Introduction

Emotion socialization is a formative process in adolescent socio-emotional development (Klimes-Dougan and Zeman 2007). Much of the extant literature on emotion socialization pertains to parents; however, friends gain increasing influence during adolescence (Rubin et al. 2009; von Salisch 2001). This narrative review will compare emotion socialization in parent–child relationships and close friendships, and how each relates to internalizing symptoms, externalizing symptoms, and social competence in adolescence. These outcomes are particularly relevant in adolescence, when youth are at heightened risk for developing psychological difficulties and face new social demands for which they may need emotional guidance (Costello et al. 2003; Garcia and Scherf 2015).

This review identified sources of emotion socialization strategies based on theoretical models (Eisenberg et al. 1998; Morris et al. 2007). Articles and chapters were then selected for inclusion when relevant to these socialization strategies (modeling, responses to emotion, emotion discussions) and focused on adolescence. Thus, inclusion was not based on a specific publication date range or keyword, but rather theoretical relevance. This strategy narrowed coverage of the vast emotion socialization literature by excluding empirical studies focused on early or middle childhood. Similarly, there is a large literature on the influence of parents and friends in adolescence, but the review focused on those studies that pertained to emotion socialization, rather than broader influence. Methodology, design, and sample characteristics for each of the articles included in this review are summarized in Table 1. The review first discusses parent–child relationships and friendships in adolescence, compared with earlier developmental periods, followed by a review of emotion socialization strategies.

Table 1 Summary of research studies

Parents and Friends in Adolescence

From a developmental perspective, the parent–child relationship is the first influential dyadic relationship in a child’s life and provides a basis from which adolescents engage with friends. First, it is a space to develop competencies to apply in future relationships, such as how to communicate (Rubin et al. 2011; Rudolph and Asher 2000). The parent–child relationship can also set the stage for future close relationships in adolescence and lead adolescents to selectively enter into particular types of friendships, referred to as niche selection (Nickerson and Nagle 2005). Friendships become particularly salient and influential during adolescence, when youth increasingly seek out friends for emotional support and explore their identity outside of how they have defined themselves in a family setting (Jobe-Shields et al. 2014; Nickerson and Nagle 2005; Rubin et al. 2011).

Adolescent friendships have several characteristics that may make friends’ emotional support desirable to adolescents. As children enter adolescence, friendships become more dyadic and intimate (Rose 2002; Rubin et al. 2011; Sullivan 1953; von Salisch 2001). Compared to the parent–child relationship, friendships tend to be characterized by similar levels of social power (von Salisch 2001). Because they perceive each other as equals and neither as responsible for regulating the behavior of the other, friends may be more likely to talk freely, express their feelings and challenge one another in a friendly context (Bukowski et al. 2007; von Salisch 2001). Friendships are also voluntary and therefore less stable than the parent–child relationship (Adams and Laursen 2001). Thus, adolescents may be more concerned with protecting their friends’ feelings to maintain the friendship or may self-disclose to friends in efforts to increase intimacy and belongingness (Doyle et al. 2009; Rose 2002; Zeman and Shipman 1997). Lastly, friendships are characterized by shared social experiences, which may facilitate emotional disclosures. Parents may offer support based on their own experiences as an adolescent, but this may not match their child’s experience (Bukowski et al. 2007).

Emotion Socialization

Adolescents learn about their emotional world through a variety of interactions with family members, teachers, and friends (Zeman et al. 2012). This transactional process, known as emotion socialization, begins early in life and continues throughout adolescence. Through emotion socialization, adolescents learn to recognize, label, and manage their emotional displays, as well as social norms for expressivity (Morris et al. 2007; Shipman et al. 2003). At this point in development, adolescents have already acquired foundational skills, but are also encountering new social and emotional challenges. Therefore, adolescents continue to learn from others in their environment regarding how to use these skills in flexible and differentiated manner. Parents are the primary agent of emotion socialization from an early age and continue to be influential; however, intimate friendships are another context for emotion socialization during adolescence (Collins and Laursen 2004; Rubin et al. 2011). Both parents and friends engage in a range of socialization strategies, such as modeling emotion expression, responding to adolescents’ emotions, and discussing emotions with adolescents (Brechwald and Prinstein 2011; Eisenberg et al. 1998; Morris et al. 2007). Though there are indirect methods of emotion socialization, such as situation selection (Fredrickson 1998), the present review will focus on the direct socialization strategies that are more widely documented in the parent and friend literatures. It should also be noted that, although cultural context affects emotion socialization processes, there is limited cross-cultural research on this topic and thus the majority of the emotion socialization research reviewed in the following sections pertains to Western cultures (Klimes-Dougan and Zeman 2007; Rubin et al. 2011).

Modeling

Both parents and friends model various emotions and how to manage expressivity in different situations, as well as with different people (Brechwald and Prinstein 2011; Morris et al. 2007). Although parents’ and friends’ modeling may transmit similar messages, they can also serve unique functions in that friends’ emotional displays provide socialization messages that adolescents may not have learned from observing their parents and vice versa. Witnessing others’ emotional displays does not ensure that the adolescent will internalize adaptive emotional experience and expression, but does inform adolescents’ perceptions of what is considered normative for the family unit or friendship (Kiuru et al. 2012). Adolescents may also seek to imitate their friends’ expressivity in efforts to increase the sense of belongingness in their friendship, or as part of a self-evaluation process when individuating from their parents (Kiuru et al. 2012). However, research demonstrates these modeling processes translate to similar outcomes across socialization agents, as both parents’ and friends’ negative expressivity are related to adolescents’ negative expressivity (Giletta et al. 2011; Klimes-Dougan et al. 2014; Luebbe and Bell 2014).

Responses

Parents and friends may also provide verbal and nonverbal responses to adolescents’ emotion expressions (Morris et al. 2007). These responses function as immediate feedback about acceptability of emotions, and thereby may increase or decrease adolescents’ expressivity. Of the few studies conducted on friends’ responses, the types of responses (supportive, unsupportive) are similar to those examined in parent emotion socialization research. Regarding supportive responses, parents and friends can encourage and validate how adolescents are feeling and expressing emotions, communicating that emotion expression is acceptable. For example, a parent or friend may empathize with the adolescent or acknowledge their emotion in the moment.

Parents and friends may also attempt to reassure adolescents or distract them from their feelings by redirecting them to something positive. For example, they may tell the adolescent not to worry, encourage them to think about something happy, or distract them with a pleasant activity. Referred to as overriding, this strategy is conceptualized as a supportive strategy in adolescence because it functions to alleviate youths’ distress and does not encourage excessive focus on their emotions, which could result in emotion dysregulation and risk for psychopathology (Brand and Klimes-Dougan 2010). Overriding may be a particularly adaptive response in adolescence, a developmental stage marked by increased emotional reactivity, rather than in early and middle childhood, when children are still gaining basic emotion knowledge and self-regulatory skills (Zeman et al. 2012). Factor analyses show that parental encouragement and overriding in adolescence are related and have similar links to psychological adjustment, offering evidence that both function as supportive responses when enacted by parents (Garside and Klimes-Dougan 2002). Whether friends’ overriding responses are supportive has yet to be determined.

Regarding unsupportive responses, parents and friends can mirror the adolescent’s emotions. Often referred to as magnifying, this response may further prolong adolescents’ emotional state, particularly with negative emotions (Moed et al. 2015; O’Neal and Magai 2005). For example, parents who respond to adolescents’ anger with their own expression of anger may promote dysregulated affect. By the same token, friends who mirror adolescents’ sadness may promote rumination and continued feelings of sadness. With parents, reciprocating adolescents’ emotions also limits parents’ ability to teach adolescents how to effectively manage their emotions because parents may focus on their own distress (Moed et al. 2015).

Parents and friends may also punish or neglect adolescents’ emotional displays (O’Neal and Magai 2005). Punishment may involve teasing or reprimanding the adolescent, whereas neglect refers to ignoring the adolescent’s expressivity. Punishing or neglecting adolescents’ emotions may discourage expression of emotion, and does not give adolescents opportunities to process emotions and develop coping skills. Overt punishment is less likely for friends to employ, in light of the horizontal structure of friendships, but may instead manifest in the form of teasing. Although less is known about the effect of friends, parental punishment and neglect of negative emotions has been linked to dysregulated affect in adolescents (Buckholdt et al. 2014b; Yap et al. 2008b). Together, research results suggest that both parents’ and friends’ responses to adolescents’ emotional displays can either facilitate or disrupt adolescents’ emotion regulation practices, depending on the nature of the response.

In both parent and friend emotion socialization literatures, supportive responses to negative emotions are more common than unsupportive responses (Jobe-Shields et al. 2014; Klimes-Dougan et al. 2014). The parent and friend emotion socialization literatures also demonstrate similar patterns in relation to adolescent emotion regulation, with supportive responses facilitating emotion regulation and unsupportive responses linked with dysregulated affect in adolescents. In comparison to parents, adolescents perceive fewer unsupportive reactions from friends (Zeman and Shipman 1997). It is possible that adolescents are more sensitive to parents’ unsupportive responses than to those of friends, perhaps because the more vertical power structure in parent–child relationships might enable more material consequences of parents’ unsupportive responses. It is also possible that friends enact fewer unsupportive reactions than parents as a result of the horizontal nature of friendships. Specifically, friends do not feel responsible for shaping and managing their friends’ emotion in the way that parents feel responsibility towards their child (von Salisch 2001). Friends may also refrain from displaying unsupportive responses for fear of dissolving the friendship, which is not a risk that parents encounter. Moreover, close dyadic friendships carry the expectation of self-disclosure and social support, which may lead to more supportive responses as an effort to solidify a friendship (von Salisch 2001).

Emotion Discussions

Parents and friends socialize adolescents’ emotions when discussing past emotional events together. In these conversations, parents and friends may help adolescents label and understand how they were feeling, process the antecedents and consequences of an emotion-eliciting event, and discuss whether adolescents demonstrated appropriate affect for the context (Morris et al. 2007). Though this is a strategy that is employed by parents and friends, there are divergent approaches to measure these behaviors and therefore a lack of consistency between the parent and friend literatures.

The parent literature has largely focused on how parents’ meta-emotion philosophies, specifically emotion coaching or emotion dismissing, influence their approach to these conversations (Gottman et al. 1996, 1997). Parents with an emotion coaching philosophy view adolescents’ emotions as valuable learning opportunities and believe that adolescents’ emotions should be expressed. Thus, they may readily engage in conversations about past emotion events and communicate the benefits of emotion expression, which may reinforce adolescents’ expressivity (Yap et al. 2008a). Parents who adopt a coaching approach also have adolescents with more proactive and detailed beliefs about emotions (Hunter et al. 2011). These conversations build adolescents’ emotion knowledge and help them learn effective coping strategies. For example, when an adolescent describes getting upset with a sibling, the parent may offer strategies to deal with anger in the future.

An emotion-dismissing philosophy reflects parents’ belief that adolescents’ emotions are dangerous and should be dampened (Gottman et al. 1996, 1997; Morris et al. 2007). Adolescents with parents who hold an emotion dismissing philosophy may not reap benefits of emotion discussions, as these parents may not enter into these conversations but instead communicate emotional avoidance. For example, an emotion-dismissing parent may change the topic or ignore their child’s prompts to discuss emotional events. When emotion dismissing parents do engage in conversations about emotional events, they may invalidate their child’s emotional experience by communicating verbally or nonverbally that emotions are overreactions or inappropriate. Research with younger children shows detrimental outcomes associated with emotion dismissing, such as poorer emotion regulation in middle childhood (Lunkenheimer et al. 2007). However, there are benefits for children’s regulatory abilities when parents evidence a diverse meta-emotion philosophy during conversations. Lunkenheimer et al. (2011) reported that when parents provided both coaching and dismissing responses during a conversation about a past negative emotional experience, children were better able to regulate their emotions. Perhaps in adolescence, continually encouraging emotions may foster excessive focus on emotional experiences, whereas some mixture of coaching and dismissing may help adolescents modulate their affect and express themselves with respect to context (Lunkenheimer et al. 2011).

Friends’ discussions of past emotional events has not been extensively examined with respect to emotion coaching and emotion dismissing, as has been done in the parent literature. The only study to date is that of Legerski et al. (2015) who observed adolescents’ conversations with close friends about past emotion-eliciting events, adapting a coaching/dismissing behavioral coding scheme used in the parent literature. This strategy is qualitatively different from friends’ responses to adolescent emotions, as these discussions may not take place “in the moment”, but rather after time has passed and the adolescent is reminiscing about a past emotion-eliciting event. Friends’ responses during these conversations were tied to adolescents’ subsequent emotional disclosures to their friends, in that adolescents receiving supportive responses (i.e., coaching) when discussing their negative emotions were more likely to disclose their negative emotions in later conversations (Legerski et al. 2015). There were no significant findings with unsupportive responses. Thus, it appears that adolescents’ self-disclosure of emotions is influenced by how friends have responded in the past.

Another phenomenon that has received considerable attention in the friend literature is co-rumination, in which friend dyads frequently discuss negative events and dwell on negative emotions (Prinstein et al. 2005; Rose 2002). One study to date has examined co-rumination in parents, though there is limited information about how co-rumination operates with both parents, or how it compares to parents’ emotion coaching/dismissing (Waller and Rose 2013). Rose et al. (2007) posit that frequently discussing problems with friends may be beneficial in that it promotes intimate friendship, but may not be an adaptive method to manage negative emotions, as these discussions may socialize a negative thinking style that leads adolescents to excessively focus on their struggles (Rose 2002; Rose et al. 2007). Participating in conversations about their friends’ struggles may heighten adolescents’ negative affect (Giletta et al. 2011). Further, this pattern may also lead adolescents to become overly reliant on emotional support from their friends to co-regulate negative emotions (Prinstein et al. 2005). Co-rumination appears to be a self-perpetuating process, as co-rumination is linked to high friendship quality, which increases the likelihood of co-rumination in best-friend dyads (Rose 2002). On the other hand, constantly seeking out friends’ emotional support may put strain on adolescents’ friendships, particularly if adolescents are oblivious about how their bids for support come across to others. For example, continual pursuit of emotional guidance may lead friends to feel ineffective and inadequate for providing support, or decrease their willingness to provide emotional support. This may lead adolescents to become socially alienated and emotionally dysregulated (Prinstein et al. 2005).

In sum, there is little overlap in the research on emotion discussions with parents and friends, as each literature examines different constructs. This may be because co-rumination does not share the same conceptual foundation as the emotion coaching and emotion dismissing constructs, as it is not grounded in an understanding of friends’ meta-emotion philosophies. However, we propose that the constructs of emotion coaching, emotion dismissing, and co-rumination can be understood with respect to one another along two dimensions: emotional engagement and guidance (Fig. 1). Co-rumination and emotion coaching are both high on emotional engagement, because both of these responses validate negative emotional experiences and encourage future emotion expression. Emotion dismissing, on the other hand, is low on emotional engagement because parents are discouraging their children’s emotions. Regarding guidance, co-rumination and emotion dismissing are similar that there is a lack of focus on finding effective coping strategies, which is a key aspect of emotion coaching. Thus, emotion coaching is high on guidance, compared to co-rumination and emotion dismissing. The divergence between emotion coaching and co-rumination may be understood in light of the nature of each relationship. With a top–down structure, parents may approach emotion-laden conversations with goals to build coping skills and mold future emotional displays. Conversely, friends have similar levels of social power and are voluntarily affiliated, thus they may provide support, rather than guidance, in attempts to strengthen their social bond.

Fig. 1
figure 1

Conceptual model of socialization responses during emotion-related discourse

Gender and Emotion Socialization

Gender plays an important role in how emotion socialization processes unfold in parent–child relationships and friendships, both in regard to adolescents’ gender and to socializers’ gender. Regarding parents, mothers are more likely to discuss and to mirror emotions with adolescents than are fathers, and mothers show more encouraging and accepting responses than fathers (Hunter et al. 2011; Klimes-Dougan et al. 2007; Zeman et al. 2012). Compared with mothers, fathers are more likely to neglect or override adolescents’ negative emotions (Klimes-Dougan et al. 2007). Fathers may be less comfortable with adolescents’ negative emotions than mothers. These differences may also be shaped by whom the adolescent approaches when distressed. Stocker et al. (2007) surmise that mothers may be more present or emotionally available than fathers, which may lead adolescents to feel more comfortable disclosing to mothers. This difference in disclosure might in turn give fathers less opportunity to practice discussing emotional topics with adolescents (Stocker et al. 2007).

There is also research to suggest that adolescents’ gender shapes how parents socialize adolescents’ emotional expressivity (Zahn-Waxler 2010). Parents discuss emotions more with daughters than sons, which may lead girls to be more comfortable discussing feelings and to seek emotional support more than boys (Fivush et al. 2000). Parents may also socialize specific emotions based on gender stereotypes for emotion expression. For example, parents may encourage sadness and discourage anger expression in daughters, whereas the inverse pattern may be present for sons (Zeman et al. 2012).

Most of the emotion socialization research on adolescent friendships focuses on same-sex friendships and shows gender differences in the nature of friendships. With the onset of adolescence, girls’ friendships become more dyadic and intimate compared to those of boys (Giletta et al. 2011; Rose and Asher 2000; Rose and Rudolph 2006). Girls are also more likely to express and discuss their emotions with friends than are boys, as well as to respond supportively to their friends’ emotions (Glick and Rose 2011; Klimes-Dougan et al. 2014; Legerski et al. 2015; Rose 2002; Rose and Rudolph 2006; Smith and Rose 2011). Boys are more punitive than girls in response to their friends’ negative emotions, perhaps because male friendships are often characterized by teasing and humor in response to emotional topics (Klimes-Dougan et al. 2014; Rose et al. 2012; Rose and Rudolph 2006). These gender differences are consistent with girls’ greater relational orientation compared with boys’, and with girls’ greater expectation of positive outcomes for disclosing their difficulties, which may lead girls to be more sensitive to other’s distress and willing to express themselves compared with boys (Prinstein et al. 2005; Rose and Rudolph 2006). Buhrmester and Furman (1987) argue that boys demonstrate intimacy in friendships through actions, such as helpful behavior towards one another, rather than emotional disclosures and conversation. Thus, it is possible that these gender differences in emotional expression in friendships may not reflect the experience of intimacy in the friendship, but rather differences in how boys and girls achieve intimacy in friendships.

Emotion Socialization and Adolescent Outcomes

Parent and friend emotion socialization has been linked to a wide range of outcomes in adolescence, such as internalizing and externalizing symptoms, as well as social competence (Stocker et al. 2007). In the childhood literature, several researchers propose a mediational model in which parents influence children’s psychological outcomes through children’s emotion regulation (i.e., Eisenberg et al. 1998; Morris et al. 2007). There has been less exploration of this model in adolescent research, though Buckholdt et al. (2014b) have some empirical findings to support this model. They reported that mothers’ invalidating responses to adolescents’ negative emotions were indirectly related to higher adolescent internalizing and externalizing symptoms through adolescent emotion dysregulation (Buckholdt et al. 2014b). Yap et al. (2010) reported similar findings, in that mother’s negative expressivity was related to higher depressive symptoms in adolescents through adolescent emotion dysregulation (Yap et al. 2010). This review will focus on evidence for direct effects of parent and friend emotion socialization on adolescents’ psychological adjustment, as there are more studies to document this link. See Table 2 for a summary of findings from this review.

Table 2 Summary of research on emotion socialization by parents and friends in relation to adolescent outcomes

Internalizing Symptoms

Parent and friend emotion socialization strategies are robustly tied to internalizing symptoms in adolescents, with some work specifying links to either anxiety or depressive disorders. We note that much of the literature on parent emotion socialization examines internalizing disorders using broad-band indices, whereas the literature on friend emotion socialization focuses on specific internalizing disorders.

Regarding modeling, parents’ negative expressivity is related to internalizing symptoms in adolescents (Luebbe and Bell 2014; Stocker et al. 2007). Parents’ low positive affect has also been linked to adolescent depression (Yap et al. 2008a). Findings in the friend literature are consistent, in that friends’ negative expressivity has also been associated with internalizing symptoms in adolescence (Dishion and Tipsord 2011; Stevens and Prinstein 2005). It has been proposed that anxious or depressed friends may model dysregulated displays of worry or sadness, which is internalized and imitated by adolescents and may put them at risk for an internalizing disorder (Brechwald and Prinstein 2011). This process could also unfold in the parent–child relationship. Witnessing a friend’s distress may also lead adolescents to engage in self-focused comforting and rumination about their role in their friend’s distress, leading to their own internalizing difficulties over time (Giletta et al. 2011, Tone and Tully 2014). Lastly, friends may also model maladaptive methods to manage their negative affect, such as engaging in self-harm, which has been linked to adolescents’ subsequent self-injurious behaviors (Brechwald and Prinstein 2011).

With respect to responses to adolescents’ emotional displays, findings for parental emotion socialization are somewhat mixed. On the one hand, when internalizing symptoms are measured as a broad-band index, research shows that parents who provide encouraging responses to their adolescents’ negative emotions have adolescents with fewer internalizing symptoms (Klimes-Dougan et al. 2007). This study was conducted with typically and atypically-developing adolescents, with a mix of internalizing and externalizing difficulties. Thus, conclusions are limited regarding how socialization responses are linked to specific diagnostic pathways. On the other hand, Schwartz et al. (2012) reported that parents of depressed adolescents were more likely than parents of typically-developing adolescents to provide encouraging responses and mirror adolescents’ sadness displays, thus potentially reinforcing dysregulated emotional displays with attention. Parental emotional support may thus, perhaps counterintuitively, promote rumination and heighten the risk for depression. Through mirroring, parents may also model depressive affect (Schwartz et al. 2012). However, Shortt et al. (2016) noted the importance of considering parent and adolescent gender, as depressed boys and girls have been found to receive different responses from their mothers and fathers. In a cross-sectional study of typically-developing adolescents, Lougheed et al. (2015) reported that parents were less likely to respond supportively to positive and negative emotions for adolescents with higher depressive symptoms. Thus, the pattern of findings is divergent when examining clinical and community samples, as well as mothers and fathers. Therefore, it appears that the impact of parental socialization responses may also vary as a function of adolescent problem status and parent gender.

In regard to unsupportive responses, when using a broad-band index of internalizing symptoms, parents’ unsupportive responses to adolescents’ negative emotions are associated with higher internalizing symptoms (Klimes-Dougan et al. 2007). Research specific to adolescent depression suggests that parents of depressed adolescents may suppress unsupportive responses to their adolescent’s dysregulated sadness in efforts to provide emotional support (Schwartz et al. 2012). However, unsupportive responses by parents may actually be beneficial, as they may signal that adolescents’ dysregulated affect is inappropriate for the context, and teach adolescents to manage their negative emotions in a situationally appropriate manner (Schwartz et al. 2012). Parents of depressed adolescents have also been shown to reciprocate their adolescents’ anger displays and provide less frequent reinforcement of positive emotions, or even respond negatively to their adolescents’ positive emotions (Yap et al. 2008a). By dampening their child’s positive emotions, parents are modeling dysphoric behaviors in a way that may reinforce depressive symptoms in teens, and might lead adolescents to replicate this behavior towards others. When their positive affect is dampened, adolescents are also less able to up-regulate their affect and thus may experience more depression (Yap et al. 2008a, b). Taken together, this socialization pattern may reinforce sadness and anger while minimizing positive emotions, which is similar to the affective profile of depression (Schwartz et al. 2012).

Very few studies have examined friends’ responses in relation to adolescent internalizing symptoms, but there are some consistencies with the parent literature. In cross-sectional study of typically-developing adolescents, Lougheed et al. (2015) reported that friends of adolescents with higher depressive symptoms were less likely to respond supportively to adolescents’ positive emotions compared with friends of adolescents with lower depressive symptoms. There are several possible interpretations for this finding, such as considering the friend’s own level of depressive symptoms. Friends may be similar in their depression symptoms such that it limits their response to positive affect. It is also possible that adolescents with higher depressive symptoms may interact with their friends in a negative or off-putting manner that decreases their supportiveness. However, Klimes-Dougan et al. (2014) report discrepant findings, noting that friends’ supportive and unsupportive responses to adolescents’ emotions were unrelated to later internalizing symptoms in a clinical sample (Klimes-Dougan et al. 2014). These two studies pose discrepant findings, which may be understood with respect to differences in sample and study design. Though more research is needed to replicate these findings, it is possible that friends’ supportive responses to negative emotions are less salient for adolescents’ internal distress because emotional support is an expectation of friendship. Further, unsupportive responses from friends may be more likely than those of parents to be interpreted as playful teasing, and thus not result in internal distress.

Lastly, socialization messages transmitted through discussions about past emotional events have been associated with adolescent internalizing symptoms. For parents, an emotion coaching approach is linked to more proactive and insightful beliefs about negative emotions in depressed, but not typically-developing, adolescents (Hunter et al. 2011). Whereas typically-developing adolescents may need less emotion coaching, depressed adolescents struggle with intense feelings of sadness, hopelessness, and sometimes anger. Thus, parental guidance about their negative emotions may help depressed adolescents accurately identify and understand these emotions, which is beneficial for later coping strategies (Hunter et al. 2011). In regard to anxiety, Suveg et al. (2008) found that parents of adolescents with an anxiety disorder discussed emotions less and discouraged their adolescent’s negative emotions more than parents with a typically-developing adolescent. Parents of anxious adolescents also expressed more negative emotions and less positive emotions than parents with a typically-developing adolescent. All in all, this pattern of socialization models negative affect and offers adolescents fewer opportunities to learn about negative emotions and adaptive coping strategies, heightening the risk for an anxiety disorder. Limited discussion of negative emotions and discouragement about expressing feelings may teach adolescents to suppress their emotional displays, which constrains their ability to process negative emotions in an adaptive manner and may result in dysregulated affect. Parental modeling of more negative affect and limited positive affect also demonstrates an affective profile similar to that of anxiety, much like the aforementioned pattern in depression (Suveg et al. 2008).

Though friends’ emotion coaching and dismissing have yet to be examined, co-rumination has been implicated as a risk factor for both anxiety and depressive symptoms in adolescents, particularly for girls. Rose et al. (2007) reported that co-rumination with friends in middle childhood was indicative of increased symptoms of anxiety and depression in early adolescence for girls, which in turn was related to increased co-rumination in later adolescence. Co-rumination was also related to increased friendship quality in female friendship dyads, which in turn predicted increased co-rumination. Although earlier depression and anxiety were related to increased co-rumination in boys, their earlier co-rumination was related only to increased friendship quality and not to increased subsequent depression and anxiety (Rose et al. 2007). Prinstein et al. (2005) reported similar findings in a longitudinal study of middle school students. They examined emotion exchanges between adolescent friends in relation to depressive symptoms over a 2-year period in a community sample. Frequently seeking out friends’ support was associated with later depressive symptoms in girls, particularly those who had lower quality friendships. Co-rumination has also been shown to interact with other risk factors to increase an adolescent’s susceptibility to depression. Bastin et al. (2015) examined co-rumination with friends as a moderator of interpersonal stress and adolescent depression. Similar to Rose et al. (2007), they found differences by adolescent gender, as interpersonal stress was associated with later depressive symptoms for girls high in co-rumination, and also for boys low in co-rumination. For girls, co-rumination and interpersonal may interact such that girls are co-ruminating with their friends about their interpersonal stressors, resulting in more distress (Bastin et al. 2015). For boys, based on the aforementioned findings from Rose et al. (2007), co-rumination may increase friendship quality and thereby buffer them from effects of interpersonal stress. Boys’ lack of co-rumination may, perhaps counterintuitively, reduce social support to cope with interpersonal stressors, increasing risk for depression.

In sum, co-rumination appears to pose more risk for girls than boys. This pattern also fits with the aforementioned gender differences in emotion socialization, such that girls’ expressivity is socialized in a manner that may increase risk for internalizing symptoms (Zahn-Waxler 2010). Girls are more likely than boys to discuss emotional difficulties and co-ruminate. Co-rumination may be socially reinforcing because adolescents receive attention from their friends during these exchanges and it potentially strengthens their friendships. However, these exchanges also reinforce excessive focus on negative emotions and depressive symptoms, thus exacerbating adolescents’ risk for psychological difficulties (Deater-Deckard 2001; Rose 2002; Rose et al. 2007). Thus, discussing problems with friends may be beneficial in that it promotes intimate friendships, but it has a trade-off in that it is also associated with increased internalizing symptoms in girls.

In one of the few studies to jointly examine co-rumination with parents and with friends, Waller and Rose (2013) found that mother–child co-rumination was indirectly related to adolescent internalizing symptoms through child–friend co-rumination (Waller and Rose 2013). Thus, it appears that mother–child co-rumination may only pose risk for internalizing symptoms if it co-occurs with friend co-rumination. Waller and Rose (2013) posit that engaging in the same communication pattern with parents and friends acts as a “double-dose” and confers risk for psychological difficulties. Parents may also be modeling communication styles that adolescents replicate with friends, thereby transmitting risk (Waller and Rose 2013). However, it is important to note that self-disclosing to friends in itself is not risky. Rather, risk comes from the way in which adolescents disclose. In their comparison of adolescents’ reports of self-disclosure and co-rumination, Waller and Rose (2013) reported that co-rumination was related to internalizing symptoms, but self-disclosure was not. Further, Rose et al. (2014) also point to the specific features of co-rumination, reporting that rehashing the details of upsetting events was not associated with internalizing symptoms. However, dwelling on negative affect associated with such events was related to increased internalizing symptoms. This suggests that discussing difficulties with a friend does not pose risk for psychopathology, but dwelling such events and the accompanying negative emotions does (Rose et al. 2014).

Not only are emotion discussions measured differently with parents and friends, but the research suggests that there are structural differences that may affect how these discussions impact adolescent outcomes. Parents and friends approach these conversations with different goals. Parents may view these conversations as a teaching opportunity, given the top–down structure of the parent–child relationship. Friends are less likely to have this perspective, because friendships are more egalitarian with a horizontal structure (von Salisch 2001). Friendships are more emotion-focused than parent–child relationships, because emotional disclosures are one method for solidifying friendships and increasing friendship quality (Rose 2002). As a result of these different approaches, friends may provide emotion-focused support that promotes co-rumination and heightens adolescent distress, whereas parents may incorporate more problem-focused support designed to ameliorate the situation and develop solutions for the future (Desjardins and Leadbeater 2011). Also, adolescents may prefer to engage in emotion-related conversations with friends because, as same-age peers, they are likely to have shared experiences or a similar perspective. However, their similarity in age also means that friends are still developing their own emotional adjustment, making them less skilled sources of emotional support than parents (Moed et al. 2015).

In sum, both parents’ and friends’ modeling of negative affect has been associated with internalizing symptoms in adolescence. Regarding supportive and unsupportive responses, parents’ unsupportive responses are associated with internalizing symptoms regardless of whether outcomes are examined with broad-band or disorder-specific indices. Research on parents’ supportive responses is mixed, with work using a broad-band index of internalizing symptoms showing benefits of supportive responses and work focusing on depression specifically showing detrimental relations of supportive responses with depressive symptoms. In the friend literature, supportive responses were less prevalent in youth with higher depressive symptoms. Lastly, there are divergent findings across socialization agents with respect to emotion discussion. Benefits of parental emotion coaching are seen for adolescents with depression and more parental emotion dismissing is seen with anxious youth compared with typically developing youth. With friends, dwelling on negative emotions (i.e., co-rumination) is associated with anxiety and depression in girls.

Externalizing Disorders

Adolescent externalizing symptoms, such as disruptive, hyperactive, or aggressive behavior, have also been linked to both parent and friend emotion socialization processes. These strategies have been examined in relation to broad-band measures of externalizing symptoms, rather than specific externalizing disorders. Regarding modeling, both parents’ and friends’ negative expressivity has been associated with externalizing symptoms in adolescents (Deater-Deckard 2001; Stocker et al. 2007). It is possible that negative expressivity by parents may elicit excessive negative affect in adolescents and lead them to act out (Stocker et al. 2007). Further, these externalizing difficulties may arise through observational learning from deviant friends, such as witnessing friends’ dysregulated affect and maladaptive coping strategies (Deater-Deckard 2001).

With respect to how parents and friends respond to adolescents’ emotional displays, supportive responses by parents and friends are linked to fewer externalizing symptoms (Klimes-Dougan et al. 2007, 2014; Shortt et al. 2010; Stocker et al. 2007). Both parents’ and friends’ supportive responses may benefit adolescents by providing support to cope with distress, thereby minimizing disruptive behavior. Regarding unsupportive responses, Klimes-Dougan et al. (2007, 2014) have demonstrated that both parents’ and friends’ unsupportive responses to negative emotions are associated with externalizing symptoms. The parent research speaks more to discrete negative emotions such as sadness and anger, whereas the friend research measures responses to broad negative emotions. Specifically, parents’ unsupportive responses to adolescent sadness were related to higher rates of externalizing symptoms in adolescents. Further, parents who punished, ignored, or mirrored their adolescent’s anger had adolescents with more externalizing symptoms (Klimes-Dougan et al. 2007). Friends’ unsupportive responses to broad negative emotions were also associated with increased externalizing symptoms (Klimes-Dougan et al. 2014). Perhaps parents’ unsupportive responses to adolescents’ negative emotions confer risk for psychopathology in a self-perpetuating manner; parents may find it taxing to handle the emotional lability of their adolescent children with psychopathology and thus respond to their negative emotions in unsupportive ways that prolong their negative affect and exacerbate adolescents’ symptoms (Klimes-Dougan et al. 2007). It is possible that friends’ unsupportive responses may lead adolescents to act out because these responses may trigger more negative affect, such as anger.

In sum, it appears that parents and friends’ modeling excessive emotionality, as well as their unsupportive responses to emotional displays, may be a risk factor for externalizing symptoms in adolescents. Thus, these strategies by each socialization agent appear to function in a similar manner in regard to adolescent externalizing symptoms. Discussion of past emotional events has not been linked to externalizing symptoms; rather, as discussed above, the majority of research studies on this strategy in adolescence have focused on internalizing outcomes.

Social Competence

As noted earlier, parental socialization strategies shape adolescents’ expectations and behavior in friendships (Collins and Laursen 2004). Research indicates that parental negative expressivity and negative reactions to adolescents’ negative emotions may pose risk for adolescents’ social competence, though there are few empirical studies to document this link in adolescents (Rudolph and Asher 2000). Adolescents who imitate their parents’ negative expressivity may struggle socially, as frequent displays of anger or sadness may interfere with their ability to have successful social interactions with same-age friends. In regard to discussion of emotions, Buckholdt et al. (2014a, b) found that maternal emotion coaching was associated with more sociability and peer respect in their adolescents, as well as a more optimistic approach to friendships and less loneliness. Moreover, maternal emotion coaching buffered the effect of poor peer relations on adolescents’ negative self-perceptions of social competence, indicating that emotion coaching operates as a protective factor (Buckholdt et al. 2014).

Despite limited empirical evidence, there is reason to believe that friend emotion socialization may also contribute to social competence in adolescents. Farley and Kim-Spoon (2014) propose that high quality friendships, such as those characterized by self-disclosure and emotional support, may promote enhanced self-regulation by modeling situationally appropriate expressivity and adaptive emotion regulation skills. Research demonstrates that boys actually reap benefits of co-rumination, in regard to increased friendship quality (Rose et al. 2007). Thus, the preliminary evidence suggests that friends’ emotion socialization may act as a risk or protective factor, depending on the dynamics of the friendship and adolescent gender. Overall, the literature on emotion socialization and adolescent social competence is in its beginning stages and is a much needed area of future study.

Future Directions

Mechanisms

More knowledge is needed regarding the mechanisms linking parent and friend emotion socialization to adolescents’ outcomes. Numerous studies indicate that parental emotion socialization is indirectly related to adolescents’ psychological outcomes through adolescents’ emotion regulation, but this model has yet to be found with friends (Morris et al. 2007). Moreover, parent and friend emotion socialization research examines different constructs, particularly in regard to emotion discussions. The parent literature largely focuses on emotion coaching and emotion dismissing, whereas co-rumination dominates the friend literature. An inclusive methodological approach is needed to delineate mechanisms and allow for a true comparison.

Contextual Differences

Researchers should also continue to consider the contexts within which emotion socialization processes occur, including gender and developmental stage. Regarding gender, future studies should examine why co-rumination operates as a stronger risk factor for girls than boys (Rose et al. 2007). Both frequency of emotional support and the nature of co-rumination may differ between boys and girls. Perhaps, compared with boys, girls engage not only in more frequent co-rumination but also more elaborate or complex co-rumination, potentially leading to more and greater likelihood of internalizing a negative thinking style. Further, it will be important to examine trajectories of change in emotion socialization throughout adolescence, due to important developmental and social transitions that may impact parent and friend emotion socialization processes.

Socialization Agents

Future studies should examine both socialization agents in the same study, to determine the relative influence of parents and friends throughout adolescence and into emerging adulthood. It would be informative to understand whether parents and friends remain equally influential on adolescents’ adjustment, or if one socialization agent begins to exert more influence on a particular socio-emotional domain with time as adolescents continue to individuate from their parents and rely on friends (Rudolph and Asher 2000). Alternatively, emotion socialization from one relationship may buffer the effects of emotion socialization from another relationship, as has been demonstrated with mothers and fathers (Lunkenheimer et al. 2007), but not yet examined with parents and friends.

Socialization agents other than parents and friends are also worthy of consideration, such as grandparents, siblings, and teachers (Zeman et al. 2012). As Zeman et al. (2012) note, there is also a lack of research with non-traditional families, such as single parents and sexual minority [lesbian/gay/bisexual/transgender (LGBT)] parents. Single parents may experience more caregiving strain, which may influence how they approach their children’s emotions, and the match or mismatch between parent and child gender in single-parent families is also worthy of consideration. Gender differences in emotion socialization may also be important to examine with LGBT parents, as gendered approaches to children’s emotions, or gender stereotyping broadly speaking, may be less evident (Sutfin et al. 2008). Lastly, romantic partners should also be examined with respect to their influence on adolescent psychological adjustment, as they are likely the next emerging source of social support in adolescence and emerging adulthood (Brechwald and Prinstein 2011).

Discrete Emotions

The literature would also benefit from more emotion-specific research, rather than examining broad negative emotion (Brand and Klimes-Dougan 2010; Zahn-Waxler 2010). Supporting an adolescent’s sadness holds a different meaning than supporting anger, and also translates to different socioemotional outcomes for adolescents. Moreover, very few studies examine socialization of positive emotions and its relevance to adolescent psychopathology, particularly in the friend literature (Suveg et al. 2008; Yap et al. 2008a, b). For example, modeling positive emotions and supporting positive affect in adolescents may operate as protective factors for psychopathology. Investigating discrete emotions is important for positive emotions as well as negative; for example, inculcating expressions of pride compared with contentment may have different effects on socioemotional competence. Curvilinear relations may also be possible, in that moderate displays of positive emotion may be associated with adaptive outcomes, with low and high (i.e., euphoric) displays of positive emotion posing risk for maladjustment (Morris et al. 2007).

Conclusions

Both parents and friends engage in emotion socialization during adolescence, a developmental stage characterized by individuation and fraught with new socio-emotional challenges that impact adolescent outcomes (Kuczynski and Parkin 2007; Nickerson and Nagle 2005). Both socialization agents teach adolescents about emotions through their own expressivity, responding to adolescents’ emotional displays, and discussing past emotional events. It appears that parents and friends are largely similar in how their socialization contributes to adolescent psychological adjustment, with the exception of discussing past emotional events. Despite this apparent consistency, addressing both parent and friend socialization is important for a more nuanced understanding of emotion socialization and its relation to adolescent psychopathology. In particular, gendered patterns of adolescent development are highlighted in regard to how adolescent boys and girls use their social relationships during a developmental period in which they are at increased risk for psychopathology. The salience of friends in adolescence also highlights the importance of a developmentally-informed approach to understanding emotion socialization, as friends take on more of an emotional support role in adolescence and thus have the power to shape adolescent emotional development. Gaps in the literature include a need for consistent measurement of how parents and friends discuss past emotional events, a relative lack of focus on adolescent social competence, and a relative lack of focus on socialization of positive emotions. Future work concentrating on mechanisms, contextual effects, varying socialization agents, and discrete emotions will advance the field.