal and agentic goals explain variation in conflict-management strategies after controlling for gender. We used hierarchical regression where gender is controlled by being entered in the first step of the equation. Using this approach, we conducted hierarchical regression analyses to test Hypotheses 3 and 4—that individuals’ goals and traits would predict strategy endorsement after controlling for gender.

Communal Strategies and Conflicts with Same-Gender Friends

In Step 1, we controlled for women’s significantly greater endorsement of communal strategies when conflicts involved a same-gender friend by regressing gender on communal strategies F(1, 331) = 10.11, p = .002 (see Table 5). In Step 2, to test Hypothesis 3, communal goals and expressive traits were entered as predictors of communal strategies after controlling for gender. The model was significant, F(3, 329) = 30.33, p < .001 (see Table 5). Inspection of the beta weights showed that in Step 2, gender no longer significantly predicted endorsement of communal strategies after goals and traits were entered (see Table 5) suggesting that the variance in strategies accounted for by gender in Step 1 was due to variance gender shared with the other two predictors. In partial support of our prediction (Hypothesis 3), communal goals accounted for a significant portion of the variance in communal strategies (see Table 5). In contrast to predictions (Hypothesis 3), expressive traits did not significantly predict communal strategies, (see Table 5). When the conflict involved a same-gender friend, greater endorsement of communal goals, but not expressive traits, predicted greater endorsement of communal strategies after accounting for gender differences in communal strategies.

Agentic Strategies: Romantic Relationships

In Step 1, we controlled for women’s greater endorsement of agentic strategies when conflicts involved a romantic partner by regressing gender on agentic strategies F(1, 332) = 19.75, p < .001 (see Table 5). In Step 2, to test Hypothesis 4, agentic goals and instrumentality were entered as predictors of the endorsement of agentic strategies. Results showed that the model was significant, F(3, 330) = 43.19, p < .001 (see Table 5). Inspection of the beta weights showed that in Step 2 gender continued to be a significant predictor of the endorsement of communal strategies (see Table 5). However gender accounted for less variance in Step 2 than in Step 1 suggesting that some of the variance in strategies accounted for by gender in Step 1 was due to variance gender shared with the other two predictors. In partial support of Hypothesis 4, agentic goals accounted for a significant portion of the variance in agentic strategies (see Table 5). However, in contrast to Hypothesis 4, instrumental traits did not significantly predict agentic strategies, (see Table 5). When the conflict involved a romantic partner, greater endorsement of agentic goals, but not instrumental traits, predicted greater endorsement of agentic strategies after accounting for gender differences in agentic strategies.

Discussion

Same-Gender Friendships vs. Romantic Relationships

Consistent with Deaux and Major’s (1987) social-constructionist model of gender, we found that the association between gender and strategy endorsement varied depending on the social context—whether the conflict was with a same-gender friend or romantic partner (see also Strough and Keener in press). Our data (see Table 4) suggest men’s strategy endorsement changes depending on the relationship context: men were more likely to endorse communal strategies and less likely to endorse agentic strategies when conflicts involved romantic relationships than when conflicts involved same-gender friendships. However, women’s endorsement of strategies tended to be more consistent across relationship contexts.

Gender differences in consistency of strategies across relationship contexts may reflect qualities of the relationships (Strough and Keener in press). Girls’ and women’s same-gender friendships are similar to romantic relationships in some respects (Sternberg 1987). For girls and women, same-gender friendships are often egalitarian, intimate, close, and cooperative. Socializing, self-disclosure, attentiveness, responsiveness, and support tend to be demonstrated (Buhrmester 1996; Leaper 1994; Zarbatany et al. 2000). However, for boys and men, the characteristics of romantic relationships are generally quite different than those of same-gender friendships. For boys and men, most same-gender friendships are hierarchical, and defined by competition and dominance. Men often demand attention, give orders, and engage in activities and games with their same-gender friends (Buhrmester 1996; Leaper 1994; Zarbatany et al. 2000). In contrast, within romantic relationships, young men are likely to be interested in maintaining the relationship and are more likely to engage in self-disclosure (Feiring 1999; Leaper et al. 1995). Thus, the level of intimacy between same-gender friendships and romantic relationships is similar for women, but different for men. This may explain why, in the present study, men’s communal and agentic conflict-management strategies appear to be more dependent on the relationship context than women’s strategies.

Gender Similarities in Conflict-Management Strategies

Our results indicate that gender similarities are more descriptive of the conflict management process than are gender differences. Specifically, men and women were equally likely to endorse agentic strategies when the conflict involved a same-gender friend. Also, men and women were equally likely to endorse communal strategies when the conflict involved a romantic partner. Although women were more likely than men to endorse communal strategies when the conflict involved a same-gender friend, we also found that the variance in strategies accounted for by gender was due to variance gender shared with communal goals. Our findings are consistent with Hyde’s (2005) gender similarities hypothesis suggesting that men and women are more similar than they are different.

Although our finding indicating that men and women were equally likely to endorse agentic strategies when conflicts involved a same-gender friend is also consistent with Hyde’s (2005) gender similarities hypothesis, it is inconsistent with our hypothesis. Based on developmental theory and research, we had hypothesized that men would endorse agentic strategies more than women in conflicts with same-gender friends. However, we based our hypothesis on research that did not focus solely on conflict situations (Suh et al. 2004) and research that focused on adolescents (Lindeman et al. 1997). Thus, our findings may be inconsistent with our hypothesis due to these methodological differences or the age of participants. Similarly, methodological differences across studies may also explain the lack of support for our hypothesis (based on Suh et al. 2004) that women would be less likely than men to endorse communal strategies with romantic partners.

In accord with our hypothesis, we found that women were more likely than men to endorse communal strategies for managing conflict with a same-gender friend. However, consistent with Berg and Strough’s (2010) contextual model of interpersonal everyday problem solving, goals accounted for additional variance in strategies beyond that accounted for by gender (see also Strough and Keener in press). That is, women and men, not just women, who endorsed the communal goal to resolve the conflict in a collaborative, mutually satisfying manner, also endorsed communal strategies such as talking to their same-gender friend to resolve the conflict.

Gender Differences in Conflict-Management Strategies

As hypothesized, women were more likely than men to endorse agentic strategies when the conflict involved a romantic partner. In addition, although agentic goals accounted for unique variance in agentic strategies after controlling for gender, gender remained a significant predictor of agentic strategies when the conflict involved a romantic partner. This was the only persistent gender difference found in the present study.

Results indicating that women were more likely than men to endorse agentic strategies for managing conflicts with a romantic partner are consistent with the extant literature on marital conflict (cf., Eldridge and Christensen 2002; Gottman, et al. 1998) and Maccoby’s interpretation of the demand-withdraw pattern of conflict negotiation based on the socialization of boys and girls in same-gender contexts. Together, research and theory indicate that wives make “demands” or use agentic strategies when managing conflict with their husbands. Eldridge and Christensen (2002) suggest that this gender-typed pattern may reflect that women are more likely to bring up the issue causing conflict. Bringing up such issues could emerge from a greater desire for change in the relationship among women (compared to men) due to differences in power within the relationship. Traditionally, relative to women, men are in power and therefore have more to lose than to gain by dealing with relationship conflict. Further, Maccoby’s (1998) theoretical work suggests that boys and men are socialized not to fight with girls and women and therefore reduce their tendency to use agentic strategies with other-gender romantic partners. It may be that socialization factors related to gender, power, and fighting with other-gender peers explain the association between gender and the endorsement of conflict-management strategies when the conflict involves a romantic partner. Studies of martial conflict examine serious martial issues (i.e., an ongoing disagreement); however, in the present study we examined everyday problems or conflicts. Thus, our findings suggest that at least some of the findings from the marital conflict literature may generalize to young adults’ everyday conflicts with romantic partners.

Gender-Typed Personality Traits

Although gender-typed traits did not explain unique variance in strategies after accounting for gender, the bivariate correlations (see Table 3) suggest that gender-typed traits tend to correspond to conflict-management in the direction predicted by Huston’s (1983) multidimensional view of gender typing. That is, for both men and women expressivity was related to communal strategies for same-gender friends and romantic partners. However, the association between instrumental traits and agentic strategies varied by relationship type and gender. For men, but not women, instrumental traits were related to agentic strategies for managing conflict with a same-gender friend, but not for conflict involving a romantic partner. This pattern of results, where the association between traits and strategies vary by gender and relationship, is consistent with Deaux and Major’s (1987) social-constructionist model of gender behavior suggesting that there is a complex interplay between individual factors such as gender and personality traits and the social context such as the relationship with whom the conflict involves. Although the hypotheses related to traits explaining variance in conflict-management strategies beyond gender were not supported, the pattern of intercorrelations suggests that future researchers should continue to examine the role of traits in the conflict-management process.

Limitations and Future Directions

Our results should be interpreted within the context of the limitations of the current study. Using hypothetical vignettes allowed for greater internal validity (i.e., all of the conflict situations were exactly the same, only the relationship type differed). However, this methodology may decrease ecological validity—findings based on hypothetical situations might not generalize to real-world situations. According to Laursen et al.’s (2001) meta-analysis, negotiation and prosocial strategies similar to the communal strategies we investigated are more likely to be reported in response to hypothetical situations than observed in response to actual conflicts. An examination of the means (see Table 2) suggests that across relationship type, participants in the present study endorsed communal strategies more than agentic strategies. Therefore, endorsement of communal strategies may be overestimated in the present study due to the use of hypothetical situations. Future research that utilizes observations of conflict-management strategies could address this issue.

It is important to consider how our findings may differ across cultures. The hypothetical conflict situations used in the present study may be specific to young adults in industrialized Western cultures. Furthermore, although the characteristics of friendships and romantic relationships are expected to have some commonalities across cultures, some variations exist. Thus, cross-cultural comparisons are needed.

It is unclear whether the findings of the present study generalize to same-gender romantic relationships. Our Peer Conflict Questionnaire allowed participants to think about either a same- or other-gender romantic partner when rating strategies and goals. That is, we did not define romantic relationships as being with a same- or other-gender person. Further, the conflict vignettes were worded in a way that allowed participants to respond to the survey regardless of sexual orientation. For example, the hypothetical situation about which concert to attend involving a romantic partner used ambiguous wording to allow all participants to read the vignette in a way that is consistent with their sexual orientation (e.g., “You and your boyfriend/girlfriend want to do something together on Saturday and you both agree that it would be fun to go to a concert.”). Thus, same-gender romantic relationships could be assessed using this measure. Only a limited number of studies have examined conflict resolution in same-gender romantic relationships (see Kurdek 2005) and further research is needed. Future research using the Peer Conflict Questionnaire should include a measure of sexual orientation to explicitly examine whether the findings of the present study generalize to same-gender romantic relationships.

When the Peer Conflict Questionnaire is used in future research, instructions should be added to clarify that the term “boyfriend/girlfriend” refers to a romantic partner to avoid potential confusion with a same-gender friend. This might be particularly important for girls who commonly refer to their same-gender friends as “girlfriends.” In our study “girlfriend” was always paired with “boyfriend” and participants were asked to indicate how they would respond to conflicts about a same–gender best friend as well as a girlfriend/boyfriend. These methodological details reduce the likelihood that participants interpreted both sets of vignettes as representing conflicts with same-gender friends. Moreover, strategies differed systematically as a function of the type of relationship and in accord with predictions derived from prior research.

As would be predicted by Hyde’s (2005) meta-analytic research, the majority of gender effects in the present study were small to moderate. Although some gender differences are small, they are often reliably found and are consistent with theory and thus warrant investigation. Future investigations should not only document such differences, but also work to identify factors that explain why differences occur.

Implications

Results of the present study add to other research on the consequences of gender-typed relationship processes (see Rose and Rudolph 2006). During adolescence and young adulthood, failure to successfully manage peer conflict has a negative impact on mental health as well as general emotional functioning (Buhrmester 1990; Feldman and Gowen 1998; Reese-Weber and Marchland 2000). The findings showing that the association between gender and conflict-management strategies depends on the relationship context could inform interventions aimed at improving conflict management within same-gender friends and romantic relationships. As suggested by Maccoby’s (1998) theoretical work, understanding the associations between the socialization occurring in same-gender contexts and relationship processes (e.g., conflict management) with same- and other-gender peers early in the life span, may contribute to our understanding of the same processes within romantic relationships in adulthood (see also Rose and Rudolph 2006).

Research aimed at understanding the developmental trajectory of conflict management with other-gender peers (friends and romantic partners) may help improve other-gender interactions. For example, researchers are currently investigating whether decreasing gender segregation by requiring children to work with other-gender peers at school starting in preschool could improve interactions between men and women and boys and girls across the life span (see the Sanford Harmony Project at http://sanford.clas.asu.edu/index.html). The present study informs this line of research by showing that the association between gender and conflict-management strategies varies by relationship type. Longitudinal or cross-sequential designs would help to address the extent to which there is developmental continuity in conflict management with other-gender peers. This knowledge is needed to inform interventions aimed at facilitating positive interactions between men and women in the workplace and at home.

Conclusion

The findings of this study have important implications for research on gender and the peer conflict-management process, which has typically emphasized gender differences rather than gender similarities and has often de-emphasized contextual factors such as relationship type. Research focusing exclusively on gender differences without considering goals for managing conflict or the social context (i.e., relationship type) may produce an incomplete description of the peer conflict-management process. The present study also cautions investigators against comparing conflict-management strategies across relationships without holding aspects of the conflict constant. An enriched understanding of the peer conflict-management process is made possible when researchers move beyond documenting the existence of gender differences and similarities to understanding why differences and similarities between men and women emerge.