Revisiting the Role of Communication in Adolescent Intimate Partner Violence

Abstract

A growing literature suggests that communication strategies can promote or inhibit intimate partner violence (IPV). Research on communication is still needed on a group ripe for early IPV intervention: high school–aged adolescents. This article revisits our previous analyses of young female reproductive clinic patients (Messinger, Davidson, & Rickert, 2011) by examining how the adolescent and young adult respondents differ. To explore replicability of the adolescent results across populations, they are compared to 487 adolescent female students sampled from four urban high schools. Across samples, all communication strategies were used more frequently within violent relationships. Multivariate analysis identified escalating strategies used and received as being positively associated with physical violence used and received in all three samples. Regarding verbal reasoning and temporary conflict avoidance, substantial differences appeared between the young adult and adolescent clinic samples, and results from the adolescent clinic sample were largely replicated with the adolescent school sample, suggesting that young adult samples in this literature are not adequate proxies for adolescents.

Keywords

dating violence domestic violence predicting domestic violence

Quantitatively-oriented scholars of intimate partner violence (IPV) have given less attention to relationship communication than other IPV predictors, perhaps due to the challenges in operationalizing communication (Lloyd, 1990; Ridley & Feldman, 2003; Sabourin, 1996). A prominent theory contends that communication strategies are often perceived as resources to “win” conflicts, such that, when verbal reasoning fails, escalating communications may be turned to, and, if necessary, violence can become a final resort (for a review, see Messinger et al., 2011). Cross-sectional data suggests verbal reasoning does help defuse conflict and the potential for violence, whereas escalating strategies—verbal aggression and controlling tactics—escalate conflicts to violence (Billingham & Sack, 1986; Bird, Stith, & Schladale, 1991; Cornelius, Shorey, & Beebe, 2010; Gryl, Stith, & Bird, 1991; Josephson & Proulx, 2008; Messinger et al., 2011; Ridley & Feldman, 2003). Furthermore, evidence indicates that temporary conflict avoidance is not associated with physical violence (Messinger et al., 2011). A limited number of studies find youth in violent relationships use both escalating strategies (Billingham & Sack, 1986; Messinger et al., 2011; Ridley & Feldman, 2003) and temporary avoidance strategies (Messinger et al., 2011) more frequently than youth in nonviolent relationships. Perhaps this is due to violent relationships encountering both more conflicts and more serious conflicts than nonviolent relationships (Gryl et al., 1991). Each type of communication strategy tends to be reciprocated in youth relationships (Messinger et al., 2011), as is physical violence (Fernandez & Fuertes, 2010; O’Leary, Smith Slep, Avery-Leaf, & Cascardi, 2008) though it is debated whether this implies mutual battering or violence in self-defense (Williams, Ghandour, & Kub, 2008).

Analyses on a range of communication strategies as predictors of physical IPV have been conducted on marital, college-aged, and middle school populations, but only one such article used a sample of high school–aged adolescents. In their recent survey of 618 female reproductive clinic patients aged 15 to 24, Messinger et al. (2011) found that physical violence used and received in opposite-sex relationships were positively associated with escalating strategies and negatively associated with verbal reasoning by either partner (with the exception of nonsignificance for reasoning used predicting violence used), while avoidance was not associated with violence in any model.

However, by pooling together such a broad age spectrum, it may be that age-related differences were aggregated and masked. The literature hints that developmental processes may affect relationship communication pathways. For example, interpersonal communication styles are known to shift during adolescence (Laursen & Collins, 1994), and some evidence suggests that high school–aged females are far more “dismissive” (i.e. avoidant) and insecurely jealous or “preoccupied” (i.e. escalating) than college-aged females, while “secure” relationship styles (likely involving more verbal reasoning) are more equally distributed across age groups (Furman & Wehner, 1997). Relationship contexts, which may affect triggers to conflict, are also known to vary by age, including the relationship length and importance, prior relationship experience, and degree of sexual intimacy (Furman & Wehner, 1997). Furthermore, age is associated with the prevalence (Rennison, 2001) and mutuality of violence (Morse, 1995), as are marital and cohabitation status (Stets & Straus, 1990), again hinting at developmental variations in IPV relationships. If indeed relationship communication pathways are affected by a developmental process, previous IPV communications research on young adults cannot be generalized to apply to adolescents. This possibility underscores the need for IPV communication research on high school–aged adolescents, a population ripe for early monitoring and intervention (O’Leary, 1999; Rennison, 2001; Rivara et al., 2009).

The present article aimed to isolate the role of relationship communication in IPV among female adolescents by examining differences between the 15- to 19-year-old female adolescents and the 20- to 24-year-old female young adults in Messinger et al.’s (2011) previous sample of female clinic patients. To explore replicability of the adolescent findings across populations, the adolescent female clinic patients are then compared to a sample of 14- to 19-year-old female adolescent students from four high schools. Based on the literature, we developed three hypotheses. First, we predicted that all communication strategies would be used and received more frequently in violent relationships. Second, we predicted that adolescent females would report using more avoidant and escalating strategies than young adult females. Third, in the absence of preexisting empirical evidence, we also predicted that each of our samples would closely follow results from our prior work in that physical violence used and received would be positively associated with escalating strategies used and received, negatively associated with verbal reasoning used and received, and not associated with avoidance used or received.

Method

To test the above hypotheses, three samples were used. The first two samples are drawn from a single larger group of 618 female reproductive health clinic patients aged 15 to 24 who had experienced an opposite-sex dating relationship in the past year. After being recruited by staff members, respondents completed an anonymous survey through ACASI (audio computer-assisted self-interviewing) and were compensated with a US$10 gift certificate (for additional study methodology, see Messinger et al., 2011). The present article categorized this group into a “young adult clinic sample”—320 female patients 20 to 24 years old—and an “adolescent clinic sample”—298 female patients 15 to 19 years old. The third sample is from a larger study of 1,454 male and female students from four urban high schools in the same city as the clinic, recruited through passive parental consent and student assent or consent. Students were given an ACASI or paper-and-pencil questionnaire version, each participant being compensated with a US$10 gift card (for a complete methodology report, see Fry, Davidson, Rickert, & Lessel, 2008). Of the 790 female students, this article analyzes the 487 adolescent females aged 14 to 19 who reported having had an opposite-sex dating relationship in the past year, termed the “adolescent school sample.” The male students were excluded from analyses both to strengthen comparisons with the all-female clinic sample and because the male students may have been referring to the same relationships as female students, thus artificially weighting results. By employing these three samples, this article aims to, first, distinguish which results are age-specific for adolescents as compared to young adults and, second, whether these adolescent findings are robust across differing population samples.

Identical variables, coding, and analyses were used for all three samples. The two dependent variables, the frequency of physically violent behaviors used and received, are each subscales drawn from the Conflict in Adolescent Dating Relationships Inventory (CADRI; Wolfe et al., 2001). Four item pairs asked if the respondent and the respondent’s partner used the following behaviors: threw something at; pushed, shoved, or shook; slapped or pulled the hair of; or kicked, hit, or punched the other partner. Possible responses were zero times (coded 0), 1 to 2 times (coded 1), 3 to 5 times (coded 2), or six or more times (coded 3). These items were summed to create scales for the use and receipt of physical violence. Also drawn from the CADRI (Wolfe et al., 2001), the six main independent variables include the amounts of verbal reasoning used and received (e.g., “offered a solution that I thought would make us both happy”), temporary conflict avoidance used and received (e.g., “put off talking until we calmed down”), and escalating strategies used and received, which is a combination of movement-restricting controlling strategies (e.g., “kept track of who s/he was with and where s/he was”) and verbally aggressive strategies (e.g., “said things just to make him/her angry”). These key independent variables are scored the same way as the physical violence dependent variables. Verbal reasoning and temporary conflict avoidance, originally a single submeasure of the CADRI, were treated as separate constructs because evidence suggests that they have very distinct impacts on violence in relationships (see Messinger et al., 2011). Cronbach’s alpha reliability coefficients were strong for violence used (young adult clinic sample, adolescent clinic sample, and adolescent school sample: α = .88, .82, and .84 respectively) and received (α = .87, .87, and .86), escalation strategies used (α = .89, .86, and .86) and received (α = .90, .88, and .85), and reasoning used (α = .86, .78, .78) and received (α = .84, .79, and .78), while reliability was more modest for avoidance used (α = .65, .61, and .52) and received (α = .66, .64, and .46). Last, several known predictors of IPV were controlled for, including experiencing child sexual abuse, relationship importance, relationship length, number of pregnancies, age, and race-ethnicity. Since the CADRI inquires only about the previous year, analyses were adjusted with an exposure time variable assessing the number of months a relationship existed within the past year.

There were no missing data in the adolescent and young adult clinic samples. In the adolescent school sample, there were missing cases for physical violence used (n = 1) and received (n = 34), reasoning used (n = 17) and received (n = 41), avoidance used (n = 7) and received (n = 38), escalation used (n = 30) and received (n = 44), experiencing child sexual abuse (n = 14), relationship importance (n = 4) and length (n = 2), times pregnant (n = 13), race-ethnicity variables (n = 59), and months exposed to the CADRI time frame (n = 12). Adolescent school data was missing at random (MAR) rather than missing completely at random (MCAR), with missingness on several of the communication variables significantly associated with lower mean scores on other communication and violence variables (t tests, p < .05). Pairwise deletion of missing cases was used for bivariate descriptive analysis of the adolescent school data. However, to account for patterns of missingness in multivariate analysis of the adolescent school sample, we employed STATA’s ICE program (Royston, 2009), where multiple imputed data sets with five iterations were generated for each model analyzed, each time using all the variables for that given model during the imputation process.

Following Messinger et al.’s (2011) approach, for multivariate analyses, the three samples were each divided into two overlapping subsamples: the first subsamples included respondents who, in the past year, used physical violence (young adult clinic subsample No. 1, n = 88; adolescent clinic subsample No. 1, n = 108; adolescent school subsample No. 1, n = 215) and, for comparison, those who neither used nor received violence (young adult clinic subsample No. 1, n = 225; adolescent clinic subsample No. 1, n = 181; adolescent school subsample No. 1, n = 239), and the second set of subsamples included respondents who, in the past year, received violence (young adult clinic subsample No. 2, n = 64; adolescent clinic subsample No. 2, n = 72; adolescent school subsample No. 2, n = 136) and, again, the same respondents from the first subsamples who neither used nor received violence. Violence users and receivers were never directly compared because of the considerable number falling into both groups (young adult clinic sample, n = 57; adolescent clinic sample, n = 63; adolescent school sample, n = 128).

Each use-receipt communication variable set was split into separate models for multivariate analysis due to strong multicollinearity within each set (within-set correlations: young adult clinic sample, r = .73-.88; adolescent clinic sample, r = .63-.89; adolescent school sample, r = .71-.87) and far lower correlation coefficients for bivariate associations across strategy types. The communication variables were assigned models to replicate Messinger et al.’s (2011) approach so as to increase the strength of cross-paper comparisons; in addition, these models reflect the literature’s assertion that escalating strategies are often perceived as a means to gain control relative to a less contentious approach like reasoning or avoidance. Specifically, negative binomial regression, suited for dependent variables measuring an event frequency for which the probability of occurrence shifts over time, was conducted on four models to see if (M1) escalating strategies used and avoidance and reasoning received affect physical violence used by the young women, if (M2) escalating strategies used and avoidance and reasoning received affect violence received, if (M3) escalating strategies received and avoidance and reasoning used affect violence used, and if (M4) escalating strategies received and avoidance and reasoning used affect violence received. Models include all aforementioned control variables and were adjusted for exposure to the CADRI’s time frame. Consistent with Messinger et al.’s (2011) analytic approach, Models M1 and M3 were run with the young adult clinic subsample No. 1, adolescent clinic subsample No. 1, and adolescent school subsample No. 1 (violence used by the women), and Models M2 and M4 were run with the young adult clinic subsample No. 2, adolescent clinic subsample No. 2, and adolescent school subsample No. 2 (violence received by the women).

Results

Demographically, the young adult clinic sample had a mean age of 22.04 years, 25% of the sample reported Latino descent, and, regarding race, 28% of the sample was Black, 32% White, and 12% Asian American. Our adolescent clinic sample (M = 17.78) was slightly older than the adolescent school sample (M = 16.02). Ethnically, 40% of the adolescent clinic sample reported Latino descent, as compared to 76% of the adolescent school sample. Racially, 40% of the adolescent clinic sample reported their race as Black, 16% Asian American, and 12% White, as compared to the adolescent school sample who reported 48% Black, 1% Asian American, and 6% White. Despite demographic differences between these samples, the means for the communication variables suggest marginal differences (Table 1). Given evidence in the literature to the contrary (see Miller et al., 2010), it is surprising that violence means were lower, albeit only slightly, for both of the clinic samples. Though the correlations (p < .05) between the communication and physical violence variables for the most part were stronger within the adolescent school sample—perhaps partially due to its higher violence means—the relative strength of the correlations between communication types was predominantly stable across samples. As predicted in our first hypothesis, in all samples, all communication strategies were used and received significantly more frequently (t test, p < .05) in violent relationships than nonviolent relationships (Table 2). In addition, for our second hypothesis we expected that adolescent females would report using more avoidant and escalating strategies than young adults. This was the case, if marginally, for escalating communication used, whereas adolescent clinic patients were actually less avoidant than young adult clinic patients. There were no significant differences between the adolescent and young adult clinic samples in mean frequencies of communication strategies and violence (p < .05).

Table 1.

Comparing Means, Standard Deviations, and Correlation Matrices for the Young Adult Clinic, Adolescent Clinic, and Adolescent School Female Samples

Variables	1	2	3	4	5	6	7	8	Sample
1. Physical used	—	0.68*	0.28*	0.33*	0.23*	0.17*	0.53*	0.57*	Clinic
2. Physical rec	0.59*	—	0.20*	0.23*	0.15*	0.02	0.43*	0.57*	20-24 yo, N = 320
3. Avoidance used	0.18*	0.12*	—	0.73*	0.60*	0.54*	0.45*	0.44*
4. Avoidance rec	0.24*	0.15*	0.63*	—	0.54*	0.48*	0.49*	0.45*
5. Reasoning used	0.16*	0.16*	0.38*	0.40*	—	0.88*	0.55*	0.50*	Clinic
6. Reasoning rec	0.19*	0.19*	0.40*	0.39*	0.89*	—	0.42*	0.36*	15-19 yo, N = 298
7. Escalating used	0.58*	0.45*	0.23*	0.41*	0.48*	0.45*	—	0.88*
8. Escalating rec	0.46*	0.55*	0.23*	0.39*	0.43*	0.39*	0.85*	—
1. Physical used	—								School
2. Physical rec	0.68*	—							14-19 yo, N = 487
3. Avoidance used	0.27*	0.21*	—
4. Avoidance rec	0.22*	0.16*	0.71*	—
5. Reasoning used	0.30*	0.25*	0.52*	0.50*	—
6. Reasoning rec	0.35*	0.22*	0.49*	0.49*	0.89*	—
7. Escalating used	0.65*	0.54*	0.37*	0.35*	0.45*	0.45*	—
8. Escalating rec	0.60*	0.61*	0.34*	0.35*	0.45*	0.39*	0.84*	—
M	0.91	0.62	2.56	2.45	9.73	9.03	8.27	7.66	Clinic
SD	2.02	1.67	1.77	1.83	4.49	4.46	6.62	6.78	20-24 yo
M	1.15	0.76	2.48	2.49	9.72	9.18	8.64	8.19	Clinic
SD	2.08	1.85	1.68	1.72	3.96	4.10	6.26	6.61	15-19 yo
M	1.71	0.95	2.54	2.41	8.68	8.47	7.09	6.20	School
SD	2.70	2.14	1.69	1.66	4.15	4.18	5.57	5.52	14-19 yo

Note: Variables listed are additive frequency scales from the CADRI. Due to differing numbers of items per scale, these variables have different score ranges (avoidance, 0-6; reasoning, 0-30; escalating, 0-18; violence, 0-12). Rec = Received; yo = years old.

p < .05.

Table 2.

Means for Avoidance, Reasoning, and Escalation Strategies Among Women Using and Receiving Violence Compared to Those in Nonviolent Relationships

	Strategies
Sample	Avoidance Used	Avoidance Received	Reasoning Used	Reasoning Received	Escalating Used	Escalating Received
Clinic, 20-24 yo
No violence	2.21*	2.03*	8.96*	8.48*	5.78*	5.10*
Violence users	3.41	3.45	11.56	10.35	14.22	13.81
Violence receivers	3.36	3.56	11.61	10.20	14.81	15.00
Clinic, 15-19 yo
No violence	2.23*	2.12*	9.09*	8.51*	5.95*	5.53*
Violence users	2.91	3.04	10.53	10.13	12.90	12.06
Violence receivers	2.78	3.17	10.78	10.22	13.96	13.99
School, 14-19 yo
No violence	2.11*	2.06*	7.58*	7.17*	4.16*	3.56*
Violence users	3.14	2.81	10.15	10.02	10.64	9.17
Violence receivers	3.04	2.76	10.46	10.01	11.62	10.88

Note: Respondent groups defined by use and receipt of physical violence in dating relationship of past year. Due to differing numbers of items per scale, strategy types have differing score ranges (avoidance, 0-6; reasoning, 0-30; escalating, 0-18); t test difference both between the no-violence group and violence users and between the no-violence group and violence receivers. yo = years old.

p < .05.

Looking at the negative binomial regression models (p < .05) detailed in Tables 3 and 4, with the young adult clinic sample, interestingly, none of the control variables were predictive of the frequency of violence used or received. Regressions of the adolescent clinic sample revealed that greater frequency of physical violence used was associated with shorter relationships and reporting Black race, while greater frequency of received physical violence was associated with experiencing more pregnancies and receipt of child sexual abuse. In the adolescent school sample, greater frequency of violence used was associated again with shorter relationships as well as with less important relationships and younger respondents, while greater frequency of violence received was associated with shorter relationships and experiencing child sexual abuse.

Table 3.

Comparison of Adolescent Women’s Use of Physical Violence With Adolescent Women Not in a Physically Violent Relationship

	Dependent Variable: Physical Violence Used by Women
	Clinic, yo 20-24^a		Clinic, 15-19 yo^a		School, 14-19 yo
Predictor	IRR	95% CI	IRR	95% CI	IRR	95% CI
Reasoning used	0.97	[0.89, 1.05]	0.96	[0.90, 1.03]	1.02	[0.97, 1.07]
Reasoning rec	0.97	[0.90, 1.04]	0.93*	[0.88, 0.99]	1.04	[1.00, 1.09]
Avoidance used	1.20*	[1.02, 1.42]	1.03	[0.90, 1.17]	1.08	[0.98, 1.19]
Avoidance rec	1.19*	[1.02, 1.39]	1.00	[0.88, 1.14]	0.98	[0.89, 1.07]
Escalating used	1.20*	[1.14, 1.26]	1.19*	[1.14, 1.24]	1.18*	[1.15, 1.22]
Escalating rec	1.18*	[1.14, 1.23]	1.13*	[1.09, 1.18]	1.15*	[1.11, 1.18]

Note: Negative binomial regressions conducted on violence users and respondents not in a violent relationship. Due to multicollinearity of used-received communication variable pairs, these pairs were divided into four models, presented here in one table for ease of interpretation. Control variables described in Methods section were included in all models. Variables listed are additive frequency CADRI scales. Rec = Received; IRR = incidence-rate ratios; CI = confidence interval; yo = years old.

Drawn from same larger sample of 15- to 24-year-old female clinic patients.

p < .05.

Table 4.

Comparison of Adolescent Women’s Receipt of Physical Violence With Adolescent Women Not in a Physically Violent Relationship

	Dependent Variable: Physical Violence Received by Women
	Clinic, yo 20-24^a		Clinic, 15-19 yo^a		School, 14-19 yo
Predictor	IRR	95% CI	IRR	95% CI	IRR	95% CI
Reasoning used	0.91*	[0.84, 0.98]	0.94	[0.87, 1.02]	0.99	[0.94, 1.05]
Reasoning rec	0.91*	[0.83, 0.99]	0.99	[0.92, 1.07]	0.99	[0.94, 1.04]
Avoidance used	1.06	[0.89, 1.27]	1.03	[0.89, 1.18]	1.08	[0.96, 1.21]
Avoidance rec	1.12	[0.93, 1.35]	0.95	[0.81, 1.11]	0.96	[0.85, 1.09]
Escalating used	1.21*	[1.14, 1.29]	1.21*	[1.15, 1.28]	1.22*	[1.18, 1.27]
Escalating rec	1.23*	[1.18, 1.29]	1.20*	[1.16, 1.25]	1.21*	[1.16, 1.25]

Note: Negative binomial regressions conducted on violence receivers and respondents not in a violent relationship. Due to multicollinearity of used-received communication variable pairs, these pairs were divided into four models, presented here in one table for ease of interpretation. Control variables described in Methods section were included in all models. Variables listed are additive frequency CADRI scales. Rec = Received; IRR = incidence-rate ratios; CI = confidence interval; yo = years old.

Drawn from same larger sample of 15- to 24-year-old female clinic patients.

p < .05.

Regarding the key predictors in these multivariate models, escalating strategies were positively associated with physical violence used and received in all three samples, wherein a one-unit increase in the amount of escalating strategies by either partner was associated with an increase in violence used by 13% to 20% and violence received by 20% to 23%. The points of departure among these samples were with reasoning and avoidance. With this in mind, several age differences were apparent between the clinic samples. For the young adult clinic patients, avoidance used and received were associated with an increase in violence used (18%-20% per unit increase of avoidance), and reasoning used and received were associated with a decrease in violence received (9% per unit increase of reasoning). Conversely, for the adolescent patients, every unit increase of reasoning received was associated with a 7% decrease in violence used. This particular adolescent clinic sample result on reasoning—in only one sample and with a weak effect size—was the only to differentiate the two adolescent samples. Beyond this, the remaining results of the adolescent clinic sample were replicated with the adolescent school sample, with only escalating strategies (in all models) predicting violence.

In reviewing these results, it is apparent our third hypothesis held across multivariate analyses of all three samples for only reports of escalating strategies used and received, which were positively associated with violence used and received as expected. While we predicted for all samples that reasoning used and received would be negatively associated with violence used and received as in our previous article (Messinger et al., 2011), instead we found reasoning was negatively associated with violence only with the two clinic samples, and the gender of the violence user differed across the two age groups. Lastly, as with our previous article merging the young adult and adolescent clinic patients, we predicted for all samples that violence would not be associated with avoidance. However, the nonsignificant negative association in our previous article was significant in our present analyses on violence used by the young adult clinic patients.

Discussion

This study addressed an important gap in the IPV literature regarding the role of relationship communication among adolescents, an age group particularly appropriate for early monitoring and intervention. Reflecting the literature, and as predicted in our first hypothesis, females in all three samples involved in violent relationships reported using and receiving all communication strategies—reasoning, avoidance, and escalating strategies—more frequently than those in nonviolent relationships. Future research could test whether this relationship is spurious with a conflict frequency measure since conflicts generate opportunities for both violence and conflict resolution communications. As to whether the literature is accurate in the posited time order for communication strategies, future research, especially longitudinal, will be invaluable. Given evidence of an increased IPV risk for clinic patients (see Miller et al., 2010), it was surprising that violence means would be lower within the clinic sample. Demographic sample differences could help explain this, but additional inquiry is needed. Importantly, contrary to what has been suggested and predicted in our second hypothesis, we found no significant differences between the adolescent and young adult clinic samples in reported mean frequencies of communication strategies and violence. This suggests that aging may affect the motivations behind and perceptions of communication strategies more so than the frequencies with which those strategies are employed.

In multivariate analysis, control variables provided both expected and unexpected results. That having experienced child sexual abuse is associated with violence receipt in the adolescent samples is consistent with prior research and suggests children can learn from parents to normalize violence victimization (e.g., Gómez, 2011; Heyman & Slep, 2002). However, the findings that shorter relationships in both adolescent samples and less important relationships in the adolescent school sample were associated with the use of violence runs contrary to prior adult and college student research on relationship seriousness (Lewis & Fremouw, 2001). Why no control variables were predictive of violence for the young adult clinic sample is not entirely clear.

Our third hypothesis regarding key variables in multivariate analysis was drawn only from prior research on young adults and adults due to the lack of similar research on adolescents, so in this regard these analyses were exploratory. It is perhaps not surprising that we found support for only a portion of our third hypothesis in that the reported use and receipt of escalating strategies were positively associated with the use and receipt of physical violence in all samples, a common result in research on adolescents (e.g., Halpern et al., 2001). Thus the key differences resided with reasoning and avoidance. When comparing the adolescent and young adult clinic patients, differences emerged that indicate our previous article (Messinger et al., 2011) may have masked age differences when combining these two groups in analysis. Our results suggest that, once one reaches young adulthood, reasoning by either partner is not associated with violence by females but is predictive of a decrease in violence by their male partners. Future research should explore the qualitative use and interpretation of reasoning in relationships, with a strong focus on gender differences. As indicated by the two clinic samples, by young adulthood, temporary avoidance used and received may become a significant trigger to violence for women, which contradicts one of the few articles on this topic (Furman & Wehner, 1997). Significant correlates of avoidance did not differ across the two clinic samples, but it is possible that the perceptions and purposes of avoidance differ with age, such as avoiding to acquire control over a partner versus avoiding due to deficits in verbal reasoning. A conflict frequency variable in future research may help determine if these significant differences between age groups are in part explained by variance in differential mean conflict frequency, though this is less likely the case for escalating strategies given its large coefficient across samples.

This comparison of the young adult and adolescent clinic samples reaffirms our suspicion that existing studies on young adults are not a sufficient proxy for data on adolescents. By focusing analyses on our adolescent clinic sample and then comparing it to the adolescent school sample, we were able to explore replicability. Our two adolescent samples differed solely in that violence used by the females was negatively associated with reasoning received from their male partners only in the adolescent clinic sample. For the most part our findings provide evidence of generalizability with far more commonalities across population type (clinic vs. school and differing ethnic groups) than differences. While future research would do well to verify our results with additional adolescent populations, several intervention implications can be drawn from our adolescent results. With a one-unit decrease in escalating strategy frequency being associated with a 13% to 22% dip in violence frequency regardless of gender, finding ways to decrease the use of escalating strategies in adolescent relationships remains a potentially fruitful violence reduction strategy. Conversely, the inhibiting effect of reasoning on violence is present neither for males as reported by females in the adolescent clinic sample, nor for both males and females as reported by females in the adolescent school sample, which raises concerns as to the potential robustness of reasoning for defusing adolescent conflict. In addition, avoidance is not associated with violence in either adolescent model, indicating that this may not be a useful avenue for decreasing violence in adolescent relationships.

Like most research, this study is not without its limitations. In particular, for feasibility reasons, this study employs a cross-sectional design, so the temporal relationship of strategies could not be tested. While relying on female self-reports from the male–female school sample enabled us to make stronger comparisons with the all-female clinic sample, it is possible that results and conclusions may have differed had males been included, particularly as evidence strongly suggests males tend to underreport physical violence in relationships relative to their female partners (Armstrong, Wernke, Medina, & Schafer, 2002). Furthermore, contextual variables like violence motivations, outcomes, and initiation were not assessed, so value labels like “abuser” and “victim” cannot be readily applied to our data (Williams, Ghandour & Kub, 2008). That said, these data push the discourse forward on communication in IPV relationships, both in verifying adolescent IPV relationship communication patterns across population types and suggesting potential points of departure between adolescents and young adults. Given these findings, future empirical and policy-oriented work should continue to explore whether indeed the use of a range of communication strategies is affected by life transitions between adolescence and young adulthood.

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 report was supported in part by the cooperative agreement 5 U49 CE000731 from the Centers for Disease Control and Prevention (CDC) with the Center for Youth Violence Prevention at Columbia University. Its contents are the sole responsibility of the authors and do not necessarily represent the official views of the CDC or Columbia University.

Bios

Adam M. Messinger, PhD, is an Assistant Professor of Justice Studies at Northeastern Illinois University. He is also continuing his research collaboration spanning the past several years with the Centers for Disease Control and Prevention’s ACE (National Academic Centers for Excellence in Youth Violence Prevention) at Columbia University. His research focuses on youth violence prevention - particularly regarding intimate partner violence, child abuse, sexual violence, and youth bullying - and its intersections with gender identity, sexual orientation, stigma, and the criminal justice system.

Vaughn I. Rickert, PsyD, is a professor of pediatrics and director of the Section of Adolescent Medicine at Indiana University School of Medicine and Riley Hospital for Children. He received his doctorate in Clinical Psychology from Central Michigan University. His research interests focus on adolescent relationship violence and contraceptive use in this population. Finally, Dr. Rickert is a past president of the Society for Adolescent Medicine or SAM, a multidisciplinary professional society devoted exclusively to the health and well-being of adolescents.

Deborah A. Fry, MA, MPH, is a research fellow at Centre for Learning in Child Protection at the University of Edinburgh. Her primary research interests are in the areas of child sexual abuse, sexual violence, children who display sexually harmful behavior, and intimate partner violence prevention and response. Prior to joining the Centre, Deborah was the Research Director at the NYC Alliance Against Sexual Assault.

Harriet Lessel, CSW, is founder and former Executive Director of The New York City Alliance Against Sexual Assault. Prior to this position, Ms. Lessel spent ten years managing the rape crisis and domestic violence program at the Long Island College Hospital (LICH) in Brooklyn, New York. In addition, she developed the first Sexual Assault Examiners program in Brooklyn, which helps improve the hospital care of rape survivors. She has been active in city and state-wide anti-sexual violence and crime victims coalitions, and, as a certified social worker, previously she was a counselor and advocate for survivors of sexual assault and abuse.

Leslie L. Davidson, MD, is a pediatrician and epidemiologist. Her primary research interests are screening, and epidemiology focusing on the prevention of accidents and violence, particularly intimate partner violence. She has also conducted substantial research in disability in children and in international child health. She is a professor of clinical epidemiology and clinical pediatrics at the Mailman School of Public Health at Columbia University and associate director of the Columbia Center for Youth Violence.

References

Armstrong

T. G.

Wernke

J. Y.

Medina

K. L.

Schafer

(2002). Do partners agree about the occurrence of intimate partner violence? A review of the current literature. Trauma, violence & abuse, 3, 181-193.

Billingham

R. E.

Sack

A. R.

(1986). Courtship violence and the interactive status of the relationship. Journal of Adolescent Research, 1, 315-325.

Bird

G. W.

Stith

S. M.

Schladale

(1991). Psychological resources, coping strategies, and negotiation styles as discriminators of violence in dating relationships. Family Relations, 40(1), 45-50.

Cornelius

T. L.

Shorey

R. C.

Beebe

S. M.

(2010). Self-reported communication variables and dating violence: Using Gottman’s marital communication conceptualization. Journal of Family Violence, 25, 439-448.

Fernandez-Fuertes

A. A.

Fuertes

(2010). Physical and psychological aggression in dating relationships of Spanish adolescents: Motives and consequences. Child Abuse & Neglect, 34, 183-191.

Fry

D. A.

Davidson

L. L.

Rickert

V. I.

Lessel

(2008). Partners and peers: Sexual and dating violence among NYC youth (A research report by the New York City Alliance Against Sexual Assault and the Columbia Center for Youth Violence Prevention, Columbia University). New York: New York City Alliance Against Sexual Assault.

Furman

Wehner

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

Collins

W. A.

(Eds.), Romantic relationships in adolescence: Developmental perspectives. New directions for child development, No. 78 (pp. 21-36). San Francisco, CA: Jossey-Bass.

Gómez

A. M.

(2011). Testing the cycle of violence hypothesis: Child abuse and adolescent dating violence as predictors of intimate partner violence in young adulthood. Youth and Society, 43, 171-192.

Gryl

F. E.

Stith

S. M.

Bird

G. W.

(1991). Close dating relationships among college students: Differences by use of violence and by gender. Journal of Social and Personal Relationships, 8, 243-264.

10.

Halpern

C. T.

Oslak

S. G.

Young

M. L.

Martin

S. L.

Kupper

L. L.

(2001). Partner violence among adolescents in opposite-sex romantic relationships: Findings from the National Longitudinal Study of Adolescent Health. American Journal of Public Health, 91(10), 1679-1685.

11.

Heyman

R. G.

Slep

A. M. S.

(2002). Do child abuse and interparental violence lead to adulthood family violence? Journal of Marriage and the Family, 64, 864-870.

12.

Josephson

W. L.

Proulx

J. B.

(2008). Violence in young adolescents’ relationships: A path model. Journal of Interpersonal Violence, 23, 189-208.

13.

Laursen

Collins

W. A.

(1994). Interpersonal conflict during adolescence. Psychological Bulletin, 115, 197-209.

14.

Lewis

S. F.

Fremouw

(2001). Dating violence: A critical review of the literature. Clinical Psychology Review, 21(1), 105-127.

15.

Lloyd

S. A.

(1990). Conflict types and strategies in violent marriages. Journal of Family Violence, 5, 269-283.

16.

Messinger

A. M.

Davidson

L. L.

Rickert

V. I.

(2011). IPV among adolescent reproductive health clinic patients: The role of relationship communication. Journal of Interpersonal Violence, 26, 1851-1867.

17.

Miller

Decker

M. R.

Raj

Reed

Marable

Silverman

J. G.

(2010). Intimate partner violence and health care-seeking patterns among female users of urban adolescent clinics. Maternal Child Health Journal, 14, 910-917.

18.

Morse

B. J.

(1995). Beyond the conflict tactics scale: Assessing gender differences in partner violence. Violence and Victims, 10(4), 251-272.

19.

O’Leary

K. D.

(1999). Developmental and affective issues in assessing and treating partner aggression. Clinical Psychology: Science and Practice, 6, 400-414.

20.

O’Leary

K. D.

Smith Slep

A. M.

Avery-Leaf

Cascardi

(2008). Gender differences in dating aggression among multiethnic high school students. Journal of Adolescent Health, 42, 473-479.

21.

Rennison

C. M.

(2001). Special report: Intimate partner violence and age of victim, 1993-99 (NCJ No. 187635). Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics.

22.

Ridley

C. A.

Feldman

C. M.

(2003). Female domestic violence toward male partners: Exploring conflict responses and outcomes. Journal of Family Violence, 18, 157-170.

23.

Rivara

F. P.

Anderson

M. L.

Fishman

Reid

R. J.

Bonomi

A. E.

Carrell

Thompson

R. S.

(2009). Age, period, and cohort effects on intimate partner violence. Violence and Victims, 24, 627-638.

24.

Royston

(2009). Multiple imputation of missing values: Further update of ice, with an emphasis on categorical variables. Stata Journal, 9, 466-477.

25.

Sabourin

T. C.

(1996). The role of communication in verbal abuse between spouses. In Cahn

D. D.

Lloyd

S. A.

(Eds.), Family violence from a communication perspective (pp. 199-217). Thousand Oaks, CA: SAGE.

26.

Stets

J. E.

Straus

M. A.

(1990). The marriage license as hitting license: A comparison of assaults in dating, cohabiting, and married couples. In Straus

M. A.

Gelles

R. J.

(Eds.), Physical violence in American families: Risk factors and adaptations to violence in 8,145 families (pp. 227-244). New Brunswick, NJ: Transaction.

27.

Williams

J. R.

Ghandour

R. M.

Kub

J. E.

(2008). Female perpetration of violence in heterosexual intimate relationships: Adolescence through adulthood. Trauma, Violence, & Abuse, 9(4), 227-249.

28.

Wolfe

D. A.

Scott

Reitzel-Jaffe

Wekerle

Grasley

Straatman

A. L.

(2001). Development and validation of the Conflict in Adolescent Dating Relationships Inventory. Psychological Assessment, 13, 277-293.