Abstract
Romantic partners’ emotions become coordinated in various ways and this may have implications for well-being (Butler (2011) Temporal interpersonal emotion systems: The “TIES” that form relationships. Personality and Social Psychology Review, 15, 367–393.). The present study uses a community sample of 44 committed heterosexual couples to examine whether cooperation, a generally beneficial relational process, is associated with emotional coordination and whether the pattern differs when men’s emotions are coordinated with their female partners’ prior emotions or vice versa. Using behavioral observations of cooperation and second-to-second measures of emotional experience during a face-to-face conversation, men showed the most positive emotional experience at high levels of mutual cooperation. As predicted, cooperation was associated with different coordination patterns for men and women, with high mutual cooperation predicting an inphase pattern for men (emotions changing in unison with their partners) and an antiphase pattern for women (emotions changing in opposite directions from their partners). Our results suggest that men and women may experience cooperation differently, despite engaging in similar behaviors.
Emotional interdependence between social partners is the defining feature of relationships and manifests as coordinated patterns of emotional experience and expression (Butler, 2011; Butler & Randall, 2013). A wide range of different forms of emotional coordination have been investigated, but they have usually been referred to by more specific terms, such as negative reciprocity (Gottman, 1979), synchrony (Saxbe & Repetti, 2010), or transmission (Larson & Almeida, 1999). Taken together, this prior research suggests that emotional coordination is a central feature of both hostile, embittered relationships (Gottman, 1980; Levenson & Gottman, 1983) and supportive, nurturing ones (Julien, Brault, Chartrand, & Bergin, 2000). Similarly, a growing body of research suggests that greater emotional coordination is associated with both putatively undesirable processes, such as marital disagreements (Gottman, Coan, Carrere, & Swanson, 1998), and desirable processes, such as interpersonal sensitivity (Schöebi, 2008). A solid understanding of emotional coordination could, therefore, inform socioemotional interventions to promote individual and relational well-being, but that will require more detailed information about the different constructive and destructive forms that emotional coordination can take as well as its associations with interpersonal emotion regulatory processes.
To contribute to a more in-depth understanding, the present study investigates associations between emotional coordination and romantic partners’ cooperation, which is known to be a component of well-functioning relationships (Sheras & Koch-Sheras, 2006). In the context of close relationships, cooperation is typically defined as a communication process involving a back-and-forth exchange of thoughts and feelings in order to reach a mutually satisfying resolution (Sheras & Koch-Sheras, 2006). Therefore, it could be suggested that cooperation in close relationships may contribute to successful interpersonal emotion regulation by enabling partners to optimize their joint emotional outcome. Therefore, by examining cooperation, the present study is relevant to understanding the interplay of interpersonal emotional regulation and emotional coordination. In addition, in the present study, we distinguish between inphase coordination (emotions changing in unison) and antiphase coordination (emotions changing in opposite directions) and assess whether the pattern differs when men’s emotions are coordinated with their female partners’ prior emotions or vice versa. By drawing these distinctions, we hope to provide a more nuanced picture of emotional coordination and its associations with cooperation, thereby contributing to the growing body of knowledge regarding the role of interpersonal emotion regulation in relationship functioning.
Interpersonal emotional coordination
Extensive theory and empirical evidence suggest that the interconnectedness of partner’s emotions (emotional coordination) is critical to survival, health, and well-being across the life span (for review see Butler, 2011). In its broadest usage, emotional coordination refers to any nonrandom, patterned, temporal covariation between relationship partners in the timing or form of emotional behaviors, internal states, or events (Butler, 2011). Emotional coordination between partners is modeled using partner’s repeated measures of emotion (expressive behavior, subjective experience, or physiology; see Butler & Randall, 2013, for a review). These measures can be assessed in second-by-second intervals (e.g., Reed, Randall, Post, & Butler, in press), in daily increments (e.g., Butner, Diamond, & Hicks, 2007), or in any other time-varying unit. The most common subtypes of emotional coordination that have been reported in the literature have been referred to as synchrony (concurrent covariation of partners’ emotions; see Butler, 2011), transmission (time-lagged covariation of partners’ emotions; see Larson & Almeida, 1999), and coupling (nonlinear crosspartner influences on emotional rates of change) (Boker & Laurenceau, 2007; Butner et al., 2007; Helm, Sbarra, & Ferrer, 2012), although many other subtypes exist as well. In the present study, we focus on the subtype often referred to as transmission, using a model of time-lagged emotional covariation (one partner’s emotions predicting the other partner’s emotions at a subsequent time point).
Emotional coordination between partners can come about in numerous ways, both consciously and unconsciously. One mechanism is jointly engaging in activities that produce similar emotions in both the partners (Rohrbaugh, Shoham, Butler, Hasler, & Berman, 2009); however, shared experiences are not necessary for emotional coordination to arise (e.g., Butner et al., 2007). Because partners tend to verbally share both positive and negative emotions with each other (Rimé, 2007), they often experience similar emotions—a relatively nonconscious and automatic process referred to as emotional contagion or empathy (Hatfield, Cacioppo, & Rapson, 1994). In addition, empathy can include a conscious aspect often referred to as perspective taking, whereby partners purposefully try to understand each other’s emotions better. Evidence that this could impact emotional coordination comes from a study in which greater interpersonal sensitivity was associated with more emotional coordination (Schöebi, 2008). Thus, empathizing with one’s partner (for both positive and negative emotional events) can produce coordination within the dyad through both conscious and unconscious mechanisms. Emotional coordination can also come about, however, when partners disagree with one another, causing shared negative emotions (Larson & Almeida, 1999). Therefore, there are a variety of interpersonal dynamics that can produce emotional coordination for both positive and negative emotions.
Different patterns of emotional coordination
Research on behavioral coordination (i.e., synchronization of physical movements) has shown that different patterns of social coordination can arise. For example, several studies have investigated unintentional synchronization of behavior by asking participants to swing pendulums in each other’s presence (Richardson, Marsh, & Schmidt, 2005; Schmidt & O’Brien, 1997). This work has shown that people gravitate toward two forms of coordination: (1) inphase, whereby partners begin to move in unison, and (2) antiphase, whereby partners begin to move in opposite directions. One study has extended this work to emotional coordination using computer simulations and found that strong interpersonal influence (e.g., direct communication) produced inphase emotional coordination, while lower levels of influence allowed antiphase patterns to appear (Vallacher, Nowak, & Zochowski, 2005). Most research focused on daily emotional coordination between partners (both concurrent and time lagged) has found inphase patterns, whereby one partner’s emotion is positively associated over time with the other partner’s same emotion (e.g., Butner et al., 2007; Saxbe & Repetti, 2010; Schöebi, 2008); however, antiphase coordination, whereby one partner’s emotions are negatively associated with their partner’s at the same time, has not yet been considered. Since these studies did not investigate the possibility of antiphase emotional coordination, it is unknown whether or when this pattern may emerge in the context of close relationships.
Implications of coordination patterns for relationship functioning
Given that inphase emotional coordination has been associated with both positive relationship qualities, such as interpersonal sensitivity (Schöebi, 2008), and negative qualities, such as marital disagreements (Gottman et al., 1998), it is clear that inphase coordination does not inherently reflect either good or bad relationship functioning. We argue that the potential role the patterning can play in relationships likely depends on the context. For example, inphase coordination in situations that produce mutual hostility (i.e., having an argument) may contribute to deleterious effects on the relationship, whereas inphase coordination of positive emotions could play a role in long-term positive effects (Gable, Reis, Impett, & Asher, 2004). Similarly, although there has not been any research on antiphase patterns in romantic relationships, we suggest that its relational implications also depend on the context. On the one hand, an antiphase pattern could contribute to couple-level emotional regulation or maintaining a jointly balanced emotional state. Take, for example, a couple that is dealing with stress. If one partner comes home from work upset and their partner attempts to help him or her deal with the stressor (see Bodenmann, 2005 for a review of dyadic coping), we might see the occurrence of an antiphase emotional coordination pattern. As an illustration, suppose Partner A comes home upset and is met by Partner B who is relatively calm and positive, then Partner B tries to help Partner A to cope with the stressor, and Partner A becomes a bit more positive, whereas Partner B becomes a bit more negative due to contagion and concern. The partners’ emotions may oscillate in this way for a while, until settling down at their usual stable state, a process that has been referred to as coregulation (Butler & Randall, 2013). This scenario would suggest that antiphase patterns, similar to an emotional regulatory system, could contribute to positive emotional functioning. Conversely, antiphase emotional coordination patterns arise due to partners misunderstanding each other or having incompatible emotional responses to some stimulus might play a role in negative outcomes for relationship functioning and well-being. Taken together, we suggest that patterns of emotional coordination may relate to relational well-being differently depending on the context in which they occur.
Cooperation between romantic partners
One way to investigate the relationship implications of emotional coordination is to consider its associations with either desirable or undesirable interpersonal processes. In the present study, we focus on desirable processes or, more specifically, cooperation. In the social psychological literature, cooperation has primarily been studied in social dilemmas, in which two individuals have to choose between their own motives or their joint motives (for reviews see Balliet, Li, Macfarlan, & Van Vugt, 2011; Kollock, 2011). In this context, cooperation has been defined as making the choice that results in a good collective outcome, as opposed to maximizing one’s own outcome. Research using this paradigm has shown that individuals’ emotional experience and expression can be both determinants and outcomes of cooperative behavior (Stouten, Ceulemans, Timmerman, & Van Hiel, 2011; Stouten & De Cremer, 2010; Wubben, De Cremer, & van Dijk, 2011), suggesting that cooperating may impact emotional patterns between social partners, but this idea has not been tested.
Surprisingly, cooperation has not been extensively studied, at least explicitly, in the context of romantic couples. In a clinical context, however, cooperation is often identified as an important communication process that is necessary for successful and long-lasting relationships (Reis & Sprecher, 2009; Sheras & Koch-Sheras, 2006). From this perspective, cooperation can include aspects of both empathy (e.g., seeing things from your partner’s point of view) and navigating disagreements (e.g., talking things out with one’s partner). Similarly, in one rare study investigating cooperation in married couples, cooperation was defined as constructive problem solving, including elements of active listening, nondefensive self-expression, and compromising (Assad, Donnellan, & Conger, 2007). Results showed that cooperation largely accounted for associations between optimism and increases in marital satisfaction over a 2-year period.
In the present study, we operationalize cooperation in a similar way as a communication process involving nondefensive, back-and-forth exchanges of thoughts in which partners appear to be listening and trying to understand each other in order to arrive at a mutually satisfying compromise. We do so in the context of committed romantic couples discussing their health-relevant lifestyle choices, such as their eating and physical activity habits. We expected this context to elicit cooperative communication because the couples had to discuss aspects of their lifestyle that potentially require cooperative solutions if both partners were to meet their health goals. For example, if one partner did not like to exercise but the other partner did and wanted the other to exercise with him or her for support (e.g., taking a walk after dinner), then they would need to cooperate with one another to develop a plan that fit both of their needs. Furthermore, while discussing the issue, they would need to engage in an interactive exchange involving each partner expressing his or her own thoughts and feelings (e.g., “I don’t like to exercise,” “Yes, but it’s important for our health and I’d like your support”), but also listening to one another and attempting to understand the other’s perspective. Thus, this discussion context provided the opportunity for a range of cooperative communication behaviors to emerge, from mild disagreement to active efforts to better understand each other and to find mutually satisfying compromises. Given some of the similarities between cooperative communication and other processes that have been either empirically associated with emotional coordination, such as interpersonal sensitivity, or that in theory may relate to it such as emotional coregulation, we hypothesized that more cooperation would be associated with greater emotional coordination.
Gender differences in emotional coordination during cooperation
Some indirect evidence suggests that we might expect to see different emotional coordination patterns (inphase vs. antiphase) associated with cooperation depending upon whether men’s emotions are coordinated with their female partner’s prior emotions or vice versa. For example, although research on gender differences in emotional expression and interpersonal emotional goals has typically focused on the individual, rather than a dyadic context (Fischer, Rodriguez Mosquera, van Vianen, & Manstead, 2004; Kring & Gordon, 1998), these literature studies indirectly suggest that cooperation may produce different patterns of emotional coordination for men and women. Fairly, extensive research has shown that when faced with potential relationship conflict, women are more emotionally expressive and confrontational, while men are more likely to try to avoid conflict and reach quick resolutions (for a review see Carstensen, Gottman, & Levenson, 1995). As a result, women’s open expression should allow them to have a fairly strong influence on their partners’ emotional states, which could produce inphase coordination, whereby his emotions begin to follow her emotions in a similar pattern. In contrast, men’s lower levels of emotional expression would represent the same sort of weak influence conditions under which antiphase patterns appeared in computer simulation research (Vallacher et al., 2005), suggesting that her emotions may oscillate out of phase with his prior emotions.
Another indirect line of evidence comes from the child and infant development literature, where Feldman (2003) found different patterns of emotional coordination for father’s and mother’s interacting with their infants. Specifically, fathers’ emotional patterns were characterized by correlated father–infant bouts of increased positive arousal (inphase coordination), whereas mothers’ emotional patterns were characterized by a more oscillating sequence of mother–infant increasing and decreasing arousal (antiphase coordination). Although this literature suggests gender differences in emotional coordination, this has not been investigated in the context of romantic relationships and so we explore this possibility in the present study.
Present study
Cooperating with one’s romantic partner seems likely to impact the overall levels of emotional experience, but we argue that it could also impact the coordination of emotional experience between partners. Research on behavioral coordination shows that coordination can be either inphase (changes in the same direction) or antiphase (changes in opposite directions) (Richardson et al., 2005), but this distinction is usually not made in the literature on emotional coordination. If the two patterns are not distinguished, it could obscure important processes, because the patterns could cancel each other out if they were simultaneously present. To address these gaps in the literature, the present study investigates two questions:
How is behavioral cooperation associated with levels of emotional experience? We expected high mutual cooperation to be associated with more positive levels of emotional experience.
How is behavioral cooperation associated with patterns of emotional coordination (inphase or antiphase)? We expected the pattern that emerged to be different when men’s emotions are coordinated with their female partners’ prior emotions compared to the reverse. Specifically, we predicted that men would show an inphase pattern with their partners’ emotions, whereas women would show an antiphase pattern with their partners’ emotions.
Methods
Participants
Participants were recruited from ads posted on Craig’s list in a midsize city in the Southwest of the US. Participating couples had to meet the following criteria: (1) both the individuals were over the age of 18 years, (2) both individuals were willing to participate, and (3) the individuals had been in a romantic relationship with each other for at least 6 weeks. Our sample included 44 heterosexual couples ranging in the age from 19 to 69 years (men: M = 32.9, SD = .7; women: M = 31.3, SD = 1.9). Approximately 44.4% of the sample was married and 54.5% of the population was cohabitating. Couples were in a relationship together from 4 months to 39 years; on average, participants reported being in a relationship with their partner for 6.2 years (SD = 7.1 years). Additional demographic information is provided in Table 1.
Descriptive statistics for all study variables.
Note. Different superscripts indicated significant differences between women and men at p < .05. Love, conflict, and ambivalence were on a scale from 0 to 8, cooperation was on a scale from 0 to 3, and emotional experience was on a scale from 0 to 5.
Procedures
Our anlaysis uses data from an initial baseline questionnaire and a laboratory session that were part of a larger study focusing on the associations between relationship processes (e.g., demand/withdraw), eating, and emotions. Eligible participants completed an online baseline questionnaire that included demographic information and a measure of relationship quality (love, conflict, and ambivalence). At the laboratory, couples completed a video-taped interaction task. Couples were instructed to have a 20-min conversation with one another based on four topics (5 min per topic). Topics were presented visually on a computer screen and audibly via a prerecorded script. If they finished a topic before the 5 min were completed, they were instructed to let the research assistant know and they were prompted with the next topic. Therefore, conversation lengths varied across the 44 couples (range = 3–20.33 min; M = 10.81; SD = 4.80). The topics were: (1) How important do you think it is to live a healthy lifestyle? In other words, how important is it to you that you eat a healthy diet, get enough exercise and sleep, and so on?; (2) How willing are each of you to make sacrifices (e.g., spend more money, take time out from other activities) in order to live a more healthy lifestyle?; (3) What are some of the things you do that have a “negative” impact on each other’s lifestyle or on each other’s attempts to be healthy? What aspects of your shared lifestyle cause problems for you when it comes to being healthy?; and (4) What are some of the things you do that have a “positive” impact on each other’s lifestyle or on each other’s attempts to be healthy? What aspects of your shared lifestyle are helpful for you when it comes to being healthy? Following the conversation, a screen was placed between the partners so they could not see one another. Then, partners were asked to watch the video of their conversation on a computer monitor and rate how they were feeling (second-to-second) during the conversation using a rating dial (see below). Upon completion of the laboratory tasks, participants were debriefed.
Measures
Relationship quality
We used the Braiker and Kelley’s (1979) Love and Relationships Scale, which includes three subscales, to control relationship quality as it relates to emotional coordination between partners. The love subscale includes 10 items such as: “To what extent do you have a sense of ‘belonging’ with your partner” and “To what extent do you love your partner at this stage?” (α = .84). The conflict subscale includes five items such as: “How often do you and your partner argue with each other” and “How often do you feel angry or resentful toward your partner?” (α = .80). Finally, the ambivalence subscale includes five items such as: “How confused are you about your feelings toward your partner?” and “How ambivalent or unsure are you about continuing in the relationship with your partner?” (α = .86). Participants rated these items on a scale of 0 = “Not at all” to 8 = “Very much.” Both men and women scored relatively high on the love scale (men: M = 6.61, SE = .18; women: M = 6.93, SE = .13) and low on both the conflict (men: M = 3.00, SE = .23; women: M = 2.90, SE = .20) and ambivalence scales (men: M = 1.79, SE = .28; women: M = 1.47, SE = .25). Men and women did not significantly differ in any of the three subscales.
Behavioral coding of cooperation during the conversation
We used the cooperation subscale of the conflict and problem-solving scales (Kerig, 1996) to create a behavioral checklist of cooperation. This scale includes the following items: (1) talked it out with their partner, (2) expressed their thoughts and feelings openly, (3) listened to their partner’s point of view, (4) tried to understand what their partner was really feeling, (5) tried to find a solution that met both of their needs, and (6) compromised. These items are consistent with the predominant view of cooperation in the literature on close relationships and dyadic coping (Bodenmann, 2005; Sheras & Koch-Sheras, 2006). A team of three trained coders viewed the videotapes of the conversations and were instructed to rate how much each partner engaged in each of the six behaviors over the entire conversation. Scales for each of these items ranged from 0 = “Never” to 3 = “Often.” The behavioral cooperation scale showed good reliability (α = .91). The six items were averaged for each rater and the intraclass correlation for this total cooperation score between the coders was .96. For analyses, we used the average of the total cooperation score across the three coders. The average rating of behavioral cooperation was 2.16 (SD = .50) and 2.25 (SD = .44) for men and women, respectively, which are not significantly different.
Second-by-second emotion experience (rating dial)
Based on the research by Levenson and Gottman (1983), we assessed emotional experience using a bipolar-rating dial. Each partner was given a rating dial clearly labeled with anchors of frowning versus smiling faces and the labels “very negative” and “very positive” on the left and right sides, respectively. Prior to the task, a research assistant demonstrated how to use the rating dial. Participants were instructed to continuously rate how they remembered feeling during their conversation (second-to-second basis). The rating dial was attached to a voltage-driving circuit that was connected to the computer that was collecting physiological measures (not relevant to the present analyses) and provided a continuous measure of emotional experience (positive to negative) in the second-by-second increments. The dial was calibrated such that it ranged from a signal of 0 = “very negative” to 5 = “very positive.” Data were recorded in 1-s increments and later averaged across 10-s segments to reduce random noise, such that the data used for analyses included average ratings for each 10-s increment in the conversation. The average rating of emotional experience across the entire conversation was 3.01 (SE = .09) and 3.14 (SE = .09) for men and women, respectively, which were significantly different, t(5396) = 31.27, p < .0001. We calculated the intraclass correlation for the emotional experience variable to test for within person variability. Our intraclass correlation was .26, which suggests that 26% of the variance was due to mean differences between people and 74% of the variance was due to within-person fluctuations over time, making this an appropriate measure for assessing between-partner associations of emotional fluctuations over time.
Data analysis
To account for nonindependence within repeated measures dyadic data, we used repeated measures versions of actor–partner models to test the associations of cooperation and: (1) average levels of emotional experience and (2) emotional coordination (Kenny, Kashy, & Cook, 2006). Figure 1 provides a conceptual diagram of the two models, and Appendix A includes SAS syntax for implementing them. To assess the associations of cooperation with the levels of emotional experience, we used a two-intercept repeated measures actor–partner model (see Figure 1 and Appendix A; Laurenceau & Bolger, 2005). We first estimated the model with random effects of actor and partner cooperation, but this model would not converge and so we included them only as fixed effects. Even with cooperation constrained to fixed effects, this model still allows for separate fixed and random intercepts for men and women, the correlations of partners’ random intercepts and autocorrelations of the within-person residuals over time, thereby appropriately dealing with various sources of nonindependence in this type of data.
Panel (a) Conceptual diagram of the model predicting emotion levels from actor and partner cooperation. Path a = woman’s actor effect, path b = man’s actor effect, path c = woman’s partner effect, path d = man’s partner effect, paths e = woman’s actor by partner interaction effect, paths f = man’s actor by partner interaction effect. Panel (b) Conceptual diagram of the model predicting emotional coordination from actor and partner cooperation. For simplicity, we only present the man’s prior emotions predicting the woman’s subsequent emotion, but the converse is also included in the model. In addition, all of the paths presented in panel (a) for the model of emotion levels are also present in the model for emotional coordination, but for clarity in the diagram, we only highlight those that are directly relevant to emotional coordination. Paths a = woman’s lagged actor emotion effect, paths b = woman’s lagged partner emotion effect (this corresponds to the coordination parameter), paths c = woman’s actor by partner cooperation interaction effect (note that this interaction moderates the coordination parameter, not emotion levels).
In order to model emotional coordination, we used an actor–partner prospective change version of the emotional covariation model outlined by Butner and colleagues (2007; see Figure 1 and Appendix A). To do so, we created lagged actor and partner emotion variables (10-s time lag) that allowed us to predict one partner’s emotional experience from the other partner’s emotion at a prior time point (10 s earlier), while controlling the first partner’s prior emotion (also 10 s earlier). We also tested models that controlled partner’s concurrent emotions as well, but the results were almost identical; therefore, we present results that do not include the partner’s concurrent emotions. Our full model included the following fixed effects: (1) actor and partner effects of the lagged emotional experience variable, (2) actor and partner effects of cooperation, (3) two-way interaction effects of actor and partner cooperation by lagged partner emotion and (4) the three-way interaction of actor cooperation by partner cooperation by lagged partner emotion. Similar to the above model, convergence failed if we allowed the effects to vary randomly and so they were constrained to fixed effects. Again, however, this model allowed us to calculate separate fixed and random intercepts for men and women as well as the between partner covariation of the random intercepts. Autocorrelation of the within-person residuals is no longer a part of the model, however, because that source of variance is included as the fixed effect of the target partner’s own prior emotion instead. To test our hypotheses regarding gender effects, we included separate estimates for men and women for all parameters. An estimate of effect size (R 2) was calculated by predicting the observed emotion scores from the predicted values of each model (Singer & Willett, 2003). SAS Proc Mixed version 9.2 was used for all analyses (SASInstitute, 2004). We conducted preliminary analyses with time as a predictor variable and found that it had neither fixed nor random effects on emotional experience; thus, it was not included as a predictor.
Control variables
It could be suggested that over time couples develop ways of dealing with one another and that either relationship length or quality could impact emotional experience and coordination. To address this, we included age, marital status, relationship length, and relationship quality (love, conflict, and ambivalence) one at a time as potential control variables in our models for both emotion levels and coordination. These controls did not alter any of our focal results and so we present results without them.
Results
Descriptive statistics of all study variables are presented in Tables 1 to 3. As seen in Table 1, the only variable in which men and women differed was age, with the men being about 1 year older on average than the women. Zero-order within-person correlations are presented in Table 2 and show that for women, the levels of emotional experience were associated positively with love and cooperation and negatively with ambivalence. None of the relationship variables were associated with emotion levels for the men, nor with cooperation for either gender. Table 3 presents zero-order between-partner correlations and shows that the relationship quality variables were inter-related as would be expected for both men and women. In addition, partners were correlated with each other for all of the variables, showing the importance of taking into account their nonindependent outcomes. Finally, men’s cooperation was negatively associated with their partners’ reports of conflict and their emotional levels were negatively associated with their partners’ ambivalence, but women’s cooperation and emotional experience were unrelated to their partners’ reports of relationship quality.
Within-person zero-order correlations of study variables.
Note. Correlations above the diagonal are women’s; under the diagonal are men’s.
*p < .05; **p < .01.
Between-partner zero-order correlations of study variables.
Note. Zero-order correlations are presented for women’s and men’s scores correlated with their partner’s score. For example, women’s love correlated with their male partner’s conflict scores (r = −.40) as well as men’s love correlated with their female partner’s conflict scores (r = −.31).
*p < .05; **p < .01.
Associations between behavioral cooperation and levels of emotional experience
We hypothesized there would be more positive emotional experience for couples in which both the partners cooperated more during a face-to-face conversation. In partial support of this, we found a significant actor by partner interaction effect of cooperation on the levels of experience for men (F(1, 5389) = 5.17, p = .023), but not for women (F(1, 5389) = .44, NS). Parameter estimates for this model are provided in Table 4. Specifically, higher levels of men’s own cooperation predicted marginally greater positive emotion when his partner was also high on cooperating (t (5389) = 1.74, p = .08). However, when women were low on cooperation, there was no effect of men’s own cooperation on their levels of emotional experience (t(5389) = −.22, NS). The R 2 for this model was .05, suggesting that the cooperation had only a small effect on emotion levels.
Parameter estimates for the model of emotional levels moderated by cooperation.
Note. Emotion variables were person-mean-centered and cooperation variables were grand-mean-centered.
Associations of behavioral cooperation and emotional coordination
We predicted that different levels of cooperation would be associated with differences in emotional coordination between partners, but that the pattern would depend upon gender. In support of this, we found significant three-way interactions of actor cooperation by partner cooperation by partner’s lagged emotion for both men (F(1, 5162) = 3.88, p = .049) and women (F(1, 5162) = 6.12, p = .013) 1 . Parameter estimates for this model are provided in Table 5. A simple slope analysis showed that, as predicted, a different pattern of coordination emerged depending upon whether men’s emotions were predicted from women’s emotions or vice versa (i.e., a different pattern of estimated regression coefficients for one partner’s emotion predicting the other partner’s subsequent emotion). For men, their own cooperation predicted greater inphase coordination (i.e., a positive regression coefficient for their partner’s lagged emotion) when their partner was also high on cooperation (b = .06, SE = .03, p = .042); however, there was no effect when she was low on cooperation (b = −.02, SE = .03, NS). For women, their own cooperation predicted greater antiphase coordination (i.e., a negative regression coefficient for their partner’s lagged emotion) when her partner was also high on cooperation (b = −.07, SE = .04, p = .04); however, there was no effect when he was low on cooperation (b = −.002, SE = .04, NS). Specifically, when both partners were high on cooperation, men showed a significantly inphase coordination pattern (b = .04, SE = .01, p < .001; see Figure 2, panels (a) and (b)). Conversely, we see a significant antiphase pattern emerging for women (b = −.05, SE = .02, p = .014; see Figure 3, panels (a) and (b)). Furthermore, these patterns were significantly different for men and women (t(5090) = 3.11, p < .01), and the full model had an R 2 of .50, indicating a medium to large effect size. This suggests that engaging in cooperation was associated with different emotional coordination patterns for men and women.
Parameter estimates for the model of emotional coordination moderated by cooperation.
Note. Emotion variables were person mean centered and cooperation variables were grand mean centered.
Panel (a) Interaction effect of actor by partner behavioral cooperation on emotional coordination for men. The y-axis represents the unstandardized regression coefficient for women’s emotions predicting men’s emotion. Positive values suggest an inphase pattern, while negative values suggest an antiphase pattern. Panel (b) Example of the inphase emotional coordination pattern observed for men at high levels of mutual partner cooperation (dyad 28). Note. * indicates the emotional coordination estimate is significantly different than zero at the p < .05 level. The pattern above represents inphase coordination at high mutual cooperation. At low mutual cooperation, couples would show a disconnected pattern of emotional experience.
Panel (b) Interaction effect of actor by partner behavioral cooperation on emotional coordination for women. The y-axis represents the unstandardized regression coefficient for men’s emotions predicting women’s emotion. Positive values suggest an inphase pattern, while negative values suggest an antiphase pattern. Panel (b) Example of the antiphase emotional coordination pattern observed for women at high levels of mutual partner cooperation (dyad 3). Note. * indicates the emotional coordination estimate is significantly different than zero at the p < .05 level. The pattern above represents antiphase coordination at high mutual cooperation. At low mutual cooperation, couples would show a disconnected pattern of emotional experience.
Discussion
Our study investigated whether cooperation between partners—a beneficial relationship process—is associated with the levels of emotional experience or interpersonal emotional coordination, and if so whether the pattern is different for men and women. In general, the results suggest that cooperation appears to be related to dyadic emotional outcomes but in ways that differ between men and women. We first discuss the results in more detail and then consider their implications for working with romantic couples.
Cooperation and interpersonal emotional levels and coordination
First, we examined the association between high mutual cooperation and levels of emotional experience and found that men’s emotional experience was dependent upon both his and his partner’s level of cooperation (an actor by partner interaction effect), such that men felt most positive at higher levels of mutual cooperation; however, this was not true for the women in our sample. Second, our results suggest that while both partners may be engaging in a similar behavior together—cooperating with one another—different emotional coordination patterns may emerge for men and women. Specifically, men showed an inphase coordination pattern, whereby their partners’ emotions from a prior time point predicted men’s emotional change in the same direction at a later time point (i.e., when she was less positive, he became less positive, or when she was more positive, he became more positive). In contrast, women showed an antiphase coordination pattern, in which her partner’s emotions predicted opposite changes in her emotions at a later time point (i.e., when he was less positive, she became more positive, or when he was more negative she became less negative). Importantly, these patterns only occurred when couples engaged in high mutual cooperation, which suggests that low cooperation, from either partner, prevented coordination from occurring. One possible reason for this is that low cooperation may be an indication of partner’s disengagement, which would undoubtedly reduce coordination, especially since both partner’s emotions contribute to emotional coordination.
A large amount of the research examining emotional coordination exists within the infant and child development literature (e.g., Feldman, 2003; Field, 1994). In this work, we see distinct gender differences in emotional coordination between infants and their mothers versus fathers. This research suggests that mothers engage in oscillating patterns (characterized by alternating high and low effect) with their infant, in a way that modulates the infant’s arousal at an optimal level, which then contributes to the development of emotion regulation strategies later in life (Feldman, 2003; Feldman, Greenbaum, & Yirmiya 1999). These findings are compatible with an antiphase form of coordination between mothers and infants, but unfortunately, the research in this area typically uses models of coordination that do not differentiate between inphase and antiphase. In contrast, research shows that father–infant interactions tend to be focused more on high arousal play, which produces shared father–infant bursts of positive effect. With respect to emotional coordination, this suggests that the father’s interaction with the infant may be modeled as a pattern of inphase coordination.
Interestingly, in the present study, we see similar gender differences in patterns of emotional coordination emerging when applied to romantic relationships. Based on our findings, we suggest that cooperation could mean different things for men and women. For men, cooperation may mean going along with their partner’s wants and moods in an attempt to avoid an argument and to reach quick resolution (Carstensen et al., 1995; Gottman & Levenson, 1988; Levenson, Carstensen, & Gottman, 1994; Levenson & Gottman, 1983). However, for women, cooperation may be more based on an interactional pattern of going back-and-forth with their partner, motivated by a desire to have both partners express their emotions and to negotiate a mutually satisfying solution.
Implications for working with couples
Both inphase and antiphase emotional coordination have implications for understanding interpersonal emotion regulation, because men’s inphase pattern could contribute to emotional escalation and de-escalation, while women’s antiphase pattern could contribute to emotional homeostasis (Butler & Randall, 2013). More broadly, the present research sheds light on the interpersonal behaviors that are associated with emotional coordination between partners. Thus, this work has relevance for therapeutic interventions that utilize aspects of cooperation, especially in dealing with everyday issues, such as stress. Assuming interdependence and satisfaction within the relationship, each partner should be motivated to help the other deal with stressful events in an effort to stabilize the partner and to reduce his or her own stress (Bodenmann, 1997, 2005; Randall & Bodenmann, 2009). Therefore, it could be concluded that a key element of dyadic coping would be the cooperation of both partners during a stressful situation and, indeed, prior research has shown that effective dyadic coping occurs when partners cooperate with one another using strategies such as joint problem solving and information seeking (Bodenmann, 2005). Couples that engage in dyadic coping show higher relationship quality, lower stress experience, and better psychological and physical well-being (Badr, 2004; Bodenmann et al., 2006). Perhaps, these positive outcomes arise because active strategies like cooperation and dyadic coping have a profound effect on couple’s shared emotions and therefore their ability to interpersonally regulate stressful situations? Confirming or disconfirming this possibility represents an important direction for future research targeted at understanding couple’s emotional dynamics and the causes and consequences of different coordination patterns (inphase versus antiphase).
Limitations
As with all research, our findings should be considered in light of several limitations. First, our sample was recruited from Craig’s list and engaged in a discussion of health behaviors, both of which limit generalizability. Similarly, our couples varied widely in age and relationship length, which suggests the need for future research to examine whether emotional coordination patterns emerge for younger couples early in their relationship or develop over the course of the relationship. Second, while we examined cooperation, research on couples and health behavior suggests that there are other dyadic patterns that could have been co-occurring, given the discussion topic of health-related issues. Specifically, one common pattern that couples can engage is in demand–withdraw pattern, whereby one partner demands change and the other withdraws from the conversation (Christensen & Heavey, 1993). It seems likely that such behavioral patterns could influence emotional coordination, but that issue awaits future research.
Another set of limitations arises from the necessity of choosing the timing interval for the measure of emotion. In the present work, we average the emotion ratings over 10 second (10-s) intervals. The decision was based upon prior work that has assessed emotional experience, behavior and autonomic physiology in 10-s segments (e.g., Gottman & Levenson, 1985; Levenson & Gottman, 1983). Using a 10-s segment of time helps smooth random noise in the measure and allowed us to observe interpersonal emotional exchanges between partners that occur across 20s, which fits with the idea that emotions rise and fall over a few minutes (e.g., Cacioppo et al., 1992). However, it would also be important to test whether these patterns of emotional coordination—inphase and antiphase—occur when emotional experience is assessed in smaller or larger time units. We also focused on a transmission model of emotional coordination due to the widespread use of that model. However, this approach does not allow for disentangling the relationship between antiphase patterns and emotional coregulation (defined as a homeostatic process) between partners (see Butler, 2011; Butler & Randall, 2013). The coupled oscillator (CO) model may be one statistical model that could be used to model emotional homeostasis between partners (e.g., Helm, Sbarra, & Ferrer, 2012). A CO model includes dampening and coupling parameters that can reflect emotional stabilizing processes, but the CO model does not include estimates of inphase versus antiphase coordination, just as the transmission model does not include estimates of stabilizing processes. Thus, a possible avenue for future research is to develop a mathematical model that includes the key theoretical terms from both the transmission and the CO model in order to establish a more comprehensive picture of the relationship between homeostasis and antiphase coordination.
Finally, because we assessed cooperation across the entire conversation, we cannot tell whether changes in cooperation were followed by subsequent changes in emotional coordination. Further research is needed that assesses behavioral cooperation and other potentially relevant couples’ behaviors in smaller time units. Such research would help us begin to understand whether changes in cooperation, or other dyadic behaviors, inhibit or foster interpersonal coordination.
Conclusions and future directions
Although our findings begin to shed light on gender differences in emotional coordination in the context of romantic relationships, we acknowledge that these findings are correlational and need to be both replicated and extended to experimental research paradigms. One direction of inquiry would be to unpack the meaning of cooperation for men and women. For example, it is possible that there were gender differences regarding which items in our scale best assessed cooperation. Specifically, we suggest future research examine cooperation based on items that are pretested to be gender neutral, as we realize the items in our measure are arguably biased toward how women may choose to cope with issues or cooperate with their partner (e.g., “talk it out”). Second, cooperation was spontaneous in our study, meaning that it is likely correlated with unassessed potential confounds such as individual agreeableness. Experimental protocols that manipulate cooperation are needed to address this issue.
Given the lack of prior research on cooperation and emotional coordination, our findings are preliminary and our interpretations somewhat speculative. Nevertheless, our results suggest several avenues for future research that could begin to clarify the role of emotional coordination in emotion regulation, health, and well-being. One issue is the potentially dynamic cycle that may exist between cooperation and emotional coordination. An important next step is to test whether increased emotional coordination generates feelings of rapport, which in turn could contribute to sustained cooperation, interpersonal emotion regulation, and relationship well-being. Resolving this temporal issue will require research that investigates the same couples during conversations and daily life and uses estimates of emotional coordination in both contexts to predict indicators of individual and relationship well-being. Such research would have potential importance for understanding how emotional coordination in close relationships contributes to well-being and for informing therapeutic interventions focused on dyadic coping and couple-level emotion regulation.
Footnotes
Appendix A
Author’s note
The authors would like to thank David Sbarra and Noel Card for their comments on an earlier version of this manuscript and all the undergraduate research assistants that helped with the behavioral coding.
Funding
Note
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
