Making Sense and Moving On: The Potential for Individual and Interpersonal Growth Following Emerging Adult Breakups

Abstract

This study assessed the key aspects of romantic relationship dissolution in emerging adulthood as predictors of future mental health and romantic qualities. It utilized a longitudinal, multiinformant, multimethod study of 160 participants with their romantic partners and close friends followed from ages 20–25, with a breakup assessed at age 22. Having control over initiating a breakup at age 22 predicted relative increases in peer-rated internalizing symptoms and autonomy-undermining interactions with a new partner at ages 23–25. Having a greater understanding of the reasons for a breakup predicted lower self-reported internalizing symptoms and relative decreases in partner-reported romantic conflict as well as relative increases in self-reported relationship satisfaction and peer-rated intimate relationship competence at ages 23–25. Predictions remained after accounting for numerous potential confounds including age 20–22 baseline relationship quality, social competence, internalizing symptoms, and gender. Implications for understanding links between breakup characteristics on emerging adult psychological and relationship functioning are discussed.

One of the primary tasks of late adolescence and emerging adulthood is learning how to manage the growing salience and intensity of romantic relationships. By the end of adolescence, the majority of teenagers report having been in at least one exclusive relationship and the prevalence and importance of romantic relationships increases during the transition into emerging adulthood (Carver, Joyner, & Udry, 2003; Furman & Shomaker, 2008; Giordano, Manning, Longmore, & Flanigan, 2009; Reis, Lin, Bennett, & Nezlek, 1993). The ability to develop and maintain intimacy within close relationships is widely recognized as a primary developmental task of early adulthood (Barry, Madsen, Nelson, Carroll, & Badger, 2009; Clark & Beck, 2010; Erikson, 1982). Partners’ needs and goals within these romantic relationships steadily change during the transition from adolescence into adulthood (Furman & Wehner, 1997), highlighting the developmental nature of engaging in such partnerships.

Recent societal changes have delayed the age at which emerging adults commit to marriage, increasing the time spent, exploring romantic involvement, and experiencing breakups (Arnett, 1998, 2000). Nearly 40% of emerging adults report one or more breakups over the course of a 20-month period (Rhoades, Kamp Dush, Atkins, Stanley, & Markman, 2011). Especially common in emerging adulthood (i.e., the period between late adolescence and early adulthood) is the initiation and then dissolution of a series of relationships of varying degrees of commitment and intimacy (Cohen, Kasen, Chen, Hartmark, & Gordon, 2003). The romantic relationships and dissolutions of emerging adults may have significant consequences for both concurrent functioning and later relationships. It is imperative to assess the significance of such early relationships to highlight the greater role that their development and dissolutions play in broader romantic and individual development. The goal of this study was to examine whether the characteristics of an emerging adult’s most recent breakup predicted future psychological health, romantic conflict management, and romantic satisfaction and competence.

Distress Following Dissolution

During emerging adulthood, individuals are striving to blend their career, personal, and romantic lives together resulting in a delayed entrance into long-term stable relationships (Shulman & Connolly, 2013). Given the increasing importance adolescents and emerging adults are placing on their romantic experiences as they age (Giordano et al., 2009; Seiffge-Krenke, 2003), it is not surprising that many psychosocial and psychological outcomes are linked to romantic functioning. Indeed, successfully navigating intimate relationships during emerging adulthood has been more closely tied to well-being than other developmental goals such as achieving financial independence or educational achievements, avoiding substance use, or maintaining close friendships (Schulenberg, Bryant, & O’Malley, 2004). Yet, despite the ubiquity of short-term relationships and breakups during this period, much research points to significant negative ramifications of romantic dissolutions including depression, posttraumatic stress, anxiety, substance abuse, poor self-esteem and self-confidence, low life satisfaction, and poor physical health (Chung et al., 2002; Fine & Harvey, 2006; Fleming, White, Oesterle, Haggerty, & Catalano, 2010; Lewandowski, Aron, Bassis, & Kunak, 2006; Monroe, Rohde, Seeley, & Lewinsohn, 1999; Rhoades et al., 2011). Sbarra and Emery (2005) found that emerging adults who recently experienced a romantic breakup reported more anger and sadness than those in a committed stable relationship. Several characteristics such as commitment, satisfaction, effort in initiating the relationship, relationship duration, time before finding a new partner, and a fearful attachment style are all related to distress at the time of the romantic dissolution (Davis, Shaver, & Vernon, 2003; Locker, McIntosh, Hackney, Wilson, & Wiegand, 2010; Sprecher, Felmlee, Metts, Fehr, & Vanni, 1998; Sweeney, 2002).

Growth Following Dissolution

Given the pervasiveness of dissolutions during this developmental stage, it is surprising that much less research has considered whether there is the potential for growth, rather than solely distress, following a breakup. A romantic breakup has been cited as one of the worst events of traumatic experiences (Frazier & Hurliman, 2001); yet a growing field assessing posttraumatic growth has emerged that has yet to be applied to romantic relationship dissolutions. According to Tedeschi and Calhoun (1996), posttraumatic growth occurs when individuals bounce back from a traumatic experience to a higher level of functioning than pretrauma. Those who exhibit benefit finding, a strategy in response to a stressor associated with posttraumatic growth, tend to have higher levels of well-being and lower levels of depression (Helgeson, Reynolds, & Tomich, 2006).

Tashiro and Frazier (2003) were pioneers in applying the idea of stress-related growth to romantic dissolution to assess both growth and distress as independent outcomes. They found individuals can report positive changes following a breakup such as feeling more self-confident, independent, stronger, and more emotionally stable. Participants most commonly reported individual and relational positive changes they can make to improve well-being and romantic relationships in the future. Lewandowski and Bizzoco (2007) found that positive emotions and personal growth can follow romantic dissolution, especially in the context of dissolution of poor relationships characterized by few opportunities for self-growth. In a qualitative study of college student dissolution, Hebert and Popadiuk (2008) found that all 11 participants reported at least one positive change, that the most important change was positive, and had overwhelmingly more positive than negative changes. These findings suggest that exploring romantic experiences, including breakups, during emerging adulthood may contribute to better conflict management and communication skills that may in turn lead to greater romantic competence and satisfaction in future relationships. Researchers have not directly tested the hypothesis that growth following dissolutions may benefit future relationship functioning. Longitudinal studies with more robust sample sizes that assess long-term positive outcomes of dissolutions are needed to augment these scarce findings.

Methodological Limitations of Prior Dissolution Research

Although there is growing evidence to suggest that emerging adults report both distress and perceived growth, the vast majority of research assesses past relationships only via self-report—suggesting that reported growth may stem from positive illusions or biases (Frazier & Kaler, 2006). Studies using only self-report data may also provide spurious results in that individuals who are reporting positive current functioning post-dissolution may be biased toward recalling the past breakup as less painful than it actually was at the time. Alternatively, those who are depressed or anxious are likely biased toward recalling past breakups as uncontrollable. In addition, studies on postdissolution growth tend to be cross-sectional and retrospective—again being subject to biases and neglecting long-term changes (Tashiro & Frazier, 2003). Therefore, these few studies lend support to the idea that growth is possible following breaking up, but the potential for positive adjustment following dissolution has been left largely unexplored via multiple reporters and repeated assessments of individual and interpersonal functioning over time that may more accurately capture adjustment postdissolution.

Attachment Theory: A Framework for Understanding Dissolution

Given that dissolutions are normal experiences during emerging adulthood, the specific context of a breakup may dictate whether it leads to positive versus negative adjustment. Prior researchers have framed romantic dissolution using Bowlby’s (1980) attachment theory. An important extension of Bowlby’s attachment theory of parent–child relationships is that adults develop similar patterns of attachment to romantic partners (Hazan & Shaver, 1994; Hazan & Zeifman, 1999). The attachment features of a secure base, proximity seeking, and safe haven are transferred from parents to romantic partners such that the partner becomes a primary attachment figure (Hazan & Shaver, 1994). Prior findings point to similar underlying interpretations and responses to separation from an attachment figure whether it is a parent or partner. Specifically, separation results in protesting, seeking the attachment figure, and symptoms of depression while evoking senses of abandonment and betrayal (Johnson & Whiffen, 1999; Kobak, 1999). Thus, a breakup can be interpreted as a loss or disrupture in attachment resulting in significant mental distress (Davis et al., 2003; Hansen & Shireman, 1986; Hazan & Shaver, 1987; Johnson, Makinen, & Millikin, 2001).

While Hazan and Shaver (1994) suggested a 2-year criterion after which partners become attachment figures, more recent literature has suggested that this process often happens more quickly (Heffernan, Fraley, Vicary, & Brumbaugh, 2012). Thus, emerging adults engaging in even short-term relationships may perceive their partner as an attachment figure. Dissolution of an attached romantic relationship may have greater consequences than casual dating relationships, similar to the greater distress children experience when separated from a primary attachment figure as compared to a casual connection (Bowlby, 1980).

Initiator Status and Dissolution Adjustment

Individuals who perceive greater sense of loss and little control over a divorce, for example, tend to suffer more negative consequences. Thus, one of the commonly studied questions surrounding romantic dissolutions is the question of whether initiators (i.e., those in control of the breakup) versus noninitiators experience different degrees of distress. Prior research has supported the idea that individuals who did not initiate a breakup tend to suffer more than those who initiated the breakup, at least in the short term. For example, Perilloux and Buss (2008) found that both men and women who did not initiate a recent breakup reported more depression, rumination, and lower self-esteem. In contrast, individuals initiating a breakup appear to suffer less in terms of psychological distress and negative emotions (Sprecher, 1994; Sprecher et al., 1998). Although there is some evidence suggesting the importance of certain breakup variables on concurrent mental health, there is a scarcity of research on the trajectories of change in mental distress (i.e., depression and anxiety) over time.

Other evidence suggests a more benign picture for those not initiating a breakup. Rhoades, Kamp Dush, Atkins, Stanley, and Markman (2011) found no differences in psychological distress based on initiator status or desire to end the relationship. Several studies have found no difference in distress between initiators and noninitiators following dissolution (Koenig Kellas & Manusov, 2003; Locker et al., 2010; Tashiro & Frazier, 2003). In addition, there is some evidence that initiators may experience more negative reactions from others than noninitiators. For example, Perilloux and Buss (2008) found that those who initiated a breakup reported feeling perceived by others as cruel. Perhaps initiators are more likely to feel guilty or experience negative backlash from friends following the decision to end a relationship. These poor reactions from others may lead to significant distress. These conflicting findings raise the possibility that initiating a breakup may in different ways be both beneficial and costly to the individual, although this possibility has not yet been explored.

Gender Differences

Prior research on romantic relationship has suggested that relationship transitions may have different implications for females versus males. Much of the literature focuses on gender differences in response to marriage and divorce finding that males benefit more from marriage and suffer more from divorce than women (Belle, 1987; Bloom, White, & Asher, 1979; Gove, 1973). Women also tend to report greater postdivorce growth than men (Colburn, Lin, & Moore, 1992; Kitson, 1992) and greater stress-related growth than men in general (Tedeschi & Clahoun, 1996). Taylor and colleagues (2000) proposed that females experience less distress following stressful events as compared to men due to their coping behaviors. Females are more likely to rely on social support and reach out to others to get through tough times while males are quick to have an amped up fight-or-flight response (Taylor et al., 2000). In general, females pay attention to close relationships more than males and they tend to experience stronger benefits from these relationships for their overall well-being as compared to men (Cross & Madsen, 1997; Saphire-Bernstein & Taylor, 2013). The ability to seek support from others, which is a strategy more commonly used by and more beneficial for women, may be a key mechanism of managing postdissolution stress contributing to gender differences.

However, more recent findings suggest that there are few sex differences in the experience of marriage and divorce than earlier results (Simon, 2002; Strohschein, McDonough, Monette, & Shawo, 2005). For example, Williams (2003) found no gender differences in well-being based on marital status, relationship transitions, or relationship qualities. In a meta-analysis, Kiecolt-Glaser and Netwon (2001) generally found mixed results for gender differences in the link between marriage and physical health perhaps reflecting societal changes of gender dynamics where men and women contribute similarly to marriages. Less research has considered gender differences in nonmarital relationships and dissolutions. Evidence suggests that men suffer more than women following a non-marital breakup, especially when men did not initiate the breakup (Helgeson, 1994; Hill, Rubin, & Peplau, 1976), although other research finds that females report greater breakup distress (Field, Diego, Pelaez, Deeds, & Delgado, 2009). Despite unclear findings on gender differences in post-dissolution distress, women report greater nonmarital postdissolution growth compared to men (Sprecher, 1994; Tashiro & Frazier, 2003). Interestingly, Helgeson (1994) found no gender differences for positive growth following dissolution for those initiating a breakup. Thus, gender will be an important consideration in assessing the context of nonmarital dissolution on future individual and interpersonal functioning.

Autonomy and Understanding in Dissolution Adjustment

Distress and adjustment following romantic dissolution have also been linked to the ability to process and understand a breakup. In a college sample, individuals who reported being certain of the reasons for a recent breakup experienced fewer problems adjusting after the dissolution (Yildirim & Demir, 2015). Although individuals may report positive changes following a dissolution, they are more likely to report problematic responses if they are unable to identify the reasons for the breakup (Tashiro & Frazier, 2003). If an individual is unsure why a relationship ended, they may ruminate on the negative feelings following a breakup and carry this distress into future relationships. Indeed, hypothetical thinking regarding reasons for dissolution is related to poor adjustment (Saffrey & Ehrenberg, 2007). Further, those lacking autonomy within their romantic relationships tend to have poor conflict management skills and lower satisfaction (Knee, Lonsbary, Canevello, & Patrick, 2005). Perceiving a lack of knowledge or control over a breakup may be more common for those in such relationship characterized by low levels of autonomy. Autonomy has been identified as an important characteristic of well-being in different contexts including in the workplace, residentially, financially, and in various types of social relationships (Allen, Hauser, Bell, & O’Connor, 1994; Kins, Beyers, Soenens, & Vansteenkiste, 2009; La Guardia, Ryan, Couchman, & Deci, 2000; O’Connor, Allen, Bell, & Hauser, 1996; Thompson & Prottas, 2006). Autonomy and connection are important for understanding close relationships, including reasons for dissolution (Baxter, 1988). One partner’s autonomous support toward the other provides mental health and relational benefits, suggesting the reciprocal nature of autonomy (Deci, Schwartz, Sheinman, & Ryan, 1981; Ryan & Connell, 1989).

In this article, we focus on the joint, dyadic autonomy that exists within couples. Prior research points to mutual influence in romantic relationships generally (Kelley & Thibaut, 1978) and specifically within romantic relationships (Knee et al., 2005). In a qualitative study of partner’s reports of autonomy within the relationship as potential roles in dissolution, the authors proposed that assessing both sides of the autonomy–connection link from both partners provides a deeper understanding than relying on one partner’s autonomous behaviors or perceptions (Sahlstein & Dun, 2008). Prior research on peer autonomy suggests dyadic autonomy (both giving and receiving support) is a robust predictor of relationship outcomes (Deci, LaGuardia, Moller, Scheiner, & Ryan, 2006). Further, their results indicate that there is unique variance in relationship quality and well-being due to autonomy at the relationship level (Deci et al., 2006). Similar to the idea of the reciprocal nature of close friendships, considering giving and receiving autonomy support within romantic relationships will be important to assess as a relational, dyadic characteristic of emerging adult relationships. During emerging adulthood, intimate relationships become central to one’s identity—a focus on dyadic autonomy versus individual autonomy seems especially important at this stage. Perhaps the mutual influence in close relationships is heightened in romantic relationships especially during emerging adulthood when maintaining intimate relationships is considered an essential developmental task.

If unsure why a relationship ended, it seems plausible that uncertainty, poor conflict negotiation strategies, and poor abilities to give and receive autonomous support can carry forward into future relationships. On the other hand, individuals who are able to gain closure and make sense of a relationship ending may experience less mental distress following a breakup. The ability to generate a coherent story regarding a recent breakup is positively correlated with overall adjustment (Koenig Kellas & Manusov, 2003). Meaning-focused coping strategies such as identifying positives and benefit finding following a stressful event are linked to positive emotions and less distress in general traumatic event literature (Davis, Nolen-Hoeksema, & Larson, 1998; Folkman & Moskowitz, 2000; Park, 2010). Samios, Henson, and Simpson (2014) found that benefit finding in the context of nonmarital dissolution was related to better adjustment including positive affect, depression, and overall satisfaction with life. Whether meaning-making, closure, or benefit finding then go on to predict functioning in future romantic relationships has not, however, been explored thus far.

In sum, research to date raises the question of whether individuals differ in their long-term outcomes following romantic dissolution depending upon whether or not they initiated the breakup and how well they were able to understand the breakup. Limitations to this area of research include a focus on divorce and on negative associations with breaking up and use of retrospective self-report data. How dissolutions of emerging adult unmarried relationships influence adjustment prospectively is less clear. In addition, no research to date has examined the ways in which dissolutions may predict future relationship functioning in key areas of relationship quality such as conflict, intimate self-confidence, or autonomy via support. By addressing specific future relationship qualities using multiple reporters and observational data, this study addresses previously neglected or biased findings.

Overview of the Research and Hypotheses

This study seeks to understand the context of emerging adult romantic breakups on later psychological and relationship functioning while addressing each of these methodological limitations in research to date. Specifically, we assess whether initiating a breakup or the level of understanding of its cause can predict both mental health (i.e., anxiety and depression) and future relationship functioning (i.e., conflict, satisfaction, autonomy, and intimate competence) in future romantic relationships. Because engaging in short-term relationships is relatively normative during emerging adulthood, we seek to assess the conditions under which there is potential for growth following dissolution of these relationships during this stage of romantic exploration.

We used self-reports, peer reports, partner reports, and observational data in a diverse community sample to examine the implications of initiator status and understanding of the reasons for a breakup during emerging adulthood using a diverse community sample followed from the ages of 20–25, with a breakup assessed at age 22. We examined the following hypotheses:

• Hypothesis 1

  Initiator status in a recent breakup will predict (a) relative changes in mental health symptoms over time, (b) future abilities to handle conflict within a romantic relationship, and (c) relative changes in romantic competence and satisfaction.

• Hypothesis 2

  Confidence in understanding the reasons for a recent breakup will predict (a) relative decreases in mental health symptoms over time, (b) improved future abilities to handle conflict within a romantic relationship, and (c) relative increases in romantic competence and satisfaction.

Method

Participants

The data were drawn from a larger longitudinal study of adolescent development focusing on social relationships with parents, peers, and romantic partners. Participants originally included 184 seventh and eighth graders (86 male and 98 female) initially recruited from a public middle school in the Southeastern United States with both urban and suburban populations. Students were first recruited via an initial mailing to all parents of students in the school and those interested in the study (presented to potential participants as the “Kids, Lives, Families, and Friends” study) were contacted by telephone. The larger data set has been used to address other topics in prior manuscripts, but none have addressed the goals of the current study.¹ Informed assent for the adolescents along with informed consent from the parents was obtained before each interview session. Informed assent and consent for close friends and romantic partners of the target participant were obtained as age appropriate as well. Participants were mailed packets of questionnaires with return envelopes so that they could complete the measures on their own time and then return them to the lab.

Follow-up data were obtained annually at ages 22–25 (M = 22.80, standard deviation [SD] = 0.96; M = 23.78, SD = 0.97; M = 24.65, SD = 0.96; M = 25.69, SD = 0.99). Demographics are reported from the first wave of data collection. The sample was diverse in terms of race, ethnicity, and socioeconomic status: 58% identified as Caucasian, 29% as African American, 8% as mixed race/ethnicity, and 5% as other. Adolescents’ parents reported a median family income around US$50,000. In addition, 63% of the teens’ mothers were married, 14.4% divorced, 9.8% single, and 13.2% reported other (separated, widowed, or living with partner). At age 22, 71% of the participants were employed and 80% of those working were employed full time. In addition, 40% of participants were current students.

Participants were also asked annually to provide contact information (i.e., phone number) for a close friend who knew them best. We approached the peers via phone calls to invite them to participate in the study with the consent of the initial participant. The same informed consent process was utilized for peers. At participant ages 22–25, participant’s close friends participated annually. Friend age was not available at the first wave of data collection. However, peers were on average aged 26 across each follow-up wave of data collection when the participant was of age 23–25 (M = 25.70, SD = 4.95; M = 25.98, SD = 4.96; M = 26.47, SD = 5.06). In addition, participants reported knowing their close friend for an average of 9 months at age 22 (M = 8.69, SD = 5.95) to 11 months at age 25 (M = 10.58, SD = 6.90). Approximately half of our participants (50.41%) had the same close friend participate across multiple waves of data collection. The majority of close friends were the same gender as the target participant (93%). A total of 149 friends participated at Wave 1; a total of 135, 148, and 128 friends participated at each annual follow-up. Peer reports at age 22 (Wave 1) provide the baseline level of our variables of interest.

If participants reported involvement in a significant heterosexual or homosexual romantic relationship (i.e., 2 months or longer) at any point during ages 20–22 and again at ages 23–25, the romantic partner was eligible to participate. Again, participants provided contact information for their partner so that we contacted the romantic partners directly to obtain consent. A similar duration criterion has been used in prior research (Collins, Welsh, & Furman, 2009; Connolly & Johnson, 1996; Hand & Furman, 2009; Madsen & Collins, 2011). Therefore, participants and their romantic partner completed assessments only once during ages 20–22 and 23–25. The data gathered during ages 20–22 provide the baseline assessment of relationship quality. A total of 80 participants (1 homosexual and 79 heterosexual) endorsed being in a romantic relationship of at least 2 months in our original sample, with 61 couples participating in Wave 1 and 73 couples participating in Wave 2. Romantic partners were on average aged 22 and 25 (age: M = 22.37, SD = 3.63; M = 24.83, SD = 4.15) at Waves 1 and 2, respectively, and relationship duration was approximately 22 months (M = 21.85, SD = 19.89; M = 21.78, SD = 20.16). It is important to note that of the 160 participants, 146 reported experiencing a recent breakup and only 22 participants (16%) had the same romantic partner at both waves of data collection and were included in the analyses, which highlight the romantic exploration typical during emerging adulthood.

Attrition Analyses

Attrition analyses examined missing data for each type of data available at baseline. Results indicate that those participants who did not complete all assessments across time points (i.e., participants at age 22 and at ages 23–25) were male (N = 8, p < .001). To best address any potential biases due to attrition in longitudinal analyses or due to missing data within waves, we use full information maximum likelihood (FIML) methods for all regression analyses, including all variables that were linked to future missing data (i.e., where data were not missing completely at random). This approach has been found to provide the least biased estimates when all available data are used for longitudinal analyses (Arbuckle, 1996). Although the original sample in our study was 184, a total of 160 participants (71 males and 89 females) and 147 close friends (64 males, 83 females) provided baseline data at age 22. A total of 158 participants (69 males and 89 females), 146 close friends (63 males and 83 females), and 87 couples participated at follow-up assessments. Therefore, the entire original sample of 160 participants was used in all subsequent analyses.

Procedure

At ages 20–22, the target participant came into the lab with their romantic partner if the participant reported being in a relationship lasting 2 months or longer. The target participant and their romantic partner each filled out a series of questionnaires about their relationship. The data collected at ages 20–22 serve as the baseline assessment for relationship qualities, which we include as covariates in all analyses. Including baseline relationship functioning as a covariate allows for the analysis of relative change over time. At age 22, the target participant completed additional questionnaires and nominated an individual as their closest peer. The closest peer also completed questionnaires at this time. The peer data collected at age 22 serve as the baseline assessment for relationship qualities and are included as a covariate in all analyses.

Finally, at ages 23–25, target participants and their closest friend completed questionnaires annually. In addition, the romantic partner and participant came into the lab to complete questionnaires about their relationship and participate in an observed behavioral task once during this 3-year period. All interviews took place in private offices in a university academic building (see Table 1 for detailed timeline of the data collection procedure including measures, total number of respondents, and descriptive statistics).

Table 1.

Descriptive Statistics and Participation Rates for All Variables of Interest.

	Wave 1 (Age 22)				Wave 2 (Age 23–25)
Measure Type	Measure	N	Mean	SD	Measure	N	Mean	SD
Predictor measures	Initiator status	147	2.17	0.80
			36 (24%) “My partner”
			50 (34%) “Both of us”
			61 (42%) “I did”
	Understanding a breakup	145	3.28	0.93
			9 (6%) Don’t know
			20 (14%) Little bit
			37 (26%) Quite a lot
			79 (54%) Exactly
Outcome measures (self-report)	State-Trait Anxiety Inventory	158	36.45	9.45	State-Trait Anxiety Inventorya	157	36.95	9.56
						158	36.62	10.11
						158	35.81	9.89
	Romantic satisfaction	128	29.98	4.22	Romantic satisfaction	173	15.06	3.05
Outcome measures (peer report)	Adult Behavior Checklist	147	5.19	5.60	Adult Behavior Checklista	135	5.79	6.31
						146	6.03	6.92
						128	5.99	6.36
	Close friendship competence	137	12.03	2.64	Intimate relationship competencea	125	12.93	2.68
						138	12.53	2.58
						124	12.57	2.74
Outcome measures (romantic partner report)	Conflict Tactics Scale	118	6.13	6.81	Conflict Tactics Scale Autonomy	101	6.44	7.36
						87	0.60	0.51

Note. SD = standard deviation.

^aThe State-Trait Anxiety Inventory, Adult Behavior Checklist, and intimate relationship competence measures were assessed annually at ages 23, 24, and 25. Therefore, we provided the descriptive statistics for each time point for these measures.

Measures

Characteristics of a Breakup

Initiating a breakup (age 22)

We assessed who initiated the most recent breakup prior to age 22 by asking a single item as follows: “Who decided to end the relationship?” with 1 = my boy/girlfriend did, 2 = both of us, and 3 = I did, such that higher scores indicate increasing reported control over initiating a breakup. Participants completed this item at Wave 1.

Understanding a breakup (age 22)

In addition, we assessed the reasons for the termination of their most recently ended romantic relationship prior to the age 22 assessment via a single item as follows: “How confident are you that you really know the reasons for your breakup?” where 1 = I don’t know at all, 2 = I know a little bit, 3 = I know quite a lot, and 4 = I know exactly why. Higher scores indicate more confidence in understanding the reasons for the breakup.

Mental Health—Internalizing Symptoms

Self-Report: State-Trait Anxiety Inventory (ages 22 and 23–25)

Internalizing symptoms were assessed annually using self-report of the State-Trait Anxiety Inventory (Speilberger, Gorsuch & Lushene, 1970) at ages 22–25. The 20-item trait subscale of the State-Trait Anxiety Inventory assesses overall stable individual differences in anxiety. Items are scored on a 4-point Likert-type scale where 1 = almost never, 2 = sometimes, 3 = often, and 4 = almost always. The overall anxiety score at age 22 provides the baseline level of anxiety. Cronbach’s αs are provided for each measure used for this study (except for our observational data) as a measurement of reliability for the proposed construct (Tavakol & Dennick, 2011). The internal consistency for the baseline anxiety measure has a Cronbach’s α of .91 and for ages 23, 24, and 25 have Cronbach’s αs of .92, .92, and.92, respectively.

Peer report: Adult Behavior Checklist (Ages 22 and 23–25)

Close friends of each participant completed the Adult Behavior Checklist (Achenbach & Rescorla, 2003) annually at participant age 22 (baseline) and again at ages 23–25. The Adult Behavior Checklist is a 122-item measure with internalizing, externalizing, substance use, attention problems, thought problems, and other problems subscales. Items were scored on a 3-point Likert-type scale where 0 = not true, 1 = somewhat or sometimes true, and 2 = very true or often true. The Internalizing subscale score on the Adult Behavior Checklist is composed of 32 items assessing anxiety, depression, withdrawal, and somatic complaints. Higher scores indicate greater internalizing symptoms. The Internalizing subscale score at age 22 provides the baseline level of mental distress and with internal consistency of a Cronbach’s α of .83. The Internalizing sub-scale scores collected annually at ages 23–25 will be used in the growth analyses to assess overall level and trajectories of change by age 25. The internal consistencies for ages 23, 24, and 25 have Cronbach’s αs of .89, .90, and .89, respectively.

Romantic Conflict Management

Romantic partner report: Conflict Tactics Scale (ages 22 and 23–25)

Conflict within the romantic relationship was assessed via romantic partner report using an adapted version of the Conflict Tactics Scale (Straus, 1979). The Conflict Tactics Scale is an 80-item measure assessing severity and frequency of conflict and conflict management techniques. We used the Psychological Aggression subscale, which includes 6 items assessing the extent to which the target participant is verbally abusive via conflict behaviors of insulting, shouting, stomping out of the room, or doing something to spite the partner. If participants were in a romantic relationship longer than 2 months at ages 20–22 or 23–25, the partner completed the Conflict Tactics Scale about the current romantic relationship. At ages 20–22, the psychological aggression internal consistency has a Cronbach’s α of .86 and serves as the baseline level of romantic conflict. At ages 23–25, the Psychological Aggression subscale has a Cronbach’s α of .88 and provides the outcome measure of romantic conflict.

Autonomy-Undermining behavior (ages 23–25)

Behaviors that undermine autonomy were assessed using an 8-min recorded observational behavior task in which couples were asked to discuss a relationship issue on which they had reported disagreement. Two trained coders used the Autonomy and Relatedness Coding System (Allen et al., 2000) to code the recorded interactions. At ages 23–25, each couple participated in the behavior task once. The Autonomy-Undermining Behavior Scale includes behaviors such as avoiding discussing the disagreement by giving into their partner or by distracting the partner away from the topic, overpersonalizing the disagreement by using personal examples as reasons, placing blame or guilt on the partner, calling upon the opinion of an outside party, or falsely characterizing the partner’s behavior in an exaggerated way, and pressuring the partner to agree by using sarcasm, condescension, frustrated or impatient body language, or repeating themselves. We took the average of both the individual and his or her partner’s negative autonomy scores to create a dyadic sum score. Interrater reliability for negative autonomy for both participant and partner has an intraclass coefficient of .85 and .81, respectively. Higher scores for overall dyadic negative autonomy indicate more behaviors that undermine autonomy within the relationship.

Romantic Relationship Functioning

Romantic relationship satisfaction (ages 20–22 and 23–25)

Between ages 20 and 22, participants in a relationship completed the Relationship Assessment Scale (Hendrick, Dicke, & Hendrick, 1998), which is a 7-item scale assessing overall relationship satisfaction including fulfilling expectations and needs within the relationship. All items are scored on a 5-point Likert-type scale where higher scores indicate greater levels of overall satisfaction. Example items include “How well does your partner meet your needs?” “In general, how satisfied are you with your relationship?” and “How good is your relationship compared to most?” The Relationship Assessment Scale total sum score serves as the baseline relationship satisfaction score. Internal consistency of the romantic satisfaction sum score has a Cronbach’s α of .84.

Between ages 23 and 25, romantic relationship satisfaction was assessed via self-report on a 5-item scale assessing overall romantic relationship satisfaction that was created for this study. Compared to those reporting using the baseline measure, all participants regardless of their relationship status completed this 5-item measure. Our team created this measure to capture satisfaction about general romantic life, not specifically tied to a current partner, and found all 5 items to have an acceptable internal consistency for determining current romantic life satisfaction. Items include “I am very satisfied with my current romantic life,” “I spend a lot of time worrying about my current romantic life,” “I would like to make significant changes to the current circumstances of my romantic life,” “I am content with the state of my romantic life,” and “The current state of my romantic life causes me a great deal of stress.” All items are scored on a 4-point Likert-type scale where 1 = strongly disagree, 2 = disagree, 3 = agree, and 4 = strongly agree, and Items 2, 3, and 5 are reverse coded. Higher scores indicate more satisfaction. Internal consistency of the outcome measure for romantic satisfaction score has a Cronbach’s α of .72.

Peer report: Intimate relationship competence (ages 23–25)

At ages 23–25, intimate relationship competence was assessed using close friend report of the 4-item Intimate Relationship subscale of the adult self-perception profile about the target participant (Harter, 1995). Items include assessing the ability to develop and establish intimate relationships and to openly communicate with others. Example items include “Some people do not find it easy to develop intimate relationships BUT other people have the ability to develop intimate relationships” and “Some people find it hard to establish intimate relationships BUT other people do not have difficulty establishing intimate relationships.” Higher scores indicate greater abilities to form and maintain close, meaningful relationships with a romantic partner. Close friends completed the Harter annually when the target participant was aged between 23 and 25 years. Each of these three waves of data collection will be included in our growth analyses. Internal consistency for ages 23, 24, and 25 has Cronbach’s αs of .77, .71, and .76, respectively.

Peer report: Close friendship competence (age 22)

Close friendship competence was assessed using the friend report of the 4-item Close Friendship subscale of the adult self-perception for adolescents about the target participant (Harter, 1988) at age 22. Higher scores reflect better developed abilities to form and maintain close and fulfilling relationships with friends including high levels of communication skills. Items include assessing the degree to which their friend has a difficult time making close friends, their knowledge of how to find a close friend, and use of trust and self-disclosure in making friends. Example items include “Some people do not know what it takes to develop a close friendship with a peer BUT other people do know what to do to form a close friendship with a peer” and “Some people find it hard to make friends they can really trust BUT other people are able to make close friends they can really trust.” Internal consistency of the Close Friendship subscale has a Cronbach’s α of .75.

Analytic Plan

Analyses were conducted to examine the associations between initiator status and confidence in understanding the reasons for a breakup with internalizing symptoms, romantic conflict management, and romantic satisfaction and competence.

Growth analyses

For those measures that were collected annually from the ages of 22–25 (i.e., self-report anxiety, close friend-report anxiety, and close friend-report intimate competence), we used an Multilevel Modeling (MLM) approach to linear growth modeling using MPLUS (Version 6; Muthén & Muthén, 2010). Thus, for those repeated measures, we examined whether initiating a breakup or understanding the reasons for a breakup predicted either intercepts or trajectories of internalizing symptoms (Hypotheses 1a and 2a) or intimate competence (romantic competence in Hypotheses 1c and 2c).

Regression analyses

For those measures that were collected at two time points (i.e., romantic partner-report conflict, observed autonomy task, and romantic life satisfaction), we used a regression model. All regression analyses were completed in SAS Version 9.4 using an FIML approach to handle any missing data.

The analytic approach of predicting the future level of a variable, such as romantic conflict, while accounting for predictions from initial levels of that variable, yields one marker of residualized change in that variable (Cohen & Cohen, 1983). Further, considering baseline levels of future behavior as a covariate eliminates the spurious effect whereby observed predictions are simply a result of cross-sectional associations among variables that are stable over time. Regression analyses were used to explore Hypotheses 1b and 2b and romantic satisfaction in Hypotheses 1c and 2c.

Results

Preliminary Analyses

Means and SDs for all variables of interest examined in the study are presented in Table 1. The number of responses and percentages in all of the response categories for the predictor variables are also provided in Table 1. Table 2 presents the simple univariate correlations among all variables of interest in the study. These correlations suggest many relationships between breakup characteristics and later reports of mental distress and relationship qualities that will be explored more fully in regression and growth curve analyses below. The t tests were conducted to examine potential gender differences in all variables of interest. Results indicate that partners of female participants report greater conflict in emerging adult romantic relationships, t(116) = 2.14, p = .03, Cohen’s d = .40, and peers of female participants report greater internalizing symptoms in emerging adulthood, t(145) = 2.19, p = .03, Cohen’s d = .36. Thus, gender was also included as a covariate in all analyses below. Potential moderating effects of gender were assessed by creating interaction terms between gender and each predictor variable. Each interaction term was included along with the predictors of interest in a separate regression analysis predicting the outcomes of interest. None of the interaction terms were significantly related to any of the outcomes assessed below at levels greater than would be expected by chance.

Table 2.

Univariate Statistics and Intercorrelations Among All Variables.

Variables	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19
1. Initiate breakup (age 22)	.12	−.05	−.03	.01	−.17	−.04	.03	−.07	.08	.11	.10	.22*	.14	−.09	−.01	.05	−.05	.31**
2. Understand reasons of breakup (age 22)	—	—.23**	.04	.04	.17	.08	−.23**	−.31***	−.19*	.12	.04	−.25**	−.25*	.29***	.21*	.12	.25**	.03
3. Internalizing (self-report; age 22)	—	—	.16	−.07	.04	−.17	.69***	.67***	.60***	.20*	.17*	.26**	.10	−.26**	−.09	−.17*	−.28**	−.03
4. Internalizing (peer report; age 22)	—	—	—	.14	.16	−.38***	.00	−.02	−.04	.41***	.34***	.09	−.09	.06	−.16	−.11	−.03	−.02
5. Conflict (partner report; age 20–22)	—	—	—	—	.13	.06	−.09	−.10	−.03	.37***	.16	.00	.27*	.14	−.19	.01	.11	.26*
6. Satisfaction (self-report; age 20–22)	—	—	—	—	—	.13	−.04	.05	.01	.06	−.06	−.01	−.03	.17	.07	.00	.20	−.03
7. Social competence (peer report; age 22)	—	—	—	—	—	—	−.08	−.09	.00	−.25*	−.12	−.05	−.12	.18*	.40***	.25**	.30**	.01
8. Internalizing (self-report; age 23)	—	—	—	—	—	—	—	.79***	.65***	.05	.13	.21*	.23*	−.50***	.00	−.15	−.19*	.00
9. Internalizing (self-report; age 24)	—	—	—	—	—	—	—	—	.74***	.01	.17*	.27**	.24*	−.52***	−.03	−.18*	−.21*	.02
10. Internalizing (self-report; age 25)	—	—	—	—	—	—	—	—	—	.00	.18*	.24**	.20	−.41***	.06	−.10	−.13	.09
11. Internalizing (peer report; age 23)	—	—	—	—	—	—	—	—	—	—	.35***	.17	.13	−.04	−.33***	−.14	−.06	.23
12. Internalizing (peer report; age 24)	—	—	—	—	—	—	—	—	—	—	—	.42***	.18	−.14	−.17	−.37***	−.26**	.10
13. Internalizing (peer report; age 25)	—	—	—	—	—	—	—	—	—	—	—	—	.12	−.25**	−.25*	−.37***	−.39***	−.01
14. Conflict (partner report; age 23–25)	—	—	—	—	—	—	—	—	—	—	—	—	—	−.29**	.00	−.17	−.14	.33**
15. Satisfaction (self-report; age 23–25)	—	—	—	—	—	—	—	—	—	—	—	—	—	—	.20*	.15	.25**	−.08
16. Intimate competence (peer report; age 23)	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	.31**	.53***	−.04
17. Intimate competence (peer report; age 24)	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	.62***	−.14
18. Intimate competence (peer report; age 25)	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	−.05
19. Negative autonomy (age 23–25)	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—

^*p ≤ .05.

^**p < .01.

^***p < .001.

In addition, family income was included as a covariate in all analyses below. Socioeconomic status has been linked to romantic functioning in prior research (Conger, Cui, Bryant, & Elder, 2001). We considered potential moderating effects of family income by creating interaction terms between income and each predictor variable. Using an identical process as with gender interactions, none of the interaction terms were significantly related to any of the outcomes assessed.

Primary Analyses

Hypothesis 1: Initiator status in a recent breakup will predict

• (a)Relative changes in mental health symptoms over time: We examined how initiating a breakup was related to initial levels and growth trajectories over time of internalizing symptoms from ages 22 to 25. Standardized predictor variables were used in these analyses to maximize ease of comparison of effects of predictors and interpretability of the model (Biesanz, Deeb-Sossa, Papadakis, Bollen, & Curran, 2004).

In models assessing self-reported internalizing symptoms, results indicated that initiator status was not significantly related to the overall level (intercept) of anxiety or to trajectories of change (slope) in anxiety. In conditional models assessing peer-reported internalizing symptoms, results revealed that initiator status was not significantly related to the overall level but was related to trajectories of change of peer-reported internalizing behaviors from ages 22 to 25 (β = .61, p = .03). The results shown in Table 3 indicated that the perception of having initiated a recent breakup predicted smaller decreases in the initiator’s internalizing symptoms as compared to those who did not initiate, as reported by their close peer; all participants internalizing symptoms decreased over time although not at a significant trend. This means that emerging adults who initiated a breakup at age 22 were perceived by close friends as experiencing smaller decreases in internalizing symptoms by age 25 as compared to those who did not initiate a breakup, but did not report any significant changes in their own perception of their internalizing symptoms.

Table 3.

Initiator Status of a Breakup Predicting Intercept and Linear Growth in Internalizing Symptoms and Intimate Competence.

	Internalizing Symptoms (Self-Report)			Internalizing Symptoms (Peer Report)			Intimate Competence (Peer Report)
Level of Model	Coefficient	SE	T	Coefficient	SE	T	Coefficient	SE	T
Intercept	36.64***	4.25	8.62***	1.75	2.58	.68	13.11***	1.26	10.38***
Linear growth rate	−0.85	1.36	−0.62	−0.84	1.20	−0.70	−0.42	0.61	−0.69
	Internalizing symptoms (self-report)			Intimate competence (peer report)			Intimate competence (peer report)
Intercept
Family income	0.14	0.41	0.34	0.01	0.25	0.03	0.20	0.13	1.61
Gender	0.47	1.55	0.31	2.43*	0.95	2.56*	0.34	0.46	0.75
Initiator status	−0.46	0.96	−0.48	−0.16	0.59	0.79	0.12	0.28	0.42
Linear change
Family income	−0.13	0.13	−1.01	0.07	0.12	0.64	−0.07	0.06	−1.16
Gender	0.40	0.50	0.80	−0.40	0.45	−0.90	0.03	0.22	0.12
Initiator status	0.37	0.31	1.18	0.61*	0.28	2.21*	−0.09	0.14	−0.62

^*p < .05.

^***p < .001.

• (b)Future abilities to handle conflict within a romantic relationship: We examined whether initiating a breakup predicted relative change in partner reports of participant’s conflict within future romantic relationships over time. Results suggest that initiating a breakup was not significantly related to the partner’s report of the emerging adult’s aggression within the relationship. We also examined the association between initiator status and future dyadic romantic relationship behaviors that undermine autonomy. Results, as shown in Table 4, suggest that control over initiating a breakup at age 22 was related to an increase in behaviors that undermine autonomy within the romantic relationship at ages 23–25 (β = .29, p = .004). This means that emerging adults who initiated a breakup were later involved in romantic relationships in which both members of the dyad engaged in behaviors that undermined autonomous functioning.
• (c)Relative changes in romantic competence and satisfaction: We examined whether initiating a breakup predicted relative change in self-reports of relationship satisfaction and in peer reports of intimate relationship competence over time. Because intimate relationship competence was not assessed in earlier waves, we used the friend report of the Close Friendship subscale of the Adult Self-Perception for Adolescents measure (Messer & Harter, 1986) at age 22 for the baseline measure of intimate competence. Regression analyses results suggest that initiating a breakup did not significantly predict any changes in self-reported satisfaction over time. Results from the conditional growth curve analysis for peer-reported intimate competence indicated that initiator status was not significantly related to the individual’s intimate competence at age 22 or to trajectories of change in intimate competence from ages 22 to 25.

Table 4.

Regression Analysis Predicting Adult Romantic Conflict, Negative Autonomy, and Relationship Satisfaction From Emerging Adult Initiator Status.

	Conflict (Partner; Age 23–25)			Undermining Autonomy (Observed; Age 23–25)			Relationship Satisfaction (Self; Age 23–25)
Outcome	β	ΔR2	Total R2	β	ΔR2	Total R2	β	ΔR2	Total R2
Step 1
Gender	.02		.01	.09		.01	−.01		.01
Step 2
Baseline functioning (age 22)	.28**	.07**	.08*	—	—	—	.17*	.02*	.03*
Step 3
Initiator status (age 22)	.12	.00	.08	.29**	.08**	.09*	−.07	.07	.03

Note. All βs reported are the final β’s for the analysis.

^*p ≤ .05.

^**p < .01.

Hypothesis 2: Confidence in understanding the reasons for a recent breakup will predict

• (a)Relative decreases in mental health symptoms over time: We next examined how confidence in understanding the reasons for a breakup was related to growth trajectories of internalizing symptoms from ages 22 to 25. In conditional growth models for self-reported internalizing symptoms, results, as shown in Table 5, indicated that understanding the reasons for a breakup at age 22 was significantly related to the overall level (intercept) of internalizing symptoms (β = −2.60, p = .001) but not to trajectories of change in internalizing symptoms from ages 22 to 25. For conditional models of peer-reported internalizing symptoms, results revealed that understanding the reasons for a breakup was not related to overall level or trajectories of change in peer reports of internalizing symptoms from ages 22 to 25. This means that emerging adults who reported greater understanding of the reasons for a recent breakup were experiencing less anxiety at age 22 compared to those who were less confident in understanding why a breakup occurred.
• (b)Improved future abilities to handle conflict within a romantic relationship: We examined whether understanding a breakup predicted relative change in partner reports of participant’s conflict within future romantic relationships over time. Results, as shown in Table 6, suggest that confidence in understanding the reasons for a breakup predicted relative decreases in aggression over time based on partner reports (β = −.30, p = .003). This suggests that the ability to make sense of a breakup by reporting confidence in understanding the reasons for the dissolution is related to less conflict in later romantic relationships with a new partner.

Table 6.

Regression Analysis Predicting Adult Romantic Conflict, Negative Autonomy, and Relationship Satisfaction from Emerging Adult Understanding Reasons for a Breakup.

Outcome	Conflict (Partner; Age 23–25)			Undermining Autonomy (Observed; Age 23–25)			Relationship Satisfaction (Self; Age 23–25)
	β	ΔR2	Total R2	β	ΔR2	Total R2	β	ΔR2	Total R2
	Step 1
Gender	.02		.01	.09		.01	−.01		.01
Step 2
Baseline functioning (Age 22)	.28**	.07**	.08*	—	—	—	.17*	.02*	.03
Step 3
Understand reasons for breakup (Age 22)	−.30**	.11**	.19**	.03	.00	.01	.28***	.07*	.10**

Note. All βs reported are the final βs for the analysis.

^*p ≤ .05.

^**p < .01.

^***p < .001.

We also examined whether confidence in understanding the reasons for a breakup at age 22 was associated with dyadic romantic relationship behaviors that undermine autonomy at ages 23–25. Results, as shown in Table 6, suggest that understanding the reasons for a breakup was not related to later behaviors that undermine autonomy within new relationships.

• (c)Relative increases in romantic competence and satisfaction: We examined whether confidence in understanding the reasons for a breakup predicted relative change in self-reports of relationship satisfaction and peer reports of intimate competence at ages 23–25. Regression analyses results are shown in Table 6 and suggest that confidence in understanding the reasons for a breakup predicted relative increases in relationship satisfaction over time (β = .28, p < .001). This means that emerging adults who were able to make sense of the reasoning for a romantic dissolution experienced an increase in relationship satisfaction in later relationships. In addition, results from the conditional growth curve analysis for peer-reported intimate competence indicated that confidence in understanding the reasons for a breakup was not significantly related to the overall level of intimate competence but was related to the trajectories of change in intimate competence from age 22 to 25 (β = .24, p = .04), as shown in Table 5. This suggests that friends of those initiating a breakup at age 22 perceived a boost in romantic competence over time.

Table 5.

Understanding Reasons for a Breakup Predicting Intercept and Linear Growth in Internalizing Symptoms and Intimate Competence.

	Internalizing Symptoms (Self-Report)			Internalizing Symptoms (Peer Report)			Intimate Competence (Peer Report)
Level of Model	Coefficient	SE	T	Coefficient	SE	T	Coefficient	SE	T
Intercept	41.97***	4.47	9.38***	0.06	2.74	0.02	13.02***	1.34	9.69***
Linear growth rate	0.38	1.51	0.25	1.34	1.40	0.96	−0.99	0.63	−1.56
	Internalizing symptoms (self-report)			Intimate competence (peer report)			Intimate competence (peer report)
Intercept
Family income	0.36	0.41	0.90	0.03	0.25	0.11	0.26*	0.13	2.07*
Gender	1.03	1.54	0.67	2.39*	0.96	2.51*	0.33	0.46	0.71
Understanding reasons	−2.60***	0.81	−3.20***	0.34	0.50	0.68	0.01	0.24	0.01
Linear change
Family income	−0.12	0.13	−0.88	0.07	0.13	0.56	−0.12*	0.06	−2.05*
Gender	0.31	0.51	0.61	0.03	0.50	0.07	−0.03	0.23	−0.15
Understanding reasons	−0.11	0.27	−0.40	−0.44	0.26	−1.73	0.24*	0.11	2.10*

^*p < .05.

^***p < .001.

Discussion

This study examined the extent to which an emerging adult’s future levels of psychological health, romantic conflict management, and romantic satisfaction and competence could be predicted from the characteristics of their most recent breakup as of age 22. Findings indicate that being in control of initiating a breakup predicted trajectories of relative increases in peer-reported internalizing symptoms in comparison to the sample as a whole, which displayed a slight average decrease in internalizing symptoms overtime. In addition, being in control of initiating a breakup predicted relative increases in formation of future romantic relationships in which both parties behave so as to undermine one another’s autonomy. In contrast, understanding the reasons for a breakup was related to a lower overall level of self-reported internalizing symptoms and to trajectories of relative increase in peer-reported intimate relationship competence as compared to the rest of the sample. Those who understood the reasons for a breakup had friends who perceived their intimate competence as improving more overtime as compared to those who did not understand dissolution reasons. Greater confidence in understanding the reasons for dissolution also predicted relative increases in relationship satisfaction and decreases in future romantic partner-reported conflict by ages 23–25.

Emerging adulthood is an important exploratory period for romantic experiences in which the end goal is to foster an intimate long-term relationship. Dissolution appears to be a normative experience during this developmental stage, yet prior findings have typically considered breaking up as a negative, problematic event. Because romantic partners become attachment figures in many relationships, how individuals cope and adjust to disruption in attachment (i.e., dissolution) may determine whether distress or growth results. Our study assessed the qualities of a romantic relationship dissolution (i.e., initiator status and understanding the reasons for the dissolution) using multi-informant reports, observational data, and longitudinal data adding to the burgeoning field of growth following romantic dissolution in unmarried emerging adult couples.

It is possible that emerging adults who are able to come to terms with romantic dissolution by understanding why and how a breakup occurred are learning more than those individuals who are left in the dark as to where the relationship went awry. Similar to the arguments made by Tashiro and Frazier (2003) in their analyses of postdissolution growth and stress, our findings suggest that individuals who struggle to identify what contributed to a breakup suffer more negative consequences. Our findings go beyond their findings of general growth and add to their argument by providing specific romantic relationship qualities that may benefit from understanding a breakup.

Cutting ties with a romantic attachment figure can be a positive or negative experience depending on the context. Prior research has indicated that attachment transference from parents to romantic partners increases as relationship length increases (Fraley & Davis, 1997; Hazan & Shaver, 1994; Hazan & Zeifman, 1994; Trinke & Bartholomew, 1997). Others have replicated the findings of attachment transference to partners but found that this happens earlier in relationships than the previous 2-year criterion with attachment features present in both casual dating relationships and appearing within the first year of dating (Eastwick & Finkel, 2008; Heffernan et al., 2012). In our study, nearly 50% of the participants reported being in a relationship with their ex-partner for longer than 1 year (16% dated for 7–12 months, 19% for 4–6 months, and 19% for 3 months or less). Thus, it is likely that many participants had formed attachment relationships with their partner leading to increased likelihood of perceiving the dissolution as an attachment loss and experiencing greater levels of distress.

Perceiving a lack of control or understanding has been linked to more negative consequences following a fracture in a romantic attachment (Hazan & Shaver, 1987, 1994; Sprecher, 1994). Therefore, certain aspects of romantic dissolution may ultimately help promote healthy adjustment despite the immediate negative impacts a split has on well-being. During a time where romantic exploration is normative, being able to make sense of the dissolutions and romantic transitions may help emerging adults learn and grow in their romantic life, fostering improved romantic experiences in the future. Our findings suggest that emerging adults who were able to make sense of and understand a breakup experienced less conflict with their future romantic partner and more satisfaction within romantic relationships in emerging adulthood. In addition, their close friends perceived an increase in intimate competence for those who were able to understand the reasons for a recent breakup. This suggests that perhaps their peers saw our participants as growing in their relationship knowledge as they explore different relationships. Although negative consequences of dissolutions have been identified in prior research, our findings shed light on the potential to grow in future relationship competence and skills if emerging adults can make sense of their breakups. There has been little empirical attention and unclear findings regarding gender differences in nonmarital postdissolution distress and growth. Our findings suggest that in emerging adult breakups, men and women tend to experience similar levels of distress and growth. We found little or no support for gender differences based on breakup characteristics on future reports of internalizing symptoms, conflict management, romantic satisfaction, and competence. Perhaps it is societally normal to experience a cycle of relationships and dissolutions during emerging adulthood as compared to in adulthood—therefore, gender differences emerge later when the importance of a stable, committed relationship is more strongly linked to health and well-being. The link between high-quality romantic relationships and healthy adjustment and mental health becomes stronger with age (Segrin, Powell, Givertz, & Brackin, 2003; Simon & Barrett, 2010). Breaking up in adulthood may be a more severe traumatic event as compared to in emerging adulthood, leading to a wider range of reactions across genders.

Not only did we expect the context of dissolutions to impact future relationships, but we also believed confidence in understanding the reasons for a romantic dissolution would improve mental health. Psychological health has been strongly tied to social relationships (Cutrona & Russell, 1987; Diener & Seligman, 2002; Kawachi & Berkman, 2001; Taylor, Doane, & Eisenberg, 2013). In adulthood, cohabitation and ultimately marriage are associated with higher levels of subjective well-being as compared to individuals who are only dating casually or rarely (Kamp Dush & Amato, 2005). Specifically, relationship qualities such as conflict, companionship, emotional security, and overall satisfaction have emerged as predictors of overall well-being (Demir, 2008, 2010). Given the importance of the developmental task of forming long-term intimate relationships in emerging adulthood, we believe the effects of romantic experiences on mental health may be potentially stronger at this stage as compared to earlier in adolescence or in later adulthood. Prior research has pointed to an increase in mental distress immediately following a breakup (e.g., increases in depression, anxiety, loneliness, and substance abuse; Fine & Harvey, 2006; Fleming et al., 2010; Monroe et al., 1999; Rhoades et al., 2011). However, it is important to distinguish between short-term distress (i.e., at the end of a relationship, people may already be unhappy, depressed, or in conflict with their partner) and long-term implications of breakups to partially eliminate the short-term confounding effects of cooccurring dissolution and distress. Our longitudinal analyses take into consideration how anxiety levels change over a 4-year period. Individuals who reported understanding the breakup also report decreasing levels of anxiety symptoms over time.

Because dissolutions are normative during this stage, perhaps being able to make sense of them leads to greater closure, less rumination, and improved mental health over time. Understanding a breakup may lead individuals to find benefits of dissolution such as learning what they can change in future relationships or what qualities they desire in a new partner. Similar to the benefit finding research by Helgeson, Reynolds, and Tomich (2006), understanding why a breakup happened may lead to better well-being and less mental distress as individuals search for the positives in gaining relationship closure. Our findings add to those of Samios and colleagues (2014) in that searching for meaning, benefits, or positives of dissolution may be linked to better adjustment. Distress is considered necessary to fuel a search for meaning (Park, 2010; Samios, Henson, & Simpson, 2014). Perhaps those who were not distressed at the time of dissolution are less likely to search for meaning to more fully understand the reasons for a breakup. Importantly, our analyses included the baseline level of distress to account for this potential confound.

These results highlight the important role of creating a coherent narrative following a romantic breakup. The ability to make sense of a difficult situation (i.e., breakdown in attachment) can increase the potential to learn and grow from prior relationship experiences, leading to more satisfied, stable, and fulfilling romantic relationships in the future. If individuals are unsure what went wrong in a relationship, they likely will not recognize their own problematic behaviors or remain blind to what needs and desires they want in future relationships. Understanding the reasons for a breakup, gaining a sense of closure, and creating a coherent story may help individuals better adjust to dissolution with less mental distress and interpersonal conflicts and more intimate competence. Our findings highlight the adaptive nature of experiencing the normative cycle of breakups during emerging adulthood. Nonmarital dissolutions provide emerging adults with time to reflect on successes and failures in past relationships, leading to self-growth, optimism, and adjustment in approaching future relationships (Weber, 1998). When emerging adults are romantically exploring their options, it appears that understanding why a relationship ended can help bolster future relationship functioning.

These findings add to the growing literature on initiator status on postdissolution adjustment. Initiator status was not associated with any changes in relationship satisfaction, romantic conflict, or intimate competence. However, our results indicate that individuals who initiated a breakup may experience greater levels of peer-rated psychological distress and observed difficulties in future romantic relationships. Friends of initiators reported an increase in initiators’ internalizing symptoms by age 25. However, the initiators were not themselves reporting more internalizing symptoms. This is consistent with prior findings that initiators may experience greater social difficulties or poorer perceptions by others following dissolution as compared to noninitiators (Perilloux & Buss, 2008). The inclusion of multireporters allowed us to more fully explore the impact of dissolution on the individual’s future functioning, finding interesting discrepancies in self versus other report that may not have been detectable in prior studies.

Being in control of initiating a breakup was also associated with observed difficulties in future romantic relationships. Individuals who had more control in initiating a breakup were more likely to be in future relationships characterized by higher levels of dyadic behaviors that undermine autonomy. We chose to focus on dyadic, rather than individual, autonomy because we consider balancing autonomy is a dance between both partners where each party plays an essential role that builds off the other. Being able to identify joint autonomy may be most useful in capturing this construct of a supportive relationship, as we consider autonomy to be a characteristic of the relationship, not the individual. Prior findings point to the mutuality of romantic relationships and begin highlighting the role of relationship-level autonomy rather than individual-level autonomy in relationship satisfaction and well-being (Deci et al., 2006; Kelley & Thibaut, 1978; Knee et al., 2005; Sahlstein & Dun, 2008). One possibility for why initiators experience poor autonomy is that those individuals initiating breakups experience guilt or remorse over their decisions and carry this negativity into their next relationship. It is also possible that initiators were prone to leave a relationship rather than work with their partner on resolving problems and difficulties or even that they were simply very poor at conflict negotiation and this continued to be reflected in future relationships. Without learning how to navigate conflict and develop healthy communication with a partner, the individual may learn maladaptive strategies such as pressuring their partner to agree, giving into their partner without discussion, becoming defensive, or using sarcasm or condescension during disagreements. Future research should explore other relationship characteristics of individuals initiating breakups to determine the extent to which initiator status is associated with future patterns of conflict and communication strategies.

There are several limitations to these findings that also warrant discussion. First, this study assessed whether romantic dissolution characteristics predicted relative changes in psychological health and important romantic relationship qualities over time during emerging adulthood. It is important to note that we assessed how the dissolution characteristics predicted relative change (i.e., shift in rank order) on romantic functioning in future relationships, not mean-level change. Given the normative nature of dissolutions in emerging adulthood, it will be enlightening to assess overall mean-level change for our sample over time, as we continue data collection across multiple waves in the future. Fortunately, for the mental distress and peer-reported variables, we assessed overall level and trajectories of change allowing a fuller exploration of how the breakup characteristics influence change over time. Even with our ability to assess trajectories of change in several of our variables of interest, longitudinal research using repeated measurement of key constructs is not sufficient to support causal claims. This study helps identify the potential role of control over and understanding of a breakup as playing a significant part of the development of mental health and relationship functioning; however, it cannot evaluate causal hypotheses or completely assess mean-level change (yet), given our data collection discrepancies across measures. Nonetheless, we build on existing literature assessing growth and distress following breakups in important methodological ways. Our study addressed limitations in prior research in that we utilized behavioral measures, multireporters, and longitudinal data as compared to cross-sectional, retrospective, self-report data (Tashiro & Frazier, 2003).

With few exceptions, research on qualities of emerging adults’ romantic relationships and in particular the consequences of romantic dissolution has been based on small sample sizes, and the current study is limited in this manner as well. We only included those adolescents who reported a breakup within the past year, and we only included their current romantic partner if they were willing to participate and were dating for more than 2 months. Because our analyses utilized observational data for emerging adult couples as well as partner and peer reports, we believe the rich data gathered at least partially offset the potential limits of the sample size. Another limitation is risk of potential volunteer bias. A study on close relationships may attract more females than males or may attract participants who have fairly positive relationships overall (Demir, Haynes, Orthel-Clark, & Ozen, 2016).

Another limitation of romantic relationship literature is the quickly changing definitions of a relationship and relationship status especially during this developmental stage. Furman and Collins (2007) point out the difficulty in operationalizing a romantic relationship and propose a 2- to 4-month duration criteria. Although we use the 2-month rule based on prior similar research and attempts to ensure that this was a somewhat significant relationship, we may not have accurately captured all definitions of romantic relationships. Because our participants may define their relationship using different words than what we presented, it is possible for multiple interpretations of our measures. Specifically, we asked participants to what extent they felt they understood the reasons for their most recent breakup. Emerging adults may endorse lack of confidence for many reasons: uncertainty why they left a partner, confusion over why a partner left them, inability to make sense of a dissolution, and so on. Differences in understanding the question may impact responses, and future research may ask more specific questions regarding the context for a breakup including reasons, motives, and so on.

Finally, a methodological limitation exists for our relationship satisfaction measures. At baseline, participants completed a measure focused on their specific current romantic relationship. However, we are more interested in romantic life in general, as we recognize that during emerging adulthood, one’s attitude about romantic life might be different than that toward your specific partner. We recognize that our two measures of romantic satisfaction are different, but we hope to capture a more global aspect of intimate satisfaction in emerging adulthood, given the measures we have in our longitudinal study.

Despite these potential limitations, these findings highlight important associations between perceived control and understanding of a romantic dissolution and later psychological and romantic functioning. They add to a limited but growing body of literature indicating the potential for positive growth following romantic dissolutions and, more importantly, suggest the conditions under which such growth is most likely to occur. At a time when emerging adults are increasingly devoting more time to exploring romantic interests prior to marriage (Arnett, 2000) and an emphasis on developing intimate relationships takes hold (Cohen et al., 2003; Shulman & Connolly, 2013), it is important to understand how certain aspects of dissolution can lead to healthier adjustment and recovery from what is typically considered a negative event. We suggest that there are certain characteristics of a breakup that can contribute to improved mental health and relationship functioning in future relationships, suggesting the normative developmental task of romantic exploration allows emerging adults to ultimately meet the goal of a long-term committed relationship successfully.

Future research should explore the role of gaining closure in a romantic relationship for future relationship functioning. Perhaps it is not only understanding the reasons for a breakup but the actual nature of that understanding that makes a difference in how one processes the dissolution. Especially, during emerging adulthood where the task of forming a long-term intimate relationship becomes salient, a deeper understanding of how individuals can grow from dissolutions may change our approach to empirical assessment of romantic experiences. Although there is a wealth of research regarding marital status and divorce, research on nonmarital relationships, especially the positives of dissolution, is relatively scarce. Our findings point to several specific qualities of nonmarital dissolution that may have beneficial outcomes on mental and interpersonal health.

Relatedly, there are few relationship education programs for nonmarital couples as compared to marriage counseling. By understanding more specifically which romantic qualities within emerging adulthood are important predictors of future relationship quality, we can begin to inform relationship-focused interventions unique to this population. An emphasis on dissolution, coping, and romantic exploration may be warranted within a program for nonmarital relationships in a way that it is not for marital relationship interventions. Most relationship education programs focus on improving likelihood of maintaining relationship stability for married couples (i.e., Prevention and Relationship Education Programs; Markman, Stanley, & Blumberg, 2010; or Within Our Reach; Stanley et al., 2006) and nonmarried couples (Davila et al., 2017). One exception is Relationship U, an adapted version of Within Our Reach for college students, which is designed for all individuals regardless of relationship status (Fincham, Stanley, & Rhoades, 2010). Yet, except for one session on helping students “learn how to break up effectively,” this program neglects post-dissolution adjustment or growth. Another exception is the Romantic Competence program, which includes discussions of postdissolution emotion regulation and learning (Davila et al., 2017). Given that romantic exploration is typical for emerging adults, relationship education may benefit from including material that aids individuals in making sense of a breakup, identifies the positives, and normalizes the process of dissolutions. By empowering individuals in this way, they may be more likely to reap the benefits associated with successfully navigating breakups (Weber, 1998). The focus of nonmarital relationship education as compared to marital programs may shift from maintenance and growth with a partner, to recovery and growth outside of a relationship.

It will be important to more closely identify the dynamic and individual processes occurring during a breakup and its recovery that lead to different relationship and psychological trajectories. Given growing evidence linking qualities of dissolutions to future relationships, it is critical to consider the complete range of romantic experiences, including breakups, in evaluating the long-term links of romantic relationship qualities and future interpersonal and psychological functioning.

Biographies

Jessica Kansky is a third-year doctoral student in clinical psychology at the University of Virginia and received her bachelor of arts from the University of Pennsylvania. Her research focuses on predictors and outcomes of romantic experiences from adolescence into adulthood. Specifically, she is interested in the role of romantic relationships in optimal interpersonal and individual development and overall well-being.

Joseph P. Allen is the Hugh P. Kelley professor of psychology and education at the University of Virginia. His research focuses on the predictors and long-term outcomes of social development processes from adolescence into adulthood. He also develops and examines socially focused interventions for adolescents designed to improve long-term academic and mental health outcomes.