PSYC questions

Journal of Family Psychology Pathways Linking Early Socioeconomic Adversity to Diverging Profiles of Romantic Relationship Dissolution in Young Adulthood Dayoung Bae and Kandauda A. S. Wickrama Online First Publication, November 5, 2018. http://dx.doi.org/10.1037/fam0000465

CITATION Bae, D., & Wickrama, K. A. S. (2018, November 5). Pathways Linking Early Socioeconomic Adversity to Diverging Profiles of Romantic Relationship Dissolution in Young Adulthood. Journal of Family Psychology. Advance online publication. http://dx.doi.org/10.1037/fam0000465

Pathways Linking Early Socioeconomic Adversity to Diverging Profiles of Romantic Relationship Dissolution in Young Adulthood

Dayoung Bae and Kandauda A. S. Wickrama University of Georgia

Although studies on romantic relationships are abundant, cumulative experiences in intimate relationship dissolution (i.e., dissolution of cohabitation and marriage) during young adulthood is not yet completely understood. Using a nationally representative sample of 9,275 young adults, we investigated heteroge- neity in timing and frequency of relationship dissolution during young adulthood, as well as its developmental precursors to dissolution. Results indicated four distinct relationship-dissolution classes that ranged from those who maintained stable romantic relationships to those who experienced multiple cohabitation dissolutions and divorces from ages 18 to 30 years. Early socioeconomic adversity predicted relationship-dissolution trajectories directly and indirectly through psychosociodemographic mechanisms in adolescence, including disrupted transitions to adulthood, conflict in dating relationships, and low levels of future orientation. Our findings highlight the heterogeneous romantic relationship trajectories of young adults and support the importance of the person-centered approach in elucidating developmental pathways underlying the longitudinal transitions in romantic relationships.

Keywords: young adulthood, romantic relationships, relationship dissolution, early socioeconomic ad- versity

Supplemental materials: http://dx.doi.org/10.1037/fam0000465.supp

Parallel to entering romantic relationships at earlier ages and delaying marriage, studies have shown that young adults typically experience more than 10 years of relationship formation and dissolution before marriage (Meier & Allen, 2009). Coupled with this trend, cohabitation has become increasingly common in the United States (Manning, 2013), with young adults in constant movement, stepping on and off the carousel of intimate partner- ships (Cherlin, 2009). As a result, more young adults are likely to experience multiple dissolutions in coliving relationships, includ- ing cohabitation dissolutions and divorces (Amato, 2010). These

societal trends indicate increased variability in romantic relation- ships among contemporary young adults.

Although the dissolution of stressful, conflictual romantic rela- tionships may improve individuals’ well-being (Wheaton, 1990), previous studies have indicated that failure to maintain a commit- ted intimate relationship is generally thought to predict physical and emotional distress as a result of disruptions in day-to-day routines and in interactions within social networks (Amato, 2010; Umberson, Crosnoe, & Reczek, 2010). Many previous studies on relationship dissolution have provided valuable insights into the experiences of single dissolution events at particular points in time (Beaujouan, 2012; Wu & Schimmele, 2005). However, those studies have failed to relate the trajectory of individuals’ multiple relationship-dissolution events and how they uniquely lead to their future relationship dissolution (Cohen & Manning, 2010). Accord- ingly, person-centered trajectories of relationship dissolution are warranted to examine heterogeneity in the timing and frequency of various dissolution events, including cohabitation dissolution, di- vorce, and perhaps most important, multiple dissolutions in those coliving relationships.

Earlier life contexts, particularly family and community socio- economic conditions, have been shown to accumulate across child- hood and adolescence, persist into young adulthood, and influence relationship functioning (Collins, Hennighausen, Schmit, & Sroufe, 1997). For instance, low family socioeconomic status and community poverty, working as relationship-inhibiting early life contexts, were observed to be directly associated with low quality and instability in adult romantic relationships (Bryant & Conger, 2002). Early contextual risk factors were also found to place youth at a distinct disadvantage in acquiring the human capital necessary

Dayoung Bae, Center for Family Research, University of Georgia; Kandauda A. S. Wickrama, Department of Human Development and Family Science, University of Georgia.

This research used data from Add Health, a program project directed by Kathleen Mullan Harris and designed by J. Richard Udry, Peter S. Bear- man, and Kathleen Mullan Harris at the University of North Carolina at Chapel Hill, and funded by Grant P01-HD31921 from the United States Department of Health and Human Services, National Institutes of Health, Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 23 other federal agencies and foundations. Information on how to obtain the Add Health data files is available on the Add Health website (http://www.cpc.unc.edu/addhealth). No direct support was received from Grant P01-HD31921 for this analysis. Parts of the work described here were presented at the National Council on Family Relations Annual Conference, Orlando, Florida, November 2017.

Correspondence concerning this article should be addressed to Dayoung Bae, Center for Family Research, University of Georgia, 1095 College Station Road, Athens, GA 30602. E-mail: dbae@uga.edu

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

Journal of Family Psychology © 2018 American Psychological Association 2018, Vol. 1, No. 999, 000 0893-3200/18/$12.00 http://dx.doi.org/10.1037/fam0000465

1

for adaptive romantic relationship functioning (Olderbak & Figueredo, 2009). Early socioeconomic disadvantages may in- crease adolescents’ psychological and sociodemographic vulnera- bility through precocious acquisition of adult roles (e.g., early employment), conflict in dating relationships, and lack of future orientation. Such vulnerability can substantially influence roman- tic relationship experiences during young adulthood.

Little is known, however, regarding the longitudinal associa- tions between early socioeconomic contexts and romantic relation- ship experiences in young adulthood (Wickrama & O’Neal, 2016). In the present study, we attempted to advance the literature by describing heterogeneity in relationship-dissolution experiences during young adulthood. To identify developmental pathways that shaped the relationship trajectories, we investigated distal predic- tors of childhood adversity and mediating roles of psychosociode- mographic characteristics during adolescence.

Background

Heterogeneity of Romantic Relationships in Young Adulthood

According to Arnett’s (2004) emerging-adulthood theory, emerging adulthood (i.e., the period from the ages of 18 to 25 years) is characterized by identity exploration and self-focus. Cou- pled with reduced pressure from society to settle down in com- mitted relationships at earlier ages, emerging adults’ selectiveness in their choices has contributed to their involvement in a series of romantic relationships (Arnett, 2004; Reifman, 2011).

Although the unstandardized and individualized life pathways of today’s young adults (Shanahan, 2000) are indicative of serial romantic relationships, prior studies have often highlighted a sin- gle relationship dissolution, implicitly focusing on the timing of the dissolution (Kamp Dush, 2011; Wu & Schimmele, 2005). The life-course perspective, however, emphasizes the timing and se- quencing of transitions rather than a single transition (Elder, 1998). Individuals’ romantic relationship experiences are interdependent and linked across the age span (Madsen & Collins, 2011; Sassler, 2010); serial relationships construct sequences of life events that give structure to an individual’s life course (Lichter, Turner, & Sassler, 2010). For instance, Lichter and Qian (2008) argued that the very experience of ending a romantic relationship could make it easier to terminate the next relationship, including marriage. Karney and Bradbury (1995) also suggested that some risk factors for divorce could be identified in premarital relationships, demon- strating that dissolution experiences of prior relationships may predict later relationship stability.

The central goal of the present study is thus to identify various trajectories in cumulative romantic relationship-dissolution expe- riences spanning young adulthood. To reflect sequential and inter- twined dissolution experiences, we considered multiple dissolu- tions in cohabitation and marriage as independent relationship- dissolution events and differentiated these events from a single cohabitation dissolution or divorce. Specifically, the occurrence and timing of three relationship-dissolution events were investi- gated with a person-centered approach.

Early Social Context and Romantic Relationship Dissolution in Young Adulthood

Researchers have previously applied the life-course perspective (e.g., Elder, 1998) to understand the enduring effects of early social contexts on individuals’ later romantic relationship experi- ences (Cui, Fincham, & Durtschi, 2011; Wickrama & O’Neal, 2016). For example, exposure to frequent relationship conflicts and higher percentages of separated or divorced parents in a disadvantaged community have been shown to influence the ro- mantic relationships of youth. Specifically, youth raised in these environments are more likely to bring learned hostile behaviors into future romantic relationships (Bryant & Wickrama, 2005). Disadvantaged communities can also weaken societal norms against hostile and violent romantic interactions, leading to failure in forming stable and satisfying close relationships (Mancini & Bowen, 2013).

Within the domain of the family, a lack of socioeconomic resources in the family of origin (e.g., economic hardship, low parental education) can also hinder young adults’ capacity to establish and maintain romantic relationships (Amato & Kane, 2011). Distressed youth from socioeconomically disadvantaged families tend to leave their families earlier in life, compared with youth from families with greater resources, and are more likely to participate in unstable intimate relationships, characterized by early marriage or cohabitation (Amato et al., 2008; Härkönen & Dronkers, 2006). On the other hand, young adults from high- resource families tend to delay marriage longer to complete col- lege and begin their careers (Bryant & Conger, 2002).

Children from low-income families are more likely to be ex- posed to a constellation of risk factors, as opposed to isolated instances of adverse circumstances (Evans, Li, & Whipple, 2013). Accordingly, investigating environmental risk factors of a single domain (either community or family) may underestimate the neg- ative developmental impacts of risky environments. Therefore, in the present study, we constructed a composite index of multiple socioeconomic risk factors in community and family contexts to understand the comprehensive influence of early adversity on later relationships (Evans & Kim, 2013).

Psychosociodemographic Factors Linking Early Contexts to Romantic Relationships

Previous studies of the life-course perspective (Elder, 1998) and the development of early adult romantic relationship model (DEARR; Bryant & Conger, 2002; Conger, Cui, Bryant, & Elder, 2000) have shown that romantic relationship pathways are influ- enced by early structural contexts as well as one’s capacity to adapt to the context (e.g., socioeconomic competencies and inter- personal skills). For instance, youth from economically disadvan- taged community and family environments tend to experience off-time transitional events during adolescence, such as dropping out of school and full-time employment, because they lack the resources to support normative transitions (Elder, George, & Sha- nahan, 1996; Wickrama, Kwon, Oshri, & Lee, 2014). Other studies have documented that community-level poverty is positively as- sociated with the risk of dating conflict among adolescents, with the increased risk attributed to a lack of social control in disorga- nized communities (Capaldi, Knoble, Shortt, & Kim, 2012; Copp,

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

2 BAE AND WICKRAMA

Kuhl, Giordano, Longmore, & Manning, 2015). Similarly, parents’ marital instability and low socioeconomic status are associated with lower levels of parental support and control of their children (Valle & Tillman, 2014), increasing the likelihood of adolescents’ involvement in poor-quality or even abusive romantic relation- ships (Aquilino, 1991). Perception of living in an uncontrollable and harmful environment, as associated with disadvantaged com- munity and family contexts, may lead to youth viewing their futures as unstable and hopeless, eroding their sense of future orientation (Donnellan, Conger, McAdams, & Neppl, 2009; Ma- digan, Atkinson, Laurin, & Benoit, 2013; Sieger, Rojas-Vilches, McKinney, & Renk, 2004). Lack of these psychological and so- ciodemographic resources stemming from early adverse life con- texts, in turn, can lower commitment to, and foster distrust in, close relationships (Bryant & Conger, 2002; Simons, Simons, Lei, & Landor, 2012).

Based on these previous findings, we chose to focus on three psychosociodemographic factors for adolescents: disrupted transi- tions to adulthood, conflict in dating relationships, and future orientation. We examined mediating roles, linking early socioeco- nomic adversity to romantic dissolution in young adulthood.

Current Study

The topic of diverse romantic relationship trajectories among today’s young adults is pertinent for family and demographic research. Also, investigating linked developmental pathways stem- ming from early community and family contexts broadens our understanding of how early life experiences shape close relation- ship dynamics. In the present study, we first hypothesized that heterogeneity in romantic relationship-dissolution trajectories in young adulthood exists and that serial patterns of relationship dissolution for young adults would be identified. We also hypoth- esized that exposure to cumulative early risk factors would be associated with higher levels of instability in coliving relation- ships, as reflected by the experience of multiple dissolutions. We also expected to find that the association between early adversity and romantic relationship-dissolution trajectories was mediated by psychosociodemographic characteristics during adolescence, in- cluding disrupted transitions to adulthood, conflict in dating rela- tionships, and decreased future orientation. Specifically, we hy- pothesized that a lack of psychological and sociodemographic resources during adolescence would be linked to known risk factors for relationship instability, such as relationship dissolution at an early age followed by multiple dissolutions.

Method

Participants and Procedures

Data for this study came from a nationally representative sample of adolescents participating in the National Longitudinal Study of Adolescent to Adult Health (Add Health). In 1995, Wave-1 (base- line) data were derived from a stratified cluster sampling of middle- and high-school students, yielding 20,745 respondents (Mage � 15.5 years) from 134 middle and high schools. In addi- tion, each respondent’s parent was interviewed using interviewer- administered questionnaires regarding her or his sociodemo- graphic characteristics and family environment. The second, third,

and fourth waves of data were collected in 1996, 2001, and 2008 (n2 � 14,738, Mage � 16.18; n3 � 15,100, Mage � 21.18; n4 � 15,701, Mage � 29.13). In the present study, we analyzed interview data from parents at Wave 1 and from young adults who partici- pated in all four waves (n � 12,026), along with 1990 United States Census-tract-level data. Sample weights were used in our analyses to ensure the representativeness of the study participants. The purpose of the present study was to investigate union- dissolution experiences, therefore, young adults who had not formed coliving relationships (n � 2,598) or who failed to provide responses to questions on romantic relationship experiences (n � 153) were excluded from the study sample, reducing our final sample size to 9,275. The sample consisted of approximately 53% White, 22% Africans American, 16% Hispanic, 5% Asian, 2% Native American, and 1% other racial/ethnic participants. About 54% of young adults were women and 1.5% were in same-sex relationships. The research was approved by the Institutional Re- view Board at the University of Georgia.

Measures

Romantic relationship-dissolution trajectories. At Wave 4 (2008), young adults were asked about the histories of all their previous romantic and sexual relationships since Wave 1 (1995). For each relationship, respondents were asked to report when the relationship started and ended. For each age from 18 to 30, a binary variable was created for three relationship-dissolution com- ponents (i.e., cohabitation dissolution, divorce, and multiple dis- solutions) indicating whether the individual experienced the dis- solution event for the first time at that age (coded 1) or had not experienced the dissolution by that age (coded 0). For cohabitation dissolution and divorce, month and year information of each relationship was used. For multiple dissolutions, month and year information for the end of the second coliving relationship was used (i.e., the second experience of at least two divorces, two cohabitation dissolutions, or one divorce and one cohabitation dissolution). To account for a relatively small percentage of indi- viduals who experienced relationship-dissolution events before age 18 (�1% divorce, �3% cohabitation dissolution), the relationship-status variable at age 18 represented whether individ- uals experienced each relationship dissolution for the first time by age 18 or younger.

Early socioeconomic adversity. In accordance with the no- tion of multiple risk exposure (Evans & Kim, 2013), an additive index for cumulative socioeconomic adversity was created by summing dichotomous indicators that captured multiple dimen- sions of socioeconomic adversity measured at Wave 1. Except for relationship stability (already a dichotomous measure), dichoto- mous indicators were created by mean splitting the measures, as described next.

Community adversity. Community adversity was assessed us- ing four indicators corresponding to census-tract-level information from the 1990 United States Census. The indicators included (a) the proportion of families living in poverty, (b) the proportion of single-parent families, (c) the proportion of adults employed in service occupations, and (d) the proportion of unemployed men (Sucoff & Upchurch, 1998). Proportions were summed to indicate the extent of community adversity (M � .41, SD � .21, � � .78).

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

3ROMANTIC RELATIONSHIP-DISSOLUTION TRAJECTORIES

Parental education. We constructed an index of parental ed- ucation by averaging both parents’ highest level of education (ranging from 1 � never went to school to 10 � professional training beyond college/university degree). For single-headed fam- ilies (n � 79) with no available data from fathers, maternal education served as the indicator of parental education (M � 5.48, SD � 1.95; 5 � high-school graduation/GED, 6 � business or vocational school after high school).

Economic hardship. Five dichotomous items (0 � no, 1 � yes) assessed whether any member of the household received the following social service benefits in the past month: social security, supplemental security income, aid to families with dependent children, food stamps, or housing subsidies. Responses to these five items were summed to create an index of economic hardship with a range of 0 –5 (M � 1.40, SD � .94).

Parental relationship stability. A binary variable was used to differentiate parents who had been consistently married to their spouses (or had been in marriage-like relationships) for at least 15 years (coded as 1) from other parents (coded as 0). Fifteen years was selected as the cut-off because the average age of adolescent respondents at Wave 1 was 15 years. Thus, for most respondents, this variable represented their parents’ relationship stability for the duration of their childhood lives. Forty-seven percent of respon- dents reported that their parents were continuously married or in marriage-like relationships.

Disrupted transitions to adulthood. Using retrospective data at Wave 3 (2001), disrupted transitions were identified by six indicators consistent with previous studies (Wickrama et al., 2014). High-school dropout was assessed by the highest level of education young adults completed at Wave 3. Respondents were coded as 0 (equal to or higher than high school graduation/GED) and 1 (lower than high school graduation/GED). Early leaving home was assessed using adolescent report on the year of moving and residential status (e.g., living in a separate house, trailer home, or group quarters). Respondents were coded as 1 if they left home before age 18, with the exception of full-time college/university students, who were not categorized as early leavers. Respondents were categorized as early full-time employment if they worked full-time during the same years they also reported attending high school. Marriage or cohabitation during adolescent years and fe- males becoming pregnant, or males fathering during the teenage years were classified as early marriage/cohabitation and teenage pregnancy, respectively. Last, the onset of sexual intercourse be- fore age 16 was categorized as early sexual activity. This cutoff was based on the average age of first sexual intercourse (Copen, Chandra, & Martinez, 2012). Scores of the six indicators were summed to indicate the extent of disrupted transition.

Conflict in dating relationships. Conflict in the dating rela- tionship was measured at Wave 2 (1996) by summing five dichot- omous items (0 � no, 1 � yes) from a short-form of the Conflict Tactics Scale (Straus, Gelles, & Steinmetz, 1980). Respondents were asked if their partners had ever done any of the following: “call you names, insult you, or treat you disrespectfully in front of others;” “swear at you;” “threaten you with violence;” “push or shove you;” or “throw something at you.”

Future orientation. Future orientation was measured at Wave 2 by four items capturing different domains (Chen & Vazsonyi, 2011; Wickrama, O’Neal, & Lee, 2016). Respondents rated their agreement to two items: “You live your life without much thought

of the future” was measured on a 5-point scale, ranging from 1 � strongly agree to 5 � strongly disagree; “You felt hopeful about the future” was measured on a 4-point scale, ranging from 0 � never or rarely to 3 � most or all the time. In addition, participants indicated their chances of living to age 35 and their chances of graduating from college on a 5-point scale, ranging from almost no chance/low to almost certain/high. A sum score was computed using standardized scores for each of the four items (� � .65).

Covariates. Our analysis included respondents’ age, gender, and race/ethnicity (White, African American, Hispanic, Asian, Native, and other) as covariates. We also controlled for depressive symptoms during adolescence (Wave 2), which were assessed with nine items (ranging from 0 � never/rarely to 3 � most/all of the time) from the Center for Epidemiological Studies Depression Scale (CES-D; Radloff, 1977). Items were summed to yield a 27-point scale (M � 5.86, SD � 3.41).

Data Analysis

To investigate the heterogeneity of romantic dissolution trajec- tories, we used the discrete-time multiple event process survival mixture (MEPSUM) model, which integrates multivariate discrete- time survival modeling and latent-class analysis (Dean, Bauer, & Shanahan, 2014). Compared with traditional discrete-time survival modeling, which allows for investigation of the occurrence and timing of a single event, the MEPSUM model enables us to examine multiple events simultaneously over time. Traditionally, risks of experiencing cohabitation dissolution and divorce have been treated as mutually exclusive events and handled with competing-risk models (Ventura, Nuechterlein, Subotnik, Hard- esty, & Mintz, 2000). In our study, however, if an individual experienced divorce for the first time after a preceding cohabita- tion dissolution, the divorce became a marker for a multiple dissolution experience. Then, divorce and multiple dissolutions were recorded as occurring in the same year. In those cases, the relationship-dissolution events were not mutually exclusive and did not necessarily occur in a sequential order. This type of data can be more correctly analyzed with the MEPSUM model than competing-risk models. Moreover, the MEPSUM model allows the identification of heterogeneity in dissolution trajectories by incorporating latent classes of hazard probabilities of event occur- rence, which is particularly pertinent to the present study, because we intentionally investigated various patterns of dissolution occur- rence over time. Using three indicators of the timing of relation- ship dissolutions (i.e., cohabitation dissolution, divorce, and mul- tiple transitions), we expected that the MEPSUM model would provide a succinct summary of individual differences in patterns of relationship dissolution.

Analyses were conducted in Mplus Version 7.0 (Muthén & Muthén, 1998 –2012). The optimal number of classes was deter- mined by the highest log-likelihood (LL) values, smallest Akaike information criterion (AIC) and Bayesian information criterion (BIC), entropy greater than 0.85, and a significant p value in the Lo–Mendell–Rubin likelihood-ratio test (LMR-LRT; Jung & Wickrama, 2008). Size and interpretability of classes were also considered (Bauer & Curran, 2003). Once relationship-dissolution trajectories were identified, the classes were compared based on study variables using multinomial logistic regression. Hypothe- sized longitudinal pathways were tested using path analysis. We

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

4 BAE AND WICKRAMA

used the TYPE � COMPLEX command to account for potential bias in standard errors and chi-square computations that arise as a result of the nested structure of the Add Health school-based design. We also employed the MODEL CONSTRAINT command, which allows for testing indirect effects in models with categorical outcomes. Full-information maximum-likelihood procedures were used to account for missing data (Enders & Bandalos, 2001).

Results

Identifying Different Romantic Relationship- Dissolution Classes

Determination of the class solution. Increases in �2LL and decreases in AIC and BIC values were observed as the number of classes increased from one to four, with the exception of the BIC, which increased slightly in the five-class solution (see Supplemen- tal Table 1). Although the five-class solution demonstrated fair entropy and significant LMR-LRT values, the five-class solution included a class that represented only 3% of the sample. Therefore, a four-class solution was determined to be the best solution to describe the underlying heterogeneity in relationship-dissolution experiences in this sample.

Characteristics of the four-class solution. Next, the hazard (ĥ) and cumulative distribution (D) probabilities for each class were examined (see Figure 1). The hazard function represents the probability of experiencing a particular relationship dissolution at a certain age. The D function displays the cumulative probability of relationship dissolution by a given age.

The first class (n � 575, 6.2%) was characterized by high risk of experiencing cohabitation dissolution and multiple dissolutions at early ages. The risk of cohabitation dissolution increased from ages 18 to 22 (ĥ18 � .27, ĥ22 � .51). The risk of multiple dissolutions was initially lower than the risk of cohabitation dis- solution alone, but caught up by age 23, at which time the risk of multiple dissolutions continuously increased and the risk of co- habitation dissolution sharply decreased. The risk of divorce in this class was consistently low. As shown in Table 1, almost all of the young adults in this group experienced cohabitation dissolution and multiple dissolutions (95.2% and 100.0%, relatively). The median age of cumulative probability of cohabitation dissolution was 19, indicating that 50% of young adults experienced cohabi- tation dissolution by age 19 (D19 � .50). The median age of cumulative probability for multiple dissolutions was 22. The me- dian age of divorce is not available because the cumulative prev- alence of divorce was less than 50%. Young adults in this class reported that they experienced 3.57 relationship dissolutions on average. Because this class was characterized by the high risk of cohabitation dissolution and multiple dissolutions during their early twenties, we labeled this group the early multiple dissolu- tions class.

The second class (n � 657, 7.1%) was characterized by a relatively low risk of relationship dissolution through the early twenties that increased during the mid- and late-twenties. The risk of cohabitation dissolution remained below .10 by age 23, with the peak risk coming at age 27 (ĥ27 � .25). The risk of divorce also remained relatively low through the ages 18 to 30, but by age 30, 67% of young adults in this group had experienced divorce (D30 � .67). The risk of multiple dissolutions remained very low until age

23 (ĥ23 � .02), then increased sharply from age 24 to 30 (ĥ23 � .07, ĥ30 � 1.00). As indicated by the median age, about 50% of young adults in this class experienced cohabitation dissolution, divorce, and multiple transitions by age 27. Due to the late in- creases in risk of relationship dissolution and the relatively high cumulative probability of all three relationship-dissolution events, this class was labeled the later multiple dissolutions class.

The third class (n � 2,345, 25.3%) was characterized by con- tinuous increases in the risk of cohabitation dissolution and con- sistently low risk of divorce and multiple dissolutions. The risk of cohabitation dissolution increased rapidly from age 23 to 28 (ĥ23 � .29, ĥ28 � 1.00). The risk of divorce remained very low, with the cumulative probability of .007 at age 30, indicating that less than 1% of young adults in this group experienced divorce by age 30. The peak risk of multiple dissolutions came at age 28 (ĥ28 � .06) with the D � .23 at age 30. Because of the high risk of cohabitation dissolution and a relatively low risk of divorce and multiple dissolutions, we labeled this group the cohabitation dis- solution class.

The hazard and cumulative distribution probabilities in the fourth class (n � 5,698, 61.4%) differed a great deal from those in the previous three classes. Risk of all three relationship-dissolution components remained low and did not increase. Peak estimated risks for cohabitation dissolution, divorce, and multiple dissolu- tions were .023, .020, and .002, respectively. On average, respon- dents reported less than one relationship-dissolution experience. Because of the low cumulative probabilities of all three relationship-dissolution components, we labeled this group the stable relationships class.

Demographic and developmental correlates of class membership. Using multinomial logistic regression, we further tested significant differences in developmental correlates across classes. As shown in the first panel of Table 2, where the early multiple dissolutions group served as the reference, the relative risk ratios (RRR) were significant and less than 1, indicating that increases in early socioeconomic adversity decreased the odds of being in the later multiple dissolutions, cohabitation dissolution, and stable relationships groups. The early multiple dissolution group also reported higher levels of disrupted transitions, conflict in romantic relationships, and lower levels of future orientation than the cohabitation dissolution and stable relationship groups. The second panel shows that the later multiple dissolutions group reported more disrupted transitions and lower levels of future orientation than the cohabitation dissolution and stable relationship groups. The cohabitation dissolution group reported higher levels of disrupted transitions and conflict in romantic relationships than the stable relationship group. household income and educational attainment were highest in the stable relationship and lowest in the early multiple dissolutions group. The highest rates of respondents in the stable relationship group and the lowest rates of respondents in the cohabitation dissolution group reported that they were living with children.

Pathways Linking Early Contexts to Relationship- Dissolution Experiences

Figure 2 presents the results of the linking pathways between early life contexts and young adults’ relationship-dissolution ex- periences through psychosociodemographic characteristics during

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

5ROMANTIC RELATIONSHIP-DISSOLUTION TRAJECTORIES

Cohabitation dissolution Divorce Multiple dissolutions Class 1: Early Multiple Dissolutions (n = 575, 6.2%)

Class 2: Later Multiple Dissolutions (n = 657, 7.1%)

0

0.2

0.4

0.6

0.8

1

18 19 20 21 22 23 24 25 26 27 28 29 30 0

0.2

0.4

0.6

0.8

1

18 19 20 21 22 23 24 25 26 27 28 29 30

0

0.2

0.4

0.6

0.8

1

18 19 20 21 22 23 24 25 26 27 28 29 30 0

0.2

0.4

0.6

0.8

1

18 19 20 21 22 23 24 25 26 27 28 29 30

Age

Pr ob

ab ili

ty

Age

Pr ob

ab ili

ty

Age

Pr ob

ab ili

ty

Pr ob

ab ili

ty

Age

Class 3: Cohabitation Dissolution (n = 2,345, 25.3%)

Class 4: Stable Relationships (n = 5,698, 61.4%)

0

0.2

0.4

0.6

0.8

1

18 19 20 21 22 23 24 25 26 27 28 29 30 0

0.2

0.4

0.6

0.8

1

18 19 20 21 22 23 24 25 26 27 28 29 30

0

0.2

0.4

0.6

0.8

1

18 19 20 21 22 23 24 25 26 27 28 29 30 0

0.2

0.4

0.6

0.8

1

18 19 20 21 22 23 24 25 26 27 28 29 30

Age Age

Pr ob

ab ili

ty

Age

Pr ob

ab ili

ty

Age

Pr ob

ab ili

ty

Pr ob

ab ili

ty

Figure 1. Hazard (left) and cumulative probabilities (right) of romantic relationship-dissolution experiences in latent classes.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

6 BAE AND WICKRAMA

adolescence. As hypothesized, early adversity was associated with negative psychological and sociodemographic characteristics dur- ing adolescence. Specifically, adolescents who experienced higher levels of early adversity reported more disrupted transitions into adulthood (� � .25, p � .001), conflict in dating relationships (� � .20, p � .001), and lower levels of future orientation (� � �.20, p � .001). In addition, a 1-SD increase in early adversity directly increased the odds of being in the early multiple dissolutions group by 20%, relative to being in the stable relation- ship group (RRR � 1.20, p � .01).

Psychosociodemographic characteristics in adolescence, in turn, were significantly associated with relationship-dissolution experi- ences. Disrupted transitions to adulthood increased the odds of being included in all three types of relationship-dissolution classes (early multiple dissolutions: RRR � 1.52, p � .001; later multiple dissolutions: RRR � 1.40, p � .001; cohabitation dissolution: RRR � 1.19, p � .001), compared with the stable relationship group. Experiencing conflict in dating relationships increased the odds of being in the early multiple dissolutions (RRR � 1.33, p � .001) and later multiple dissolutions (RRR � 1.28, p � .05) groups. Future orientation was only associated with the early multiple dissolutions group, in which a 1-SD increase in future orientation decreased the odds of being in the early multiple dissolutions group by 7% (RRR � .93, p � .01).

Indirect-effect analysis revealed that all linking pathways stem- ming from early adversity through psychosociodemographic char- acteristics in adolescence were significant. Specifically, the effects of early adversity were indirectly associated with the early multi- ple dissolutions group through disrupted transitions (b � .09, p � .001, effect size � .13), conflict in dating relationships (b � .02,

p � .05, effect size � .04), and future orientation (b � .04, p � .01, effect size � .07). Early adversity predicted the later multiple dissolutions group via disrupted transitions (b � .07, p � .001, effect size � .26), and conflict in dating relationships (b � .02, p � .05, effect size � .08). The indirect effects of early adversity on the cohabitation dissolution group via disrupted transitions was also significant (b � .04, p � .01, effect size � .25).

Discussion

According to the stress-relief hypothesis (Wheaton, 1990), not every relationship-dissolution event is associated with negative developmental outcomes. Individuals in overt and prolonged con- flict, in particular, may benefit from relationship dissolution and experience personal growth and psychological well-being (Cohen, Klein, & O’Leary, 2007). However, the accumulation of dissolu- tions, as compared with a single dissolution event, may be stress- ful, resulting in emotional and physical problems. Although tran- sitions in and out of multiple unions have become widespread among young adults in the United States (Cherlin, 2009), the heterogeneity in timing and frequency of relationship-dissolution events and developmental antecedents are not well-understood. Using three distinct, yet potentially interdependent, relationship- dissolution experience markers, we investigated the cumulative history of romantic relationship dissolutions, adopting a person- centered approach called MEPSUM.

Our findings clearly illustrate the variability in young adults’ romantic relationship-dissolution patterns by identifying four unique trajectories. These trajectories varied in the occurrence and timing of relationship-dissolution events. The most prevalent

Table 1 Prevalence and Means of Union-Dissolution Events, Developmental Precursors, and Demographic Characteristics as a Function of Class Membership

Characteristics Class 1

(n � 575) Class 2

(n � 657) Class 3

(n � 2,345) Class 4

(n � 5,698)

Total (n � 9,275)

M (SD)/% Range

Cumulative prevalence at age 30 Cohabitation dissolution 95.2% 81.8% 100.0% 12.5% 44.7% 0–1 Divorce 22.7% 66.6% .7% 14.5% 15.2% 0–1 Multiple dissolutions 100.0% 100.0% 23.2% .5% 19.5% 0–1

Median age of union dissolutiona

Cohabitation dissolution 19 26 22 — — 18–30 Divorce — 27 — — — 18–30 Multiple dissolutions 22 27 — — — 18–30

Mean number of union dissolutions 3.57 2.94 1.40 .71 1.20 (1.35) 0–21 Developmental precursors

Early socioeconomic adversity (Wave 1) 2.05 1.86 1.76 1.75 1.78 (1.10) 0–4 Disrupted transitions (Wave 2) 1.16 1.13 .90 .84 .90 (.85) 0–6 Conflict in romantic relationships (Wave 2) .25 .19 .12 .10 .15 (.45) 0–8 Future orientation (Wave 2) �.91 �.73 �.31 �.17 �.29 (2.69) �10.77–4.20

Demographics (Wave 4) Female 57.2% 52.7% 54.8% 55.4% 55.1% 0–1 Age 28.9 29.8 29.3 29.6 29.5 (1.65) 18–30 Educational attainmentb 4.60 4.86 5.27 5.46 5.32 (2.14) 1–13 Household income per year $52,957 $54,719 $59,424 $63,581 $61,212 ($37,083) $2,500–$150,000 Living with children 53.0% 50.5% 44.8% 57.0% 53.2% 0–1

Note. Class 1 � early multiple dissolutions; Class 2 � later multiple dissolutions; Class 3 � cohabitation dissolution; Class 4 � stable relationships. a Median age is not available in cases in which the cumulative prevalence of relationship dissolution was less than 50%. b Educational attainment variable ranges from 1 (eighth grade or less) to 13 (completed post-baccalaureate professional education). Value 4 indicates some vocational/technical training after high school, 5 indicates completed vocational/technical training after high school, and 6 indicates some college.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

7ROMANTIC RELATIONSHIP-DISSOLUTION TRAJECTORIES

group, stable relationships (61.4%), was characterized by low risks of cohabitation or marital dissolution throughout young adulthood. The pursuit of higher educational attainment and job status among today’s young adults (Arnett, 2004) may delay their involvement in committed relationships during their twenties, especially cohab- itation and marriage. As Cherlin (2009) asserted, committed rela- tionships today often represent the last step into adulthood, not the first. Therefore, the low risks of cohabitation dissolution and divorce throughout young adulthood may reflect the fact that these young adults entered committed relationships at older ages than their counterparts in other groups. Our findings indeed showed that the age at first coliving relationship of the stable relationship group was relatively late compared with the two multiple dissolutions groups (see Supplemental Table 2). In addition, the stable rela- tionship group showed the highest educational attainment and household income levels. However, as shown in Figure 1, the cumulative prevalence of experiencing relationship dissolution slightly increased in the late twenties; at age 30, approximately 13% of young adults in the stable relationships group experienced cohabitation dissolution, whereas 15% experienced divorce. Lon- gitudinal follow-up on the stable relationships group, therefore, may shed light on whether early stability in committed relation- ships can be maintained in later years.

The cumulative prevalence showed that all young adults in the early multiple dissolutions and later multiple dissolutions groups experienced multiple disruptions in cohabitation and divorce by age 30. These groups represented over 13% of the entire sample of the study. Young adults who experienced multiple dissolutions may be more likely to undergo a chain of relationship instabilities with their future partners (Lichter et al., 2010; Lichter & Qian, 2008). The phenomenon of multiple partnerships is an important research topic in family studies and family policy (Cherlin, 2008) because of the potential links with psychological and physical health problems (Barrett, 2000), economic difficulties (Amato, 2010), and children’s low psychological well-being and behavioral problems (Cavanagh & Huston, 2008). Given the importance of establishing a stable intimate relationship, it is critical to under- stand what kinds of early life experiences are associated with risk of dissolution in later romantic relationships.

Our findings revealed that early socioeconomic adversity was directly associated with inclusion in the early multiple dissolutions group, suggesting that higher levels of socioeconomic hardship faced in community and family environments may result in in- creased risk of relationship instability at earlier ages. This finding is consistent with an escape-from-stress perspective, which posits that, in cases in which youth are exposed to increased levels of stress because of adverse life contexts, they perceive cohabitation or marriage as opportunities to escape their stressful family envi- ronments, as they hope to establish intimacy and supportive rela- tionships with their partners (Amato & Kane, 2011). However, forming a coliving relationship in early adulthood is a known risk factor for relationship instability (Lichter et al., 2010), which then may be linked to a chain of relationship-dissolution experiences.

The results of the current study also elucidate psychosociode- mographic mechanisms that link early socioeconomic adversity to young adults’ romantic relationship experiences. First, early ad- versity predisposed adolescents to subsequent negative life expe- riences in transitional events to adulthood. Precocious and dis- rupted transitions to adulthood, such as dropping out of school orT

ab le

2 M

u lt

in o m

ia l

L o g is

ti c

R eg

re ss

io n

o f

C la

ss M

em b er

sh ip

R eg

re ss

ed o n

D ev

el o p m

en ta

l P

re cu

rs o rs

a n d

D em

o g ra

p h ic

C h a ra

ct er

is ti

cs

C om

pa ri

so n

cl as

s

E ar

ly so

ci oe

co no

m ic

ad ve

rs it

y D

is ru

pt ed

tr an

si ti

on s

C on

fl ic

t in

ro m

an ti

c re

la ti

on sh

ip s

F ut

ur e

or ie

nt at

io n

A ge

E du

ca ti

on al

at ta

in m

en t

H ou

se ho

ld in

co m

e L

iv in

g w

it h

ch il

dr en

R el

at iv

e ri

sk ra

ti o

(9 5%

C I)

E ar

ly m

ul ti

pl e

di ss

ol ut

io ns

vs .

L at

er m

ul ti

pl e

di ss

ol ut

io ns

.8 5

(. 74

,. 98

)� .9

8 (.

83 ,1

.1 5)

.8 6

(. 71

,1 .0

4) 1.

02 (.

97 ,1

.0 7)

1. 40

(1 .2

9, 1.

52 )�

� �

1. 07

(1 .0

1, 1.

13 )�

1. 01

(. 96

,1 .0

5) .8

4 (.

66 ,1

.0 6)

C oh

ab it

at io

n di

ss ol

ut io

n .7

9 (.

70 ,.

88 )�

� �

.7 3

(. 64

,. 83

)� � �

.7 5

(. 65

,. 87

)� � �

1. 08

(1 .0

4, 1.

13 )�

� �

1. 17

(1 .1

0, 1.

25 )�

� �

1. 17

(1 .1

2, 1.

23 )�

� �

1. 05

(1 .0

2, 1.

09 )�

� .6

6 (.

55 ,.

80 )�

� �

S ta

bl e

re la

ti on

sh ip

s .7

8 (.

70 ,.

87 )�

� �

.6 7

(. 59

,. 76

)� � �

.6 2

(. 54

,. 71

)� � �

1. 10

(1 .0

6, 1.

15 )�

� �

1. 31

(1 .2

3, 1.

39 )�

� �

1. 22

(1 .1

7, 1.

28 )�

� �

1. 13

(1 .0

9, 1.

16 )�

� �

1. 09

(. 91

,1 .3

0) L

at er

m ul

ti pl

e di

ss ol

ut io

ns vs

. C

oh ab

it at

io n

di ss

ol ut

io n

.9 2

(. 83

,1 .0

3) .7

4 (.

67 ,.

85 )�

� �

.8 7

(. 74

,1 .0

3) 1.

06 (1

.0 2,

1. 10

)� �

.8 4

(. 79

,. 89

)� � �

1. 10

(1 .0

5, 1.

15 )�

� �

1. 05

(1 .0

1, 1.

08 )�

� .8

0 (.

66 ,.

95 )�

S ta

bl e

re la

ti on

sh ip

s .9

2 (.

83 ,1

.0 2)

.6 8

(. 61

,. 77

)� � �

.7 2

(. 61

,. 84

)� � �

1. 08

(1 .0

4, 1.

12 )�

� �

.9 3

(. 88

,. 99

)� 1.

15 (1

.1 0,

1. 20

)� � �

1. 12

(1 .0

9, 1.

16 )�

� �

1. 30

(1 .1

0, 1.

54 )�

�

C oh

ab it

at io

n di

ss ol

ut io

n vs

. S

ta bl

e re

la ti

on sh

ip s

.1 0

(. 94

,1 .0

5) .9

2 (.

85 ,.

99 )�

.8 2

(. 74

,. 91

)� � �

1. 02

(. 99

,1 .0

4) 1.

11 (1

.0 8,

1. 15

)� � �

1. 04

(1 .0

2, 1.

07 )�

� �

1. 07

(1 .0

5, 1.

09 )�

� �

1. 64

(1 .4

9, 1.

80 )�

� �

N o te

. E

ac h

cl as

s w

as co

m pa

re d

w it

h th

e om

it te

d co

m pa

ri so

n cl

as s.

A ll

de ve

lo pm

en ta

l pr

ec ur

so rs

an d

de m

og ra

ph ic

fa ct

or s

w er

e in

cl ud

ed in

ea ch

an al

ys is

. B

ec au

se re

sp on

de nt

s’ ge

nd er

di d

no t

sh ow

an y

si gn

if ic

an t

as so

ci at

io ns

w it

h cl

as s

m em

be rs

hi p,

it s

co ef

fi ci

en ts

ar e

no t

sh ow

n to

si m

pl if

y th

e pr

es en

ta ti

on .

� p

� .0

5. � �

p �

.0 1.

� � �

p �

.0 01

.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

8 BAE AND WICKRAMA

full-time employment during teenage years, indicate off-time events that deviate from the anticipated timing for adult roles (Elder et al., 1996). Structural constraints and resource limitations embedded in early community and family environments may not provide enough resources for adolescents to pursue normative educational and career pathways. Results also suggest that expo- sure to disadvantaged early life contexts may lead to the develop- ment or maintenance of coercive interactions with their partners (Simons et al., 2012). Distressed youth from adverse environments may also view their future as hopeless and unpredictable, reflected by low future orientation in the present study.

These psychosociodemographic characteristics, in turn, were significantly associated with romantic relationship outcomes in young adulthood, suggesting mediating pathways linking early life contexts to later romantic relationships. Disrupted transitions to adulthood, in particular, were associated with all three relationship-dissolution trajectories. For most individuals, these disrupted transitions represented chronically stressful life situa- tions that placed excessive demands on those who were not pre- pared for the responsibilities of adult roles. Emotional, social, and financial problems stemming from these early transitions may further drive youth toward more stressful life experiences in young adulthood. Low income and financial strains have consistently predicted low relationship quality, with distressed young adults more likely to reject intimate relationships (Wickrama & O’Neal, 2016).

Conflict in dating relationships was associated with the early multiple dissolutions and later multiple dissolutions groups. Hos- tile interactions with partners and exposure to abusive dating relationships during adolescence may lead to an acquisition of cynical relationship schemas and negative beliefs about the de- pendability of romantic partners. Such expectations are likely to contribute to conflict in and instability of romantic relationships in

young adulthood. We find it interesting that the cohabitation dissolution group was not associated with conflict in dating expe- riences, which may suggest that instability of romantic relation- ships during young adulthood is better reflected by multiple dis- solutions than a single dissolution event.

In addition, future orientation was associated with the early multiple dissolutions group. Future orientation is an important psychological resource that influences transitional events by in- forming the expectation of positive future selves and motivating current behaviors (Clinkinbeard, 2014; Greene & DeBacker, 2004). Young adults with low levels of future orientation may not invest time and resources needed for long-term stable relationships because they do not anticipate positive future consequences of romantic relationships. All three mediators were associated with young adults in the early multiple dissolutions group, indicating that experiencing early multiple relationship dissolutions is the riskiest pattern of dissolution among the three dissolution trajec- tories, although it showed the lowest prevalence in our sample.

Several issues potentially limit the scope and the generalizabil- ity of our results. First, given that latent-class analysis is sample- dependent, our findings should be interpreted with caution and need to be replicated with different samples. Second, the quality of relationships was not considered in the present study because of a lack of information. The level of relationship quality with a current partner can affect the stability of the relationship, independent of a history of previous dissolution experiences. The inclusion of this variable may lend more insight into the dynamics of young adults’ romantic relationships. Also, enhanced measures would increase the confidence that can be placed in these findings. Most measures used in the present study relied on self-report data, and potential biases would be alleviated by using more objective measures in future research. Last, our four trajectory classes could be linked to variation in levels of psychological and physical distress in later

Early socioeconomic adversity

Disrupted transitions

Future orientation

Early Multiple

Dissolutions

Later Multiple

Dissolutions

Cohabitation Dissolution

.25***

.57 (1.19)***

.28 (1.20)**

.20***

-.20***

.50 (1.52)***

.61 (1.40)*** .19 (1.33)***

-.27 (.93)**

.26 (1.28)*

.07***

-.04**-.21***

Conflict in romantic

relationships

Figure 2. Model predicting pathways to relationship-dissolution classes. Standardized coefficients are shown. RRRs of the relationship-dissolution classes are in parenthesis. Reference is the stable relationship group. Respondents’ age, gender, race/ethnicity, and depressive symptoms during adolescence were controlled. � p � .05. �� p � .01. ��� p � .001.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

9ROMANTIC RELATIONSHIP-DISSOLUTION TRAJECTORIES

years; however, this was outside the scope of the present study. Future researchers should further explore the impact of different characteristics of relationship-dissolution experiences on later health.

These limitations notwithstanding, the present study contributes to the extant literature by identifying heterogeneity in romantic relationship experiences across young adulthood when the pursuit of intimacy is a critical developmental task. As suggested by the life-course perspective, our study provides useful insight on the various pathways (i.e., timing and sequences of multiple transi- tional events) of young adults’ romantic relationship experiences, especially with the inclusion of multiple intimate relationship dissolutions being an important marker of relationship instability. We were also able to investigate an array of developmental pre- cursors, describing how early socioeconomic life contexts, medi- ated by psychosociodemographic characteristics, shape the diverse course of romantic relationship development in young adulthood. Our findings underscore the utility of understanding early family and community environments in exploring youths’ life-course development. In addition, the finding that young adults exposed to conflict with romantic partners during adolescence were at greater risk of experiencing multiple dissolutions suggests that such indi- viduals might benefit from prevention efforts designed to address risks in intimate relationship conflict and communication.

References

Amato, P. R. (2010). Research on divorce: Continuing trends and new developments. Journal of Marriage and Family, 72, 650 – 666. http://dx .doi.org/10.1111/j.1741-3737.2010.00723.x

Amato, P. R., & Kane, J. B. (2011). Parents’ marital distress, divorce, and remarriage: Links with daughters’ early family formation transitions. Journal of Family Issues, 32, 1073–1103. http://dx.doi.org/10.1177/ 0192513X11404363

Amato, P. R., Landale, N. S., Havasevich-Brooks, T. C., Booth, A., Eggebeen, D. J., Schoen, R., & McHale, S. M. (2008). Precursors of young women’s family formation pathways. Journal of Marriage and Family, 70, 1271–1286. http://dx.doi.org/10.1111/j.1741-3737.2008 .00565.x

Aquilino, W. S. (1991). Family structure and home-leaving: A further specification of the relationship. Journal of Marriage and Family, 53, 999 –1010. http://dx.doi.org/10.2307/353003

Arnett, J. J. (2004). Emerging adulthood: The winding road from the late teens through the twenties. New York, NY: Oxford University Press.

Barrett, A. E. (2000). Marital trajectories and mental health. Journal of Health and Social Behavior, 41, 451– 464. http://dx.doi.org/10.2307/ 2676297

Bauer, D. J., & Curran, P. J. (2003). Distributional assumptions of growth mixture models: Implications for overextraction of latent trajectory classes. Psychological Methods, 8, 338 –363. http://dx.doi.org/10.1037/ 1082-989X.8.3.338

Beaujouan, E. (2012). Repartnering in France: The role of gender, age and past fertility. Advances in Life Course Research, 17, 69 – 80. http://dx .doi.org/10.1016/j.alcr.2012.03.001

Bryant, C. M., & Conger, R. D. (2002). An intergenerational model of romantic relationship development. In A. L. Vangelisti, H. T. Reis, & M. A. Fitzpatrick (Eds.), Stability and change in relationships (pp. 57– 82). New York, NY: Cambridge University Press. http://dx.doi.org/ 10.1017/CBO9780511499876.005

Bryant, C. M., & Wickrama, K. A. S. (2005). Marital quality among African Americans: Contextual perspectives. In K. A. Dodge, V. McLoyd, & N. Hill (Eds.), Emerging issues in African-American family

life: Context, adaptation, and policy (pp. 111–134). New York, NY: Guilford Press.

Capaldi, D. M., Knoble, N. B., Shortt, J. W., & Kim, H. K. (2012). A systematic review of risk factors for intimate partner violence. Partner Abuse, 3, 231–280. http://dx.doi.org/10.1891/1946-6560.3.2.231

Cavanagh, S. E., & Huston, A. C. (2008). The timing of family instability and children’s social development. Journal of Marriage and Family, 70, 1258 –1270. http://dx.doi.org/10.1111/j.1741-3737.2008.00564.x

Chen, P., & Vazsonyi, A. T. (2011). Future orientation, impulsivity, and problem behaviors: A longitudinal moderation model. Developmental Psychology, 47, 1633–1645. http://dx.doi.org/10.1037/a0025327

Cherlin, A. (2008). Multiple partnerships and children’s wellbeing. Family Matters, 80, 33–36.

Cherlin, A. J. (2009). The marriage-go-round: The state of marriage and the family in America today. New York, NY: Random House.

Clinkinbeard, S. S. (2014). What lies ahead: An exploration of future orientation, self-control, and delinquency. Criminal Justice Review, 39, 19 –36. http://dx.doi.org/10.1177/0734016813501193

Cohen, J., & Manning, W. (2010). The relationship context of premarital serial cohabitation. Social Science Research, 39, 766 –776. http://dx.doi .org/10.1016/j.ssresearch.2010.04.011

Cohen, S., Klein, D. N., & O’Leary, K. D. (2007). The role of separation/ divorce in relapse into and recovery from major depression. Journal of Social and Personal Relationships, 24, 855– 873. http://dx.doi.org/10 .1177/0265407507084187

Collins, W. A., Hennighausen, K. C., Schmit, D. T., & Sroufe, L. A. (1997). Developmental precursors of romantic relationships: A longitu- dinal analysis. New Directions for Child and Adolescent Development, 1997, 69 – 84. http://dx.doi.org/10.1002/cd.23219977807

Conger, R. D., Cui, M., Bryant, C. M., & Elder, G. H., Jr. (2000). Competence in early adult romantic relationships: A developmental perspective on family influences. Journal of Personality and Social Psychology, 79, 224 –237. http://dx.doi.org/10.1037/0022-3514.79.2 .224

Copen, C. E., Chandra, A., & Martinez, G. (2012). Prevalence and timing of oral sex with opposite-sex partners among females and males aged 15–24 years: United States, 2007–2010. National Health Statistics Re- ports, 56, 1–14.

Copp, J. E., Kuhl, D. C., Giordano, P. C., Longmore, M. A., & Manning, W. D. (2015). Intimate partner violence in neighborhood context: The roles of structural disadvantage, subjective disorder, and emotional distress. Social Science Research, 53, 59 –72. http://dx.doi.org/10.1016/ j.ssresearch.2015.05.001

Cui, M., Fincham, F. D., & Durtschi, J. A. (2011). The effect of parental divorce on young adults’ romantic relationship dissolution: What makes a difference? Personal Relationships, 18, 410 – 426. http://dx.doi.org/10 .1111/j.1475-6811.2010.01306.x

Dean, D. O., Bauer, D. J., & Shanahan, M. J. (2014). A discrete-time multiple event process survival mixture (MEPSUM) model. Psycholog- ical Methods, 19, 251–264. http://dx.doi.org/10.1037/a0034281

Donnellan, M. B., Conger, K. J., McAdams, K. K., & Neppl, T. K. (2009). Personal characteristics and resilience to economic hardship and its consequences: Conceptual issues and empirical illustrations. Journal of Personality, 77, 1645–1676. http://dx.doi.org/10.1111/j.1467-6494.2009 .00596.x

Elder, G. H., Jr. (1998). The life course as developmental theory. Child Development, 69, 1–12. http://dx.doi.org/10.1111/j.1467-8624.1998 .tb06128.x

Elder, G. H., Jr., George, L. K., & Shanahan, M. J. (1996). Psychosocial stress over the life course. In H. G. Kaplan (Ed.), Psychosocial stress: Perspectives on structure, theory, life course, and methods (pp. 247– 292). Orlando, FL: Academic Press.

Enders, C. K., & Bandalos, D. L. (2001). The relative performance of full information maximum likelihood estimation for missing data in struc-

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

10 BAE AND WICKRAMA

tural equation models. Structural Equation Modeling, 8, 430 – 457. http://dx.doi.org/10.1207/S15328007SEM0803_5

Evans, G. W., & Kim, P. (2013). Childhood poverty, chronic stress, self-regulation, and coping. Child Development Perspectives, 7, 43– 48. http://dx.doi.org/10.1111/cdep.12013

Evans, G. W., Li, D., & Whipple, S. S. (2013). Cumulative risk and child development. Psychological Bulletin, 139, 1342–1396. http://dx.doi.org/ 10.1037/a0031808

Greene, B. A., & DeBacker, T. K. (2004). Gender and orientations toward the future: Links to motivation. Educational Psychology Review, 16, 91–120. http://dx.doi.org/10.1023/B:EDPR.0000026608.50611.b4

Härkönen, J., & Dronkers, J. (2006). Stability and change in the educa- tional gradient of divorce. A comparison of seventeen countries. Euro- pean Sociological Review, 22, 501–517. http://dx.doi.org/10.1093/esr/ jcl011

Jung, T., & Wickrama, K. A. S. (2008). An introduction to latent class growth analysis and growth mixture modeling. Social and Personality Psychology Compass, 2, 302–317. http://dx.doi.org/10.1111/j.1751- 9004.2007.00054.x

Kamp Dush, C. M. (2011). Relationship-specific investments, family chaos, and cohabitation dissolution following a nonmarital birth. Family Relations: An Interdisciplinary Journal of Applied Family Studies, 60, 586 – 601. http://dx.doi.org/10.1111/j.1741-3729.2011.00672.x

Karney, B. R., & Bradbury, T. N. (1995). The longitudinal course of marital quality and stability: A review of theory, method, and research. Psychological Bulletin, 118, 3–34. http://dx.doi.org/10.1037/0033-2909 .118.1.3

Lichter, D. T., & Qian, Z. (2008). Serial cohabitation and the marital life course. Journal of Marriage and Family, 70, 861– 878. http://dx.doi.org/ 10.1111/j.1741-3737.2008.00532.x

Lichter, D. T., Turner, R. N., & Sassler, S. (2010). National estimates of the rise in serial cohabitation. Social Science Research, 39, 754 –765. http:// dx.doi.org/10.1016/j.ssresearch.2009.11.002

Madigan, S., Atkinson, L., Laurin, K., & Benoit, D. (2013). Attachment and internalizing behavior in early childhood: A meta-analysis. Devel- opmental Psychology, 49, 672– 689. http://dx.doi.org/10.1037/a0028793

Madsen, S. D., & Collins, W. A. (2011). The salience of adolescent romantic experiences for romantic relationship qualities in young adult- hood. Journal of Research on Adolescence, 21, 789 – 801. http://dx.doi .org/10.1111/j.1532-7795.2011.00737.x

Mancini, J. A., & Bowen, G. L. (2013). Families and communities: A social organization theory of action and change. In G. W. Peterson & K. R. Bush (Eds.), Handbook of marriage and the family (pp. 781– 813). New York, NY: Springer. http://dx.doi.org/10.1007/978-1-4614-3987- 5_32

Manning, W. D. (2013). Trends in cohabitation: Over twenty years of change, 1987–2010 (Family Profiles FP-13–12). Bowling Green, OH: National Center for Family & Marriage Research. Retrieved from https://www.bgsu .edu/content/dam/BGSU/college-of-arts-and-sciences/NCFMR/documents/ FP/FP-13-12.pdf

Meier, A., & Allen, G. (2009). Romantic relationships from adolescence to young adulthood: Evidence from the National Longitudinal Study of Adolescent Health. The Sociological Quarterly, 50, 308 –335. http://dx .doi.org/10.1111/j.1533-8525.2009.01142.x

Muthén, L. K., & Muthén, B. O. (1998 –2012). Mplus user’s guide (7th ed.). Los Angeles, CA: Author.

Olderbak, S., & Figueredo, A. J. (2009). Predicting romantic relationship satisfaction from life history strategy. Personality and Individual Dif- ferences, 46, 604 – 610. http://dx.doi.org/10.1016/j.paid.2008.12.019

Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385– 401. http://dx.doi.org/10.1177/014662167700100306

Reifman, A. (2011). Romantic relationships in emerging adulthood: Con- ceptual foundations. In F. D. Fincham & M. Cui (Eds.), Romantic relationships in emerging adulthood (pp. 15–26). New York, NY: Cam- bridge University Press.

Sassler, S. (2010). Partnering across the life course: Sex, relationships, and mate selection. Journal of Marriage and Family, 72, 557–575. http://dx .doi.org/10.1111/j.1741-3737.2010.00718.x

Shanahan, M. J. (2000). Pathways to adulthood in changing societies: Variability and mechanisms in life-course perspective. Annual Review of Sociology, 26, 667– 692. http://dx.doi.org/10.1146/annurev.soc.26.1.667

Sieger, K., Rojas-Vilches, A., McKinney, C., & Renk, K. (2004). The effects and treatment of community violence in children and adoles- cents: What should be done? Trauma, Violence, & Abuse, 5, 243–259. http://dx.doi.org/10.1177/1524838004264342

Simons, R. L., Simons, L. G., Lei, M. K., & Landor, A. (2012). Relational schemas, hostile romantic relationships, and beliefs about marriage among young African American adults. Journal of Social and Personal Relation- ships, 29, 77–101. http://dx.doi.org/10.1177/0265407511406897

Straus, M. A., Gelles, R. J., & Steinmetz, S. K. (1980). Behind closed doors: Violence in the American family. Garden City, NJ: Anchor Press.

Sucoff, C. A., & Upchurch, D. M. (1998). Neighborhood context and the risk of childbearing among metropolitan-area black adolescents. Amer- ican Sociological Review, 63, 571–585. http://dx.doi.org/10.2307/ 2657268

Umberson, D., Crosnoe, R., & Reczek, C. (2010). Social relationships and health behavior across the life course. Annual Review of Sociology, 36, 139 –157. http://dx.doi.org/10.1146/annurev-soc-070308-120011

Valle, G., & Tillman, K. H. (2014). Childhood family structure and romantic relationships during the transition to adulthood. Journal of Family Issues, 35, 97–124. http://dx.doi.org/10.1177/0192513X12463555

Ventura, J., Nuechterlein, K. H., Subotnik, K. L., Hardesty, J. P., & Mintz, J. (2000). Life events can trigger depressive exacerbation in the early course of schizophrenia. Journal of Abnormal Psychology, 109, 139 – 144. http://dx.doi.org/10.1037/0021-843X.109.1.139

Wheaton, B. (1990). Life transitions, role histories, and mental health. American Sociological Review, 55, 209 –223. http://dx.doi.org/10.2307/ 2095627

Wickrama, K. A., O’Neal, C. W., & Lee, T. K. (2016). The health impact of upward mobility: Does socioeconomic attainment make youth more vulnerable to stressful circumstances? Journal of Youth and Adoles- cence, 45, 271–285. http://dx.doi.org/10.1007/s10964-015-0397-7

Wickrama, K. A. S., Kwon, J. A., Oshri, A., & Lee, T. K. (2014). Early socioeconomic adversity and young adult physical illness: The role of body mass index and depressive symptoms. Journal of Adolescent Health, 55, 556 –563. http://dx.doi.org/10.1016/j.jadohealth.2014.04 .006

Wickrama, K. K. A. S., & O’Neal, C. W. (2016). The socioeconomic pathways leading to romantic relationship outcomes: A genetically in- formed early life course investigation. Journal of Research on Adoles- cence, 26, 492–508. http://dx.doi.org/10.1111/jora.12207

Wu, Z., & Schimmele, C. M. (2005). Repartnering after first union dis- ruption. Journal of Marriage and Family, 67, 27–36. http://dx.doi.org/ 10.1111/j.0022-2445.2005.00003.x

Received September 20, 2017 Revision received June 7, 2018

Accepted July 6, 2018 �

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

11ROMANTIC RELATIONSHIP-DISSOLUTION TRAJECTORIES