Introduction

Increases in intimacy and conflict characterize the course of romantic development during adolescence (Chen et al. 2006; Furman and Buhrmester 1992; Laursen et al. 2001). These normative changes in the nature of romantic relationships are accompanied by a growing preference for compromise and negotiation to resolve conflict with romantic partners (Laursen et al. 2001). This shift in conflict behavior is presumed to reflect a growing awareness that some conflict is expectable as well as an underlying goal orientation that favors relationship needs over individual domination. Yet research to support these assumptions is notably absent. We actually know very little about individuals’ beliefs regarding the meaning of conflict in romantic relationships. Equally scant is information about the particular goals individuals pursue during romantic relationship conflict and their links to conflict behavior. In the current study, we draw upon social information processing theory (Crick and Dodge 1994; Huesmann 1988) to examine how beliefs about conflict in romantic relationships are associated with social goals for conflict, conflict behavior, and romantic relationship quality during late adolescence.

Beliefs About Romantic Conflict

Social information processing theory posits that general knowledge about social behavior affects how individuals process and respond to social situations (Crick and Dodge 1994; Huesmann 1988). Research supports a model in which latent knowledge structures guide social information processing which, in turn, influences behavioral responses. Various types of knowledge structures have been examined in relation to youths’ peer relationships, including normative beliefs about aggression (Huesmann and Guerra 1997), beliefs about social rejection (Downey et al. 1998), emotion understanding (Dodge et al. 2002), and general conceptions about peers (Burks et al. 1999). For example, individual differences in youths’ normative beliefs about aggression predict concurrent and future aggression with peers and romantic partners, and these effects are at least partially mediated through the intervening effects of normative beliefs on social information processing (Dodge et al. 2002; Huesmann and Guerra 1997; Kinsfogel and Grych 2004; Werner and Nixon 2005).

The current study builds on this literature by examining a different type of knowledge structure—beliefs about conflict in romantic relationships. By late adolescence, individuals should possess general knowledge about the meaning of romantic relationship conflict. During mid to late adolescence, intimacy becomes more central to romantic relationships and conflict becomes more frequent (Chen et al. 2006; Furman and Buhrmester 1992). Disagreements between partners are often not relationship threatening and can even provide a means for clarifying positions, enhancing mutual understanding, and strengthening bonds (Hartup 1992; Laursen and Collins 1994). Nonetheless, conflict also has the potential to provoke relationship difficulties and for approximately 35–40% of adolescents to become physically or emotionally abusive (Malik et al. 1997; Wolfe et al. 2001).

These developmental shifts in romantic experiences should be accompanied by corresponding shifts in knowledge about the meaning of conflict. By late adolescence, conflict should be viewed as somewhat expectable with more constructive than destructive potential. However, individuals’ beliefs about romantic conflict will likely differ as a function of direct experiences in romantic and other close relationships and indirect experiences observing parents’, siblings’ and peers’ romantic relationships (Crick and Dodge 1994; Furman and Simon 1999; Simon and Furman, under review). Individual differences in social knowledge structures promote individual differences in social behavior (Dodge et al. 2002; Huesmann and Guerra 1997). Accordingly, we expected that beliefs about the constructive value of conflict would be associated with the use of negotiation and compromise during conflict with romantic partners. Beliefs about the destructive value of conflict were expected to motivate destructive conflict behaviors, such as aggression or excessive compliance. Although compliance may be effective for reducing conflict, it is maladaptive for relationships because it is ineffective for managing inevitable disagreements and undermines the development of intimacy (Foscoe et al. 2007).

Social Goals During Romantic Conflict

As noted above, latent knowledge structures are believed to influence social behavior through their effect on social information processing (Crick and Dodge 1994; Dodge et al. 2002; Huesmann and Guerra 1997). One phase of information processing that appears important to conflict behavior is the selection of social goals (Rose and Asher 1999). Social goals are the objectives individuals strive for in a given social situation, and individuals must choose from the array of goals available in a given social situation. Once a goal is chosen, behavioral responses are selected, evaluated, and enacted according to what will maximize the desired goal (Crick and Dodge 1994; Emmons 1996).

The selection of social goals during peer conflict is linked to conflict behavior. Relationship-oriented goals (e.g., relationship maintenance) are associated with prosocial conflict strategies, whereas control goals (e.g., having control over one’s possessions, space, or activities) predict the use of hostile and coercive conflict strategies (Chung and Asher 1996; Delveaux and Daniels 2000; Murphy and Eisenberg 2002; Ojanen et al. 2007; Renshaw and Asher 1983; Rose and Asher 1999). To date, research on conflict goals and strategies has focused on children’s peer group relationships. A notable exception is Rose and Asher’s (1999) study of friendships, which found that children’s conflict goals with close friends were associated with both conflict strategies and friendship quality.

The current study extends the literature on conflict goals by examining conflict goals in romantic relationships during late adolescence. The consistency of extant findings linking conflict goals to strategies among youth of various ages across contexts suggested that conflict goals would be important predictors of conflict behavior in romantic relationships. However, testing this hypothesis required identifying developmentally sensitive conflict goals that were appropriate to romantic relationships. Beginning in adolescence, learning to balance autonomy and relatedness needs becomes an important developmental task, and close dyadic relationships are an important context for working through this task (Allen et al. 1994; Grotevant and Cooper 1988; Hill and Holmbeck 1986). Conflict with romantic partners requires partners to balance needs of the relationship with those of the self and the partner (Laursen et al. 2001; Shulman 2003). Accordingly, we hypothesized four types of romantic conflict goals that varied according to their focus on relationship, self, or partner needs: relationship-oriented goals, self-focused goals, revenge goals, and partner-focused goals.

Consistent with prior research (e.g., Rose and Asher 1999), we expected that relationship-oriented goals would be most commonly endorsed, as most youth value close peer relationships. The salience of relationship goals for late adolescents is also underscored by the centrality of romantic partners in their social network and their preference for using negotiation to resolve conflict (Furman and Buhrmester 1992; Laursen et al. 2001). The characteristics of romantic relationships at this age suggested a range of possible relationship-oriented goals. Whereas children’s relationship goals are primarily limited to relationship preservation, late adolescents may also pursue intimacy, mutuality, and respect for autonomy during conflict with romantic partners (Allen et al. 2006; Collins and Sroufe 1999; Shulman 2003).

Unlike relationship-oriented goals, individual goals center on the needs of oneself or the partner. Self-focused goals include having the situation work out in one’s own favor (e.g., getting one’s way and “winning”), whereas partner-focused goals represent a desire to keep one’s partner content and include having the situation work out in the partners’ favor or making sure the partner is not upset. Revenge or retaliation is typically conceptualized as a distinct goal (Rose and Asher 1999); however, it is self-focused to the extent that it serves a personal desire to hurt the partner.

According to social information processing theory, the selection of social goals is influenced by extant knowledge structures (Crick and Dodge 1994; Huesmann and Guerra 1997). Hence we expected constructive conflict beliefs to promote the selection of relationship-oriented conflict goals and destructive conflict beliefs to promote the selection of individual focused goals. In addition, patterns of information processing, including the selection of social goals, are believed to be a mechanism through which latent knowledge structures influence social behavior (Crick and Dodge 1994; Huesmann and Guerra 1997). Accordingly, we expected that romantic conflict goals would mediate the hypothesized associations between conflict beliefs and conflict behavior. Specifically, we hypothesized that (1) relationship-oriented goals would mediate the association between constructive conflict beliefs and the use of negotiation; (2) self-focused and revenge goals would mediate the association between destructive conflict beliefs and aggression; and (3) partner-focused goals would mediate the association between destructive conflict beliefs and compliance.

Conflict and Intimacy in Romantic Relationships

Thus far, we have focused on hypothesized links between beliefs about conflict, conflict goals, and conflict behavior in romantic relationships. A second goal of this study was to examine whether conflict beliefs and goals were associated with romantic relationship quality, including the degree of intimacy and conflict in relationships. Although intimacy and conflict may appear to be divergent social tasks, the two become increasingly intertwined during adolescent romantic development (Shulman 2003). As couples spend more time together and relationships become more intimate, conflict is more likely (Laursen et al. 2001). Conflict has the potential to provoke relationship difficulties, but it also provides a means to clarify roles, strengthen connections, and improve relationships (Hartup 1992). As such, managing conflict constructively becomes critical for establishing and maintaining intimacy in romantic relationships (Collins and Sroufe 1999; Laursen et al. 1996). Well-handled conflict mitigates its disruptive potential whereas poorly managed conflict undermines intimacy and breeds further conflict (Laursen and Collins 1994; Shulman et al. 2006).

We expected to replicate these findings in the current sample, such that individual differences in romantic conflict behavior would be reflected in the intimacy and conflict in adolescents’ romantic relationships (Laursen et al. 1996; Shulman 2003). Moreover, individual differences in conflict beliefs and goals were also expected to be associated with relationship intimacy and conflict. If individual differences in conflict behavior reflect variations in social cognition, then social cognition about romantic conflict should also predict conflict and intimacy in adolescents’ romantic relationships. Accordingly, we expected that constructive conflict beliefs, relationship-oriented goals, and the use of negotiation would each predict greater intimacy and less conflict. Destructive beliefs, individual-focused goals, aggression, and compliance were each expected to predict lower levels of intimacy and higher levels of conflict.

The Present Study

The current study sought to establish the relevance of social cognition about romantic relationship conflict for late adolescents’ conflict behavior with romantic partners. Our first aim was to assess whether beliefs about romantic relationship conflict are associated with late adolescents’ goals and behaviors during conflicts with romantic partners. Drawing upon social information processing theory and research, we hypothesized that beliefs about conflict in romantic relationships would be meaningfully related to conflict behavior. We expected constructive beliefs to be associated with negotiation and destructive beliefs to be associated with aggression and excessive compliance. Social information processing theory suggested that these associations would be mediated by the social goals individuals pursue during romantic relationship conflict (Crick and Dodge 1994; Huesmann and Guerra 1997). Accordingly, we predicted that relationship-oriented goals would mediate the association between constructive conflict beliefs and the use of conflict negotiation; self-focused and revenge goals would mediate the association between destructive conflict beliefs and aggression; and partner-focused goals would mediate the association between destructive conflict beliefs and compliance. The second goal was to assess whether individual differences in latent knowledge about conflict, romantic conflict goals, and romantic conflict behavior were associated with differences in romantic relationship quality. We predicted that constructive conflict beliefs, relationship-oriented goals, and the use of negotiation would each predict more intimate and less conflictual romantic relationships. Destructive beliefs, individual-focused goals, aggression, and compliance were each expected to predict less intimate and more conflictual romantic relationships.

Methods

Participants

Participants included 494 undergraduate students at a mid-western state university who were between 18 and 21 years old (M = 19.3, S.D. = 1.14). The sample was predominately female (78%). To qualify for participation, interested students must have been involved in a dating relationship of at least three months duration during the prior year and not be married, engaged to, or living with a romantic partner. Romantic relationships ranged from 3 to 84 months (M = 20.3 months, S.D. = 14.5 months), and 290 participants (59%) were in a romantic relationship at the time of their participation. All identified their sexual orientation as heterosexual, though this was not a criterion for study eligibility. The ethnic composition of the sample was similar to that of the university student population and surrounding community. Eighty-six percent of participants were Caucasian, 7% African American, 3% Latino or Hispanic American, 1% Asian American, 0.5% Native American, and 2.5% biracial or “other”.

Procedure

Participants were recruited from psychology, sociology, and education classes across campus via classroom announcements and flyers containing information about study requirements and eligibility. Participation was completed through computer-administered questionnaires about conflict, intimacy, and satisfaction in family, friends, and romantic relationships. Information was also obtained about relationship histories and psychosocial functioning. Participating students received class credit and a raffle entry for their participation.

Measures

Beliefs About Romantic Relationship Conflict

The Beliefs About Conflict Inventory (BACI; Simon and Kobielski 2006) is an 18-item questionnaire that measures adolescents’ beliefs about conflict in romantic relationships. The BACI contains 3 scales that tap views about the normalcy and meaning of conflict in romantic relationships. Six items assess beliefs about the Normalcy of romantic relationship conflict. Sample items include: “Conflicts or disagreements are a normal part of being in a romantic relationship”; “Conflicts or disagreements are bound to come up when romantic partners spend a lot of time together”; and “Conflicts or disagreements should not really happen in a romantic relationship if two people are really compatible” (reverse scored). Six items assess Constructive beliefs about romantic relationship conflict. Sample items include: “Conflicts or disagreements can be a healthy way to work out differences in a romantic relationship”; “Conflicts or disagreements can improve communication in a relationship”; and “Conflicts or disagreements are a way for romantic partners to work toward a better understanding”. Six items assess Destructive beliefs about romantic conflict. Sample items include: “Conflicts or disagreements are a threat to continuing the dating relationship”; “Conflicts or disagreements mean that this could be the end of your romantic relationship”; and “Conflicts or disagreements make it likely that someone will end up feeling bad or getting hurt”. Parallel items were included for relationships with parents and close friends, but they are not reported here. For each item, respondents indicate agreement on a 5-point scale, with higher ratings indicating greater agreement. Mean scores were computed for each of the three scales. Simon and Kobielski (2006) reported that a factor analysis of the BACI items produced a structure that replicated the conceptual structure of the measure. Results from a principal axis factor analysis with oblimin rotation indicated a three-factor solution that explained 68.15% of the variance. Item loadings for the current sample ranged from 0.56 to 0.74, and all cross-loadings were less than 0.35. Cronbach’s alphas for the Normalcy, Constructive, and Destructive scales in the current sample were 0.74, 0.80, and 0.86 respectively.

Romantic Relationship Conflict Goals

The Goals for Intimate Peer Conflict Inventory (GIPCI; Simon and Martin 2006) is a 34-item questionnaire that measures the extent to which respondents pursue relationship-oriented and individual-focused goals during conflict with romantic partners. Item development was guided by prior research on children’s conflict goals with friends (e.g., Rose and Asher 1999) and by developmental research on intimacy development, which stresses the importance of balancing each partner’s individual needs with those of the relationship during relationship conflict (Shulman 2003). Parallel items were included for close friends, but are not reported here. The GIPCI asks participants to rate the importance of each goal when they are having a disagreement with their romantic partner using a five point Likert scale (1 = not important to 5 = most important), with higher scores indicating greater importance of the conflict goal. The Relationship-Oriented, Self-Focused, Partner-Focused, and Revenge goal scales of the GIPCI are included in this study. The Relationship-Oriented scale includes 11 goals that highlight relationship needs and include items relating to relationship maintenance, fairness, and mutual understanding. Sample items include “maintaining the relationship”, “being fair”, “becoming closer with my partner”, and “understanding each other better”. The Self-Focused scale includes 7 goals that center on adolescents’ own needs, such as “getting my way”, “being in charge of the situation”, and “getting my partner to agree with me”. The Partner-Focused scale includes 4 goals that revolve around partners’ needs, such as “having my partner get his/her way”, “making sure the argument goes my partner’s way”, and “keeping my partner from getting upset.” The Revenge scale includes 7 goals that focus on hurting the partner, such as “wanting payback or revenge”, “making sure my partner feels hurt”, and “making my partner jealous”. Mean scores were created for each of the four scales. Simon and Martin (2006) reported that factor analyses of the GIPCI produced a factor structure that was consistent with the conceptual structure of the measure. Item loadings for the current sample ranged from 0.42 to 0.88, and all cross-loadings were less than 0.32. The internal consistencies of the scales for the current study were adequate, alpha for Relationship-Oriented goals = 0.87, alpha for Self-Focused goals = 0.86, alpha for Partner-Focused goals = 0.68, and alpha for Revenge goals = 0.87.

Conflict Behavior

Adolescents’ conflict behavior with romantic partners was assessed with two measures. The Conflict Resolution Styles Inventory (CRSI; Kurdek 1994) was used to assess participants’ use of negotiation and excessive compliance during romantic conflict. Each of these strategies was measured by four items indicating how often the respondent uses a given conflict strategy (1 = never to 5 = always). Mean scores were calculated for each scale, with higher scores indicating greater use of the conflict strategy. The alpha coefficient was 0.80 for the Negotiation scale and 0.79 for the Excessive Compliance scale.

The Conflict in Adolescent Dating Relationships Inventory (CADRI; Wolfe et al. 2001) was used to assess participants’ use of aggression during romantic relationship conflict. The CADRI includes scales for physical aggression, threatening behavior, relational, and verbal emotional aggression. In the current study, an overall aggression score was created by averaging across scales, with higher scores indicating more aggression. The alpha coefficient for the summary score was 0.85.

Romantic Relationship Quality

The Negative Interactions factor score from the Network of Relationships Inventory (NRI; Furman and Buhrmester 1985) was used to assess the degree of conflict in respondents’ romantic relationships. The NRI contains 24 five-point Likert items that tap respondents’ perceptions of support and negative interaction in various relationships. Participants were instructed to answer the NRI questions about their “most important” romantic relationship in the past year. The Negative Interactions factor includes scales assessing relationship conflict and annoyance. Higher scores indicate more conflictual relationships. Cronbach alpha for the Negative Interactions factor in the current sample was 0.87.

The Emotional Closeness, Respect, and Balanced Relatedness subscales of the Intimacy Questionnaire developed by Shulman et al. (1997) were used to assess intimacy in romantic relationships. These scales were selected because they appear to be most important to intimate peer relationships in late adolescence (Shulman et al. 1997). Participants were instructed to answer the intimacy questions for the same “most important” romantic relationship they referenced in completing the NRI. Items were rated on a five-point Likert scale, and an overall score was computed by averaging across items. Higher scores indicate more intimate romantic relationships. Cronbach alpha for the overall intimacy score was 0.83.

Results

Preliminary Analyses

Before the central analyses were conducted, all variables were examined for skew and kurtosis. Only the summary score for the CADRI was positively skewed. This score was transformed by taking the natural logarithm after adding a constant of one. The transformation reduced the skew of the measure, and the transformed variable was used in descriptive statistics and all analyses.

Next, the associations between potential confounding variables and the primary study variables were examined to test their relevance to the study hypotheses. For example, significant associations for participants’ age might suggest developmental variation in the proposed associations. However, age was not associated with conflict beliefs, goals, or behaviors or with intimacy or conflict in romantic relationships (all ps > 0.10). Differences in participants’ relationship status were examined to assess whether the central findings might vary for those reporting on a current versus past romantic relationship. Participants in a current relationship might view current relationships or partners in a more positive light and thus confound the planned analyses. No differences were found between students who were and were not in a romantic relationship at the time of the study on any variables (all ps > 0.10). Lastly, differences in participants’ romantic relationship experiences were examined to rule out the possibility that less experienced participants differ in their conflict beliefs, conflict goals, conflict behavior, or relationship quality. Neither the number of romantic relationships nor the average length of these relationships was significantly related to any of the primary study variables (all ps > 0.10).

Descriptive Information for Conflict Beliefs, Goals, and Behavior

Means and standard deviations for each measure are presented in Table 1. The measures showed good variability. Scores for the conflict belief scales indicate that late adolescents view conflict as a fairly predictable part of romantic relationships that tends to be more constructive than destructive in nature, t(492) = 22.79, p < 0.001. Of the four romantic conflict goals, relationship-oriented goals were more strongly endorsed than any other: t(490) = 30.03, p < 0.001 for self-focused, t(489) = 45.47, p < 0.001 for partner-focused, and t(490) = 54.99, p < 0.001 for revenge goals. Self-focused goals received the second strongest endorsement and were more highly rated than either partner-focused, t(490) = 33.19, p < 0.001 or revenge goals, t(490) = 6.38, p < 0.001. Partner-focused goals ranked third, and were endorsed more strongly than revenge goals, t(490) = 24.21, p < 0.001. The relative frequency of romantic conflict behaviors was consistent with prior studies with negotiation being more frequent than compliance, t(490) = 29.78, p < 0.001 or aggression, t(490) = 57.42, p < 0.001 (Laursen et al. 2001). Compliance received the second highest rating and was more frequent than aggression, t(490) = 21.47, p < 0.001.

Table 1 Means and standard deviations for study variables
Full size table

Associations between Conflict Beliefs, Conflict Goals, Conflict Behavior, and Romantic Relationship Quality

Correlations between the primary study constructs are presented in Table 2. The association between the constructive and destructive conflict belief scales was modest, suggesting that these beliefs are not opposite ends of a single continuum. Constructive conflict beliefs were significantly associated with higher endorsement of relationship-oriented goals and more frequent use of negotiation with romantic partners. Destructive conflict beliefs were associated with lower endorsement of relationship-oriented goals and greater endorsement of individual-focused (self and partner) and revenge goals. They were also related to less use of negotiation and greater use of destructive conflict behavior, including compliance and aggression. Beliefs about conflict were associated with romantic relationship quality. Constructive beliefs were associated with more intimate romantic relationships, whereas destructive conflict beliefs were associated with less intimate and more conflictual relationships (see Table 2).

Table 2 Correlations between conflict beliefs, conflict goals, conflict behavior, and romantic relationship quality
Full size table

Conflict goals and behaviors in romantic relationships were also meaningfully related (see Table 2). Stronger endorsement of relationship-oriented goals was most highly related to greater use of negotiation, though it was also related to lesser use of aggression and compliance. Stronger endorsements of self-focused and revenge goals were most highly related to greater use of aggression and lesser use of negotiation. Stronger endorsement of partner-focused goals was associated with more compliant conflict behavior. Conflict goals were also significantly associated with romantic relationship quality (see Table 2). Endorsement of relationship-oriented goals was associated with greater intimacy and decreased conflict with romantic partners. Both self-focused and revenge goals were associated with decreased intimacy and increased conflict.

Predicting Conflict Behavior from Conflict Beliefs and Goals

A series of regression analyses were conducted to test whether romantic conflict goals mediated the proposed associations between beliefs about romantic relationship conflict and conflict behavior with romantic partners. Relationship-oriented goals were expected to mediate links between constructive conflict beliefs and the use of negotiation during romantic relationship conflict. Self-focused and revenge goals were each expected to mediate associations between destructive conflict beliefs and the use of aggression. Partner-focused goals were expected to mediate associations between destructive conflict beliefs and the use of compliance. For each of the four hypothesized models, mediation was tested with a series of three regression analyses (Baron and Kenny 1986; Kenny et al. 1998). First, the conflict belief was regressed on the conflict behavior to establish a significant association between the dependent and independent variables. Next, the conflict goal was regressed on the conflict belief to establish a significant association between the independent variable and the mediator. The third regression predicted conflict behavior from the hierarchical equation of conflict belief scores followed by conflict goal scores. When conflict goals were significant predictors of conflict behavior after controlling for conflict beliefs, Sobel tests were conducted to test for significant mediation effects (Preacher and Hayes 2004; Sobel 1982).

Table 3 presents the standardized betas from the set of three regressions calculated for the four hypothesized models. In each of the four models, conflict beliefs (the independent variable) were significantly associated with conflict behavior (the dependent variable) and conflict goals (the mediator). Although conflict beliefs remained significant predictors of conflict behavior after controlling for the effect of conflict goals, results of the Sobel tests were significant in each of the four models, indicating partial mediation (see Table 3). Specifically, the association between constructive conflict beliefs and the use of negotiation was partially mediated by the endorsement of relationship-oriented goals. The association between destructive conflict beliefs and aggression was partially mediated by self-focused and revenge goals. The association between destructive conflict beliefs and the use of compliant conflict strategies was partially mediated by partner-focused goals.

Table 3 Mediation of conflict beliefs and conflict strategies by conflict goals
Full size table

Predicting Romantic Relationship Quality from Conflict Beliefs, Goals, and Behavior

Hierarchical regression analyses were conducted to examine the additive contributions of conflict beliefs, conflict goals, and conflict behavior to romantic relationship quality. Dependent variables included the degree of conflict and intimacy in romantic relationships. For each of the two regressions, scores for constructive and destructive conflict beliefs were entered in the first step. The four conflict goals were entered in the second step, followed by the three conflict behaviors. Results of the regression analyses are presented in Tables 4 and 5.

Table 4 Regressions predicting romantic relationship conflict from conflict beliefs, conflict goals, and conflict behavior
Full size table
Table 5 Regressions predicting romantic relationship intimacy from conflict beliefs, conflict goals, and conflict behavior
Full size table

Destructive but not constructive beliefs about conflict predicted the degree of romantic relationship conflict (see Table 4). When conflict goals were entered into the equation the beta for destructive goals decreased but remained significant. The self-focused and revenge goals significantly predicted more conflictual relationships while partner-focused goals predicted less conflictual relationships. Relationship-oriented goals did not predict the degree of romantic relationship conflict. When conflict behaviors were entered into the equation, destructive conflict beliefs along with self-focused and partner-focused goals continued to predict more conflictual relationships. The beta for revenge goals dropped to marginal significance. Greater use of aggressive conflict behavior and lesser use of negotiation each significantly predicted more conflictual romantic relationships.

Destructive conflict beliefs predicted less intimacy in romantic relationships (see Table 5). When conflict goals were entered into the equation the beta for destructive goals decreased but remained significant. Of the four conflict goals, only the relationship-oriented goal was significant, with greater endorsement predicting more intimacy. When conflict behaviors were entered into the equation, the betas for destructive conflict beliefs and relationship-oriented goals decreased but remained significant. Greater use of negotiation and lesser use of compliant strategies each significantly predicted more intimate romantic relationships.

Discussion

This study examined social cognition about conflict in romantic relationships and its associations with conflict behavior and relationship quality among a sample of late adolescents. Beliefs about conflict in romantic relationships and specific conflict goals were uniquely related to conflict behavior and relationship quality. Overall, these results point to the relevance of social cognitive processes for understanding individual differences in late adolescents’ use of constructive and destructive conflict strategies. Associations between social cognition and relationship quality suggest that social cognition about conflict is important to intimacy as well as conflict in romantic relationships.

Conflict Beliefs, Goals, and Behavior

Our results indicate that older adolescents view romantic relationship conflict as a relatively ordinary occurrence that is more constructive than destructive. This finding is consistent with empirical reports of age-related increases in romantic relationship conflict whose outcomes are more often relationship enhancing or neutral than negative (Chen et al. 2006; Furman and Buhrmester 1992; Laursen and Collins 1994). Together, these data suggest that adolescents’ romantic experiences may help to shape cognitive schema about the meaning of relationship conflict. Longitudinal research is required to examine how developmental changes in the nature of romantic relationships affect and are affected by shifts in beliefs about relationship conflict.

Individual differences in beliefs about conflict were meaningfully related to conflict goals and behavior with romantic partners. Furthermore, the pattern of associations for constructive and destructive beliefs differed. Constructive beliefs were only associated with relationship-oriented goals and behavior whereas destructive beliefs were linked to both the absence of relationship-oriented processes and the presence of destructive processes. The modest correlation between constructive and destructive beliefs and the differential predictions for each indicate that these beliefs are distinct and do not represent opposite ends of a single continuum. Their associations with conflict goals and behavior suggest that individual differences in beliefs about conflict represent meaningful variations in general knowledge about romantic relationship conflict.

Social goals for romantic relationship conflict were also related to conflict behavior. Consistent with prior studies, prosocial goals were related to prosocial strategies while self-focused and revenge goals were related to coercive strategies. The measurement of romantic goals required the creation of developmentally sensitive goals centered on key features of late adolescents’ romantic relationships. Goals for peer conflict are likely to be somewhat different than goals for conflict in close dyadic relationships. Older adolescents’ close relationships are voluntary relationships involving emotional commitment to the partner and the relationship. As such, any tension in the selection of conflict goals is more likely to involve competing desires for intimacy and personal power (self or partner) than competing desires for reputation, morality, and group status (Allen et al. 2006; Shulman 2003). With these considerations in mind, we constructed a series of goals reflecting relationship needs, self-needs, partner-needs, and revenge. Only revenge goals mapped directly onto those used in prior studies. Items for relationship-oriented goals included desires to maintain the relationship, intimacy, mutuality, and respect for autonomy. Individual-focused goals included objectives centered narrowly on the needs and power of one person. Each of these four goals was linked to conflict behavior in meaningful ways. Relationship-oriented goals were associated with conflict negotiation; self-focused and revenge goals were associated with aggression; and partner-focused goals were associated with excessive compliance.

Consistent with social information processing theory, goals partially mediated the associations between conflict beliefs and conflict behavior (Crick and Dodge 1994). The consistency of this finding across the analyses suggests that one way by which conflict beliefs may help to shape conflict behavior is by rendering some goals more salient. However, even when accounting for the contributions of specific goals, constructive and destructive conflict beliefs remained significant predictors of conflict behavior. Hence, beliefs about romantic relationship conflict may influence conflict behavior both indirectly and directly. Alternatively, other factors, such as earlier phases of social information processing that precede goal selection or affective intensity (e.g., sadness, anger and fear), may also mediate associations between conflict beliefs and behavior (Murphy and Eisenberg 2002; Stein and Albro 2001).

Social Cognition about Conflict, Conflict Behavior, and Romantic Relationship Quality

Conflict beliefs, goals, and behavior were each associated with the quality of late adolescents’ romantic relationships. When considered simultaneously, stronger endorsement of destructive conflict beliefs, self-focused goals, and aggressive behavior each predicted more conflictual romantic relationships. These findings are consistent with those of Rose and Asher (1999) who reported that children with more conflictual friendships endorsed conflict goals that were non-relational (i.e., control and revenge) and conflict strategies that were hostile or coercive. Interestingly, conflict goals and strategies were unrelated to positive friendship quality in the Rose and Asher study, but they were significant predictors of romantic relationship intimacy in the current study. For late adolescents, the absence of destructive conflict beliefs and behaviors (e.g., compliance) along with the presence of relationship-oriented goals and behavior (e.g., negotiation) predicted greater intimacy in romantic relationships.

It seems unlikely that the findings for relationship conflict and intimacy are redundant. Conflict and intimacy were correlated, but not so highly as to suggest a single underlying construct (r = −0.43). Moreover, different combinations of cognition and behavior predicted relationship conflict and intimacy. We believe these findings are better understood as reflecting the developmental dynamics of romantic relationships. As conflict and intimacy each increase during adolescence, constructive conflict skills become critical for establishing and maintaining intimacy in romantic relationships (Collins and Sroufe 1999; Laursen et al. 1996). The current findings extend this work by highlighting the importance of cognitions about conflict for intimacy. Additional research is needed to articulate the interrelations between social cognition about conflict and social cognition about intimacy. Both should be important to relationship representations, and both should be linked to experiences with conflict and intimacy in romantic relationships (Furman and Simon 2006). Including measures of social cognition about conflict and intimacy in future studies would help to clarify their unique and combined contributions to experiences in romantic relationships.

Finally, associations between partner-focused goals, compliant conflict behavior, and relationship quality are worth noting. Few studies have examined the subordination of personal needs in connection with romantic development in adolescence. In the adult relationship literature, those who subordinate personal needs during romantic relationship conflict frequently evaluate their behavior as inauthentic (Neff and Harter 2002). In the current study, partner-focused conflict goals predicted less conflict in romantic relationships, and submissive conflict behavior predicted less intimacy. Together, these findings suggest that patterns of unhealthy self-subjugation may have important implications for adolescents, whose romantic relationships provide an important context for identify development (Furman and Shaffer 2003). Further research is needed to determine if the decreased conflict attained through pursuing partner-focused goals comes at the cost of relationship intimacy. Greater attention to this dynamic would enrich our understanding of romantic development, but it will also require distinguishing self-sacrificing cognitions and behaviors that emanate from authentic care from those that reflect an inauthentic mode of relating designed to avoid conflict (Neff and Harter 2002).

Limitations and Conclusions

Several methodological limitations of the current study should be considered when interpreting the findings from this study. First, the exclusive reliance on self-report measures may have inflated associations among constructs. Future studies would benefit from the addition of observational methods to assess relationship functioning as well as partner reports of conflict behavior. The inclusion of romantic partners would allow researchers to examine the interdependence of participants’ and partners’ conflict cognitions and behaviors and their implications for romantic relationship quality (Kenny et al. 2006). Second, we have discussed the findings in ways that suggest that beliefs about romantic relationship conflict affect conflict goals and, ultimately, behavior. This interpretive framework is consistent with social information processing theory, but the data are cross-sectional and correlational; hence, the effects may be in the other direction, bi-directional, or reflect a third unmeasured variable. Longitudinal and experimental designs are required to establish causal relations. Finally, the restricted diversity of the sample limits the external validity of the current study. The current sample consisted primarily of Caucasian heterosexual females. As such, it is difficult to assess how gender, sexual orientation, or cultural norms may have affected the current findings, though these are clearly important issues for future research.

Despite these limitations, the current study is among the first to demonstrate the relevance of social cognition about romantic relationship conflict for conflict behavior and relationship quality in romantic relationships during late adolescence. Beliefs about the constructive value of conflict were associated with the pursuit of relationship-oriented goals during romantic conflict, which, in turn, predicted late adolescents’ use of negotiation to resolve romantic conflict. Beliefs that conflict is destructive predicted the pursuit individual-focused and revenge goals, which, in turn, predicted greater use of destructive conflict behavior in romantic relationships. Furthermore, destructive conflict beliefs, goals, and behavior were associated with more conflictual and less intimate romantic relationships. These findings suggest that social cognitive processes may be important for understanding individual differences in older adolescents’ conflict behavior and for promoting healthy romantic relationships.