Abstract
A functionalist approach to emotion regulation (ER) emphasizes the strategies people use to regulate their emotions and their goals for doing so. In the present work 90 mother–young adult child dyads (51 daughters; 39 sons) reported on their perceptions about relationship quality and the extent to which they employed one ER strategy (expressive suppression) and two ER goals (self- and other-oriented) when controlling anger toward the partner. Suppression and both ER goals predicted one’s own reports of relationship quality; however, only the two ER goals predicted partner reports of relationship satisfaction. Relations between ER processes and relationship quality were comparable for both generations. Results are discussed from a functionalist perspective in which ER is conceptualized as an integrated interpersonal system.
Emotion regulation (ER) has enjoyed a surge of interest among researchers during the past two decades. The functionalist approach to emotion suggests that ER encompasses “the extrinsic and intrinsic processes responsible for monitoring, evaluating, and modifying emotional reactions, especially their intensive and temporal features, to accomplish one’s goals” (Thompson, 1994, pp. 26–27). Within this framework, ER-related processes are defined as including both the strategies employed to modulate emotional responses, and the goals that underlie decisions to do so. Though ER can refer to efforts to initiate new emotions or to amplify emotions currently being experienced or expressed, the term is more typically used to describe attempts to avoid the experience or expression of emotion (Gross & Thompson, 2007).
A great deal of ER research has focused on strategies that are employed to avoid emotional displays, and extant literature supports the influence of such strategies on interpersonal functioning (Butler et al., 2003). In contrast, our understanding of the goals that motivate individuals to downregulate emotion is limited. Furthermore, although ER strategies and goals may both impact on interpersonal relations (Martini & Busseri, 2010), the majority of ER research has been intrapersonal, and studies related to social relationships have been limited. This neglect of social context is noteworthy, given that emotions are most often directed toward other people, and controlled when others are present (Fischer & van Kleef, 2010).
Consistent with the view that emotions are inherently relational (van Kleef, 2009), the present work addresses recent calls by emotion researchers (e.g., Fischer & van Kleef, 2010) to pay closer attention to emotion-related processes in social contexts. Drawing on the functionalist approach to ER, we also extend previous research by focusing jointly on why people seek to control emotion (ER goals) and how they opt to do so (ER strategies). Specifically, the present study investigates how an ER strategy (suppression), and two types of ER goals (self- and other-oriented) directed toward the avoidance of anger relate to relationship quality in mother–young adult child dyads.
ER strategies
With respect to how people avoid displaying felt emotion, researchers (e.g., Parkinson & Totterdell, 1999) have elucidated numerous ER strategies including substitution (showing one emotion while feeling another) and removing oneself physically from the emotion-eliciting situation. Among the most commonly studied strategies for avoiding emotional displays is expressive suppression, defined as the process by which ongoing outward signs of emotion are inhibited (i.e., maintaining a neutral expression rather than showing one’s genuine feelings; Gross, 2002).
In an interpersonal context, suppression is most often employed in the service of downregulating negative emotion, particularly anger. Although anger can damage relationships (Butler & Gross, 2004), engaging in suppression to avoid its expression also appears to disrupt social interactions and communication, and inhibit relationship formation between unfamiliar partners (Butler et al., 2003). These poor relationship outcomes may stem from the fact that suppression is cognitively taxing, and its use may render people less responsive to their partners (Gross, 2002). Such negative interpersonal effects may also result from suppressors feeling that their partners do not have an accurate impression of them. Further, relationship partners may be negatively impacted if they view the suppressor as disengaged or disinterested in the relationship. In summary, the use of suppression to avoid expressions of anger in interpersonal contexts may have negative implications for both suppressors and their partners.
ER goals
In contrast to the growing body of research focused on ER strategies such as suppression, relatively little research has investigated why individuals avoid displaying negative emotions (i.e., ER goals). For many years, psychologists assumed that ER was motivated exclusively by hedonics (i.e., that people want to feel positive emotions and avoid feeling negative emotions). However, this long-standing view may be overly simplistic, as it does not take into account the fact that people are sometimes motivated to feel emotions that are useful to them, regardless of whether they are pleasant or not (Tamir, 2009; Tamir & Ford, 2009). Though this more recent instrumental perspective has been useful in terms of drawing attention to the importance of ER goals, it is situated within an intrapersonal approach to emotion. In contrast, both empirical findings and theoretical formulations have underscored the importance of the social context in guiding ER goals (Fitzsimons & Finkel, 2010). Indeed, recent research supports the idea that ER goals may differ depending on the characteristics of one’s social partner (Martini, 2011). Studying ER goals within an interpersonal framework, then, requires explicit consideration of both members of the dyad.
In keeping with an interpersonal approach to the study of ER goals, developmental psychologists have distinguished between self-oriented goals, aimed at preventing negative consequences for the self (e.g., getting fired or getting into trouble) and other-oriented goals, motivated by the desire to maintain or enhance the wellbeing of the partner (Hastings & Grusec, 1998). In his affective model of parent–child relationships, Dix (1991) indicated that this distinction is important in helping to understand how goals are linked to relationship-based outcomes. Given that individuals typically bring numerous motives and concerns to an interaction (many of which have little to do with their partners’ wishes or wellbeing), the extent to which dyad members are aware of, and responsive to, the impact of their own actions (including, for example, expressing or controlling anger) on their partner is important in fostering ongoing cooperative exchanges. Specifically, when partners believe their feelings and concerns are being respected and promoted, they are far more likely to respond in kind, thus contributing to the development of a mutually responsive relationship (Martini & Busseri, 2010).
With regard to their implications, research suggests that self-oriented goals are linked to a lessened likelihood of helping (Eisenberg et al., 1989), and more negative affect and less sympathy for the relationship partner (Hastings & Grusec, 1998). In contrast, other-oriented goals are associated with more positive and less negative affect within the relationship, as well as greater satisfaction and more supportive behavior toward the partner (Dix, Gershoff, Meunier, & Miller, 2004; Martini & Busseri, 2010). In summary, whereas avoidance of anger in social contexts that is motivated by self-oriented ER goals may have negative implications for both relationship partners, other-oriented ER goals may be positively associated with relationship functioning for both members.
ER as an integrated interpersonal system
According to the functionalist perspective, focusing solely on either why people avoid anger (i.e., ER goals) or how they typically do so (i.e., ER strategies) provides an incomplete picture of ER processes. Such an approach would not inform, for example, how these processes relate to each other as an integrated system or to interpersonal outcomes such as relationship quality. Each of these issues is discussed in turn.
The integration of ER processes
With respect to the integrated nature of ER processes, studies have found that self- and other-oriented ER goals are not mutually exclusive, and both can be pursued within a single social interaction (Dix, 1991). Indeed, past studies have reported both negative (Hastings & Grusec, 1998) and positive (Martini & Busseri, 2010) correlations between these types of goals, as well as developmental differences in the pattern of relations between these variables (Martini & Busseri, 2010).
To understand the diversity of these patterns, it is worth noting that endorsement of ER strategies and goals is likely to be closely linked to the relationship context. In the Emotion as Social Information (EASI) model (van Kleef, 2009; van Kleef, DeDreu, & Manstead, 2010), emotional expression is seen as a means of influencing the behavior of social partners through its effects on both the partners’ inferential processes and their affective reactions. In long-term relationships, it might be expected that the outcomes of repeated interactions will enable an individual to predict with some accuracy what their partner’s inferences, affective reactions, and subsequent behaviors will be in response to displays of various emotions. Informed by this well-developed understanding of their partner’s cognitive, emotional, and behavioral response tendencies toward particular emotional displays, it is possible that individuals create relatively stable interpersonal schemas that are unique to each relationship, and include information that is emotion-specific (Baldwin, 1992; Fehr, Baldwin, Collins, Patterson, & Benditt, 1999). Thus, emerging knowledge about a partner’s likely response to displays of anger may shape stable patterns in an individual’s preferred strategies for regulating that emotion and their motivation (i.e., goals) for doing so in the context of that relationship. Such patterns may differ, however, for the two members of the dyad or for emotions other than anger.
The view that ER processes (i.e., strategies and goals) are guided by relatively stable interpersonal schemas has two implications. First, this view suggests that the sequential models of emotion activation (e.g., goals --> strategies, or strategies --> goals) often tested in intrapersonal studies (e.g., Urry, 2009) may not hold in close personal relationships, where ER strategies and goals are potentially activated simultaneously as part of an implicit interpersonal schema built up through repeated interactions and reciprocal shared experiences. A second implication of this view is that there may be no single pattern of association among ER processes that typifies most relationships, or most emotion-related encounters within a single relationship. In the case of ER directed toward the avoidance of anger, commonality among ER strategies and goals – reflected in the tendency for individuals to report higher levels of suppression and both types of ER goals – may reflect a broader belief system that includes the perception that anger is an emotion with destructive consequences whose expression should be controlled or avoided (Gottman, Katz, & Hooven, 1996). However, such a pattern may not be observed for all individuals, or even within the same dyad in diverse circumstances. For example, divergence among ER processes – reflected in inconsistent or non-significant associations – may be observed when contextual cues are influential in determining an individual’s anger-related responses (e.g, anger with the partner that is experienced in public vs. private situations).
ER processes and relationship quality
Regardless of the degree of commonality or independence among ER strategy- and goal-related processes, unique relationship outcomes associated with the avoidance of anger may be predicted by each type of strategy and goal. For example, irrespective of the reasons why one suppresses the expression of anger to one’s relationship partner, expressive suppression may lead to negative interpersonal outcomes, as reviewed above. Alternatively, whether one’s ER goals are predominantly self- versus other-oriented may be a unique, if not more primary, indicator of relationship functioning than the strategies employed to achieve this goal (Martini & Busseri, 2010). Despite growing recognition of the importance of attending both to strategies and goals in evaluating ER in interpersonal relations (e.g., Thompson, 1994), researchers have yet to examine these factors jointly with respect to relationship functioning. Consequently, the combined and unique predictive roles of suppression as an ER strategy and self- and other-oriented ER goals have yet to be evaluated with respect to interpersonal outcomes associated with the avoidance of anger.
Relationships between mothers and young adult children
Despite the interpersonal nature of ER and its implications for relationship functioning, research concerning ER in relationship dyads is limited, and has focused largely on interactions between strangers, peers, or romantic partners (e.g., Butler et al., 2003; Gross & John, 2003; Carstensen, Gottman, & Levenson, 1995). In contrast, the present study examined ER in relationships between young adult children and their mothers. These relationships are noteworthy because they are long-standing (unlike strangers) and are not discretionary (unlike peers or romantic partners). There is also a power differential that is often not present in other adult relationships.
The present research focuses on the stage when children are considered to be “emerging adults” (the developmental period occurring between the ages of 18 and 25; Arnett, 2000). We see this developmental period as one in which ER is important because numerous transitions make it an emotion-laden point in the parent–child relationship. Some of these transitions are likely to bring about positive emotional experiences, including increased mutuality and reciprocity (Morgan, Thorne, & Zurbriggen, 2010). Further, filial maturity, that is, the tendency to see the world from one’s parents’ perspective and to view one’s parents as adults with their own concerns and limitations (Lawton, 2001; Rook, Sorkin, & Zettel, 2004), also increases, particularly with respect to mothers (Birditt, Fingerman, Lefkowitz, & Dush, 2008).
In spite of these positive developments, features that characterize emerging adulthood, including instability and identity exploration, can also be sources of tension in the parent–child relationship (Arnett, 2000). Indeed, these relationships are often marked by intense negative emotions such as anger and frustration (Arnett, 2000). Because these negative emotions can have a detrimental impact on relationships (Newsom, Rook, Nishishiba, Sorkin, & Mahan, 2005), ER strategies and goals may be important in maintaining a positive emotional tone between young adult children and their parents. To date, however, we know little about ER strategies and goals related to the avoidance of anger, or their associated outcomes, in this relationship context.
Developmental research examining generational differences in ER processes and associated outcomes is rare. Whereas some evidence suggests that the use of suppression declines with age (John & Gross, 2004), the only published study examining generational differences in self- and other-oriented ER goals reported non-significant findings (Martini & Busseri, 2010). In contrast, numerous studies have indicated that parents often report better relationship quality than their children (Martini, Grusec, & Bernardini, 2001; Rossi & Rossi, 1990). With respect to the potential linkages among these processes, findings suggest generational congruence in the outcomes associated with ER strategies and goals (e.g., John & Gross, 2004; Martini & Busseri, 2010). That is, for both generations, suppression and self-oriented goals are typically associated with poorer relationship quality while other-oriented goals have been linked to better relationship quality. In summary, notwithstanding potential generational differences in the frequency of ER-related behavior, ER directed toward the avoidance of anger may have comparable implications for both dyad members in relationships between mothers and their young adult children.
A dyadic approach to studying ER in interpersonal relationships
In the present work, we employ a dyadic approach based on the assumption that each member of the mother-child relationship influences the functioning and outcomes for both members of the dyad (Cook & Kenny, 2005). In the present context, a dyadic approach treats the ER strategy and goals employed by mothers and their young adult children as joint predictors of their own and their partners’ beliefs about relationship quality. We believe this approach to be an important one given that prior research has largely neglected the social implications of emotions, even though the reciprocity that characterizes emotional exchanges in close relationships is critical to an understanding of ER (Fischer & van Kleef, 2010).
The actor-partner interdependence model (APIM) is an analytic approach for investigating dyadic systems, and has been used extensively in studying parent–child relationships across the lifespan (Martini & Busseri, 2010; Pesonen, Raikkonen, Heinonen, Jarvenpaa & Strandberg, 2006). This approach addresses the statistical interdependence between dyad members (Cook & Kenny, 2005; Kenny, 1996) by examining “actor effects”, that is, the statistical effects of each member’s predictor variable(s) on their own outcome(s) and “partner effects”, that is, the statistical effects of one member’s predictor variable(s) on the other member’s outcome(s). The statistical equivalency of these effects can also be tested across dyad members, thus permitting an assessment of whether the observed actor and partner effects differ significantly between dyad members. In the present context, testing this type of equivalency informs whether the predictive associations involving ER and relationship quality are consistent across generations, or instead vary between generations in magnitude and/or direction.
The present study
Drawing on the functionalist approach to ER, the present work addresses recent calls by emotion researchers to pay closer attention to emotion-related processes in social contexts, and extends previous research by focusing both on why people seek to control emotion (ER goals) and how they opt to do so (ER strategies), as well as the implications of these ER processes for relationship functioning. Though ER can relate to both up- and downregulation of emotion, the present study focuses specifically on the control of anger, given its well-established negative consequences for relationship quality. Our main goal was to investigate how ER processes related to the avoidance of anger – including the ER strategy of suppression, and self- and other-oriented ER goals – were linked with relationship quality in a sample of mother–young adult child dyads. Developmental differences were also evaluated for each of the variables and predictive associations were assessed.
In light of the significance of angry displays in social relationships, some degree of commonality (i.e., association) among suppression and ER goals was expected. However, given that a consistent pattern has yet to emerge from the few studies examining this issue to date (Hastings & Grusec, 1998; Martini & Busseri, 2010), no specific hypotheses were made concerning the direction or magnitude of the correlations among ER processes.
Despite the potential commonality among ER processes, we anticipated that the “how” and the “why” components of the ER system would have unique implications for relationship quality. Consistent with evidence linking expressive suppression with less adaptive interpersonal functioning (e.g., Butler et al., 2003; Gross & John, 2003), we hypothesized that less use of expressive suppression would predict more positive evaluations of relationship quality. Further, consonant with emerging theory and research suggesting the importance of self- and other-oriented ER goals as distinct, and possibly opposing, predictors of relationship outcomes (Martini & Busseri, 2010), we hypothesized that, independent of suppression, weaker endorsement of self-oriented ER goals would predict more positive evaluations of relationship quality, as would stronger endorsement of other-oriented goals. All of these hypotheses were expected to hold with respect to actor and partner effects (i.e., within as well as between dyad members) and for both members of the dyad.
Based on previous research addressing generational differences (e.g., John & Gross, 2004; Martini et al., 2001; Rossi & Rossi, 1990), we predicted that mothers would report less suppression and greater relationship quality than their young adult children; however, given the paucity of research concerning ER goals, no specific predictions were made concerning developmental differences in the use of self- and other-oriented ER goals.
Method
Participants and procedure
Young adult participants (80 female, 80 male) were randomly selected from a database of undergraduates who had indicated a willingness to take part in research for credit that would also involve their parents’ participation. Those who agreed to take part (80 female, 66 male; M age = 20.31 years; SD = 2.28) completed a self-report survey in small group settings in exchange for course credit. These participants also provided their mothers’ mailing addresses but were told there would be no penalty should their mothers decline to participate. Mothers received a mailed introductory letter, a consent form (indicating there was no penalty to their child if they did not participate), study materials, and a return envelope for completed measures. Of the 160 mothers contacted, 62% responded (n = 90, 51 mothers of females and 39 mothers of males; M age = 47.21 years; SD = 4.13). Of these 90 mother–young adult child dyads, 79% of mothers and 90% of their children were born in North America. Ninety-eight percent of the young adult children were unmarried, and 66% resided with their mother. The percentage of young adult participants residing at home was comparable to that of the university as a whole (69% living at home vs. 31% not at home). Among the mothers, 91% were married, and 82% were employed outside the home. Mothers reported having an average of 2.5 children (range 1–5). All participants received written debriefing after the completed study materials were returned.
Measures
Items were developed to address both how participants avoided the expression of anger with the other dyad member (in this case, the extent to which they used suppression), as well as why they elected to do so (self- vs. other-oriented ER goals).
ER strategy – suppression
On a scale ranging from 1 (never/rarely) to 7 (almost always), participants rated their use of suppression to control anger toward the partner using a single item (“When you are angry with your mother/daughter/son, how often do you try not to let him/her see that you are angry by keeping the expression on your face neutral?”). This item was derived from a review of extant literature (e.g., Parkinson & Totterdell, 1999).
ER goals – self-oriented and other-oriented
On the same 7-point scale, participants indicated the extent to which self-oriented ER goals (e.g., “I avoid showing anger to my mother/daughter/son because I want to avoid negative consequences for myself”; 9 items; αmothers = .94; αchildren = .85) and other-oriented ER goals (e.g., “I avoid showing anger to my mother/daughter/son because I want to avoid hurting his/her feelings”; 3 items; αmothers = .87; αchildren = .84) came to mind when controlling anger toward the partner. These items were derived from previous research on ER goals (e.g., Martini, 2011).
Relationship quality
As an overall measure of relationship quality, participants completed the 25-item Quality of Relationships Inventory (Pierce, Sarason, & Sarason, 1991) using 5-point scales ranging from 1 (not at all) to 5 (very much) to assess relationship conflict (e.g., “To what extent is your mother/daughter/son critical of you?”), social support (e.g., “To what extent could you count on your mother/daughter/son for help with a problem?”) and relationship depth (e.g., “How significant is your relationship with your mother/daughter/son in your life?”). Ratings were averaged across all items after reverse-scoring those that were related to conflict, such that higher scores indicated higher relationship quality; αmothers = .84; αchildren = .92.
Results
Preliminary analyses
Demographic differences
The main effect of young adult children’s gender on suppression, ER goals, and relationship quality was non-significant in independent sample t-tests (all ps > .05), and the covariance matrices comprising all study measures did not differ between sexes (p > .05). Consequently, young adult children’s gender was not considered in subsequent analyses. In contrast, the main effect of young adult children’s residence location (at home vs. not, coded as 1 and 2, respectively) was significant for mothers’ reports of suppression and both ER goals (ps < .05), but not relationship quality (p > .05); as well as young adult children’s reports of both ER goals (ps < .05), but not suppression or relationship quality (p > .05). Compared to dyads in which the young adult child did not live at home, in dyads in which the young adult child lived at home, mothers reported less suppression and lower endorsement of self- and other-oriented ER goals, and young adult children reported lower endorsement of self- and other-oriented ER goals. However, the covariance matrices did not differ between dyads as a function of the young adult children’s residence location (p > .05). Given the mean group differences, however, young adult children’s residence location was treated as a covariate in subsequent analyses.
Generational differences
Means, standard deviations, and correlations among study measures are shown in Table 1. The main effect of generation (mother vs. young adult child) on suppression, ER goals, and relationship quality was evaluated using paired-sample t-tests. On average, mothers reported less use of suppression, self-oriented ER goals, and other-oriented ER goals, as well as higher relationship quality than their young adult children did (ps < .05). Note that after controlling for young adult child residence location, mothers reported less use of self- and other-oriented ER goals (ps < .001) whereas differences in mean levels of suppression and relationship quality were non-significant (ps > .05). Results concerning generational differences in the predictive relations between ER processes and relationship quality are presented with the APIM results below.
Means, standard deviations, and correlations among study measures
Note. N = 90.
* p < .05.
Bivariate associations among ER processes and with relationship quality
As shown in Table 1, correlations among suppression and both types of ER goals were positive and significant for mothers and for young adult children. That is, as anticipated, within both generations there was a significant degree of commonality among ER processes such that individuals reporting greater use of suppression to avoid the expression of anger also tended to report greater self- and other-oriented goals for this avoidance. With respect to pairwise associations between ER processes and relationship quality for mothers, greater relationship quality was associated with less use of suppression and less endorsement of self-oriented ER goals; the correlation with other-oriented ER goals was non-significant. For young adult children, all three ER processes were associated with relationship quality such that greater relationship quality was associated with less use of suppression, less endorsement of self-oriented ER goals, and greater endorsement of other-oriented ER goals. Note that these patterns were unchanged after statistically controlling for young adult children’s dwelling location (at home vs. not).
Predicting relationship quality
The pairwise associations between ER processes and relationship quality do not inform the unique predictive relations involving each ER process independent of the other processes, nor do they inform potential partner effects. To address both issues, an APIM was estimated in which mothers’ and young adult children’s relationship quality scores were predicted by their endorsement of suppression together with self- and other-oriented ER goals. To estimate the APIM, correlations were specified between each pair of predictors and between residual variance terms for mothers’ and young adult children’s relationship quality scores; young adult children’s dwelling location was also included as a covariate in each model (correlated with each predictor). Analyses were conducted using AMOS software (version 17.0) and maximum likelihood estimation. Multiple criteria were employed to evaluate goodness of model fit: model chi-square, CFI, SRMR, RMSEA, and residual covariances (see Kline, 1998).
In the base version of the model, actor and partner effects were freely estimated, resulting in a saturated path model (i.e., df = 0). To test the statistical consistency of the actor and partner effects between mothers and young adult children, a constrained APIM was estimated in which corresponding actor effects (e.g., mothers’ use of suppression predicting their reports of relationship quality, and young adult children’s use of suppression predicting their reports of relationship quality) and corresponding partner effects (e.g., mothers’ use of suppression predicting their young adult children’s reports of relationship quality, and young adult children’s use of suppression predicting their mothers’ reports of relationship quality) were constrained to be equal between mothers and young adult children based on the corresponding unstandardized path coefficients. The predictive paths from young adult children’s residence location to relationship quality were also constrained between mothers and young adult children.
The constrained model provided good overall fit (χ2 = 12.56, df = 7, p = .09; CFI = .97; SRMR = .04; RMSEA = .09, p for close fit = .17). Further, the model chi-square statistic did not differ significantly from the unconstrained model and the residual covariance matrix revealed no significant discrepancies between the unconstrained and constrained models. These results suggest that the assumption of consistency between mothers and young adult children in corresponding actor effects and corresponding partner effects in predicting relationship quality was tenable. Stated differently, the actor and partner effects of ER processes on relationship quality were not significantly moderated by generation (mother vs. young adult child). Results from the constrained model are presented in Table 2.
Results from APIMs predicting mother’s and young adult children’s relationship quality from suppression and ER goals simultaneously
Note. N = 90. Standardized path coefficients are shown by predictor (row variable) for each criterion (column variable). All corresponding paths were constrained to equality between mothers and young adult children based on the unstandardized path coefficients. RQ = relationship quality.
*p < .05
The model explained 34% and 23% of the variances, respectively, in mothers’ and young adult children’s relationship quality scores. For mothers and young adult children, unique significant actor effects were found for suppression and both self- and other-oriented ER goals, wherein higher relationship quality was predicted by less use of supression (βs = −.19 and −.15 for mothers and young adult children, respectively; ps < .05), less endorsement of self-oriented goals (βs = −.51 and −.33 for mothers and young adult children, respectively; ps < .05) and greater endorsement of other-oriented goals (βs = .50 and .33 for mothers and young adult children, respectively; ps < .05) by the self. Partner effects for mothers and young adult children were significant for self- and other-oriented goals, wherein lower endorsement of self-oriented goals (βs = −.15 and −.18 for mothers and young adult children, respectively; ps < .05) and greater endorsement of other-oriented goals (βs = .19 and .23 for mothers and young adult children, respectively; ps < .05) uniquely predicted greater relationship quality for one’s relationship partner. In contrast, partner effects for suppression were non-significant.1, 2
Each of these patterns was consistent when suppression and ER goals (self- and other-oriented) were examined as predictors of relationship quality in two separate APIMs. In the constrained APIM comprising suppression as the predictor, the R 2 values for mothers’ and young adult children’s relationship quality were .08 and .05, respectively. In the constrained APIM comprising ER goals as the predictor, the R 2 values were .31 and .21, respectively.
Discussion
Drawing on functionalist and interpersonal perspectives on ER, the present research investigated one ER strategy (expressive suppression) and two ER goals (self- and other-oriented) for the avoidance of anger as predictors of relationship quality in mother–young adult child dyads. All three ER processes were positively related, such that greater use of suppression to avoid expressing anger to one’s partner was associated with greater endorsement of self- and other-oriented ER goals. This finding may reflect the fact that both measures overlapped in their reference to avoiding expressions of anger. Alternatively, it might reflect the extent to which anger is experienced in a relationship, with individuals who experience more anger activating ER goals and strategies such as suppression more frequently. In addition, we found that suppression and both types of ER goals were unique predictors of one’s own reports of relationship quality; however, only the two ER goals predicted the partner’s reports of relationship quality. The relations between ER processes and relationship quality were comparable for mothers and their young adult children, suggesting that the processes that underlie these relationships may be similar for both generations.
ER as an integrated system
Consistent with Thompson (1994), we undertook a joint examination of ER strategies and goals in order to provide a fuller picture of the integrated nature, and interpersonal implications, of ER. In the present work we observed positive correlations among the ER strategy under study (suppression) and both types of ER goals (see also Martini & Busseri, 2010). This pattern of relations is consistent with the contention that repeated dyadic interactions over many years may result in the creation of stable interpersonal schemas, which include ER processes that are specific to avoiding the expression of anger (Baldwin, 1992; Fehr et al., 1999). Further, to the extent that anger is viewed as an emotion that should be controlled rather than expressed (e.g., Fehr et al., 1999), the positive correlations among ER processes may reflect a generalized response aimed at avoiding anger through multiple means.
Our findings also support the idea that ER strategies and goals may be activated simultaneously as part of these implicit schemas. However, we do not see this perspective as excluding the possibility that a different pattern of relations among ER processes might emerge in other research, such as an inverse relation between self- and other-oriented goals (Hastings & Grusec, 1998), or a temporal-sequential model (e.g., Urry, 2009). Based on our findings, we suggest that the pattern of relations between ER processes may vary according to whether they are being studied from an intrapersonal or an interpersonal perspective, and depending on the nature of the relationship and type of emotion under investigation.
Suppression
Consistent with predictions based on previous research (e.g., Gross & John, 2003), greater use of expressive suppression was associated with one’s own report of lower relationship quality. Research suggests that suppression is linked to feelings of inauthenticity (Gross & John, 2003), and individuals who feel that they are not conveying their true thoughts and feelings to the partner may experience a diminished sense of relationship quality because they do not feel as though their partner has a true image of who they are. This reasoning may help to explain why individuals report using suppression often in some relationships while simultaneously indicating that they would prefer to be able to express the emotions that they feel more openly (Martini, 2011). Gross and John (2003) have also noted that suppression is associated with greater attachment avoidance. It may be the case, therefore, that individuals who favor suppression as an ER strategy are more comfortable with relationships that are emotionally distant. They may have relatively low expectations with regard to the support that they can expect from others and may perceive (erroneously, perhaps) that their suppression of negative emotion helps to minimize conflict in the relationship.
Greater use of suppression as an ER strategy did not predict partner perceptions of relationship quality, even when ER goals were excluded from the predictive model. While this finding may have emerged because the single-item measure of suppression was less reliable than the measure of ER goals, it may also reflect the fact that suppression is not always obvious to the partner. Lab-based studies (e.g., Gross & Levenson, 1997) suggest that participants who are instructed to suppress their feelings are typically able to do so without great difficulty, often eliminating all evidence of emotional experience from their facial expression and body posture. Individuals who are skilled at suppressing negative emotion may use this strategy so successfully that partners are completely unaware of the feelings that are being concealed from them. It is also plausible that some partners are unable to detect the use of suppression even when the strategy is being poorly executed and traces of felt negative emotion are apparent. Studies suggest that there are significant individual differences in the sensitivity to emotional displays, particularly those that are more subtle (Halberstadt, Denham, & Dunsmore, 2001). Either of these circumstances may play a role in explaining the non-significant relationship between the use of suppression and partner reports of relationship quality.
While we have interpreted our findings in terms of ER strategies influencing relationship outcomes, the correlational nature of this study makes it necessary to consider that the direction of the effect may be the opposite of that which we have proposed. Specifically, it may be the case that individuals are more likely to employ suppression in lower-quality relationships in which they feel less supported or experience greater conflict.
It would be valuable to evaluate whether the significant actor effects and non-significant partner effects of suppression observed here would be found in other cultures or relationship types (e.g., work colleagues, spouses), for additional facets of relationship quality or other types of relationship outcomes, and for emotions other than anger. Across cultures, the interpersonal function of ER is typically to maintain positive relationships with others and to aid in managing the impressions that others have of us (Mesquita & Albert, 2007). However, the different values that are placed on various emotions across cultures means that norms governing the endpoints of ER, and the most effective ER strategies to achieve those endpoints, may vary. For example, while the negative effects of suppression have been well established in North American samples (Gross, 2002), cross-cultural research suggests that expressive suppression is considered to be a valuable way of controlling less desirable emotions in some cultures, and is emphasized by parents when socializing their young children’s affective responses (Chen et al., 1998). Clark and her colleagues (e.g., Clark & Taraban, 1991) have also noted that in non-communal relationships (e.g., work-based environments) people prefer reduced levels of emotional expression, particularly negative emotion, and suppression may be an appropriate means of creating such working conditions. Finally, it has been noted that expressive suppression may be a sound ER strategy to employ when the intensity of anger is routinely very high, as it may help to avoid the escalation of conflict within the relationship (van Kleef & Côté, 2007). In any of these relationship contexts, the effect of suppression may be significant and possibly even positive, particularly as a predictor of desirable context-dependent relationship outcomes (e.g., the ability to complete tasks efficiently). Other related issues that deserve attention include determining the effects of suppression when positive (e.g., happiness) or negative emotions other than anger (e.g., sadness, fear) are experienced in the context of relationships.
ER goals
ER goals were consistent predictors of relationship quality for both dyad members as well as their partners. As hypothesized, endorsement of other-oriented goals was positively associated with reports of relationship quality for both dyad members and their partners, while the reverse pattern was observed for self-oriented goals. Such findings are consonant with the suggestion that ER goals serve to guide other interpersonal processes that are important determinants of relationship quality. While goals are typically viewed as intrapersonal, they may be communicated to the partner either explicitly or implicitly. Further, the overall degree to which behaviors are, or appear to be, motivated by self- versus other-oriented ER goals may communicate something to the partner about the value that has been placed on their needs and wishes. This observation may help to explain why ER goals, even though they are not necessarily observable by the partner, may impact on partner perceptions of relationship quality.
In addition, Dix (1992) has proposed that when other-oriented goals are active during negative interpersonal situations, they may influence attention and cognition in ways that promote positive relationships, including prompting individuals to scan the environment for factors likely to affect the partner’s wellbeing, and increasing sensitivity to the partner’s displays of affect. Other-oriented goals are also thought to render the partner’s thoughts and feelings more accessible to consciousness, and to increase the likelihood that benign appraisals will be drawn during emotionally ambiguous situations. Conversely, ER that is motivated by self-oriented goals appears to have detrimental effects for the relationship partner (e.g., Dix et al., 2004; Eisenberg et al., 1989). For example, parents of younger children who are motivated by self-oriented goals appear to be more likely to engage in punitive and controlling behavior during conflict, and less likely to negotiate, or display warmth and acceptance (Hastings & Grusec, 1998).
With respect to actor effects, results were parallel for both mothers and their children. Self-oriented goals were negatively associated with relationship quality for the self, whereas the opposite was true of other-oriented goals. It is possible that some of the mechanisms purported to explain relations between ER goals and partner perceptions of relationship quality are also relevant with regard to self-reported relationship quality. For example, the benign appraisals that are associated with other-oriented motives during emotionally ambiguous circumstances are likely to prompt behaviors that are more palatable to the partner than those that would emerge from hostile attributions. Such benign appraisals are also likely to factor favorably into one’s own overall estimation of the partner, rendering a more positive assessment of the relationship more generally.
Though ER processes are typically viewed as preceding relationship quality in the literature (e.g., Butler et al., 2003; Gross, 2002), it is important to consider that the direction of these associations might also be reversed, with individuals in higher-quality relationships being more likely to adopt relatively stronger other-oriented (vs. self-oriented) concerns, and those in lower-quality relationships being relatively more apt to be motivated by self-oriented goals. Alternately, in keeping with our views about interpersonal schemas (Baldwin, 1992), it may be the case that ER processes such as strategies and goals are tightly bound to perceptions of relationship quality in long-standing close relationships, and that these constructs influence one another reciprocally over time. Evidence does suggest that ER processes and perceptions of relationship quality change across the lifespan (Birditt et al., 2008; John & Gross, 2004); however, more sophisticated designs are needed to evaluate issues related to directionality and reciprocal influence, including longitudinal dyadic studies examining relations among these processes over time, or studies comparing dyads with relationship histories of varying length and experiences.
Generational differences and consistency in ER processes
In the present study, mothers reported less use of suppression, and less endorsement of self- and other-oriented goals than their children. These findings may reflect the fact that in parent–child relationships, children are more prone than their parents to engage in ER, and thus the use of suppression and ER goals may be more salient to the younger generation. It is also possible that such a tendency may stem from the fact that the young adult children in this sample experienced more anger in the relationship than did their mothers, resulting in an increased need to engage in ER and, consequently, greater activation of suppression and ER goals. In future studies, examination of specific emotional experiences (e.g., anger, happiness) alongside corresponding emotion-specific ER strategies and goals would provide a valuable extension of the present work. It is also important to note that in confining ourselves to the study of suppression (assessed using only a single-item indicator), we assessed only one of the many ER strategies that can be deployed to avoid the expression of anger (Parkinson & Totterdell, 1999).
Similar to several other researchers (Rossi & Rossi, 1990), we also observed that mothers reported higher relationship quality than did their young adult children. Such a result is consonant with theorizing about the generational stake (Bengston & Kuypers, 1971), but it is noteworthy that this generational difference (along with the generational difference in the use of suppression) was no longer significant after partialling out the effects of co-residence. In measuring generational differences in perceptions of relationship quality it is rare to control for physical proximity or frequency of contact, and it may be the case that future research should be more sensitive to the effects of such variables. Stated differently, apparent generational effects may, in some cases, stem from structural characteristics of the relationship, including proximity and shared residence.
Given the proposed importance of effective ER to interpersonal functioning in adulthood (e.g., Butler et al., 2003; Martini & Busseri, 2010), we anticipated that the patterns of predictive relations between ER and relationship quality would be observed for both mothers and their children. Present findings supported this hypothesis and, furthermore, suggested that the direction and magnitudes of the actor and partner effects did not differ significantly across generations. It is unclear, however, whether a similar pattern of results would be found in parent–child relationships across the lifespan. For example, research indicates that the use of suppression decreases with age (John & Gross, 2004), while lifespan theories (e.g., Carstensen, 2006; Erikson, 1982) suggest that other-oriented goals may increase. Both of these age-related trends may have implications for the unique roles of suppression and ER goals, and the generational consistency of these predictive effects, in relationship dyads involving at least one older adult.
Limitations
The present findings are limited by the sample, and may not replicate in groups of people from different cultural groups or age cohorts. Additional research is also needed to determine whether the present pattern of results would hold in different types of relationships, for other measures of relationship quality or relationship functioning, and for emotions other than anger. This study was also limited by the single-item measure of expressive suppression, and associated concerns about the reliability of this measure. While a more comprehensive measure of suppression (e.g., a measure that assesses other dimensions of suppression, including those of a cognitive nature, in addition to behavioral forms of emotion suppression) and other ER strategies would be desirable in future research, prior literature supports the use of single-item measures to assess the suppression of negative emotion (Nezlek & Kuppens, 2008). Further, the present research relied on self-report data and is therefore subject to concerns regarding the validity of responses. We note, however, that prior empirical evidence suggests that participants are able to respond with some accuracy about their use of ER strategies (e.g., Gross & John, 2003) and their expression of emotion (e.g., Kring, Smith, & Neale, 1994).
Finally, although we treated ER strategies and goals as predictors of relationship quality, the cross-sectional design prevents claims concerning causality or temporal order. Additional research using longitudinal or experimental research designs is needed to test the directionality of these effects. Further, longitudinal approaches may also reveal predictive effects of suppression and ER goals which, when evaluated over time, differ in magnitude or direction from those observed based on cross-sectional findings.
In spite of its weaknesses, this study also had a number of strengths, including the dyadic nature of the data and a relatively balanced number of male and female young adults. Moreover, the present work examined ER in a context that has received little attention: mothers and young adult children. As such, this research complements social psychological research that has focused on ER in nonsocial settings, or during interactions between peers or strangers. By studying emerging adults, the present work also extends the developmental literature, much of which has been focused on ER either in childhood (e.g., Cole, Martin, & Dennis, 2004), or later in life (Carstensen, 2006).
Summary and conclusions
We examined expressive suppression as an ER strategy and ER goals (self- and other-oriented) for avoiding anger with one’s relationship partner as joint predictors of relationship quality in a dyadic context that has received relatively little attention: mothers and young adult children. Consistent with a functionalist view of ER as an integrated system, positive co-activation was found among all three ER processes. Despite this commonality, unique actor and partner effects were found for self- and other-oriented ER goals as predictors (negative and positive, respectively) of relationship quality; unique actor (but not partner) effects of suppression were also observed. Further, associations involving ER processes did not differ significantly across dyad members, suggesting generational consistency in the relevance of effective ER to relationship quality. There is clearly still much to be learned about ER strategies and goals, and it is hoped that the present research will stimulate interest in further exploring ER as an integrated system (rather than examining strategies or goals in isolation), as well as links between ER and the quality of social relationships and associated interpersonal experiences.
Footnotes
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.