Dyadic Dynamics in Young Couples Reporting Dating Violence

Abstract

This study uses a combination of observational methods and dyadic data analysis to understand how boyfriends’ and girlfriends’ perpetration of dating violence (DV) may shape their own and their partners’ problem-solving communication behaviors. A sample of 39 young heterosexual couples aged between 15 and 20 years (mean age = 17.8 years) completed a set of questionnaires and were observed during a 45-min dyadic interaction, which was coded using the Interactional Dimension Coding System (IDCS). Results suggest that neither boyfriends’ nor girlfriends’ own perpetration of DV was related to their display of positive and negative communication behaviors. However, estimates revealed significant partner effects, suggesting that negative communication behaviors displayed by girls and boys and positive communication behavior displayed by girls were associated to their partner’s DV but not to their own. Such results confirm the need to shift our focus from an individual perspective to examining dyadic influences and processes involved in the couple system and the bidirectionality of violent relationships.

Keywords

dyadic interaction patterns adolescence dating violence

Dating violence (DV) is highly prevalent among adolescents and emerging adults, with estimates of reported perpetrated DV ranging between 14% and 82% for psychological violence, 11% and 41% for physical violence, and 3% and 10% for sexual violence (Foshee & Matthew, 2007). The scope of the problem as well as the adverse consequences it can incur on adolescents’ physical and mental well-being have made DV a major public health concern. There has been accumulating evidence that a large proportion of violence in adolescent dating relationships is bidirectional (perpetrated by both partners). For example, a recent review of the literature showed that, on average, 57.5% of intimate partner violence reported across samples (population, community, school and clinical samples) could be categorized as bidirectional (Langhinrichsen-Rohling, Selwyn, & Rohling, 2012). According to the results of a meta-analytic review by Stith, Smith, Penn, Ward, and Tritt (2004), male-to-female physical violence victimization is a strong predictor of the victim using violence toward her partner. Furthermore, the results from a study on the co-occurrence of partner and parent physical violence found that one partner’s aggression against the other increased the risk of female-to-male aggression by 5 times and of male-to-female aggression by 10 times (Slep & O’Leary, 2005). However, authors point out that the presence of bidirectional violence does not mean that all aspects of female and male violence are symmetrical as many gender differences can still be evident. For example, girls aged 12 to 17 have been found to be more likely to experience physically injurious or fear-inducing incidents of DV (Hamby & Turner, 2013).

To take better account of bidirectionality, it is important to adopt a dyadic approach when exploring DV. In fact, researchers have recently developed conceptual models of intimate partner violence that focus on interactional processes and highlight the importance of dyadic context when understanding violence (e.g., Bartholomew & Cobb, 2011; Capaldi & Kim, 2007; Langhinrichsen-Rohling, 2010). To date, research has largely focused on males and females separately. We cannot, however, expect data from two individuals in a relationship to be independent from each other because, as Bartholomew and Cobb (2011) appropriately point out, “relationships are inherently interactional.” For this reason, recent conceptual models of couple violence seek to integrate individual and dyadic factors. For example, Bartholomew and Cobb refer to four sets of factors: (a) background and dispositional characteristics (e.g., family background, personality, psychopathology, and interpersonal abilities), (b) individual and dyadic factors related to the relationship context (e.g., attachment security, power imbalance, relationship discord, communication), (c) factors related to the situational context (e.g., precipitating factors such as partner provocation and relationship threats, dyadic interaction, inhibition of aggression), and (d) patterns of partner violence (e.g., severity, mutuality, consequences of partner violence). The model argues that partners may influence each other at any of the four stages of the model. Bartholomew and Cobb maintain that partners in mutually satisfying relationships should not be at risk of experiencing violence regardless of their individual characteristics.

While research on adolescent DV is limited in comparison with adult partner violence, there has been a recent surge of interest in this population. The onset of adolescence is an important developmental life stage that involves numerous physical, emotional, and sexual changes. During this time, the emergence of romantic relationships plays a pivotal role in adolescent social development (Connolly & McIsaac, 2009). Such experiences can greatly influence the formation of identity, attachment patterns, and interactional patterns that can persist into adulthood.

Only a handful of scholars have examined dyadic influences in adolescent relationships. Among these, we note the work of O’Leary and Slep (2003) whose longitudinal study of dating relationships was, to our knowledge, one of the first to demonstrate the importance of dyadic influences on DV. More specifically their findings indicated how perpetration of physical violence in high school boys and girls is predicted by the victimization they reported 3 months earlier. Verbal aggression, jealous behavior, and controlling behavior were also found to be predictors of physical DV, both concurrently and longitudinally (O’Leary & Slep, 2003). Although this study provided valuable insight into the interactive nature of DV, the fact that it was based solely on individual and not couple data is a definite shortcoming. Indeed, the data used for the study were not from both partners but instead relied on one partner’s rating of his own and his or her partner’s violence and may have introduced a potential rater bias. While also using individual data, other studies have documented dyadic risk factors of DV. For example, communication difficulties (Schumacher, Feldbau-Kohn, Slep, & Heyman, 2001), power imbalance (Connolly, Nocentini, et al., 2010), and greater relationship conflicts (Connolly, Friedlander, Pepler, Craig, & Laporte, 2010; O’Keefe, 2005) have consistently been linked with perpetration of violence against a dating partner.

Previous research with adult couples have repeatedly demonstrated how dyadic interactions observed during problem-solving discussions can differentiate between violent and non-violent couples (e.g., Burman, Margolin, & John, 1993; Cordova, Jacobson, Gottman, Rushe, & Cox, 1993; Margolin, John, & Gleberman, 1988). Furthermore, studies have clearly confirmed the importance of communication in marital interaction for both partners’ relationship satisfaction (Woodin, 2011). In their study on the effects of alcohol on marital interactions, Leonard and Roberts (1998) found that couples in which husbands had perpetrated intimate partner violence exhibited significantly more negative behaviors and higher rates of negative reciprocity than couples without a history of husband-to-wife aggression. Similar results were reported by Gottman and Notarius (2000) in their review of existing observational research on marital interactions. These authors proposed that one distinct feature differentiating violent from non-violent couples is how violent couples appear to lack the ability to exit situations of “either reciprocated or escalating hostility” (Gottman & Notarius, 2000). It remains to be determined whether such observations can be found in adolescents.

This study seeks to add to the small but growing body of work on dyadic influences in DV. Specifically, we use a combination of observational methods and dyadic data analysis to understand how boyfriends’ and girlfriends’ perpetration of DV may shape their own and their partners’ problem-solving communication behaviors. The objective is to move beyond individual factors, improve our comprehensive understanding of the relational dynamics at play, and offer cues for the development of more efficient prevention and intervention efforts targeted at adolescents.

Method

Participants

Thirty-nine young couples of French–Canadian background aged between 15 and 20 years old were recruited in the Greater Montreal area. To be eligible, couples had to be in a committed heterosexual relationship for more than 2 months. Partners could neither be cohabitating nor be the responsible caretaker of a dependent child. The demographic and relationship characteristics for boys and girls are presented in Table 1. Most respondents reported attending CEGEP (i.e., junior/technical College in Quebec; 69.2%) while 14.1% were not in school at the time of the study. Most adolescents had been dating their partner more than a year (average length 15.3 months; ranging between 3 and 41 months).

Table 1.

Means and Standard Deviations of Participants’ Demographic and Relationship Characteristics.

Characteristics	Total (N = 78)	Boyfriends (n = 39)	Girlfriends (n = 39)
M age (SD)	17.8 (1.2)	18.0 (1.1)	17.6 (1.4)
Current academic level
High school	10 (12.8%)	3 (7.7%)	7 (17.9%)
College or vocational school	54 (69.2%)	27 (69.2%)	27 (69.2%)
University	3 (3.8%)	1 (2.6%)	2 (5.1%)
Not currently in school	11 (14.1%)	8 (20.5%)	3 (7.7%)
Currently working	47 (60.3%)	24 (61.5%)	23 (59.0%)
If yes, more than 10 hr per week	34 (43.6%)	16 (41.0%)	18 (46.2%)
Length of dating relationship in months (M and SD)	15.3 (10.2)	15.3 (10.2)	15.3 (10.3)
First dating relationship	12 (15.4%)	6 (15.4%)	6 (15.4%)
More than 1 sexual partner in past 12 months	20 (25.6%)	7 (17.9%)	13 (33.3%)

Procedures

Young couples were recruited through youth organizations from the Greater Montreal area (e.g., schools and community organizations) and directly from their living environments (e.g., schools, community organizations, parks, and libraries). Potential participants were first screened for eligibility by telephone. Eligible couples came to our laboratory for a 2-hr session. After being greeted by two experimenters, the study protocol was presented in detail to ensure adequate comprehension of study goals, procedure, risks, and benefits. Ethical considerations were also discussed (e.g., voluntary and confidential participation, mandatory reporting to authorities when the safety of the participant is believed to be compromised). Upon consent, each member of the couple completed a set of questionnaires and an individual semi-structured interview on their own. The couple was then reunited to participate in a videotaped interaction session consisting of a 3-min warm-up activity followed by two 7-min discussions. The warm-up activity (i.e., planning an activity to do together with a limited budget) was used to allow the couples to familiarize themselves with their surroundings and the testing situation. The couples then discussed a topic of disagreement previously selected independently by each partner from the Adolescent Couples’ Issues Checklist (ACIC; Welsh, Grello, Dickson, & Harper, 2001). A counter-balanced approach was used to ensure that an equal number of dyads discussed the male or female issue first. No observer was present during the videotaped discussions. Individual debriefing was provided following the interaction and a list of psychosocial resources was handed out to all participants. Each member of the couple was given a CAD$20 compensation for their time. This study received the approval of the University’s institutional review ethics board.

Measures

Perpetration of DV

DV was assessed using the 80-item VIFFA (“Violence faite aux filles dans les Fréquentations à l’adolescence”; Lavoie & Vézina, 2001). Items on the VIFFA assess the number of times respondents and their current partners have been the victim and/or the perpetrator of 40 specific violent acts, including psychological, sexual, and physical acts of DV. In the current study, we focused specifically on perpetration of DV reported by the respondent against his or her current dating partner. Examples of items include “I used physical force to make him/her have sexual contact with me” and “I pushed or shook him/her in a moment of anger or frustration.” Items are rated on a 4-point scale ranging from 1 (never) to 4 (more than 10 times). Internal consistency for the perpetration subscale was satisfactory with an alpha score of .85. DV perpetration was treated as a dichotomous score (1 = yes and 0 = no), indicating if participants had endorsed perpetrating at least one act of physical DV, one act of sexual DV, or two acts of psychological DV more than twice.

Disagreement topics

To identify two topics of disagreement for the interactions, both partners were individually instructed to rate, on a 4-point Likert-type scale ranging from 1 (never) to 4 (very often), the extent to which topics of disagreement from a modified version of the ACIC (Welsh et al., 2001) were a problem in their relationship. This checklist includes 17 common issues of disagreement between adolescent romantic couple members, as well as an option to write issues not on the list. Researchers selected the most highly rated issue of each member of the couple to discuss during the interaction session.

Observational coding

The Interactional Dimension Coding System (IDCS; Julien, Chartrand, Markman, & Lindahl, 1991; Julien, Markman, & Lindahl, 1989) was used to code key dimensions of the couples’ two videotaped discussions of problem areas in their relationship. This global coding system was developed to better assess the quality of problem-solving behaviors. It provides seven individual dimensions (i.e., each partner receives a separate code by dimension) and four dyadic dimensions (i.e., the couple is rated as a whole). Table 2 provides a brief overview of the definitions of the seven individual observed variables used in the current analysis. Individual dimensions are rated from 1 to 9, with higher scores indicating greater intensity of the behavior. Each discussion was used as a unit of observation for the seven individual dimensions of this study. In addition, Positive interaction and Negative interaction composite scores were created for the present study. The subscale for Positive communication behaviors averaged positive affect, problem-solving skills, support/validation, and communication skills dimensions (α for boys was .80; for girls, it was .73). The composite score for Negative communication behaviors averaged negative affect, withdrawal, and conflict dimensions (α for boys = .64; α for girls = .76). Two graduate students, who received more than 80 hr of training, coded the interactions. Intercoder agreement for the IDCS dimensions was assessed by independently coding 30% of randomly selected interactions. Interrater agreement coefficient ranged from .88 to .95. Preliminary analyses (correlations, t tests) did not support the need to use the observed variables from both discussions as scores of the first and second 7-min discussions were highly correlated and no significant differences were identified between scores. On the basis of these results, and for matters of parsimony, only the scores from the second videotaped problem-solving discussion were used for the present analyses.

Table 2.

Definition of the Observed Variables.

Construct	Definition
Positive communication behaviors
Communication skill	Refers to an individual’s ability to convey thoughts and feelings in a clear, constructive manner (e.g., self-disclose emotions; expressing emotions clearly).
Support/validation	Refers to positive listening and speaking skills that an individual uses to demonstrate support and understanding to his or her partner (e.g., being attentive when the partner is talking; expression of concern and sympathy for the partner).
Problem solving	Refers to an individual’s ability to define a problem and work toward a mutually satisfactory solution for the problem. It is based on the individual’s ability to try to solve the problem, not on whether the problem was actually solved (e.g., recognizing the existence of the problem; negotiating compromises).
Positive affect	Refers to positivity expressed through facial expressions, body positioning, and tone of voice (e.g., relaxed posture; smiling).
Negative communication behaviors
Conflict	Refers to behaviors that encourage arguing. For example, the level of tension, hostility, oppositionality, antagonism, and negative affect displayed (e.g., tension in the body; negative comments toward the partner).
Withdrawal	Refers to attempting to avoid an interaction or discussion through body language or stating a desire not to discuss a topic. It is actively rejecting the existence of the disagreement or denying personal responsibility for the problem (e.g., avoiding visual contact; changing the topic).
Negative affect	Refers to negativity expressed through facial expressions, body positioning, and tone of voice (e.g., frowning; angry voice).

Source. Julien, Chartrand, Markman, and Lindahl (1991); Kline et al. (2004, pp. 116-117).

Data Analysis

We examined the associations between perpetrated DV and communication behaviors using the Actor–Partner Interdependence Model (APIM; Cook & Kenny, 2005; Kenny, Kashy, & Cook, 2006). Given that the current study assumes the presence of mutual influence between boyfriends and girlfriends, we needed an analytic approach that takes into account the interdependence of dyadic data. The APIM allows this by assessing both actor and partner effects and, therefore, simultaneously accounting for both individual and dyadic factors. More specifically, the two types of effects calculated by the APIM consists of (a) the actor estimate, a measure of the relationship between an individual’s own scores on an independent variable and an outcome variable (e.g., boyfriends’ own self-reported DV and communication behaviors) and (b) the partner estimate, a measure of the relationship between a partner’s score on an independent variable and the individual’s own outcome variable (e.g., boyfriends’ self-reported DV and girlfriends’ communication behaviors).

Figure 1 represents the current study’s actor effects by two paths labeled a, whereas partner effects are represented by paths labeled p. Two different models were estimated; one for positive communication behaviors and one for negative communication behaviors. Both models were examined with the software package Mplus-Version 6.0 (Muthén & Muthén, 1998-2007) using maximum likelihood estimation (MLE). To compare estimates within each model, we specified equality constraints and tested the differences using the chi-square difference test as described by Cook and Kenny (2005). With this procedure, it is possible to test, for example, whether boyfriends and girlfriends have the same influence on each other. To do so, the chi-square of the initial fully saturated model is compared with the chi-square of the same model but with the two partner effects forced to be equal. If the difference between the two models is found to be non-significant, the two estimates are considered equivalent. However, a significant Δχ² indicates that the more constrained model has worsened the fit. In our example, the two partner effects would therefore be determined significantly different from each other, meaning that one partner has more influence on the other.

Figure 1.

Actor–partner interdependance model of self-reported violence against a dating partner on communication behaviors.

Results

Out of a total of 39 couples, there were 23 boys (59.0%) and 22 girls (56.4%) who indicated having inflicted at least one form of violence against their dating partner in the past 12 months. In terms of specific DV experiences, McNemar’s chi-square tests for related samples (Siegel, 1956) indicated that the boyfriends had significantly higher rates of self-reported sexual aggression than their girlfriends (30.8% vs. 0.0%; χ² = 12.00, p < .000). No such difference was found for psychological (38.5% vs. 48.7%; χ² = 1.00, p = .317) or physical violence (28.2% vs. 30.8.7%; χ² = .08, p = .782). For 55.2% of the 29 couples reporting DV, the violence perpetrated was self-reported by both partners and could be categorized as bidirectional. Regarding the composite scores that were created from the seven individual communication dimensions from the IDCS, mean scores on the Positive scale were 4.41 for boyfriends (SD = 1.37; ranging from 1.50 to 7.25) and 4.45 for girlfriends (SD = 1.18; ranging from 1.75 to 6.75). Mean scores on the Negative scale were 3.70 for boyfriends (SD = 1.61; ranging from 1.00 to 6.67) and 3.74 for girlfriends (SD = 1.56; ranging from 1.33 to 7.67). Results of paired t tests indicated that there were no significant differences between boyfriends and girlfriends on their levels of communication behaviors. To assess differences within each gender and determine which, between positive and negative communication behaviors, were exhibited the most, t tests were conducted. For both boys, t(38) = 1.606, p = .117, and girls, t(38) = 1.774, p = .084, results were insignificant, which suggests that positive and negative behaviors were equally exhibited when interacting with their partners.

Preliminary analyses of the relationship between perpetrated DV and Communication Behaviors were conducted using Pearson correlations. Table 3 presents the intercorrelations among all variables for boyfriends and girlfriends. A boyfriend’s use of violence against his dating partner was significantly correlated with both his own and his partner’s negative behaviors, while being negatively correlated to only his partner’s positive communication behaviors. A girlfriend’s use of DV was only significantly correlated to her partner’s negative behaviors. To determine the presence of non-independence between outcome variables, an essential condition for dyadic analyses, we also examined within-dyad correlations. Significant results indicated a moderate positive association between boyfriends and girlfriends display of negative communication behaviors (r = .488; p = .002). Given that Cook and Kenny (2005) recommend that a more liberal test (p = .20, two-tailed) should be used when testing for non-independence, boyfriends’ and girlfriends’ positive communication behaviors were also considered significantly related (r = .243; p = .137).

Table 3.

Intercorrelations for Positive and Negative Communication Behaviors and Self-Reported Violence Against a Dating Partner (n = 39 Couples).

	Boyfriends			Girlfriends
Measure	1	2	3	1	2	3
Boyfriends
1. Positive behaviors	1.00
2. Negative behaviors	−.75***	1.00
3. Dating violence	−.19	.33*	1.00
Girlfriends
1. Positive behaviors	.24	−.22	−.49**	1.00
2. Negative behaviors	−.16	.49**	.51***	−.70***	1.00
3. Dating violence	−.28^†	.39*	.32*	−.11	.29^†	1.00

†

p < .10. *p < .05. **p < .01. ***p < .001.

The relationships between perpetrated DV and communication behaviors were then examined using APIM (Cook & Kenny, 2005; Kenny et al., 2006). As presented in Figure 1, the basic APIM model includes actor and partner effects. It is, by definition, a saturated model (i.e., df = 0) which produces a perfect model fit (Kenny et al., 2006). Therefore, it was not relevant to examine traditional fit indices. Separate models were estimated for negative communication behaviors and for positive communication behaviors. The estimates pertaining to both models are presented in Table 4. Results suggest that for neither boyfriends nor girlfriends actor effects were statistically significant, indicating that their own perpetration of DV is not related to their display of positive or negative communication behaviors. Regarding crossed effects of DV on negative communication behaviors, estimates revealed significant partner effects for both boyfriends (β = .468, p = .000) and girlfriends (β = .317, p = .028). Boyfriend’s self-reported DV was also found to be negatively linked with their partners’ positive behaviors (β = −.506, p = .000). The first model accounted for 19.6% of the total variance in boyfriends’ negative behavior and 27.8% of the total variance in girlfriends’ negative behavior. The second model accounted for 9.1% and 24.2% of the variance in boys’ and girls’ positive communication behaviors.

Table 4.

Actor and Partner Effects of Dating Violence on Communication Behaviors (n = 39 Couples).

Measure	Actor Effects (β)	Partner Effects (β)
Negative behaviors
DV_b	.224	.468***
DV_g	.137	.317*
Positive behaviors
DV_b	−.107	−.506***
DV_g	.049	−.250

Note. DV_b = boyfriend’s dating violence perpetration; DV_g = girlfriend’s dating violence perpetration. Actor and partner effects are standardized regression coefficient.

p < .05. **p < .01. ***p < .001.

Furthermore, to explore whether one dating partner had more influence on the other, we tested, as specified by Cook and Kenny (2005), sets of equality constraints using the chi-square difference test. In both models, partner effects on boyfriends’ and girlfriends’ communication behaviors were similar in magnitude, Δχ²(1) = 0.412, p = .521 and Δχ²(1) = 0.803, p = .370, which suggests that there are no substantial differences between the influence of boyfriends’ and girlfriends’ self-reported DV on each other’s communication behaviors. We tested an additional set of equality constraints to compare the actor and the partner effects predicting communication behaviors displayed by each partner. In the model with positive communication behaviors as the outcome variables, the partner effect was significantly greater in size than the corresponding non-significant actor effects on girl’s positive behaviors, Δχ²(1) = 5.097, p = .024. This result suggests that the positive communication behaviors displayed by girls were affected more by her partner’s DV than her own. When we compared the size of the actor and the partner effects on both the boy’s and the girl’s display of negative behaviors, no significant differences were found Δχ²(1) = 0.138, p = .710 and Δχ²(1) = 1.995, p = .158.

Discussion

The present study used a combination of observational methods and dyadic data analysis to understand how boyfriends’ and girlfriends’ perpetration of DV may shape their own and their partners’ problem-solving communication behaviors. In doing so, our study expands on prior knowledge by providing insight into interinfluences within the couple system and the bidirectionality of violent relationships.

Out of a total of 39 couples, 59.0% of boys and 56.4% of girls reported having inflicted at least one form of violence against their dating partner in the previous year. These proportions are consistent with those of previous findings and confirm how violence is alarmingly prevalent in the context of dating relationships among adolescents and emerging adults (Foshee & Reyes, 2011). Perpetration of violence was found to be self-reported by both partners in 55.2% of couples reporting DV. This rate of bidirectional violence is very comparable with what was found in Langhinrichsen-Rohling and colleagues’ (2012) comprehensive review (57.5% of bidirectional violence across samples). Until recently, most studies that have examined reciprocal DV have done so using only unilateral reporting of participants own perpetration of violence and that of their partners. The current findings address this gap in knowledge and suggest that bidirectionality is also evident when collecting violence data directly from both partners. Such reciprocal violence in adults samples is often referred to as situational violence (Johnson, 1995), which according to Johnson (1995) is the most common type of partner violence. It is believed to arise from couple conflicts that “turn into an argument that escalates into violence” (Johnson, 2008). While it is often considered to be a relatively milder form of partner violence, Johnson argues that the nature of Situational Couple Violence varies and that it can escalate into chronic and/or severe violence with serious injuries (Johnson, 2008). Moreover, the mutually aggressive dyads have been shown to be particularly at risk of experiencing deleterious consequences (Capaldi & Kim, 2007). It may be difficult to overlap existing adult typologies such as Johnson’s on adolescents given that adult development has generally stabilized, whereas adolescents are still in formative stages, rapidly developing. Further research in this area will therefore be necessary to develop similar typologies of adolescent to classify bidirectional violent dating relationships. Such typologies could help personalize prevention and intervention efforts to optimize their impact.

In addition to the high prevalence of bidirectional DV, the interdependence of boyfriends’ and girlfriends’ observed communication behaviors confirmed the importance of using a dyadic approach to continue our investigation. A notable finding from the couple level analysis was that neither boyfriends’ nor girlfriends’ own perpetrations of DV were found to be related to their display of positive or negative communication behaviors. However, estimates speak to the dyadic nature of DV and revealed significant partner effects, suggesting that negative communication behaviors displayed by girls and boys and positive communication behavior displayed by girls were associated to their partner’s DV but not to their own. These results seem to suggest that adolescents’ observed behaviors were more reactive, that is, more influenced by circumstances and events from the relationship, then reflective of their own inner emotions and dispositions. This result is not surprising considering that, for a majority of participants, this was their very first romantic relationship. The novelty of romantic love elicits intense physiological, cognitive, and emotional responses from adolescents who have not yet developed the skills to cope with these new feelings (Larson, Clore, & Wood, 1999) or to self-reflect on their involvement. The observed association between one partner’s DV and the other partner’s negative communication behavior may signal that these couples have a hard time solving their problems constructively. Indeed, as outlined by Connolly and McIsaac (2009), adolescents can develop coercive patterns of “criticism, contempt, defensiveness, and withdrawal” within their romantic relationships which can create distress for both partners. More longitudinal research designs are needed to investigate how both partners’ behaviors may influence one another, escalate to physical aggression and then lead to repeated occurrences.

The present study has some limitations which provide additional directions for future research. First, given our modest sample size (n = 39 couples), the chronicity and frequency of DV could not be taken into account. As suggested in research on existing typologies of adult intimate partner violence (e.g., Holtzworth-Munroe & Stuart, 1994; Johnson, 1995), it is possible that specific patterns of association between DV and dyadic dynamics could emerge when considering the severity of DV. In addition, although our findings provide clear insights into the importance of dyadic influences, a larger sample would allow the inclusion of more variables to the model. For instance, adding family background characteristics, such as violence and maltreatment in the family of origin, could significantly improve our understanding of the factors contributing to dyadic adjustment. Second, for the current study, recruitment relied on couples who volunteered and were interested in participating in a study that would address “communication and conflict resolution.” This may have introduced a self-selection bias in the composition of our sample as, for instance, youths experiencing more severe violence may have been less favorable to participate in a study on conflict resolution. Third, future research would greatly benefit from a longitudinal approach to explore how the observed interaction patterns evolve over time. This is especially important for research with adolescents and emerging adults given their rapid development, the time-limited nature of their relationships, and the risk of aggressive behaviors being repeated in subsequent romantic relationships. In addition, a longitudinal design could also provide better evidence of causality and help determine the correct temporal sequence in terms of who initiates and who responds to the violent behavior. Finally, because our sample included older adolescents (age = 15-20), the results cannot be generalized to younger populations without any reservations. Future research should verify our findings with more diverse samples.

Despite these limitations, our study is an important first step toward validating how improving our understanding of dyadic influences in couples experiencing DV is crucial for the development of more efficient prevention and intervention efforts. The majority of existing prevention programs of DV elaborated in the past decade have been based on a universal approach implemented in school settings and target changes in individual risk factors. As argued by an increasing number of researchers (Langhinrichsen-Rohling & Capaldi, 2012; Reyes, Foshee, Bauer, & Ennett, 2012; Stith, McCollum, Amanor-Boadu, & Smith, 2012) and as advocated by Tharp (2012), a greater focus of prevention programs for intimate partner violence on interactions between individual and dyadic factors is likely to be more effective than the current individual approaches which have limited success in reducing incidence of victimization (O’Leary & Slep, 2012). Indeed, available data suggest that most current universal primary prevention programs are associated with positive changes concerning their proximal objectives (i.e., changes in knowledge, awareness of the phenomenon, attitudes), but results are more ambiguous regarding distal objectives (i.e., reduction of incidence of victimization; Hébert, Daigneault, & Van Camp, 2012). Empirical reports underscore that DV is clearly a multidetermined phenomenon. In this regard, the findings reported in this study provide strong support for targeting relational patterns in prevention efforts and reinforce recent prevention efforts that address both individual and relationship risk factors. An example of a more comprehensive approach is the Dating Matters™: Strategies to Promote Healthy Teen Relationships, a DV prevention initiative by the Centers for Disease Control and Prevention (CDC; Tharp et al., 2011). Dating Matters™ focused on 11- to 14-year-old youth and targets individual, peer, partner, parent, as well as neighborhood factors. According to Tharp, the promotion of healthy relationship behaviors is critical to the prevention of DV in adolescence. For this reason, Dating Matters™ programs focus on dyadic processes aiming at promoting healthy relationship skills (e.g., communication, conflict resolution).

In summary, our study provides clear initial evidence of how boyfriends’ and girlfriends’ perpetration of DV shape their partners’ problem-solving communication behaviors. Such results confirm the need to move our focus from an individual perspective to examining dyadic influences and processes by which characteristics from both partners interact with each other and contribute to the development or persistence of violent behaviors in dating relationships. Furthermore, results like those found in this study are crucial to improve our understanding of the relational dynamics at play especially given that previous research have shown that interaction patterns in early romantic relationships may become embedded and repeated in later adult relationships (Furman & Wehner, 1997; Simon, Bouchey, & Furman, 2000). In the long run, improving prevention and intervention strategies will reduce the social costs associated with violence among dating adolescents and help provide teenagers with means to experience healthy romantic relationships.

Footnotes

Acknowledgements

The authors wish to thank adolescents and young adults who participated in this project and partner organizations.

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 was supported by a grant from the SSHRC (410-2008-1807) awarded to Mylène Fernet and from CIHR (103944) awarded to Martine Hébert.

Author Biographies

Alison Paradis (PhD in psychology) is an assistant professor in the Department of Psychology at Université du Québec à Montréal (UQAM). Her research focuses on the development of comprehensive models to understand couple and family violence. Her current research concerns mutual dating violence and the impact of child maltreatment on later interpersonal functioning.

Martine Hébert (PhD in psychology) is a full professor in the Department of Sexology at UQAM. Her research interests focus on the diversity of outcomes in children and adolescents victims of interpersonal trauma and trajectories of resilience. She has also conducted evaluative studies of both intervention and prevention programs.

Mylène Fernet (MA in sexology and PhD in public health) is a full professor in the Department of Sexology at UQAM. Her research interests focus on romantic relationships and sexuality development during adolescence. Her current research program is on adolescents and emerging adults who are vulnerable to interpersonal violence.

References

Bartholomew

Cobb

R. J.

(2011). Conceptualizing relationship violence as a dyadic process. In Horowitz

L. M.

Strack

(Eds.), Handbook of interpersonal Psychology (pp. 233-248). Hoboken, NJ: John Wiley.

Burman

Margolin

John

R. S.

(1993). America’s angriest home videos: Behavioral contingencies observed in home reenactments of marital conflict. Journal of Consulting and Clinical Psychology, 61, 28-39.

Capaldi

D. M.

Kim

H. K.

(2007). Typological approaches to violence in couples: A critique and alternative conceptual approach. Clinical Psychology Review, 27, 253-265.

Connolly

Friedlander

Pepler

Craig

Laporte

(2010). The ecology of adolescent dating aggression: Attitudes, relationships, media use, and socio-demographic risk factors. Journal of Aggression, Maltreatment & Trauma, 19, 469-491.

Connolly

McIsaac

(2009). Romantic relationships in adolescence. In Lerner

R. M.

Steinberg

(Eds.), Handbook of adolescent psychology, Vol. 2: Contextual influences on adolescent development (3rd ed., pp. 104-151). Hoboken, NJ: John Wiley.

Connolly

Nocentini

Menesini

Pepler

Craig

Williams

T. S.

(2010). Adolescent dating aggression in Canada and Italy: A cross-national comparison. International Journal of Behavioral Development, 34, 98-105.

Cook

W. L.

Kenny

D. A.

(2005). The Actor-Partner Interdependence model: A model of bidirectional effects in developmental studies. International Journal of Behavioral Development, 29, 101-109.

Cordova

J. V.

Jacobson

N. S.

Gottman

J. M.

Rushe

Cox

(1993). Negative reciprocity and communication in couples with a violent husband. Journal of Abnormal Psychology, 102, 559-564.

Foshee

V. A.

Matthew

R. A.

(2007). Adolescent dating abuse perpetration: A review of findings, methodological limitations, and suggestions for future research. In Flannery

D. J.

Vazonsyi

A. T.

Waldman

I. D.

(Eds.), The Cambridge handbook of violent behavior and aggression (pp. 431-449). New York, NY: Cambridge University Press.

10.

Foshee

V. A.

Reyes

H. L. M.

(2011). Dating abuse prevalence, consequences, and predictors. In Brown

B. B.

Prinstein

M. J.

(Eds.), Encyclopedia of adolescence (Vol. 3, pp. 119-126). San Diego, CA: Academic Press.

11.

Furman

Wehner

E. A.

(1997). Adolescent romantic relationships: A developmental perspective. In Shulman

Collins

W. A.

(Eds.), Romantic relationships: Developmental perspectives (pp. 21-36). San Francisco, CA: Jossey-Bass.

12.

Gottman

J. M.

Notarius

C. I.

(2000). Decade review: Observing marital interaction. Journal of Marriage and the Family, 62, 927-947.

13.

Hamby

Turner

(2013). Measuring teen dating violence in males and females: Insights from the national survey of children’s exposure to violence. Psychology of Violence, 3, 323-339.

14.

Hébert

Daigneault

Van Camp

(2012). Agression sexuelle et risque de revictimisation à l’adolescence: Modèles conceptuels et défis liés à la prévention [Sexual abuse and risk of revictimization in adolescence: Conceptual models and challenges for prevention]. In Hébert

Cyr

Tourigny

(Eds.), L’agression sexuelle envers les enfants, Tome II [Sexual abuse against children, Vol. II] (pp. 171-223). Québec, Canada: Presses de l’Université du Québec.

15.

Holtzworth-Munroe

Stuart

G. L.

(1994). Typologies of male batterers: Three subtypes and the differences among them. Psychological Bulletin, 116, 476-497.

16.

Johnson

M. P.

(1995). Patriarchal terrorism and common couple violence: Two forms of violence against women. Journal of Marriage and the Family, 57, 283-294.

17.

Johnson

M. P.

(2008). A typology of domestic violence: Intimate terrorism, violent resistance, and situational couple violence. Boston, MA: Northeastern University Press.

18.

Julien

Chartrand

Markman

H. J.

Lindahl

K. M.

(1991). Système d’observation des dimensions d’interaction [Interactional dimensions observation system]. Unpublished coding system, Department of Psychology Université du Québec à Montréal.

19.

Julien

Markman

H. J.

Lindahl

K. M.

(1989). A comparison of a global and a microanalytic coding system: Implications for future trends in studying interactions. Behavioral Assessment, 11, 81-100.

20.

Kenny

D. A.

Kashy

D. A.

Cook

W. L.

(2006). Dyadic data analysis. New York, NY: Guilford Press.

21.

Kline

G. H.

Julien

Baucom

Hartman

Gilbert

Gonzales

. (2004). The interactional dimensions coding system: A global system for couple interactions. In Kerig

Baucom

(Eds.), Couple observational coding systems. Mahwah, NJ: Lawrence Erlbaum.

22.

Langhinrichsen-Rohling

(2010). Controversies involving gender and intimate partner violence in the United States. Sex Roles, 62, 179-193.

23.

Langhinrichsen-Rohling

Capaldi

D. M.

(2012). Clearly we’ve only just begun: Developing effective prevention programs for intimate partner violence. Prevention Science, 13, 410-414.

24.

Langhinrichsen-Rohling

Selwyn

Rohling

M. L.

(2012). Rates of bidirectional versus unidirectional intimate partner violence across samples, sexual orientations, and race/ethnicities: A comprehensive review. Partner Abuse, 3, 199-230.

25.

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.

26.

Lavoie

Vézina

(2001). Violence faite aux filles dans le contexte des fréquentations à l’adolescence: Élaboration d’un instrument (VIFFA) [Violence committed against girls in the context of adolescent relationships: Elaboration of an instrument]. Revue Canadienne de santé mentale, 20(1), 153-171.

27.

Leonard

K. E.

Roberts

L. J.

(1998). The effects of alcohol on the marital interactions of aggressive and nonaggressive husbands and their wives. Journal of Abnormal Psychology, 107, 602-615.

28.

Margolin

John

R. S.

Gleberman

(1988). Affective responses to conflictual discussions in violent and nonviolent couples. Journal of Consulting and Clinical Psychology, 56, 24.

29.

Muthén

L. K.

Muthén

B. O.

(1998-2007). Mplus user’s guide (5th ed.). Los Angeles, CA: Author.

30.

O’Keefe

(2005). Teen dating violence: A review of risk factors and prevention efforts. National Electronic Network on Violence Against Women, pp. 1-13. Retrieved from http://www.unajauladeoro.com/cd/documentos/AR_TeenDatingViolence.pdf.

31.

O’Leary

K. D.

Slep

A. M. S.

(2003). A dyadic longitudinal model of adolescent dating aggression. Journal of Clinical Child and Adolescent Psychology, 32, 314-327.

32.

O’Leary

K. D.

Slep

A. M. S.

(2012). Prevention of partner violence by focusing on behaviors of both young males and females. Prevention Science, 13, 329-339.

33.

Reyes

H. L. M.

Foshee

V. A.

Bauer

D. J.

Ennett

S. T.

(2012). Heavy alcohol use and dating violence perpetration during adolescence: Family, peer and neighborhood violence as moderators. Prevention Science, 13, 340-349.

34.

Schumacher

J. A.

Feldbau-Kohn

Slep

A. M. S.

Heyman

R. E.

(2001). Risk factors for male-to-female partner physical abuse. Aggression and Violent Behavior, 6, 281-352.

35.

Siegel

(1956). Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill.

36.

Simon

V. A.

Bouchey

H. A.

Furman

(2000). The social construction of adolescent’s representation of romantic relationships. In Larose

Tarabulsy

G. M.

(Eds.), Attachment and development (Vol. 2, pp. 301-326). Québec, Canada: Les Presses de l’Université du Québec.

37.

Slep

A. M. S.

O’Leary

S. G.

(2005). Parent and partner violence in families with young children: Rates, patterns, and connections. Journal of Consulting and Clinical Psychology, 73, 435.

38.

Stith

S. M.

McCollum

E. E.

Amanor-Boadu

Smith

(2012). Systemic perspectives on intimate partner violence treatment. Journal of Marital and Family Therapy, 38, 220-240.

39.

Stith

S. M.

Smith

Penn

Ward

Tritt

(2004). Intimate partner physical abuse perpetration and victimization risk factors: A meta-analytic review. Aggression and Violent Behavior, 10, 65-98.

40.

Tharp

A. T.

(2012). Dating matters™: The next generation of teen dating violence prevention. Prevention Science, 13, 398-401.

41.

Tharp

A. T.

Burton

Freire

Hall

D. M.

Harrier

Latzman

N. E.

. . . Vagi

K. J.

(2011). Dating matters™: Strategies to promote healthy teen relationships. Journal of Women’s Health, 20, 1761-1765.

42.

Welsh

D. P.

Grello

C. M.

Dickson

J. W.

Harper

M. S.

(2001). The Adolescent Couples’ Issues Checklist. Unpublished questionnaire, Department of Psychology, University of Tennessee, Knoxville.

43.

Woodin

E. M.

(2011). A two-dimensional approach to relationship conflict: Meta-analytic findings. Journal of Family Psychology, 25, 325-335.