Abstract
Intimate partner violence (IPV) is a major public health concern. Previous studies have consistently shown that IPV is tied by to a variety of detrimental consequences for affected individuals, including negative mental health outcomes. However, the differential impact of gender and perpetrator-victim role (i.e., whether an individual is the perpetrator or victim of violence or both) remains largely understudied in the academic literature. Thus, the purpose of the present study was to describe a variety of mental health outcomes and risk behaviors among men and women experiencing no violence, perpetration only, victimization only, and bidirectional violence. Data from Waves 3 and 4 of the National Longitudinal Study of Adolescent Health (N = 7187) were used. Participants provided information on their perpetrator-victim role and on a variety of factors related to mental health (depression, suicidality, alcohol use, illegal drug use, and relationship satisfaction). For all outcomes, prevalence and severity generally tended to be highest among individuals affected by bidirectional IPV and lowest among individuals not affected by any violence (independent of gender). The present findings highlight that IPV and negative mental health outcomes and risk behaviors should be addressed as co-occurring problems in research, prevention, and treatment. In addition, all gender-role combinations should be addressed in order to better understand and address all potential effects of IPV. According to the present findings, couples affected by bidirectional violence are at particularly high risk for developing mental health disorders. Thus, policy makers and clinicians should predominantly target couples as well as individuals, who are not only the victims but also the perpetrators of IPV, and pay particular attention to potential signs of mental health distress these individuals might exhibit.
Keywords: Intimate partner violence, mental health, gender, perpetrator-victim role
Intimate partner violence (IPV) is a major public health concern that has been tied to a variety of detrimental consequences for affected individuals, including poor mental health (Pico-Alfonse et al., 2006), alcohol and drug problems (e.g., Coker et al., 2002), and decreased relationship satisfaction (e.g., Lawrence & Bradbury, 2001). However, the majority of previous studies have focused on the negative consequences associated with male-to-female violence (e.g., Cascardi & O’Leary, 1992; Cascardi, O’Leary, Lawrence, & Schlee, 1995), while female-to-male violence has been largely understudied. In addition, most previous studies examining negative consequences of IPV have solely focused on IPV victimization. Despite findings indicating that more than half of IPV victims also have perpetrated IPV (Anderson, 2002), studies examining outcomes among perpetrators of IPV and individuals affected by bidirectional IPV are rather limited. Thus, it remains largely unclear whether the association between IPV and mental health outcomes and risk behaviors might differ based on individuals’ gender and their perpetrator-victim role (i.e., whether they are the perpetrator, the victim, or both). Therefore, the aim of the present study was to examine the differential impact of gender and perpetrator-victim role on the association between IPV and associated outcomes, including depression, suicidality, alcohol and drug use, and relationship satisfaction.
An association between IPV and mental health is well established in the academic literature (e.g., Pico-Alfonso et al, 2006). More specifically, studies show that individuals affected by IPV are at increased risk for experiencing symptoms of depression (e.g., Filson, Ulloa, Runfola, & Hokoda, 2010), for attempting and committing suicide (Pico-Alfonse et al., 2006), for engaging in alcohol and drug use and abuse (e.g., Cunradi, Caetano, & Schafer, 2002), and for experiencing decreased relationship satisfaction (Lawrence & Bradbury, 2001). In a meta-analysis conducted by Devries et al. (2013), IPV was found to have an effect on depression for both men and women and an effect on suicide attempts for women. As these and other studies using differing samples and methodology (e.g., Coker et al., 2002; Kim & Lee, 2013) underscore, IPV may negatively affect individuals, in that it may increase individuals’ risk for developing a variety of mental disorders and maladaptive risk behaviors tied to mental health.
Research on gender differences in prevalence rates of IPV perpetration and IPV victimization traditionally has focused on female victims of violence perpetrated by abusive male partners, resulting in the perception that rates of perpetration are higher among men while rates of victimization are higher among women. However, more recent research suggests that women are equally likely to be the perpetrators of IPV and that men are equally likely to be the victims of IPV (Anderson, 2002). Other research suggests that men may in fact be more likely to be the victims of IPV than women (Kaura & Lohman, 2007). While findings from research conducted with community samples indicate that both men and women consistently engage in partner violence (O’Leary, Slep, & O’Leary, 2007), findings from research conducted with clinical samples (e.g., women recruited from battered women’s shelters, men recruited from aggression management programs) indicate that men may be more likely to perpetrate violence and women may be more likely to be victimized (Holtzworth-Munroe, 2010).
Although IPV appears to be largely gender-symmetric, the negative consequences that follow from IPV generally are more severe for women than for men (e.g., Archer, 2000; Vivian & Malone, 1997). Coker et al. (2002) used data from the National Violence Against Women Survey (NVAWS) to examine the relationship between physical and psychological IPV and physical and mental health for both male and female victims. Analyses indicated that both men and women experienced negative outcomes. However, women were more likely than men to report poor physical and mental health.
Existing research indicates that the majority of couples affected by IPV experience bidirectional violence, in which both partners are perpetrators and victims of IPV (Cascardi & Vivian, 1995). Nevertheless, most previous studies have focused on the negative mental health consequences among victims of IPV only. There are relatively few studies examining the correlates of perpetration and of bidirectional IPV. In one notable exception, Melander, Noel, and Tyler (2010) examined demographic predictors of bidirectionally-violent, unidirectionally-violent, and non-violent partnered young adults and concluded that depressive symptoms, alcohol use, and lower partner education were found to be predictive of bidirectional violence. Stith, Smith, Penn, Ward, and Tritt (2004) conducted a meta-analysis of 85 articles examining risk factors of both IPV perpetration and IPV victimization. They identified several different mental health and related factors that were associated with IPV. These included strong associations with emotional abuse, forced sex, illicit drug use, attitudes condoning marital violence, using violence toward her partner, and marital satisfaction, as well as moderately sized associations with traditional sex-role ideology, anger/hostility, history of partner abuse, alcohol use, depression, and career/life stress. However, the differential impact of bidirectional violence as compared to unidirectional violence was not examined in this meta-analysis.
Studies examining the impact of perpetration, victimization, and bidirectional violence for men and for women are even scarcer and findings from these few existing studies remain largely inconclusive. For example, in a study using a sample of 14,063 adults from the GENACIS (Gender Alcohol and Culture: An International Study) Canada survey, it was found that for men, prevalence of depression was highest among those reporting perpetration only, followed by those reporting bidirectional violence, and lowest among those reporting victimization only. For women, prevalence of depression was highest among those reporting bidirectional violence or victimization only and was lowest among those reporting perpetration only (Graham et al., 2012). In a similar study using a sample of 7,295 couples from Wave 1 of the National Survey of Families and Households (NSFH-1), perpetration, victimization, and bidirectional violence were found to be positively associated with depression for both men and women. In addition, the link between bidirectional violence and depression was found to be stronger for women (Anderson, 2002). Finally, results from a study using a sample of 1,136 couples, who participated in the second round of data collection of a five-year national longitudinal study, showed that for men, the likelihood of experiencing depression was the same for perpetrators and victims of violence, while for women, the likelihood of experiencing depression was greater for perpetrators of violence than for victims (Vaeth, Ramisetty-Mikler, & Caetano, 2010). Clearly, additional research is needed to describe the impact of gender and perpetrator-victim role on different mental health outcomes and risk behaviors associated with IPV.
The Present Study
The goal of the present study was to examine differences based on gender and perpetrator-victim role of a variety of factors related to mental health. Specifically, we examined depression, suicidality, alcohol use, illegal drug use, and relationship satisfaction among a nationally representative community sample of men and women experiencing no violence, unidirectional violence (perpetration-only or victimization-only), and bidirectional violence. As outlined above, research on differences in mental health outcomes among male and female perpetrators and victims of IPV as well as individuals experiencing bidirectional violence is limited. Thus, additional research is clearly warranted. To our knowledge, this study is among the first of its kind to describe differences in such a comprehensive set of mental health outcomes based on gender (men versus women) and perpetrator-victim role (no violence versus perpetration-only versus victimization-only versus bidirectional violence) in a single sample, thereby allowing for comparison of the resulting groups. Based on the few existing studies examining these differences, we predicted that individuals affected by bidirectional violence would generally suffer the most in regards to the negative outcomes associated with IPV and individuals affected by no violence would suffer the least. No specific hypotheses were proposed in regards to differences between perpetrators-only and victims-only as well as differences between men and women, because the few existing studies examining these differences have largely yielded inconclusive findings.
Method
Participants and Procedure
The present study used archival data from the National Longitudinal Study of Adolescent Health (Add Health) data set. In the original Add Health study, individuals between the ages of 11 and 34 completed in-home interviews at four separate time points (see Udry & Bearman, 2002 for study design). In the present study, analyses were conducted using a sub-sample of 7187 individuals who were participants in Waves 3 (collected in 2001–2002) and 4 (collected in 2008) of Add Health and who were involved in one current relationship at Wave 3. This final sample for analysis included 2876 men and 4311 women, ranging in age from 18 to 28 years (M = 22.38, SD = 1.80) at Wave 3 and from 25 to 34 years (M = 29.12, SD = 1.74) at Wave 4. The majority of participants were White (74.0%).
Materials
All variables, with the exception of relationship satisfaction, were assessed at both Wave 3 and Wave 4. Relationship satisfaction was not measured at Wave 3 of the original Add Health study.
Intimate partner violence.
A modified version of the Conflicts Tactics Scale (CTS; Straus, 1979) asked participants if they had experienced psychological and/or physical abuse over the past 12 months. The Add Health data set contains a total of eight items assessing intimate partner violence (four items assessing perpetration and four items assessing victimization). However, due to the low inter-item correlations of the two sexual violence items with the other IPV items (rs = .19 to .20 for perpetration and rs = .21 to .29 for victimization at Wave 3; rs = .08 to .17 for perpetration and rs = .05 to .17 for victimization at Wave 4), these two items were excluded from the analysis. Thus, there was a remaining total of three items assessing perpetration and three items assessing victimization. Examples include, “How often (have/did) you (slapped/slap), hit, or (kicked/kick) {initials}?” and, “How often (has/did) {initials} (slapped/slap), hit or (kicked/kick) you?” Items were rated on a 7-point scale from 0 (never) to 6 (more than 20 times). Perpetration and victimization scores on the individual three items were summed to yield overall summed scores with higher scores indicating higher levels of IPV (α = .73 for perpetration and α = .80 for victimization at Wave 3; α = .74 for perpetration and α = .79 for victimization at Wave 4). In the present sample, at Wave 3, summed scores ranged from 0 to 15 for perpetration (M = 0.32, SD = 1.17) and from 0 to 18 for victimization (M = 0.51, SD = 1.54). At Wave 4, summed scores ranged from 0 to 15 for perpetration (M = 0.32, SD = 1.17) and from 0 to 15 for victimization (M = 0.51, SD = 1.54). Using these summed perpetration and victimization scores participants were divided into four categories based on their perpetrator-victim role. Participants who reported zero acts of perpetration and zero acts of victimization were coded into the non-violent group. Participants who reported one or more acts of perpetration and zero acts of victimization were coded into the perpetration-only group. Participants who reported zero acts of perpetration and one or more acts of victimization were coded into the victimization-only group. Finally, participants who reported one or more acts of perpetration and one or more acts of victimization were coded into the bidirectionally-violent group. In the present sample, at Wave 3, there were 4502 (62.6%) participants in the non-violent group, 484 (6.7%) participants in the perpetration-only group, 296 (4.1%) participants in the victimization-only group, and 884 (12.3%) in the bidirectionally-violent group. At Wave 4, there were 5237 (72.9%) participants in the non-violent group, 252 (3.5%) participants in the perpetration-only group, 542 (7.5%) participants in the victimization-only group, and 475 (6.6%) in the bidirectionally-violent group. Thus, the percentage of individuals in the non-violent group overall slightly increased from Wave 3 to Wave 4; the percentage of individuals in the perpetration-only group decreased; the percentage of people in the victimization-only group increased; and the percentage of individuals in the bidirectionally-violent group decreased. The correlation between IPV role at Wave 3 and IPV role at Wave 4 was r = .18 (p < .001).
Depression.
A modified version of the Center of Epidemiological Studies Depression Scale (CES-D; Radloff, 1977), which contained a total of nine items at Wave 3 and ten items at Wave 4, asked participants whether they had experienced depressive symptoms in the past seven days (the item, “You felt happy.” was not asked of participants at Wave 3). Examples include, “(In the past seven days) You were bothered by things that usually don’t bother you.” and, “(In the past seven days) You felt depressed.” Items were rated on a 4-point scale from 0 (never or rarely) to 3 (most of the time or all of the time). After reverse-coding three of the items, scores of the individual ten items were averaged to yield a mean depression score, with higher scores indicating higher levels of depression (α = .80 at Wave 3; α = .84 at Wave 4). In the present sample, at Wave 3, mean depression scores ranged from 0 to 2.78 (M = 0.49, SD = 0.49) and at Wave 4, mean depression scores ranged from 0 to 3 (M = 0.60, SD = 0.46).
Suicidality.
Three individual items asked participants about suicidality. The first item asked, “During the past 12 months, have you ever seriously thought about committing suicide?” For this item, participants responded either no (0) or yes (1). In the present sample, at Wave 3, 380 (5.3%) participants responded that they had thought about committing suicide in the past 12 months and at Wave 4, 452 (6.3%) participants responded that they had thought about committing suicide in the past 12 months. The second item asked, “During the past 12 months, how many times have you actually attempted suicide?” This item was dichotomized so that answers reflected whether participants had attempted suicide zero times (0) or one or more times (1) in the past 12 months. In the present sample, at Wave 3, 100 participants (1.4%) responded that they has attempted suicide one or more times and at Wave 4, 103 (1.4%) participants responded that they had attempted suicide one or more times. The third item asked, “Did any attempt result in an injury, poisoning, or overdose that had to be treated by a doctor or nurse?” For this item, participants responded either no (0) or yes (1). In the present sample, at Wave 3, 34 (0.5%) participants indicated that a suicide attempt had resulted in an injury, poisoning, or overdose that had to be treated by a doctor or nurse and at Wave 4, 30 (0.4%) participants indicated that a suicide attempt had resulted in an injury, poisoning, or overdose that had to be treated by a doctor or nurse.
Alcohol use.
Two individual items asked participants about alcohol use during the past 12 months. The first item asked, “During the past 12 months, on how many days did you drink five or more drinks in a row?” (at Wave 4, it was specified, “5 or more drinks (for men) or 4 or more drinks (for women)”). The second item asked, “During the past 12 months, on how many days have you been drunk or very high on alcohol?” Both items were rated on 7-point scales ranging from 0 (none) to 6 (every day or almost every day). In the present sample, at Wave 3, scores ranged from 0 to 6 for both items (M = 1.50, SD = 1.55 for the first item and M = 1.43, SD = 1.39 for the second item) and at Wave 4, scores ranged from 0 to 6 for both items (M = 1.51, SD = 1.54 for the first item and M = 1.27, SD = 1.31 for the second item).
Illegal drug use.
Individual, dichotomous items were used to assess participants’ use of four different drugs – steroids, marijuana, cocaine, and crystal meth. At Wave 3, participants were asked, “In the past year, have you used (drug)?” At Wave 4, participants were asked, “Have you ever used any of the following drugs?” Participants responded either no (0) or yes (1) to these items. In the present sample, at Wave 3, 72 (1.0%) participants reported that they had used steroids, 2147 (29.9%) participants reported that they had used marijuana, 394 (5.5%) participants reported that they had used cocaine, and 180 (2.5%) reported that they had used crystal meth in the past year and at Wave 4, 132 (1.8%) participants reported that they had ever used steroids, 3937 (54.8%) participants reported that they had ever used marijuana, 1295 (18.0%) participants reported that they had ever used cocaine, and 642 (8.9%) reported that they had ever used crystal meth.
Relationship satisfaction.
Seven items were used to assess participants’ satisfaction with their relationship. These items were only asked of participants at Wave 4 of the Add Health study. Examples include, “We enjoy doing even ordinary, day-to-day things together.” and, “I am satisfied with the way we handle our problems and disagreements.” Items were rated on a 5-point scale from −2 (strongly disagree) to +2 (strongly agree). Relationship satisfaction scores of the individual seven items were averaged to yield a mean satisfaction score (α = .89). In the present sample, mean satisfaction scores ranged from −2 to +2 (M = 1.09, SD = 0.82).
Results
Since Levene’s test indicated that variances in the groups were heterogeneous, one-way analyses of variance (ANOVAs) using robust F-tests (F* Brown-Forsythe) were performed to assess means differences in outcome variables among the four perpetrator-victim role groups. In an attempt to control for Type 1 error in the multiple analyses conducted, we utilized a Bonferroni adjustment for the family-wise alpha level. Since the family of comparisons was made up of 3 analyses (overall sample, men, and women) at each wave, the Bonferroni-adjusted alpha level was set at .017.
Depression by Perpetrator-Victim Role
Symptoms of depression among non-violent, perpetrator-only, victim-only, and bidirectionally-violent individuals at Waves 3 and 4 are depicted in Table 1.
Table 1.
Depression (Modified CES-D) among Men and Women Experiencing No Violence, Unidirectional Violence, and Bidirectional Violence at Waves 3 and 4.
| Depression | ||||||||
|---|---|---|---|---|---|---|---|---|
|
| ||||||||
| Wave 3 | Wave 4 | |||||||
|
| ||||||||
| Role in IPV | n | M (SD) | 95% CI | F* | n | M (SD) | 95% CI | F* |
|
| ||||||||
| Overall | 82.43* | 51.59* | ||||||
| Non-violent | 4502 | 0.44 (0.41) | 0.43-0.45 | 5237 | 0.55 (0.44) | 0.54-0.56 | ||
| Perpetrator-only | 484 | 0.57 (0.45) | 0.53-0.61 | 252 | 0.76 (0.55) | 0.70-0.83 | ||
| Victim-only | 296 | 0.53 (0.43) | 0.48-0.57 | 542 | 0.66 (0.46) | 0.62-0.69 | ||
| Bidirectional | 884 | 0.69 (0.51) | 0.66-0.73 | 475 | 0.81 (0.56) | 0.76-0.86 | ||
| Men | 25.12* | 23.01* | ||||||
| Non-violent | 1883 | 0.38 (0.37) | 0.37-0.40 | 2053 | 0.50 (0.41) | 0.48-0.51 | ||
| Perpetrator-only | 69 | 0.48 (0.47) | 0.36-0.59 | 29 | 0.82 (0.41) | 0.67-0.98 | ||
| Victim-only | 204 | 0.46 (0.38) | 0.40-0.51 | 353 | 0.61 (0.43) | 0.57-0.66 | ||
| Bidirectional | 286 | 0.62 (0.46) | 0.57-0.67 | 171 | 0.72 (0.50) | 0.65-0.80 | ||
| Women | 49.17* | 32.07* | ||||||
| Non-violent | 2619 | 0.48 (0.44) | 0.46-0.49 | 3184 | 0.58 (0.45) | 0.57-0.60 | ||
| Perpetrator-only | 415 | 0.58 (0.44) | 0.54-0.62 | 223 | 0.76 (0.56) | 0.68-0.83 | ||
| Victim-only | 92 | 0.68 (0.48) | 0.58-0.78 | 189 | 0.74 (0.49) | 0.67-0.81 | ||
| Bidirectional | 598 | 0.73 (0.54) | 0.68-0.77 | 304 | 0.86 (0.58) | 0.79-0.92 | ||
p < .017;
Robust F* Brown-Forsythe is reported.
Severity of depressive symptoms significantly differed by perpetrator-victim role for the overall sample at Wave 3 (F* (3, 1711.84) = 82.43, p < .001) and Wave 4 (F* (3, 1130.42) = 51.59, p < .001), as well as for men at Wave 3 (F* (3, 356.28) = 25.12, p < .001) and Wave 4 (F* (3, 256.69) = 23.01, p < .001), and for women at Wave 3 (F* (3, 660.65) = 49.17, p < .001) and Wave 4 (F* (3, 768.81) = 32.07, p < .001). Consistently, depression was found to be highest among individuals in the bidirectionally-violent group and lowest among individuals in the non-violent group (for all group means, standard deviations, and 95% confidence intervals see Table 1).
Suicidality by Perpetrator-Victim Role
Prevalence of suicidality (thoughts about suicide, one or more suicide attempts, and attempts that resulted in an injury, poisoning, or overdose treated by a doctor or nurse) for non-violent, perpetrator-only, victim-only, and bidirectionally-violent individuals at Waves 3 and 4 are depicted in Tables 2 and 3. Chi-square analyses were performed to assess differences in prevalence of suicidality by individuals’ perpetrator-victim role.
Table 2.
Suicidality among Men and Women Experiencing No Violence, Unidirectional Violence, and Bidirectional Violence at Wave 3.
| Role in IPV | Thoughts about Suicide | Suicide Attempts | Treatment of Injury | ||||||
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| %Yes (n) | Total | χ2 | %Yes (n) | Total | χ2 | %Yes (n) | Total | χ2 | |
|
| |||||||||
| Overall | 93.11* | 1.46 | 1.24 | ||||||
| Non-violent | 3.9 (173) | 4424 | 25.0 (43) | 172 | 32.6 (14) | 43 | |||
| Perpetrator-only | 4.2 (20) | 479 | 35.0 (7) | 20 | 14.3 (1) | 7 | |||
| Victim-only | 5.5 (16) | 290 | 31.3 (5) | 16 | 40.0 (2) | 5 | |||
| Bidirectional | 11.8 (102) | 866 | 29.7 (30) | 101 | 34.5 (10) | 29 | |||
| Men | 54.03* | 2.11 | 0.47 | ||||||
| Non-violent | 3.1 (57) | 1847 | 17.9 (10) | 56 | 60.0 (6) | 10 | |||
| Perpetrator-only | 1.5 (1) | 68 | 0.0 (0) | 1 | --- | --- | |||
| Victim-only | 6.0 (12) | 201 | 33.3 (4) | 12 | 50.0 (2) | 4 | |||
| Bidirectional | 12.5 (35) | 280 | 26.5 (9) | 34 | 44.4 (4) | 9 | |||
| Women | 44.37* | 0.66 | 1.05 | ||||||
| Non-violent | 4.5 (116) | 2577 | 28.4 (33) | 116 | 24.2 (8) | 33 | |||
| Perpetrator-only | 4.6 (19) | 411 | 36.8 (7) | 19 | 14.3 (1) | 7 | |||
| Victim-only | 4.5 (4) | 89 | 25.0 (1) | 4 | 0.0 (0) | 1 | |||
| Bidirectional | 11.4 (67) | 586 | 31.3 (21) | 67 | 30.0 (6) | 20 | |||
p < .017.
Table 3.
Suicidality among Men and Women Experiencing No Violence, Unidirectional Violence, and Bidirectional Violence at Wave 4.
| Role in IPV | Thoughts about Suicide | Suicide Attempts | Treatment of Injury | ||||||
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| %Yes (n) | Total | χ2 | %Yes (n) | Total | χ2 | %Yes (n) | Total | χ2 | |
|
| |||||||||
| Overall | 102.23* | 55.89* | 0.48 | ||||||
| Non-violent | 4.8 (253) | 5225 | 1.0 (50) | 5226 | 28.0 (14) | 50 | |||
| Perpetrator-only | 11.6 (29) | 251 | 4.8 (12) | 251 | 25.0 (3) | 12 | |||
| Victim-only | 8.0 (43) | 537 | 2.0 (11) | 537 | 27.3 (3) | 11 | |||
| Bidirectional | 15.4 (73) | 473 | 4.2 (20) | 473 | 35.0 (7) | 20 | |||
| Men | 37.48* | 17.42* | 3.59 | ||||||
| Non-violent | 4.4 (90) | 2042 | 0.9 (19) | 2043 | 36.8 (7) | 19 | |||
| Perpetrator-only | 14.3 (4) | 28 | 7.1 (2) | 28 | 0.0 (0) | 2 | |||
| Victim-only | 6.6 (23) | 348 | 1.1 (4) | 348 | 25.0 (1) | 4 | |||
| Bidirectional | 14.7 (25) | 170 | 3.5 (6) | 170 | 66.7 (4) | 6 | |||
| Women | 66.12* | 42.83* | 0.36 | ||||||
| Non-violent | 5.1 (163) | 3183 | 1.0 (31) | 3183 | 22.6 (7) | 31 | |||
| Perpetrator-only | 11.2 (25) | 223 | 4.5 (10) | 223 | 30.0 (3) | 10 | |||
| Victim-only | 10.6 (20) | 189 | 3.7 (7) | 189 | 28.6 (2) | 7 | |||
| Bidirectional | 15.8 (48) | 303 | 4.6 (14) | 303 | 21.4 (3) | 14 | |||
p < .017.
Prevalence of thoughts about suicide significantly differed by perpetrator-victim role, in the overall sample at Wave 3 (χ2 (3) = 93.11, p < .001) and Wave 4 (χ2 (3) = 102.23, p < .001), as well as for men at Wave 3 (χ2 (3) = 54.03, p < .001) and Wave 4 (χ2 (3) = 37.48, p < .001), and for women at Wave 3 (χ2 (3) = 44.37, p < .001) and Wave 4 (χ2 (3) = 66.12, p < .001). Prevalence of thoughts about suicide was consistently highest in the bidirectionally-violent group at both Wave 3 and Wave 4. At Wave 4, prevalence of thoughts about suicide was consistently lowest in the non-violent group. Individuals in the perpetrator-only group consistently had higher prevalence of thoughts about suicide than individuals in the victim-only group (for all group percentages and sample sizes see Tables 2 and 3).
Prevalence of one or more suicide attempts significantly differed by perpetrator-victim role at for the overall sample at Wave 4 (χ2 (3) = 55.89, p < .001), as well as for men at Wave 4 (χ2 (3) = 17.42, p = .001), and for women at Wave 4 (χ2 (3) = 42.83, p < .001). There were no other significant results. In the overall sample as well as for men, prevalence of one or more suicide attempts was highest among individuals in the perpetrator-only group and lowest in the non-violent group. Prevalence of one or more suicide attempts was second highest in the bidirectionally-violent group. For women, prevalence of one or more suicide attempts was highest in the bidirectionally-violent group and lowest in the non-violent group. Prevalence of one or more suicide attempts was second highest in the perpetrator-only group (for all group percentages and sample sizes see Tables 2 and 3).
Prevalence of suicide attempts that resulted in an injury, poisoning, or overdose did not significantly differ by perpetrator-victim role.
Alcohol Use by Perpetrator-Victim Role
Alcohol use (number of times participants had consumed five or more drinks (five or four or more drinks at Wave 4; see methods) in the past 12 months and number of times participants had been drunk or very high on alcohol in the past 12 months) for non-violent, perpetrator-only, victim-only, and bidirectionally-violent individuals at Waves 3 and 4 is depicted in Table 4.
Table 4.
Alcohol Use (Last 12 Months) among Men and Women Experiencing No Violence, Unidirectional Violence, and Bidirectional Violence at Waves 3 and 4.
| 5/4 or More Drinks | Drunk or Very High on Alcohol | 5/4 or More Drinks | Drunk or Very High on Alcohol | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||||||||
| Wave 3 | Wave 4 | |||||||||||||||
|
| ||||||||||||||||
| Role in IPV | n | M (SD) | 95% CI | F* | n | M (SD) | 95% CI | F* | n | M (SD) | 95% CI | F* | n | M (SD) | 95% CI | F* |
|
| ||||||||||||||||
| Overall | 2.37 | 0.83 | 3.38 | 7.20* | ||||||||||||
| Non-violent | 3275 | 1.46 (1.53) | 1.41-1.51 | 3275 | 1.37 (1.37) | 1.32-1.41 | 3826 | 1.46 (1.53) | 1.41-1.51 | 3829 | 1.22 (1.29) | 1.18-1.26 | ||||
| Perpetrator-only | 353 | 1.39 (1.51) | 1.23-1.55 | 354 | 1.33 (1.34) | 1.19-1.47 | 198 | 1.54 (1.52) | 1.32-1.75 | 199 | 1.30 (1.25) | 1.13-1.48 | ||||
| Victim-only | 215 | 1.73 (1.63) | 1.51-1.95 | 214 | 1.46 (1.38) | 1.28-1.65 | 402 | 1.61 (1.52) | 1.46-1.76 | 402 | 1.43 (1.32) | 1.30-1.56 | ||||
| Bidirectional | 651 | 1.49 (1.57) | 1.37-1.61 | 645 | 1.43 (1.44) | 1.32-1.54 | 367 | 1.69 (1.62) | 1.53-1.86 | 368 | 1.50 (1.46) | 1.35-1.65 | ||||
| Men | 2.28 | 3.06 | 2.65 | 3.21 | ||||||||||||
| Non-violent | 1432 | 1.94 (1.65) | 1.86-2.03 | 1429 | 1.69 (1.51) | 1.61-1.76 | 1600 | 1.85 (1.64) | 1.77-1.93 | 1602 | 1.57 (1.42) | 1.50-1.64 | ||||
| Perpetrator-only | 51 | 2.16 (1.75) | 1.67-2.65 | 51 | 1.84 (1.62) | 1.39-2.30 | 23 | 1.96 (1.89) | 1.14-2.78 | 23 | 1.87 (1.22) | 1.34-2.40 | ||||
| Victim-only | 157 | 2.10 (1.66) | 1.84-2.36 | 156 | 1.73 (1.42) | 1.51-1.96 | 276 | 1.72 (1.47) | 1.55-1.90 | 276 | 1.56 (1.30) | 1.40-1.71 | ||||
| Bidirectional | 227 | 2.23 (1.67) | 2.01-2.45 | 223 | 2.02 (1.59) | 1.81-2.23 | 138 | 2.22 (1.71) | 1.93-2.51 | 139 | 1.91 (1.49) | 1.66-2.16 | ||||
| Women | 3.34 | 3.20 | 3.37 | 5.23* | ||||||||||||
| Non-violent | 1843 | 1.08 (1.32) | 1.02-1.14 | 1846 | 1.12 (1.19) | 1.06-1.17 | 2226 | 1.18 (1.38) | 1.12-1.24 | 2227 | 0.97 (1.14) | 0.93-1.02 | ||||
| Perpetrator-only | 302 | 1.26 (1.43) | 1.10-1.42 | 303 | 1.24 (1.28) | 1.10-1.39 | 175 | 1.48 (1.47) | 1.26-1.70 | 176 | 1.23 (1.23) | 1.04-1.41 | ||||
| Victim-only | 58 | 0.72 (1.01) | 0.46-0.99 | 58 | 0.74 (0.91) | 0.50-0.98 | 126 | 1.36 (1.59) | 1.08-1.64 | 126 | 1.26 (1.31) | 0.93-1.39 | ||||
| Bidirectional | 424 | 1.09 (1.36) | 0.96-1.22 | 422 | 1.12 (1.24) | 1.00-1.24 | 229 | 1.38 (1.49) | 1.18-1.57 | 229 | 1.25 (1.38) | 1.07-1.43 | ||||
p < .017;
Robust F* Brown-Forsythe is reported.
Number of five (or four) or more drinks did not significantly differ by perpetrator-victim role for any of the groups examined either at Wave 3 or Wave 4 (for all group means, standard deviations, and 95% confidence intervals see Table 4).
Number of times being drunk or high on alcohol significantly differed by perpetrator-victim role for the overall sample at Wave 4 (F* (3, 1032.71) = 7.20, p < .001) and for women at Wave 4 (F* (3, 550.47) = 5.23, p = .001). There were no other significant results (for all group means, standard deviations, and 95% confidence intervals see Table 4).
Drug Use by Perpetrator-Victim Role
Prevalence of drug use (steroids, marijuana, cocaine, and crystal meth) for non-violent, perpetrator-only, victim-only, and bidirectionally-violent individuals at Wave 3 (in the past year) and Wave 4 (ever) is depicted in Tables 5 and 6. Chi-square analyses were performed to assess differences in drug use by individuals’ perpetrator-victim role.
Table 5.
Illegal Drug Use (Last 12 Months) among Men and Women Experiencing No Violence, Unidirectional Violence, and Bidirectional Violence at Wave 3.
| Role in IPV | Steroids | Marijuana | Cocaine | Crystal Meth | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||||
| %Yes (n) | Total | χ2 | %Yes (n) | Total | χ2 | %Yes (n) | Total | χ2 | %Yes (n) | Total | χ2 | |
|
| ||||||||||||
| Overall | 10.42* | 6.36 | 3.73 | 2.58 | ||||||||
| Non-violent | 0.9 (40) | 4474 | 65.3 (1201) | 1838 | 58.3 (208) | 357 | 41.4 (82) | 198 | ||||
| Perpetrator-only | 0.6 (3) | 479 | 66.7 (144) | 216 | 63.6 (28) | 44 | 40.9 (9) | 22 | ||||
| Victim-only | 2.4 (7) | 295 | 65.8 (100) | 152 | 46.3 (19) | 41 | 56.7 (17) | 30 | ||||
| Bidirectional | 1.7 (15) | 876 | 71.5 (338) | 473 | 53.0 (70) | 132 | 44.7 (38) | 85 | ||||
| Men | 10.31* | 3.55 | 13.75* | 3.30 | ||||||||
| Non-violent | 1.6 (30) | 1869 | 67.9 (602) | 886 | 67.0 (126) | 188 | 45.5 (51) | 112 | ||||
| Perpetrator-only | 0.0 (0) | 67 | 64.9 (24) | 37 | 33.3 (2) | 6 | 16.7 (1) | 6 | ||||
| Victim-only | 3.4 (7) | 203 | 63.5 (73) | 115 | 37.1 (13) | 35 | 54.5 (12) | 22 | ||||
| Bidirectional | 3.9 (11) | 282 | 73.2 (139) | 190 | 55.7 (34) | 61 | 52.3 (23) | 44 | ||||
| Women | 2.00 | 6.69 | 10.45* | 3.08 | ||||||||
| Non-violent | 0.4 (10) | 2605 | 62.9 (599) | 952 | 48.5 (82) | 169 | 36.0 (31) | 86 | ||||
| Perpetrator-only | 0.7 (3) | 412 | 67.0 (120) | 179 | 68.4 (26) | 38 | 50.0 (8) | 16 | ||||
| Victim-only | 0.0 (0) | 92 | 73.0 (27) | 37 | 100.0 (6) | 6 | 62.5 (5) | 8 | ||||
| Bidirectional | 0.7 (4) | 594 | 70.3 (199) | 283 | 50.7 (36) | 71 | 36.6 (15) | 41 | ||||
p < .017.
Table 6.
Illegal Drug Use (Ever) among Men and Women Experiencing No Violence, Unidirectional Violence, and Bidirectional Violence at Wave 4.
| Role in IPV | Steroids | Marijuana | Cocaine | Crystal Meth | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||||
| %Yes (n) | Total | χ2 | %Yes (n) | Total | χ2 | %Yes (n) | Total | χ2 | %Yes (n) | Total | χ2 | |
|
| ||||||||||||
| Overall | 4.44 | 48.50* | 52.47* | 44.57* | ||||||||
| Non-violent | 1.7 (89) | 5233 | 52.0 (2720) | 5227 | 16.0 (838) | 5229 | 7.6 (400) | 5232 | ||||
| Perpetrator-only | 1.6 (4) | 252 | 64.7 (163) | 252 | 22.6 (57) | 252 | 9.1 (23) | 252 | ||||
| Victim-only | 3.0 (16) | 541 | 59.6 (321) | 539 | 25.9 (140) | 540 | 12.6 (68) | 540 | ||||
| Bidirectional | 1.9 (9) | 474 | 64.8 (307) | 474 | 24.1 (114) | 474 | 15.4 (73) | 474 | ||||
| Men | 2.57 | 15.33* | 26.33* | 8.60 | ||||||||
| Non-violent | 3.6 (74) | 2051 | 59.2 (1212) | 2048 | 20.9 (428) | 2048 | 10.5 (215) | 2050 | ||||
| Perpetrator-only | 0.0 (0) | 29 | 69.0 (20) | 29 | 20.7 (6) | 29 | 10.3 (3) | 29 | ||||
| Victim-only | 4.3 (15) | 352 | 64.9 (227) | 350 | 30.8 (108) | 351 | 12.5 (44) | 351 | ||||
| Bidirectional | 5.3 (9) | 171 | 72.5 (124) | 171 | 32.7 (56) | 171 | 17.5 (30) | 171 | ||||
| Women | 8.83 | 38.89* | 25.55* | 41.81* | ||||||||
| Non-violent | 0.5 (15) | 3182 | 47.4 (1508) | 3179 | 12.9 (410) | 3181 | 5.8 (185) | 3182 | ||||
| Perpetrator-only | 1.8 (4) | 223 | 64.1 (143) | 223 | 22.9 (51) | 223 | 9.0 (20) | 223 | ||||
| Victim-only | 0.5 (1) | 189 | 49.7 (94) | 189 | 16.9 (32) | 189 | 12.7 (24) | 189 | ||||
| Bidirectional | 0.0 (0) | 303 | 60.4 (183) | 303 | 19.1 (58) | 303 | 14.2 (43) | 303 | ||||
p < .017.
Prevalence of steroid use significantly differed by perpetrator-victim role in the overall sample at Wave 3 (χ2 (3) = 10.42, p = .015) as well as for men at Wave 3 (χ2 (3) = 10.31, p = .016). There were no other significant results (for all group percentages and sample sizes see Tables 5 and 6).
Prevalence of marijuana use significantly differed by perpetrator-victim role in the overall sample at Wave 4 (χ2 (3) = 48.50, p < .001), as well as for men at Wave 4 (χ2 (3) = 15.33, p = .002), and for women at Wave 4 (χ2 (3) = 38.89, p < .001). There were no other significant results. With one exception (highest marijuana use prevalence among women at Wave 4), prevalence of marijuana use was consistently highest among individuals in the bidirectionally-violent group and lowest among individuals in the non-violent group. In addition, prevalence of marijuana use was consistently found to be higher among individuals in the perpetrator-only group than among individuals in the victim-only group (for all group percentages and sample see Tables 5 and 6).
Prevalence of cocaine use significantly differed by perpetrator-victim role in the overall sample at Wave 4 (χ2 (3) = 52.47, p < .001), as well as for men at Wave 3 (χ2 (3) = 13.75, p = .003) and Wave 4 (χ2 (3) = 26.33, p < .001), and for women at Wave 3 (χ2 (3) = 10.45, p = .015) and Wave 4 (χ2 (3) = 25.55, p < .001). There were no other significant results (for all group percentages and sample see Tables 5 and 6).
Prevalence of crystal meth use significantly differed by perpetrator-victim role in the overall sample at Wave 4 (χ2 (3) = 44.57, p < .001), as well as for women at Wave 4 (χ2 (3) = 41.81, p < .001). There were no other significant results. Prevalence of crystal meth use was generally highest in the bidirectionally-violent group and lowest in the non-violent group. In addition, prevalence of crystal meth use was generally higher among individuals in the victim-only group than among individuals in the perpetrator-only group (for all group percentages and sample see Tables 5 and 6).
Relationship Satisfaction by Perpetrator-victim role
Relationship satisfaction for non-violent, perpetrator-only, victim-only, and bidirectionally-violent individuals at Wave 4 only (this construct was not assessed at Wave 3 of the Add Health study) is depicted in Table 7.
Table 7.
Relationship Satisfaction among Men and Women Experiencing No Violence, Unidirectional Violence, and Bidirectional Violence at Wave 4.
| Relationship Satisfaction | ||||
|---|---|---|---|---|
|
| ||||
| Wave 4 | ||||
|
| ||||
| Role in IPV | n | M (SD) | 95% CI | F* |
|
| ||||
| Overall | 160.57* | |||
| Non-violent | 5232 | 1.24 (0.71) | 1.23–1.26 | |
| Perpetrator-only | 252 | 0.91 (0.79) | 0.81–1.01 | |
| Victim-only | 542 | 0.73 (0.95) | 0.65–0.81 | |
| Bidirectional | 475 | 0.47 (0.93) | 0.39–0.55 | |
| Men | 37.94* | |||
| Non-violent | 2050 | 1.22 (0.71) | 1.19–1.25 | |
| Perpetrator-only | 29 | 1.02 (0.68) | 0.77–1.28 | |
| Victim-only | 353 | 0.89 (0.82) | 0.80–0.97 | |
| Bidirectional | 171 | 0.70 (0.86) | 0.57–0.83 | |
| Women | 129.12* | |||
| Non-violent | 3182 | 1.26 (0.72) | 1.24–1.29 | |
| Perpetrator-only | 223 | 0.89 (0.81) | 0.79–1.00 | |
| Victim-only | 189 | 0.44 (1.10) | 0.28–0.60 | |
| Bidirectional | 304 | 0.34 (0.94) | 0.23–0.45 | |
p < .017.
Robust F* Brown-Forsythe is reported.
Relationship satisfaction significantly differed by perpetrator-victim role, in the overall sample F* (3, 1335.18) = 160.57, p < .001), as well as for men (F* (3, 301.83) = 37.94, p < .001), and for women (F* (3, 655.45) = 129.12, p < .001). Relationship satisfaction was consistently highest in the non-violent group, followed by the perpetrator-only group, then the victim-only group, and lowest in the bidirectionally-violent group (for all group means, standard deviations, and 95% confidence intervals see Table 7).
Discussion
This study provided information on the prevalence and severity of a variety of mental health outcomes and risk behaviors within a nationally representative community sample of men and women who experienced no IPV, IPV perpetration only, IPV victimization only, or bidirectional IPV. Information was obtained about symptoms of depression, suicidality, alcohol use, illegal drug use, and relationship satisfaction at two different time points.
Across time points, findings consistently showed that individuals experiencing bidirectional violence were generally at highest risk for negative mental health outcomes and individuals experiencing no violence were generally at lowest risk. Individuals experiencing the two forms of unidirectional violence – perpetration-only and victimization-only – often were at similar risk for experiencing negative mental health outcomes. These findings are concordant with previous research showing that negative mental health outcomes tend to be more likely and more severe when violence is bidirectional than when violence is unidirectional (e.g., Temple, Weston, & Marshall, 2005).
There were minor exceptions to the general pattern of results. For alcohol use, no significant group differences emerged at Wave 3. However, there were tendencies for this pattern at Wave 4. These findings indicate that there might be additional factors beyond gender and perpetrator-victim role that have a larger impact on whether an individual decides to use and abuse alcohol or not.
For illegal drug use, the pattern described above did not clearly emerge for most of the illegal drugs examined. At Wave 3, few clear group differences were detected while at Wave 4, group differences were detected for most sub-samples and drugs examined. These differences in findings might be due to the difference in questions asking about drug use at Waves 3 and 4 of the Add Health study. While at Wave 3, participants reported drug use within the past year, at Wave 4, participants reported whether they had ever used any of these drugs. For most drugs at Wave 4, findings indicated that individuals experiencing bidirectional violence or perpetration-only were at highest risk for abusing illegal substances. Individuals experiencing no violence consistently were found to be at lowest risk.
These patterns were generally similar for men and women. However, some interesting findings emerged that warrant mentioning: For depression, it should be noted that rates of depression were generally higher for women than for men (concordant with previous research; e.g., Coker et al., 2002). Interestingly, at Wave 4, men in the perpetration-only group showed higher levels of depression than men in the victimization-only group, while no similar differences were found for women. It may be that male victims do not experience as severe injuries as a result of IPV and thus do not feel as depressed, while male perpetrators, who observe the negative impact their violent behavior has on their partners, may feel depressed as a result.
For suicidal thoughts, a clear difference between the perpetration-only and the victimization-only groups emerged for men, with more victims than perpetrators thinking about suicide at Wave 3 and the reverse at Wave 4, while there was no similar difference for women. However, because the pattern of these findings was inconsistent across the two waves, interpretation may be limited.
For drug use, it is interesting to highlight that 100% of women in the victimization-only group indicated that they used cocaine in the past year at Wave 3, while less than half of them men in the victimization-only group indicated that they used cocaine. However, due to the very small number of participants in the groups (6 and 13, respectively), the conclusions that can be drawn are again limited. Furthermore, no similar pattern emerged at Wave 4, where participants indicated whether they had ever used these drugs.
Finally, for relationship satisfaction, men indicated they had higher levels of satisfaction than women in all perpetrator-victim groups, except for the non-violent group, where men’s and women’s mean levels of satisfaction were relatively close to one another. It can be speculated that in non-violent relationships, men and women tend to be similarly satisfied with their relationships, while in violent relationships, women suffer more from the experience of IPV in terms of their levels of satisfaction than do men (independent of their perpetrator-victim role). A possible reason behind these findings, which is concordant with previous research, may be that women generally experience more negative outcomes and more severe injuries as a result of IPV than do men (e.g., Archer, 2000).
The present findings reveal that perpetrator-victim roles may have important implications for the outcomes one may experience. Although patterns were not always consistent across all types of mental health outcomes and risk behaviors, in general, individuals affected by bidirectional violence tended to suffer the most. These results are particularly disturbing since in most cases, IPV may be bidirectional in that it may reflect an outgrowth of conflict between both partners in which each actively contributes to the escalation of violence (Cascardi & Vivian, 1995).
Strengths and Limitations
First and foremost, research on gender and perpetrator-victim role differences in mental health outcomes of IPV is limited and thus, the present study is one of only a few studies that examines these variables in this context. In addition, the present study examined a variety of variables allowing for comparison of various IPV-mental health associations in one sample. Second, the present study included a large and nationally representative sample, which allows for high power and the detection of existing effects as well as higher generalizability to the general population. Third, data were collected at two different time points, allowing for comparison of whether observed group differences would hold across time. Finally, robust F-tests were used to account for violations of the homogeneity of variances assumption.
Despite these strengths, several limitations temper the conclusions that can be drawn. First, the sample used for the present study was largely composed of individuals experiencing relatively low levels of IPV. In addition, participants were divided into the four violence groups based on whether they had perpetrated and/or been victimized by one or more acts of violence. Thus, although the sample used was large and representative, this sample was not a clinical sample and thus, clinical implications remain unclear. In addition, base rates for several of the mental health variables under examination (e.g., depression, suicidal thinking) were relatively low across all four IPV groups, including the bidirectionally-violent group. Nevertheless, at several significant group differences were detected. Second, the present study relied on self-report data, which might bias the present findings. Also, we were limited by the design and measures used in the original Add Health study. As such, some of this study’s outcome variables were based on single-item measures, which might restrict the present findings’ conclusions. In addition, one outcome – relationship satisfaction – was not assessed at Wave 3 of the Add Health study. Therefore, we only have findings describing gender and perpetrator-victim role differences in relationship satisfaction at Wave 4. Finally, although the present study made use of two waves of data, these data were not examined longitudinally (i.e. we did not use Wave 3 data as predictors and Wave 4 data as outcomes), because the time span that separated the two waves was too large (6–7 years) to draw causal inferences.
Replication of the present study in a clinical sample would elucidate whether the current findings generalize to individuals experiencing more frequent and severe forms of violence as well as more negative mental health symptomatology. In addition, researchers should attempt to distinguish between mild, moderate, and severe levels of violence when examining the three violence groups in which IPV is present (perpetration-only, victimization-only, and bidirectional violence). Using a controlled design, in which sample sizes are equal and of adequate size, might validate the present results. Finally, using mental health and mental heath-related measures comprised of several items and using measures other than self-report in order to assess mental health outcomes (observational measures of IPV are not advisable due to ethical concerns) might be advantageous.
Implications
The study results suggest that more work to understand why bidirectional violence is relatively detrimental is indeed warranted. It is possible that because, independently, perpetration and victimization each have negative mental health and risk behavior implications, when combined, there is an additive effect on these outcomes. Additionally, it is possible that bidirectional violence is the result of, or leads directly to, an escalation of the severity of IPV. It seems logically feasible that those who experience bidirectional IPV experience relatively worse outcomes because the violence they experience is more severe. In fact, some research indicates that, if both partners perpetrate IPV, this might lead to an escalation of violence and to more severe aggression (e.g., Cascardi & Vivian, 1995) and thus to an increased risk of negative mental health and related outcomes (Graham et al., 2012). Because the current study is cross-sectional in nature it is not possible to draw conclusions about causality, however, our results suggest that it is possible that there are common predictors that put individuals at risk for bidirectional IPV, and that those risk factors are also associated with more severe negative outcomes that mirror our results.
Given that bidirectional violence appears to be the group with the most significant issues, our findings clearly point to the importance of examining this group of individuals more closely in future research. Previous studies have repeatedly shown that the prevalence of bidirectional violence among couples affected by IPV is high (e.g., Cascardi & Vivian, 1995). However, only few studies have examined this group and compared it to the other perpetrator-victim groups in terms of the negative outcomes individuals experience (e.g., Anderson, 2002). It is important to note that this bidirectional form of violence as described in the current sample most frequently occurs in couples affected by mild to moderate violence, such as the participants in the present sample (often referred to as “common couple violence;” Johnson, 2004) while IPV in clinical samples tends to be more gender-asymmetric. However, since common couple violence still appears to have detrimental impacts, we call for researchers and practitioners to give more attention to this group.
.In terms of practical implications of this research, it is important to recognize the link between IPV and mental health outcomes in order to develop effective prevention and intervention strategies. It is important to recognize that the co-occurrence of IPV and mental health disorders is influenced by the interplay of gender and perpetrator-victim role. The present findings lend support for the conclusion that policy makers and prevention and intervention developers should predominantly target couples affected by bidirectional violence when designing large-scale prevention and intervention programs.
In terms of clinical practice, although clearly more research on bidirectional violence is warranted, the present study highlights the importance for clinicians to assess what role an individual affected by IPV plays in the violent act in order to adjust treatment accordingly. Clinicians working with violent couples may want to be aware of the exact dynamics of violence in their clients’ relationships because it may have implications for how to treat them. It is important to keep in mind that base rates, not only of IPV but also of several of the mental health variables examined in the present study, were relatively low and, thus, further interpretation of the current findings should await replication of this study in a clinical sample.
Conclusions
Despite the aforementioned limitations, this study provides important information on a relatively understudied area in the intimate partner violence literature. This study extends previous research on the prevalence and severity of a variety of mental health outcomes among individuals experiencing no violence, perpetration only, victimization only, and bidirectional violence. Relatively clear patterns emerged, indicating which violence group tended to be most negatively impacted and which violence group tended to be least negatively impacted by IPV. Given the complexity and specificity of mental health outcomes that may be associated with the experience of IPV, future research endeavors are needed to replicate the present findings, to elucidate potential gender differences in the role-mental health association, and to examine how prevalence and severity of additional mental health disorders may differ by gender and perpetrator-victim role.
Acknowledgments
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. Information on how to obtain the Add Health data files is available on the Add Health website (http://www.cpc.unc.edu/addhealth). No direct support was received from grant P01-HD31921 for this analysis.
References
- Anderson KL (2002). Perpetrator or victim? Relationships between intimate partner violence and well-being. Journal of Marriage and Family, 64, 851–863. [Google Scholar]
- Archer J (2000). Sex differences in aggression between heterosexual partners: A meta-analytic review. Psychological Bulletin, 126, 651–680. [DOI] [PubMed] [Google Scholar]
- Cascardi M, & O’Leary KD (1992). Depressive symptomatology, self-esteem, and self-blame in battered women. Journal of Family Violence, 7(4), 249–259. [Google Scholar]
- Cascardi M, O’Leary K, Lawrence EE, & Schlee KA (1995). Characteristics of women physically abused by their spouses and who seek treatment regarding marital conflict. Journal Of Consulting And Clinical Psychology, 63(4), 616–623. [DOI] [PubMed] [Google Scholar]
- Cascardi M, & Vivian D (1995). Context for specific episodes of marital violence: Gender and severity of violence differences. Journal of Family Violence, 10, 265–293. [Google Scholar]
- Coker AL, Davis KE, Arias I, Desai S, Sanderson M, Brandt HM, & Smith PH (2002). Physical and mental health effects of intimate partner violence for men and women. American Journal of Preventive Medicine, 23(4), 260–268. [DOI] [PubMed] [Google Scholar]
- Cunradi CB, Caetano R, & Schafer J (2002). Alcohol-related problems, drug use, and male intimate partner violence severity among US couples. Alcoholism: Clinical and Experimental Research, 26(4), 493–500. [PubMed] [Google Scholar]
- Devries KM, Mak JY, Bacchus LJ, Child JC, Falder G, Petzold M, Astbury J, & Watts CH (2013). Intimate partner violence and incident depressive symptoms and suicide attempts: A systematic review of longitudinal studies. PLoS Medicine, 10(5), e1001439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Filson J, Ulloa E, Runfola C, & Hokoda A (2010). Does powerlessness explain the relationship between intimate partner violence and depression?. Journal Of Interpersonal Violence, 25(3), 400–415. [DOI] [PubMed] [Google Scholar]
- Graham K, Bernards S, Flynn A, Tremblay PF, & Wells S (2012) Does the relationship between depression and intimate partner violence vary by gender, victim-perpetrator role, and aggression severity? Violence and Victims, 27(5), 730–743. [DOI] [PubMed] [Google Scholar]
- Holtzworth-Munroe A (2005). Male versus female intimate partner violence: Putting controversial findings into context. Journal of Marriage and Family, 67, 1120–1125. [Google Scholar]
- Johnson MP (2004). Patriarchal Terrorism and Common Couple Violence: Two Forms of Violence Against Women. In Reis HT, Rusbult CE (Eds.) , Close relationships: Key readings (pp. 471–482). Philadelphia, PA, US: Taylor & Francis. [Google Scholar]
- Kaura SA, & Lohman BJ (2007). Dating violence victimization, relationship satisfaction, mental health problems, and acceptability of violence: A comparison of men and women. Journal Of Family Violence, 22(6), 367–381. [Google Scholar]
- Kim J, & Lee J (2013). Prospective study on the reciprocal relationship between intimate partner violence and depression among women in Korea. Social Science & Medicine, 9942–48. [DOI] [PubMed] [Google Scholar]
- Lawrence E, & Bradbury TN (2001). Physical aggression and marital dysfunction: A longitudinal analysis. Journal of Family Psychology, 15(1), 135–154. [DOI] [PubMed] [Google Scholar]
- Maxwell SE, & Delaney HD (2004) (2nd edition). Designing experiments and analyzing data. A model comparison perspective. Mahwah, NJ: Lawrence Erlbaum. [Google Scholar]
- Melander LA, Noel H, & Tyler KA (2010). Bidirectional, unidirectional, and nonviolence: A comparison of the predictors among partnered young adults. Violence and Victims, 25(5), 617–630. [DOI] [PubMed] [Google Scholar]
- O’Leary KD, Slep AM, O’Leary SG (2007). Multivariate models of men’s and women’s partner aggression. Journal of Consulting and Clinical Psychology, 75(5), 752–764. [DOI] [PubMed] [Google Scholar]
- Pico-Alfonso MA, Garcia-Linares M, Celda-Navarro N, Blasco-Ros C, Echeburúa E, & Martinez M (2006). The Impact of physical, psychological, and sexual intimate male partner violence on women’s mental health: Depressive symptoms, posttraumatic stress disorder, state anxiety, and suicide. Journal of Women’s Health, 15(5), 599–611. [DOI] [PubMed] [Google Scholar]
- Silverman J, Raj A, Mucci L, & Hathaway J (2001). Dating violence against adolescent girls and associated substance use, unhealthy weight control, sexual risk behavior, pregnancy, and suicidality. JAMA: The Journal of the American Medical Association, 286(5), 572–579. [DOI] [PubMed] [Google Scholar]
- Stith SM, Smith DB, Penn CE, Ward DB, & Tritt D (2004). Intimate partner physical abuse perpetration and victimization risk factors: A meta-analytic review. Aggression and Violent Behavior, 10(1), 65–98. [Google Scholar]
- Temple JR, Weston R, & Marshall LL (2005). Physical and Mental Health Outcomes of Women in Nonviolent, Unilaterally Violent, and Mutually Violent Relationships. Violence and Victims, 20(3), 335–359. [DOI] [PubMed] [Google Scholar]
- Udry JR, & Bearman PS (2002). New methods for new perspectives on adolescent sexual behavior. In Jessor R (Ed.), New perspectives on adolescent sexual behavior. Cambridge, UK: Cambridge University Press. [Google Scholar]
- Vaeth PC, Ramisetty-Mikler S, & Caetano R (2010). Depression among couples in the United States in the context of intimate partner violence. Journal of Interpersonal Violence, 25(5), 771–790. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Vivian D, & Malone J (1997). Relationship factors and depressive symptomatology associated with mild and severe husband-to-wife physical aggression. Violence and Victims, 12(1), 3–18. [PubMed] [Google Scholar]