Adolescent Dating Violence Victimization and Relationship Dissolution

Abstract

We integrate literature on gender and adolescent friendships to examine the association between adolescent dating violence victimization (ADVV) and relationship dissolution. In particular, we test whether ADVV increases the hazard of relationship dissolution among adolescent romances, and whether a number of friendship dynamics alter the association between ADVV and relationship dissolution. Using discrete time event history models from 5,787 romantically involved youth from the National Longitudinal Study of Adolescent to Adult Health (Add Health), results indicated, on average, ADVV was not associated with the hazard of relationship dissolution for girls or boys. However, the positive effect of ADVV was stronger for girls who did not withdraw from their friendships over the course of their romantic relationships. This study highlights the importance of peer groups and gender in shaping youths’ decisions to exit abusive relationships.

Keywords

teen dating violence peer relationships relationship dissolution

Introduction

Adolescent dating violence victimization (hereafter ADVV) is a significant public health concern. Estimates from nationally representative samples of high-school-aged youth in the United States indicate 35% of youth reported experiencing dating violence (either psychological or physical) victimization (Haynie et al., 2013). While research has focused on the causes and consequences of exposure to ADVV, much less attention has focused on identifying factors associated with the dissolution of violent intimate partner relationships. That is, what factors affect the likelihood adolescents exit violent relationships?

This study explored the association between ADVV and relationship dissolution. In particular, we drew from sociological literature on adolescent peer contexts and gender to theorize the role friendships play in informing the association between ADVV and the subsequent likelihood of dissolving romantic relationships. Using data from the National Longitudinal Study of Adolescent to Adult Health (hereafter Add Health), we tested the association between ADVV and relationship dissolution. Specifically, we focused on physical ADVV perpetrated by the respondent’s partner. We also examined whether friendship processes alter the association between ADVV and relationship dissolution. In particular, we tested whether the association between ADVV and relationship dissolution varied by whether youth withdraw from friendships over the course of the romance, whether relationship partners were friends prior to relationship onset, whether partners shared mutual friends prior to the relationship onset, and the strength of one’s friendship attachments prior to relationship onset. Finally, we examined whether these processes differed by gender. Our results underscore the importance of attending to gender and peer groups when examining the association between ADVV and relationship dissolution.

ADVV and Relationship Dissolution

Numerous studies highlight why some adults remain in abusive relationships, while others exit. For instance, fettered access to economic or social capital helps explain why some remain in abusive relationships (Anderson, 2007; Kim & Gray, 2008). Other research suggests the presence of children often keeps parents in abusive relationships (Rhodes, Cerulli, Dichter, Kothari, & Barg, 2010). Still, other studies focus on the strength of relationships and the amount of time and emotional investment that characterizes relationships to explain why some exit or remain in abusive relationships (Rhatigan & Axson, 2006; Rhoades, Stanley, Kelmer, & Markman, 2010).

While prior research helps explain why adults leave or remain in abusive relationships, it is doubtful these findings fully extend to adolescents. For instance, economic inequality among partners is unlikely to affect leave/stay decisions among adolescents because most adolescents are not dependent on their partners for their own economic well-being. Adolescents are also less likely than adults to have children with romantic partners. Thus, parental considerations likely do not factor into the association between dating violence and relationship dissolution among adolescents. Finally, although youth romances may involve strong emotions (Larson et al., 1999), they tend to be shorter in duration when compared with adult intimate relationships. Thus, relationship length—which informs adult leave/stay decisions—is not likely to be a strong factor in adolescents’ decisions to stay or leave relationships.

Emergent research on dating violence and relationship dissolution among youth has focused on the potential for peer networks to shape decisions to leave or stay in abusive relationships, for good reason (Helm, Baker, Berlin, & Kimura, 2015). First, compared with adults and younger children, adolescents spend more time socializing with peers, and are more strongly influenced by peers’ behavior and attitudes (Giordano, 2003). Adolescents’ romantic relationships also often begin as friendships or emerge from overlapping peer groups (Connolly, Craig, Goldberg, & Pepler, 2004), and friends influence adolescent romantic relationship dynamics in numerous ways (Brown, 1999; Connolly, Furman, & Konarski, 2000; Furman, 1999). Below, we theorize four key ways in which peers and friendship processes may factor into the association between ADVV and relationship dissolution. We focus specifically on whether respondents withdrew from friendships during their relationship, respondents’ friendship strength prior to the initiation of their romantic relationship, whether respondents were friends with their partners before relationship onset, and whether the respondents and partners shared mutual friends.

Peers and Leaving Violent Relationships

Peers are integral to adolescence (Coleman, 1961). In early adolescence, youth establish independence from parents while peer relationships take on an increasingly important role in their lives (Steinberg, 2001). Close friendships are critical to the development of interpersonal skills including intimacy, empathy, and perspective-taking (Sullivan, 1953), and provide a primary source of social support and informal social control (Brown, 1999). Thus, adolescents emphasize socializing with peers and are highly susceptible to peer influence during this period (Brown & Larson, 2009).

Research suggests peer relationships are both shaped and altered by adolescent romantic experiences. Early romantic relationships often emerge from peer relationships (Connolly et al., 2000; Dunphy, 1963). Research also indicates the nature and quality of adolescent romantic relationships are related to features of youths’ close friendships and larger peer groups (Furman & Wehner, 1994). Others argue youth encounter pressures from friends and peers to form and maintain relationships with socially acceptable others (Brown, 1999). Thus, peers exercise control over the dynamics of youth relationships, particularly early on in romantic relationships. However, adolescents often withdraw from friends (i.e., sacrifice time spent with friends) as relationships progress (Brown, 1999; Laursen & Williams, 1997). Thus, friendship withdrawal is a normative part of relationship development and may reflect increasing levels of commitment and attachment to a romantic partner (Kuttler & La Greca, 2004).

Withdrawing from friendships over the course of romances may influence decisions regarding whether to stay/leave abusive relationships. Adolescents’ and emerging adults’ decisions to leave violent relationships are influenced by peers’ opinions of the partner (Martsolf, Druacker, & Brandau, 2013), and peers’ negative appraisals of the romantic partner in particular (Copp, Giordano, Longmore, & Manning, 2015). Research also suggests friends play a pivotal role in informing youths’ stay/leave decisions by providing social support following instances of relationship violence (Helm et al., 2015). Furthermore, adolescents often discuss relationship matters with peers, which shapes attitudes about acceptable relationship behaviors, as well as norms of romantic interaction more generally (Connolly & Friedlander, 2009).

Adolescents who withdraw from friendships may lack the support needed to leave a relationship exhibiting violence and may perceive fewer quality alternatives to the relationship (Young & Furman, 2013). Youth who withdraw from friendships over the course of a relationship also have fewer opportunities to discuss relationship events (e.g., partner violence) with friends. Thus, withdrawing from friendships limits friends’ knowledge of the inner workings of one’s relationship and inhibits friends’ capacity to form negative appraisals of one’s abusive partners. Such disconnects also decrease the likelihood friends can or will advise or encourage youth to exit abusive relationships (Brown, 1999). As such, withdrawing from friendships may lessen the odds of relationship dissolution among youth with ADVV exposure.

Friendship strength prior to the onset of a relationship may also influence decisions regarding whether to exit abusive relationships. First, as close friendships often entail intimate self-disclosure, peers may play an important role in stay/leave decisions by communicating messages of disapproval regarding the romantic partner. Thus, even if youth withdraw from friendships as their romantic relationship progresses, friends tied through strong bonds prior to the relationship onset will likely have more access to information surrounding the relationship and be better equipped to provide support and exercise control over relationships compared with those who were weakly tied to friends prior to the relationship onset. In this sense, youth tied through strong bonds may be made more acutely aware of partner violence in their friends’ relationship and in turn encourage relationship dissolution or express negative appraisals of friends’ abusive partners. Thus, strong friendship attachment prior to relationship onset may heighten the odds of relationship dissolution in violent relationships.

Research on the progression of adolescent romances suggests romantic relationships often emerge from overlapping friendship groups (Connolly et al., 2004). As a result, adolescents often begin dating in group settings. Researchers have indicated group dating may present a “developmental advantage,” as relationship dynamics tend to develop alongside friends’ supportive interactions (Foshee et al., 2013). However, as scholars tend to agree supportive peers are protective in many ways, in-group romantic relationships may be particularly difficult to dissolve as they may threaten friendships and group cohesion (Brown, 1999). Exiting a violent relationship may thus be more difficult in this situation as friends will have divided loyalties, resulting in less support provided to the victim. Thus, being friends with one’s romantic partner prior to relationship onset may lessen the odds of relationship dissolution among adolescents with ADVV. In addition, having overlapping friendships with one’s partner prior to relationship onset may lessen the odds of relationship dissolution among those experiencing ADVV.

Girls and Women Leaving Abusive Relationships

There are reasons to anticipate gender alters the association between peer withdrawal, ADVV, and relationship dissolution. For instance, due to gender socialization, girls place more importance on having and maintaining romantic relationships than boys (Eder, Evans, & Parker, 1995; Gilligan, 1982; Soller, 2014; Steffensmeier, Schwartz, & Roche, 2013). In this case, girls may need more convincing from friends to dissolve a violent or abusive relationship compared with boys. Related, girls are more likely to discuss relationship matters with friends including experiences of partner violence than are boys (Sylaska & Edwards, 2014). Without these friendships in place, girls may lack the support necessary to end a romantic relationship. Moreover, abusive male partners may strategically isolate female partners by limiting or ending their friendships with others, in essence removing their ability to turn to others for aid when abuse occurs (Johnson, 2010). Therefore, we hypothesize withdrawal from friendships will be associated with a lower chance of girls leaving an abusive relationship than it will be for boys. There is considerable evidence to suggest girls are more attached to their peers, display higher levels of trust, and a greater willingness to engage in intimate self-disclosure (Giordano, Longmore, & Manning, 2006; Gorrese & Ruggieri, 2012). Furthermore, girls are more likely to draw on peers for social support (Ma & Huebner, 2008). Taken together, these findings suggest girls may be more likely to disclose ADVV to peers, giving peers the opportunity to communicate disapproval of the partner, and potentially influencing the dissolution of relationships characterized by violence. Thus, we hypothesize girls will be even more likely than boys to exit abusive partner relationships when they have strong friendships in place.

The association between friendship overlap, DVV, and relationship dissolution may also vary by gender. When friendships are precursors to romantic relationships, it is likely partners’ friendship networks will share more overlap. Because girls are socialized to be more attendant to and emotionally invested in interpersonal relationships (Gilligan, 1982; Steffensmeier, Schwartz, & Roche, 2013), they may be more concerned about how the dissolution of their romantic relationship will affect their friendships, especially when those friendships are common to both partners. Thus, being friends with one’s partner prior to relationship onset may attenuate the positive effect of ADVV on the odds of relationship dissolution to a greater extent for girls compared with boys. In addition, having overlapping friendships with one’s partner prior to relationship onset may attenuate the positive effect of ADVV on the odds of relationship dissolution more strongly for girls compared with boys.

The Present Study

We examined the association between ADVV and romantic relationship dissolution. We first tested whether ADVV was associated with girls’ and boys’ hazard of relationship dissolution. We also examined whether friendship processes altered the association between ADVV and relationship dissolution. In particular, we tested whether the effect of ADVV on dissolution varies according to whether a youth withdrew from friends over the course of the relationship, the friendship attachment prior to relationship onset, whether one was friends with the partner prior to relationship onset, and whether youth shared mutual friends with his or her partner after accounting for potential confounding factors. Finally, drawing from research on gender and interpersonal relationships, we tested whether the association between ADVV, friendship dynamics, and relationship dissolution varies by gender.

Method

Sample

Add Health comprises a nationally representative sample of youth from the United States who were in middle or high school at the time of the Wave 1 interview, which was conducted in 1995. Roughly 15,000 respondents participated in the first two waves of data collection. Our analysis focuses on youth who indicated at Wave 2 that they formed a new romantic relationship after the date of the Wave 1 interview (the Wave 2 interview took place roughly 1 year after Wave 1). Among those respondents, we exclude respondents who were missing survey weights or information on the date of relationship dissolution (if broken up) or dates of relationship aggression. In total, our sample includes 5,787 respondents. We display descriptive statistics for all variables in our analysis in Table 1. In total, 54% of our sample were female, 57% were White, 21% were Black, 16% were Latina/o, and 5% were of another race/ethnicity. Among our female respondents, 58% were White, 22% were Black, 15% were Latina, and 5% were of another race/ethnicity. Among boys in our sample, 57% were White, 20% were Black, 17% were Latino, and 5% were of another race/ethnicity. The mean age among all respondents is 15.9 years, while the mean age among girls is 15.77 years and 16.05 among boys.

Table 1.

Descriptive Statistics (N = 5,787).

Variable	Total sample		Girls		Boys
Variable	M/proportion	(SD)	M/proportion	(SD)	M/proportion	(SD)
Relationship has ended	0.47		0.50		0.43
Dating violence victimization	0.07		0.06		0.08
Peer processes
Friendship withdrawal	0.44		0.43		0.46
Friendship attachment	0.00	(0.74)	0.05	(0.75)	−0.05	(0.72)
Shared mutual friends	0.40		0.45		0.35
Friends before relationship	0.40		0.43		0.37
Personal characteristics
Age	15.90	(1.50)	15.77	(1.48)	16.05	(1.50)
Female	0.54
Race/ethnicity
White (reference category)	0.57		0.58		0.57
Black	0.21		0.22		0.20
Latina/o	0.16		0.15		0.17
Other	0.05		0.05		0.05
Socioeconomic status	0.02	(0.87)	−0.02	(0.87)	0.06	(0.87)
Single parent household	0.29		0.30		0.29
Violent delinquency	0.23	(0.36)	0.15	(0.27)	0.33	(0.42)
Relationship characteristics
Same race	0.26		0.27		0.26
Age dissimilarity
Same age (reference category)	0.74		0.71		0.79
Partner is older	0.16		0.26		0.04
Respondent is older	0.10		0.03		0.17
Had sexual intercourse	0.41		0.41		0.41

Dependent Variable

We focus on the interactive effects of experiencing partner violence and friendship dynamics on the hazard of romantic relationship dissolution using discrete time event history models (we provide more information on our modeling strategy below). At Wave 2, respondents indicated whether they formed a “special romantic relationship” with someone after the Wave 1 interview. Those answering “yes” provided detailed information on up to three “special romantic relationships,” including the month and year of relationship onset and dissolution if the relationship ended. Those who did not identify a special romantic relationship were asked whether they told another person they liked/loved him or her, kissed, or held hands with someone after Wave 1. Those who engaged in all of these behaviors identified up to three “liked” relationship partners and provided detailed information for each partnership. We focus on the first “special” or “liked” romantic relationship that occurred after the date of the Wave 1 interview to measure our dependent variable. Our dependent variable is binary and indicates the respondent’s relationship status at observation period t (0 = still together, 1 = broken up). Those who were still in a relationship with their partner at the date of the Wave 2 interview were right censored, and are coded 0 for all observation months.

Independent Variables

Our first independent variable captures whether the respondent experienced dating violence victimization in the relationship at month t. At Wave 2, respondents answered questions on dating violence victimization within each relationship from a subset of items from the Conflict Tactics Scale–Revised (CTS-R; Straus, Hamby, Boney-McCoy, & Sugarman, 1996). Respondents indicated whether each liked or special romantic partner had ever (a) thrown something at them that could hurt them or (b) pushed or shoved them. Respondents then provided the month and year that each item first occurred. Using this information, we constructed a time-varying measure of dating violence victimization, which indicates whether the respondent experienced any form of physical violence during or prior to month t in the relationship. As we are concerned with how the onset of dating violence victimization relates to the hazard of relationship dissolution, and changes in the level of violence victimization over time were not measured, dating violence victimization is included as a binary variable.

We constructed four measures capturing respondents’ friendship dynamics before the relationship onset and throughout the relationship. Our first measure was friendship withdrawal, which was assessed using responses to the following item assessed at Wave 2, indicating whether the respondent “saw less of other friends so you could spend more time with your partner” (0 = no, 1 = yes). Our second friendship measure captured the strength of friendship ties prior to the relationship onset. This measure friendship attachment was measured at Wave 1 and is based on the following items: “during the past week, how many times did you just hang out with friends” (responses ranged from 0 = not at all to 3 = 5 or more times) and “how much do you feel that your friends care about you?” (responses ranged from 1 = not at all to 5 = very much). Our measure of prior friendship attachment represents the mean of the standardized items.

We also focus on the friendship patterns of respondents and their partners prior to the relationship onset. Our third friendship measure, friends before relationship, was measured during the Wave 2 interview. This measure is binary and indicates whether the respondent reported being friends with his or her partner prior to the onset of the relationship (0 = no, 1 = yes). Our final friendship measure, shared mutual friends, was also measured at Wave 2 and indicates whether the respondent and his or her partner had mutual friends prior to the relationship onset (0 = no, 1 = yes).

All models adjust for control variables and potential confounding factors. We include a number of variables that capture features of individuals and relationships that may influence both partner aggression and relationship dissolution. First, sexual activity within romantic relationships has been associated with ADVV (Kaestle & Halpern, 2005) and may confound the relationship between ADVV and relationship dissolution. Therefore, we construct a measure of had sexual intercourse. We use information regarding the month and year in which respondents first had sexual intercourse with the partner to construct a binary, time-varying measure of sexual intercourse indicating whether the respondent had sexual intercourse in the relationship by month t (0 = no, 1 = yes).

We include three variables that capture demographic differences between respondents and partners as they may increase the risk of romantic relationship dissolution. First, relationship dissolution might be more likely among couples who are not members of the same racial/ethnic group (Wang, Kao, & Joyner, 2006). Our measure of same race/ethnicity is binary and indicates whether the respondent and partner are members of the same race/ethnic group (0 = no, 1 = yes). Given the association between adolescent involvement with an older partner and risk behavior, we assess age discordance with two binary measures (Halpern, Spriggs, Martin, & Kupper, 2009). First, partner is older is binary and indicates whether the respondent is 2 or more years younger than his or her partner (0 = no, 1 = yes). The second measure, respondent is older is binary and indicates the respondent is 2 or more years older than the partner (0 = no, 1 = yes). Same age (i.e., the respondent and partner are within 2 years of age of each other) serves as the reference category.

Respondents were not asked about ADV perpetration during the Wave 2 interview. Therefore, we are unable to measure the association between adolescents’ own dating violence and relationship dissolution. We did include a measure of personal violence, which was based on seven items capturing respondents’ involvement in violent acts, such as fighting and pulling a knife or gun on someone throughout the past year (measured at Wave 1). Our measure of personal violence captured the proportion of items the respondent performed prior to Wave 1 (alpha = .785).

All models included demographic variables, including age (in years at relationship onset) and race/ethnicity, which was captured by a series of dummy variables indicating whether the respondent is Black, Latino/a, or other (0 = no, 1 = yes; White is the reference category). We included a measure indicating the respondent resides in a single parent household (0 = no, 1 = yes) and the household socioeconomic status, which represents the standardized means of the highest parental education and highest parental occupational status.

Analytic Approach

We utilize multilevel discrete time event history models to test the association between ADVV, friendship dynamics, and relationship dissolution (Browning, Leventhal, & Brooks-Gunn, 2005). Importantly, we estimate gender-interacted effects for all independent variables in our analysis, except for the dummy variables capturing duration dependence. Doing so allows us to measure the strengths of the interactive associations between friendship processes, DVV, and relationship dissolution among girls and boys. Using models of relationship dissolution regressed on DVV, friendship withdrawal, and control variables, our Level 1 may be expressed as

\begin{array}{l} l o g (\frac{P_{t i j}}{1 - P_{t i j}}) = D_{F i j} (π_{F 0 i j} + π_{F 1 i j} D V V_{t i j} + π_{F 2 i j} S E X_{t i j}) \\ + D_{M i j} (π_{M 0 i j} + π_{M 1 i j} D V V_{t i j} + π_{M 2 i j} S E X_{t i j}) + \sum_{t = 3}^{T} α_{t} M O N T H_{t i j}, \end{array}

where D_Fij and D_Mij are binary variables that respectively indicate whether the respondent is female or male (0 = no, 1 = yes), π_F_0ij and π_M_0ij are female- and male-specific person-level intercepts (respectively), π_F_1ij and π_M_1ij are the female- and male-specific effects of dating violence victimization for person i in school j, and DVV_tij is a binary variable indicating whether ADVV has occurred by month t (0 = no, 1 = yes). Coefficients π_F₂_ij and π_M2ij capture the effects of having had sexual intercourse for girls and boys (respectively), SEX_tij is a binary variable indicating whether the respondent has had sexual intercourse with his or her partner by month t, α_t captures the effect of the relationship being in month t (i.e., adjusts for relationship duration), and MONTH_tij is a dummy variable indicating whether the observation refers to month t.

The person-level intercepts are modeled as follows:

\begin{array}{l} π_{F 0 i j} = D_{F i j} (β_{F 00 j} + β_{F 01 j} W D_{i j} + \sum_{q = 2}^{Q} β_{F 0 q j} X_{q i j} + r_{F 0 i j}) \\ π_{M 0 i j} = D_{M i j} (β_{M 00 j} + β_{M 01 j} W D_{i j} + \sum_{q = 2}^{Q} β_{M 0 q j} X_{q i j} + r_{M 0 i j}) \end{array}

where β_F_00j and β_M_00j are gender-specific school intercepts, β_F_01j and β_M_01j, respectively, capture the effects of friendship withdrawal among girls and boys, and WD_ij is a binary variable indicating whether the respondent withdrew from her or his friendships over the course of the relationship. Negative values for β_F_01j and β_M_01j would indicate friendship withdrawal decreases the hazard of relationship dissolution. Coefficients β_F_0qj and β_M_0qj capture the effects of covariate q on dissolution among girls and boys (respectively), and X_qij is the value of covariate q for respondent i in school j. Finally, r_F_0ij and r_M_0ij are gender-specific person-level random effects.

The final Level 2 equations test whether friendship withdrawal alters the gender-specific DVV slopes and test our primary hypotheses:

\begin{array}{l} π_{F 1 i j} = D_{F i j} (β_{F 10 j} + β_{F 11 j} W D_{i j}) \\ π_{M 1 i j} = D_{M i j} (β_{M 10 j} + β_{M 11 j} W D_{i j}) \end{array}

where β_F_10j and β_M_10j are gender-specific school intercepts, β_F_11j and β_M_11j capture the effects of friendship withdrawal among girls and boys on the DVV slope. Negative values on β_F_11j and β_M_11j indicate withdrawal from friendships is associated with a lower chance of leaving an abusive relationship among girls and/or boys (respectively).

Missing values on independent variables were imputed, using ICE in Stata13 (Royston, 2004). Respondents with missing values on relationship dissolution were excluded from our statistical models. Models were estimated based on 10 imputed data sets, using the multiple imputation procedure in HLM7. We also applied individual- and school-level survey weights that account for Add Health’s complex survey design. Finally, we centered each control variable around its grand mean to aid in interpretation of results.

Results

We present results from discrete time event history models of relationship dissolution in Table 2. We first examined the effects of friendship processes and dating violence victimization on the hazard of relationship dissolution in Model 1. We found sharing mutual friends and being friends with one’s partner before the relationship onset is negatively associated with the hazard of relationship dissolution for girls (b = −0.17, p = .011 and b = −0.26, p = .018, respectively). Conversely, being friends prior to relationship onset was positively associated with the hazard of relationship dissolution for boys (b = 0.20, p = .041). We also found DVV was not significantly associated with the hazard of relationship dissolution for girls or boys in Model 1.

Table 2.

Multilevel Discrete Time Event History Models of Relationship Dissolution.

	Model 1		Model 2		Model 3		Model 4		Model 5
	Girls	Boys	Girls	Boys	Girls	Boys	Girls	Boys	Girls	Boys
Friendship withdrawal	.09	.03	.13	−.01	.09	.02	.09	.03	.09	.03
Friendship withdrawal	(.10)	(.10)	(.10)	(.10)	(.10)	(.10)	(.10)	(.10)	(.10)	(.10)
Friendship attachment	−.01	.03	−.02	.02	−.02	.01	−.02	.03	−.01	.03
Friendship attachment	(.08)	(.06)	(.08)	(.06)	(.07)	(.06)	(.08)	(.06)	(.08)	(.06)
Shared mutual friends	−.17*	−.06	−.18**	−.05	−.18**	−.06	−.21**	−.05	−.17*	−.06
Shared mutual friends	(.07)	(.12)	(.07)	(.12)	(.07)	(.12)	(.07)	(.12)	(.07)	(.12)
Friends before relationship	−.26*	.20*	−.26*	.21*	−.26*	.21*	−.25*	.20*	−.28**	.19
Friends before relationship	(.11)	(.10)	(.11)	(.10)	(.11)	(.10)	(.11)	(.10)	(.11)	(.11)
DVV	.05	.29	.91*	−.30	.07	.29	−.52	.38	−.28	.15
DVV	(.36)	(.27)	(.40)	(.38)	(.31)	(.27)	(.64)	(.34)	(.49)	(.27)
DVV × Friendship withdrawal			−1.39*	.92
DVV × Friendship withdrawal			(.58)	(.49)
DVV × Friendship attachment					.39	.46
DVV × Friendship attachment					(.50)	(.32)
DVV × Shared mutual friends							1.00	−.15
DVV × Shared mutual friends							(.67)	(.45)
DVV × Friends before relationship									.90	.32
DVV × Friends before relationship									(.66)	(.48)
Age	−.11**	−.05	−.11**	−.04	−.11**	−.04	−.11**	−.05	−.11**	−.05
Age	(.03)	(.04)	(.03)	(.04)	(.03)	(.04)	(.03)	(.04)	(.03)	(.04)
Black	−.05	−.33	−.04	−.32	−.05	−.33	−.04	−.33	−.07	−.32
Black	(.26)	(.20)	(.27)	(.20)	(.27)	(.20)	(.27)	(.20)	(.26)	(.20)
Latina/o	−.29	−.09	−.27	−.10	−.28	−.09	−.28	−.09	−.31	−.09
Latina/o	(.21)	(.19)	(.21)	(.19)	(.21)	(.19)	(.21)	(.19)	(.21)	(.19)
Other	−.10	−.41	−.09	−.42	−.11	−.41	−.12	−.41	−.09	−.41
Other	(.16)	(.24)	(.16)	(.24)	(.16)	(.24)	(.16)	(.24)	(.16)	(.24)
Socioeconomic status	−.09*	−.12	−.09	−.12	−.09	−.12	−.09	−.12	−.09	−.12
Socioeconomic status	(.05)	(.07)	(.05)	(.07)	(.05)	(.07)	(.05)	(.07)	(.05)	(.07)
Single parent household	−.09	.18	−.08	.18	−.09	.19	−.09	.18	−.08	.18
Single parent household	(.11)	(.17)	(.11)	(.17)	(.10)	(.17)	(.11)	(.17)	(.10)	(.17)
Violent delinquency	.00	−.22	.00	−.22	.00	−.22	−.01	−.21	.00	−.22
Violent delinquency	(.18)	(.14)	(.18)	(.14)	(.18)	(.14)	(.18)	(.14)	(.18)	(.14)
Same race partner	.09	.25	.06	.25	.08	.26	.08	.25	.11	.25
Same race partner	(.29)	(.17)	(.30)	(.16)	(.30)	(.16)	(.30)	(.17)	(.29)	(.17)
Partner is older	−.15	−.04	−.14	−.03	−.14	−.06	−.15	−.04	−.14	−.04
Partner is older	(.11)	(.26)	(.11)	(.26)	(.11)	(.26)	(.11)	(.26)	(.11)	(.26)
Respondent is older	.46	.75**	.47*	.75**	.47*	.75**	.46*	.74**	.46*	.75**
Respondent is older	(.24)	(.15)	(.24)	(.15)	(.23)	(.15)	(.23)	(.15)	(.23)	(.15)
Had sexual intercourse	.26*	.45**	.29**	.45**	.26*	.44**	.26*	.45**	.26*	.45**
Had sexual intercourse	(.11)	(.15)	(.10)	(.15)	(.10)	(.15)	(.11)	(.15)	(.11)	(.15)
Intercept	−2.25**	−2.67**	−2.27**	−2.66**	−2.25**	−2.68**	−2.24**	−2.68**	−2.25**	−2.67**
Intercept	(.08)	(.10)	(.08)	(.11)	(.08)	(.10)	(.08)	(.10)	(.08)	(.11)

Note. DVV = dating violence victimization.

p < .05. **p < .01 (two-tailed tests).

Model 2 introduced an interaction between DVV and friendship withdrawal. Note, the coefficients for DVV capture the effects of DVV for boys and girls who did not withdraw from their friendships over the course of the relationship when this interaction is included. Among girls, the coefficient for DVV was positive and significant in Model 2 (b = 0.91, p = .021), indicating DVV was positively associated with the hazard of relationship dissolution for girls who did not withdraw from their friendships over the course of the relationship. Conversely, the DVV*Friendship Withdrawal coefficient for girls was negative and significant (b = −1.39, p = .018), which suggests withdrawing from friendships attenuates the positive association between DVV and girls’ hazard of relationship dissolution. We did not find DVV is associated with the hazard of relationship dissolution for boys, or that friendship withdrawal moderated this association. Notably, the 95% confidence interval for DVV*Friendship Withdrawal coefficient for girls (95% CI [−2.52, −0.24]) did not overlap with the 95% confidence interval for DVV*Friendship Withdrawal coefficient for boys (95% CI [−0.04, 1.88]), indicating that the difference in the interactive association between DVV, friendship withdrawal, and the hazard of relationship dissolution among girls and boys was significant at the .05 level (Schenker & Gentleman, 2001).

Model 3 omitted the interaction between DVV and friendship withdrawal and introduced an interaction between DVV and prior friendship attachment. The nonsignificant interaction coefficients introduced in Model 3 indicated that prior friendship attachment did not moderate the association between DVV and relationship dissolution for boys or girls.

In Model 4, we omitted the interaction between DVV and friendship attachment and introduced an interaction between DVV and shared mutual friends. The nonsignificant interaction coefficients indicated that sharing mutual friends before the onset of the relationship did not moderate the association between DVV and the hazard of relationship dissolution for girls or boys. In Model 5—which is our final model—we omitted the interaction between DVV and shared mutual friends and introduced an interaction between DVV and friends before relationship. Again, the nonsignificant coefficients for the interactions introduced in Model 5 indicated that the effects of DVV on the hazard of relationship dissolution for girls and boys did not vary according to whether respondents were friends before the onset of the relationship.

To summarize our key results, we did not find that on average DVV was associated with the hazard of relationship dissolution for girls or boys. However, the association between DVV and relationship dissolution among girls varied according to whether they withdrew from their friendships. Specifically, DVV was positively associated with the hazard of relationship dissolution for girls who maintained their friendships over the course of the relationship, but withdrawing from one’s friendships attenuated the positive effect of DVV on the hazard of relationship dissolution among girls.

Discussion and Conclusion

We examined the association between adolescent dating violence victimization and relationship dissolution. Highlighting linkages between adolescent peer groups and romantic relationship dynamics, we tested whether ties between youth and their friends prior to romantic relationship onset and during the relationship alter the association between teen DVV and relationship dissolution. We also considered whether the association between DVV and relationship dissolution varied according to whether respondents and their partners shared mutual friends or were friends prior to the relationship onset. Last, we tested whether the association between friendship processes, DVV, and relationship dissolution varied by gender.

Using Add Health data, our models indicated that on average, DVV was not associated with the hazard of relationship dissolution for girls or boys. However, our models revealed that for girls, DVV increased the hazard of relationship dissolution for those who did not withdraw from their friends over the course of the relationship. Conversely, withdrawing from friendships attenuated the association between DVV and relationship dissolution for girls suggesting that girls who withdraw from friends are less likely to leave abusive relationships. We did not find that the association between DVV, friendship withdrawal, and the hazard of relationship dissolution varied for boys.

Our results shed light onto the association between dating violence and the longevity of youth romances. To date, few studies have examined whether DVV matters for adolescent relationship longevity. Several studies that have examined leave/stay decisions among adults who experience partner violence focused on how economic dependency and the presence of children discourage adults from exiting abusive relationships. However, these factors likely do not influence adolescents’ leave/stay decisions because youth are typically not economically dependent on, or have children with, their romantic partners. Our results suggest DVV matters for adolescents’ relationship longevity, but how it affects relationship longevity depends on gender and whether friendship withdrawal occurred.

We situate our findings with those from recent studies that highlight the relationship between peer processes and adolescents’ decisions to terminate abusive relationships. For instance, Helm and colleagues’ (2015) qualitative study of youth in Hawai`i suggests social support from friends is an important factor shaping adolescents’ decisions to leave abusive relationships. Similarly, Copp and colleagues’ (2015) quantitative study of adolescents in Toledo suggests peers’ capacity to express negative appraisals inform their decisions to leave or remain in abusive relationships. While these studies are instrumental for understanding the mechanisms through which friends shape whether adolescent dating violence results in relationship dissolution, it has remained unclear whether these processes are generalizable to the larger population of romantically involved youth in the United States. Our results corroborate findings from these studies focusing on dating violence and relationship longevity. Importantly, we highlight a key process—friendship withdrawal—through which friends matter for the longevity of physically abusive relationships.

Theory and research highlight key linkages between peer relationships and early romances (Connolly et al., 2004). But little is known about which friendship processes shape the association between ADVV and relationship longevity. Our study contributes to the theoretical understanding of which friendship processes matter for the likelihood that youth exit abusive relationships. In particular, friendship withdrawal likely reduces opportunities to disclose abuse to friends and limits friends’ ability to express negative appraisals of abusive partners and exert informal social control over friends’ romantic relationships. Youth who withdraw from friendships throughout the course of the relationship may also perceive less support and feel less empowered to leave abusive relationships. Our findings—coupled with prior theory and research on friendship dynamics and the longevity of abusive relationships among youth—underscore the importance of friendship withdrawal in shaping youths’ decisions to exit or remain in abusive relationships, particularly for girls.

Our study opens avenues for research on the association between partner violence and relationship longevity. Our findings—coupled with research on adults’ decisions to leave or remain in abusive relationships—suggest different factors matter for the association between relationship violence victimization and relationship dissolution throughout the life course. Research that follows individuals over life course stages (e.g., adolescence, emerging adulthood, etc.) can assess exactly when particular economic and social factors affect whether individuals exit abusive relationships. Events such as the Great Recession may have made economic considerations more salient in individuals’ decisions to leave or remain in abusive relationships. Alternatively, women’s increasing employment, coupled with delayed childbirth, may mean that economic dependence and children are less salient factors in adults’ decisions to leave abusive relationships, while friendships play an increasingly important role in informing young adults’ decisions to end or remain in abusive relationships today.

Our study has implications for youth development programs. Adolescence is a time of increasing romantic involvement. While early romances are often important for youth and can inform healthy relationships during adulthood, many youth experience abuse in early romances. Alarmingly, our results indicated that DVV was not on average associated with the hazard of relationship dissolution, indicating that many youth remain in abusive relationships. Programs that empower youth to exit relationships when physical abuse first occurs may decrease the chances that youth experience repeated victimization within their relationships. Second, our results suggest girls who maintain their friendships over the course of a romantic relationship are more likely to end abusive relationships compared with those who withdrew from their friendships. Programs that encourage youth—particularly girls—to maintain friendships over the course of their romantic relationships may help youth retain sources of social support and informal control that can embolden youth to exit abusive relationships.

Although our study helps understand how friendships and gender shape the association between adolescent dating violence victimization and relationship dissolution, it has limitations. First, data limitations precluded us from determining the conditions under which the relationship ended (e.g., which partner ended the relationship, etc.). Add Health also did not measure respondents’ partner violence perpetration. Accordingly, we cannot assess whether and how dating violence perpetration factor into the relationship between DVV and relationship dissolution. In addition, respondents only reported on a narrow set of relatively minor acts of physical aggression. Therefore, we likely underestimated respondents’ DVV. Future research that gathers detailed information on a wider variety of more severe DVV items, the timing of violence perpetration, and who initiated breakups may shed more light onto how DVV affects relationship dissolution. Our data were also collected in the mid-1990s. Thus, it remains unclear whether the processes observed in our study continue to operate today. Although we have no reason to suspect that our results are period-specific, data collection efforts that build on Add Health’s scope and focus can further assess whether and how friends factor into the association between ADVV and relationship dissolution among today’s youth. Add Health also only included youth who were in school at Wave 1. Thus, our results may not be generalizable to romantically involved youth who have dropped out of school and may be particularly vulnerable to ADVV. Future studies that include out-of-school respondents in their sample may help assess whether our findings reflect the experiences for such youth. Finally, although Add Health gathered rich information on friendship dynamics, no information was collected regarding friends’ attitudes or norms surrounding dating or gendered violence, which likely affect teen dating violence (Reed, Silverman, Raj, Decker, & Miller, 2011) and may also inform youths’ decisions to leave abusive relationships. Future research that gathers information on peer norms and attitudes surrounding dating violence may shed more light onto how friends matter for dating relationships among youth.

Despite these limitations, our study contributes to the understanding of how friendships, gender, and dating violence factor into relationship dissolution among youth. We hope future research builds on our study to further understand the conditions that empower youth to exit abusive romances and form healthy and emotionally satisfying romances in adolescence and throughout the life course.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research uses data from Add Health (), a program project directed by Kathleen Mullan Harris and designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris at the University of North Carolina at Chapel Hill, and funded by Grant P01-HD31921 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 23 other federal agencies and foundations. Special acknowledgment is due Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. No direct support was received from Grant P01-HD31921 for this analysis.

Author Biographies

Brian Soller, PhD, is an assistant professor in the Department of Sociology at the University of New Mexico. His research centers on relationship and contextual effects on mental health and behavioral well-being, with a focus on adolescence.

Jennifer E. Copp, PhD, is an assistant professor in the College of Criminology and Criminal Justice at Florida State University. Her research focuses on crime and other problem behaviors during adolescence and young adulthood, with a particular emphasis on IPV.

Dana L. Haynie, PhD, is a professor of Sociology and director of the Criminal Justice Research Center at The Ohio State University. Her research examines social networks, crime and delinquency, gender, and incarceration/re-entry.

Alena Kuhlemeier, MA, is a graduate student in Sociology at the University of New Mexico. Her research interests include adolescent relationships and identity development, and the ways in which gender shapes these experiences.

References

Anderson

K. L.

(2007). Who gets out? Gender as structure and the dissolution of violent heterosexual relationships. Gender & Society, 21, 173-201.

Brown

B. B.

(1999). “You’re going out with who?” Peer group influences on adolescent romantic relationships. In Furman

Brown

B. B.

Feiring

(Eds.), The development of romantic relationships in adolescence (pp. 291-329). New York, NY: Cambridge University Press.

Brown

B. B.

Larson

(2009). Peer relationships in adolescence. In Steinberg

(Ed.), Handbook of adolescent psychology, Vol. 2: Contextual influences on adolescent development (3rd ed., pp. 74-103). Hoboken, NJ: John Wiley.

Browning

C. R.

Leventhal

Brooks-Gunn

(2005). Sexual initiation in early adolescence: The nexus of parental and community control. American Sociological Review, 70, 758-778.

Coleman

J. S.

(1961). The adolescent society: The social life of the teenager and its impact on education. Westport, CT: Greenwood Publishing.

Connolly

Craig

Goldberg

Pepler

(2004). Mixed-gender groups, dating, and romantic relationships in early adolescence. Journal of Research on Adolescence, 14, 185-207.

Connolly

Friedlander

(2009). Peer group influences on adolescent dating aggression. The Prevention Researcher, 16, 8-11.

Connolly

Furman

Konarski

(2000). The role of peers in the emergence of heterosexual romantic relationships in adolescence. Child Development, 71, 1395-1408.

Copp

J. E.

Giordano

P. C.

Longmore

M. A.

Manning

W. D.

(2015). Stay-or-leave decision making in nonviolent and violent dating relationships. Violence and Victims, 30, 581-599.

10.

Dunphy

D. C.

(1963). The social structure of urban adolescent peer groups. Sociometry, 26, 230-246.

11.

Eder

Evans

Parker

(1995). School talk: Gender and adolescent culture. New Brunswick, NJ: Rutgers University Press.

12.

Foshee

V. A.

Benefield

T. S.

Reyes

H. L. M.

Ennett

S. T.

Faris

Chang

. . . Suchindran

C. M.

(2013). The peer context and the development of the perpetration of adolescent dating violence. Journal of Youth and Adolescence, 42, 471-486.

13.

Furman

(1999). Friends and lovers: The role of peer relationships in adolescent romantic relationships. In Collins

W. A.

Laursen

(Eds.), Relationships as developmental contexts: The Minnesota symposia on child psychology (Vol. 30, pp. 133-154). Mahwah, NJ: Erlbaum.

14.

Furman

Wehner

E. A.

(1994). Romantic views: Toward a theory of adolescent romantic relationships. In Montemayor

Adams

G. R.

Gullotta

(Eds.), Personal relationships during adolescence (pp. 168-195). Thousand Oaks, CA: SAGE.

15.

Gilligan

(1982). In a different voice: Psychological theory and women’s development. Cambridge, MA: Harvard University Press.

16.

Giordano

P. C.

(2003). Relationships in adolescence. Annual Review of Sociology, 29, 257-281.

17.

Giordano

P. C.

Longmore

M. A.

Manning

W. D.

(2006). Gender and the meanings of adolescent romantic relationships: A focus on boys. American Sociological Review, 71, 260-287.

18.

Gorrese

Ruggieri

(2012). Peer attachment: A meta-analytic review of gender and age differences and associations with parent attachment. Journal of Youth and Adolescence, 41, 650-672.

19.

Halpern

C. T.

Spriggs

A. L.

Martin

S. L.

Kupper

L. L.

(2009). Patterns of intimate partner violence victimization from adolescence to young adulthood in a nationally representative sample. Journal of Adolescent Health, 45, 508-516.

20.

Haynie

D. L.

Farhat

Brooks-Russell

Wang

Barbieri

Iannotti

R. J.

(2013). Dating violence perpetration and victimization among U.S. adolescents. Journal of Adolescent Health, 53, 194-201.

21.

Helm

Baker

C. K.

Berlin

Kimura

(2015). Getting in, being in, staying in, and getting out: Adolescents’ descriptions of dating and dating violence. Youth & Society, 49, 318-340.

22.

Johnson

M. J.

(2010). A typology of domestic violence: Intimate terrorism, violent resistance, and situational couple violence. Boston, MA: Northeastern University Press.

23.

Kaestle

C. E.

Halpern

C. T.

(2005). Sexual intercourse precedes partner violence in adolescent romantic relationships. Journal of Adolescent Health, 36, 386-392.

24.

Kim

Gray

K. A.

(2008). Leave or stay? Battered women’s decision after intimate partner violence. Journal of Interpersonal Violence, 23, 1465-1482.

25.

Kuttler

A. F.

La Greca

A. M.

(2004). Linkages among adolescent girls’ romantic relationships, best friendships, and peer networks. Journal of Adolescence, 27, 395-414.

26.

Larson

R. W.

Clore

G. L.

Wood

G. A.

(1999). The emotions of romantic relationships: Do they wreak havoc on adolescents? In Furman

Brown

B. B.

Feiring

(Eds.), The development of romantic relationships in adolescence (pp. 19-49). New York, NY: Cambridge University Press.

27.

Laursen

Williams

V. A.

(1997). Perceptions of interdependence and closeness in family and peer relationships among adolescents with and without romantic partners. New Directions for Child Development, 78, 3-20.

28.

C. Q.

Huebner

E. S.

(2008). Attachment relationships and adolescents’ life satisfaction: Some relationships matter more to girls than boys. Psychology in the Schools, 45, 177-190.

29.

Martsolf

D. S.

Draucker

C. B.

Brandau

(2013). Breaking up is hard to do: How teens end violent dating relationships. Journal of the American Psychiatric Nurses Association, 19, 71-77.

30.

Reed

Silverman

J. G.

Raj

Decker

M. R.

Miller

(2011). Male perpetration of teen dating violence: Associations with neighborhood violence involvement, gender attitudes, and perceived peer and neighborhood norms. Journal of Urban Health, 88, 226-239.

31.

Rhatigan

D. L.

Axson

D. K.

(2006). Using the investment model to understand battered women’s commitment to abusive relationships. Journal of Family Violence, 21, 153-162.

32.

Rhoades

G. K.

Stanley

S. M.

Kelmer

Markman

H. J.

(2010). Physical aggression in unmarried relationships: The roles of commitment and constraints. Journal of Family Psychology, 24, 678-687.

33.

Rhodes

K. V.

Cerulli

Dichter

M. E.

Kothari

C. L.

Barg

F. K.

(2010). “I didn’t want to put them through that”: The influence of children on victim decision-making in intimate partner violence cases. Journal of Family Violence, 25, 485-493.

34.

Royston

(2004). Multiple imputation of missing values. The Stata Journal, 4, 227-241.

35.

Schenker

Gentleman

J. F.

(2001). On judging the significance of differences by examining the overlap between confidence intervals. The American Statistician, 55, 182-186.

36.

Soller

(2014). Caught in a bad romance: Adolescent romantic relationships and mental health. Journal of Health and Social Behavior, 55, 56-72.

37.

Steffensmeier

D. J.

Schwartz

Roche

(2013). Gender and twenty-first-century corporate crime: Female involvement and the gender gap in Enron-era corporate frauds. American Sociological Review, 78, 448-476.

38.

Steinberg

(2001). We know some things: Parent-adolescent relationships in retrospect and prospect. Journal of Research on Adolescence, 11, 1-19.

39.

Straus

M. A.

Hamby

S. L.

Boney-McCoy

Sugarman

D. B.

(1996). The Revised Conflict Tactics Scale (CTS2): Development and preliminary psychometric data. Journal of Family Issues, 17, 283-316.

40.

Sullivan

H. S.

(1953). The interpersonal theory of psychiatry. New York, NY: W.W. Norton.

41.

Sylaska

K. M.

Edwards

K. M.

(2014). Disclosure of intimate partner violence to informal social support network members: A review of the literature. Trauma, Violence, & Abuse, 15, 3-21.

42.

Young

B. J.

Furman

(2013). Predicting commitment in young adults’ physically aggressive and sexually coercive dating relationships. Journal of Interpersonal Violence, 28, 3245-3264

43.

Wang

Kao

Joyner

(2006). Stability of interracial and intraracial romantic relationships among adolescents. Social Science Research, 35, 435-453.