|Home | About | Journals | Submit | Contact Us | Français|
Previous studies have found high rates of help-seeking among women who are abused by intimate partners. However, little research has investigated the help-seeking patterns of women who are aggressive in their relationships. Understanding the help-seeking of violent women may aid in designing interventions that will reduce the women’s violent behavior, as well as reducing the victimization they may be experiencing from their partners. This study examines the resource utilization of a sample of 108 women who had used violence against a male intimate partner in the previous six months. Ninety-four percent of the women also experienced physical violence from their partners. The study found that almost all of the women utilized community resources in an attempt to manage the violence in their relationships. In addition, a path model revealed that greater resource utilization predicted a decrease in women’s violence against their partners. The path model also revealed that women’s victimization did not directly predict greater resource utilization; rather, women who experienced higher levels of victimization were more likely to have a self-defensive motive for using violence, which increased the likelihood of seeking resources. Women who use violence to defend themselves may realize that they cannot handle their partner’s violence on their own, and so seek help. Women with more symptoms of posttraumatic stress symptoms also reported using more resources. Social support networks were the most widely used resource, and social support coping also contributed to increased resource utilization.
The resources abused women use to deal with their partners’ violence, such as domestic violence services, counseling, criminal justice system-related services, and social support from family and friends, have become a topic of increasing interest to researchers in the field of intimate partner violence. The motive underlying these studies is to ascertain how women respond when their partners are abusive, and what helps or does not help them to achieve safety for themselves and their children. Understanding what women do to maximize their safety is essential in designing interventions to aid battered women (Goodman, Dutton, Weinfurt, & Cook, 2003).
The current study examines resource utilization among a sample of women who used violence against male intimate partners. The study is based on the theory that women who use violence are, in many ways, not very different from those who have been identified in previous research as victims of domestic violence. For example, many women who use violence also experience violence from their partners. Depending on the study, the percent of domestically violent women who also experienced violence from their partners ranges from 64% in the large, nationally representative National Family Violence Survey (Straus & Gelles, 1990); 86% in a diverse Texas community sample (Temple, Weston, & Marshall, 2005); to greater than 90% in a diverse Connecticut community sample (Swan, Snow, Sullivan, Gambone, & Fields, 2004) and two studies with women who had been arrested for domestic violence (Hamberger & Guse, 2002; Stuart, Moore, Gordon, Hellmuth, Ramsey, & Kahler, in press). For some abused women, violence is a strategy used to deal with their partners’ violence (Campbell, Rose, Kub, & Nedd, 1998; Goodman et al., 2003). Furthermore, like women who are identified as victims, women who use violence have demonstrated high prevalence rates of posttraumatic stress, depression and physical injuries (Anderson, 2002; Sullivan, Meese, Swan, Mazure, & Snow, 2005; Swan & Snow, 2003).
The present study is also based on the principles of survivor theory, which posits that, contrary to a learned helplessness model (Walker, 1979), battered women use a number of different strategies to combat abuse and to try to achieve safety for themselves and their children (Gondolf & Fisher, 1988) The study seeks to address the following questions: Do women who use violence seek resources to try to protect themselves and their children, as has been found in studies of abused women (Goodman et al., 2003)? How does women’s violent behavior, as well as their own victimization, relate to their use of resources?
A central goal of the present study is to examine if use of community resources is associated with less violence among a sample of women who are violent towards their partners. A study conducted in California of the impact of domestic violence shelter-based organizations on the incidence of spousal killings found that domestic violence resources reduced the number of African American men killed by female partners (Wells & DeLeon-Granados, 2005; unfortunately, the study did not find that domestic violence resources contributed to a decline in African American women being killed by male partners.) The authors of the study hypothesized that perhaps the availability of resources provided battered women with alternatives to killing their partners to protect themselves (Wells & DeLeon-Granados, 2005). These findings suggest that use of domestic violence resources may reduce the extent to which women in abusive relationships use violence.
In the present study, we take a broad and comprehensive approach to examining resource utilization and its correlates. We examine women’s use of domestic violence services, such as shelters, hotlines, and support groups for battered women; and we also examine women’s use of social support networks; criminal justice related resources, such as police, victim advocates, and protective orders; professional counseling, self help groups, and substance use treatment; housing resources, such as services that aid low income individuals in finding affordable housing; and services for the women’s children, such as school-based services and counseling. Thus, our examination of resources goes beyond services commonly thought of as related to domestic violence. Path models are used to explore the relationship between women’s victimization from their partners and their use of resources, and the extent to which use of resources lowers the frequency of women’s violence. Variables that may indirectly affect the relationship between victimization and resource use are explored, including self-defensive motivations for using violence, coping, and symptoms of posttraumatic stress. The theoretical rationale for the model is developed in the following literature review. Much of the literature examining resource utilization that is cited in this paper has been conducted with women identified as victims of domestic violence. Given that almost all of the domestically violent women examined in the current study are also victims, we argue that this literature is appropriate for providing a theoretical framework for this study.
As violence becomes more frequent and severe, abused women increase their efforts to stop the violence (Campbell, Rose, Kub, & Nedd, 1998; Gondolf & Fisher, 1988; Goodman et al., 2003; McFarlane et al., 1997). In some cases, these efforts include using violence, either in self-defense or to convey a message that the abuser’s violence will not be tolerated (Burke, Gielen, McDonnell, O’Campo, & Maman, 2001). For example, in a study of resources used by 406 urban women seeking help for intimate partner violence, 82% fought back physically (Goodman et al., 2003). Further evidence is derived from a longitudinal study of the course of abusive relationships, which found that some women used violence against their partners as a strategy to keep themselves safe. At times the violence was in self-defense but at other times, it was initiated by the women (Campbell, Rose, Kub, & Nedd, 1998). Another study examined the applicability of the transtheoretical model of behavior change as applied to women’s ending of abusive relationships. “She fought back by hitting her abuser” was classified in the action stage of change, in which the woman is actively engaged in making changes to end the abuse (Burke, Gielen, McDonnell, O’Campo, & Maman, 2001). These studies found evidence that women’s violence may be one strategy among many to end the abuse or ameliorate its impact.
Self-defense is a common motive for women’s violence against male partners (Bowker, 1993; Fernandez-Esquer & McCloskey, 1999; Swan & Snow, 2003). In a previous paper examining women’s motives for using violence, which was conducted with the same sample used in the present paper, we found that 75% of women said that self-defense was a motive for their use of violence in their current relationship at least sometimes (Swan & Snow, 2003). We expect that women whose violence is primarily in self-defense will be more likely to turn to outside resources for help. Indirect support for this expectation is found in studies examining the relationship between frequency of abuse and resource utilization among victimized women. For example, Raghavan, Swan, Snow, and Mazure (2005) found that women who experienced more frequent abuse also were more likely to seek help from community resources. Similarly, Gondolf and Fisher (1988) compared domestic violence victims who used a shelter and those who did not, and found that women who used the shelter had experienced more frequent abuse than women who did not use the shelter. Although neither study specifically examined women’s use of violence in self defense, a self-defensive motive for violence may be correlated with frequency of abuse. That is, women who are abused more frequently may be more likely to use violence in self-defense to protect themselves from the increasing violence, and to seek resources to help them deal with the violence. This is consistent with the suggestion that active forms of coping (such as fighting back in self-defense and seeking help from community resources) are more likely when frequency of abuse increases (Waldrop & Resick, 2004).
In the general stress and coping literature, coping is often grouped into three types: avoidant, problem solving, and support seeking (Amirkhan, 1990). Avoidant coping refers to trying not to think about a stressor, distracting oneself, and withdrawing from others; problem-solving coping involves actively trying to resolve a stressor by planning and carrying out a course of action; and support seeking coping includes talking to family and friends about a stressor for support and advice. Studies relating coping to a variety of psychological and physical health outcomes have found that avoidant strategies are generally related to poorer outcomes, such as depression (Billings & Moos, 1981); while problem solving and support seeking are usually related to positive outcomes (Ingledew, Hardy, & Cooper, 1997; Shinn, Wong, Simko, & Ortiz-Tomes, 1989; Snow, Swan, Raghavan, Connell, & Klein, 2003). Among intimate partner violence victims, avoidance coping has been related to the development of psychological difficulties (Foa, Cascardi, Zoellner, & Feeny, 2000). The relationship between avoidance coping and resource utilization has not, to our knowledge, been examined. However, avoidance coping has been found to decrease the likelihood of seeking general counseling services (Vogel & Wester, 2003). Likewise, a small study of abused women found that women who used more avoidance coping had less social support (i.e., fewer supporters and less contacts with friends and family; Mitchell & Hodson, 1983). Therefore, we expect that women who use high levels of avoidance coping will be less likely to seek resources.
Social support is both a means of coping, and a resource contributing to the availability of other forms of coping (Aspinwall & Taylor, 1997). Waldrop & Resick (2004) suggest that greater social support (i.e., more individuals in the social support network; empathic and helpful responses from those in the network) may encourage women in violent relationships to engage in problem-solving kinds of coping strategies (see also Mitchell & Hodson, 1986). We expect that women in the present study who engage in more social support coping will be more likely to seek resources.
The level of posttraumatic stress disorder among battered women is estimated to be “extremely high” (Foa et al., 2000); this has been confirmed in a number of studies, many of which document the prevalence rate at approximately 40% (Dansky, Byrne, & Brady, 1999; Kocot & Goodman, 2003). No studies that we are aware of have examined the relationship between intimate partner violence-related posttraumatic stress symptoms and resource utilization. However, some literature exists regarding help-seeking and service utilization comparing women with and without PTSD (without a specified referent traumatic event). Findings show that women with PTSD are more likely to seek formal and informal help than women without PTSD (Amaya-Jackson et al., 1999; Lewis et al., 2005) including greater use of mental health services (Amaya-Jackson et al., 1999; Switzer et al., 1999). Based on these studies, we predict that women with greater symptoms of posttraumatic stress will be more likely to seek resources, perhaps as a function of being motivated to relieve the uncomfortable and debilitating effects of posttraumatic stress symptoms.
This section briefly reviews research on resources used by women experiencing domestic violence. Social support from informal networks plays a critical role in the lives of women dealing with intimate partner violence. Supportive family members and friends may provide material support, such as money and a place to stay; they may guide women towards other resources, such as legal aid and counseling; and they can be a long-term form of assistance (as opposed to time-limited contact with advocates and therapists; Bell et al., 1996). For example, in a study of African American women seeking hospital care, social support mediated the relationship between victimization from partners and psychological distress: experiencing partner abuse predicted greater levels of distress; however, the more social support the women reported, the less their distress symptoms (Thompson et al., 2000). In Goodman et al.’s (2003) study of the strategies battered women used to deal with their violent relationships, social support was an essential resource: the majority of the women talked to family members about the abuse, most were helped by them, and more than half stayed with family or friends.
Use of the criminal justice system has become the primary way domestic violence is addressed in the United States. By 1988, 90% of U.S. cities had implemented mandatory or preferred arrest policies in cases of domestic violence (Sherman, 1992). The National Violence Against Women Survey, conducted with a nationally representative sample of 8,000 women, found that 27% of women who indicated that they had been physically assaulted by an intimate partner reported the incident to the police (Tjaden & Thoennes, 2000b). Of those who reported the incident, approximately one-third stated that the attacker was arrested or detained. Fewer women (16%) obtained a restraining order, and half of those that did reported that the order was violated. Police officers are not only agents of social control responsible for stopping the violence and for arresting the offender; increasingly officers also are serving as referral agents for victims (Miller & Krull, 1997). In the National Violence Against Women survey, 25% of the women who reported their assaults to the police said the police referred them to services (Tjaden & Thoennes, 2000b). Goodman and colleagues (2003) found that the majority of their sample of women seeking help for intimate partner violence used criminal justice related resources, and many found them helpful.
A few studies provide information about women’s use of various types of counseling to help them cope with intimate partner violence. The National Violence Against Women survey found that over one-fourth of the female victims in their survey received mental health counseling (Tjaden & Thoennes, 2000a). Goodman and colleagues (2003) found that almost one-third of the women in their study called a mental health counselor, and most found this helpful; and one quarter tried to get services for alcohol/substance use issues, with half finding this helpful.
Domestic violence specific services, such as advocates, shelters, support groups, and hotlines, can be very beneficial to women dealing with intimate partner violence. For example, Sullivan and Bybee (1999) conducted a longitudinal study of sheltered women randomly assigned to either an advocacy or a no advocacy condition. At the two-year follow-up, the women who worked with the advocates experienced less violence, reported higher quality of life and social support, and were better able to obtain the resources they needed. Likewise, in the array of resources examined in the Goodman et al. (2003) study, the resource with the highest percentage of women stating that it was helpful was domestic violence programs (79%).
While the negative consequences of exposure to domestic violence for children has been documented in a number of studies (Groves, 2002; Holden, Geffner, & Jourlies, 1998; Margolin, 1998; Rossman, Hughes, & Rosenberg, 2000), very few have examined battered women’s seeking of resources specifically for their children. The impact of children on women’s decisions to seek resources is complex. One qualitative study of mothers using victim services found that more than half sought help because of their children. On the other hand, some women delayed seeking assistance because of their children’s attachment to the abusive partner (Zink, Elder, & Jacobson, 2003). In a study of women participating in a health survey, 16% of those who reported intimate partner violence in the past year sought help for their children (Hathaway et al., 2000). Some women fear that they will lose custody of the children if anyone learns that there is violence in the home (Koss et al., 1994).
Housing is a significant need for many women dealing with intimate partner violence, particularly for women with low incomes who leave their partners. Sullivan and Rumptz’s (1994) longitudinal study of sheltered women found that six months after their shelter stay, over half had moved at least once, and over one-third were dissatisfied with their housing situations. Similarly, 79% of women in the Goodman and colleagues (2003) study left their homes to get away from the abuser.
Based on the literature reviewed above, we developed the following hypotheses. A path model portraying the hypothesized relationships among variables is shown in Figure 1.
Women were recruited from four locations in a moderate-sized New England city: a court-mandated domestic violence program; a large inner-city health clinic for low income residents; a division of family court that provides services for people with domestic violence, divorce, and child custody cases; and a local domestic violence shelter.
The criterion for entry into the study was that the woman had to have used some form of physical violence against a male intimate partner within the previous six months, determined via a phone screen with items from the Conflict Tactics Scale-2 (Straus, Hamby, Boney-McCoy and Sugarman, 1996; CTS-2). Assessments were administered in face-to-face interviews.
The majority of the 108 women in the sample (73%) were patients of the inner-city health clinic. Seventy-one percent of the participants were African-American; 14% were White, 10% were Latina, and 5% used other categories to describe themselves (two bicultural, three “Other”). The majority of participants (62%) were between the ages of 25–40, 18% were below age 25, and 17% were over 40. Twenty-six percent of the sample had not completed high school, 42% did complete high school; 8% graduated from a vocational school, 18% had attended some college, and 6% had a college or graduate degree. Overall, the income range of the women was quite low, with 68% earning less than $10,000 per year, 19% earning between $10,000 and $19,999, and 13% earning $20,000 or more. Seventy-five percent of the sample was unemployed, 11% had some part-time work, and 14% worked full-time. Eighty-four percent had children, while 65% had children living with them. Almost half of the sample had been with their partners between one to five years, with 38% together between five and 20 years, and 27% together less than one year.
The two dimensions of partner abuse assessed in the current study were physical abuse (e.g. punch or hit with something that could hurt) and sexual coercion (e.g insist on sex when partner did not want to) using items from the CTS-2 (Straus, et. al, 1996). A referent time period of six months was used to assess a participant’s commission of an abusive behavior toward her partner, and the partner’s commission of each behavior toward the woman. The response scale ranged from never, once, twice, 3–5 times, 6–10 times, to more than 10 times in the past six months. Items were recoded so that the value of the variable was the midpoint of the frequency range (i.e., never = 0, once = 1, twice = 2, 3–5 times = 4, 6–10 times = 8, and, more than 10 times was conservatively coded as 11.
The Conflict Tactics Scale has been used in hundreds of studies since 1972 with over 70,000 participants of diverse cultural and ethnic backgrounds around the world (Straus, et al., 1996). The CTS-2 has been shown to have good internal consistency and validity (Straus, et al., 1996). For the purposes of this study, women’s victimization and women’s violence variables each were composite scores created by summing the physical abuse and sexual coercion scales. In the present study, the reliability of the CTS-2 items for women’s violence was α = .80 and for women’s victimization was α = .86.
Coping with relationship stress was measured using the 33-item Coping Strategy Indicator (Amirkhan, 1990). A factor analysis conducted by Amirkhan (1990) revealed three factors: problem solving coping (e.g., I tried to carefully plan a course of action rather than acting on impulse); seeking support coping (e.g., I confided my fears and worries to a friend or relative); and avoidance coping (e.g., I avoided being with people in general). Participants were asked to rate each item in relation to an identified relationship problem on a scale from 1 (not at all) to 3 (a lot). Amirkhan (1990) has found that the measure has good reliability and validity. In the present study, each coping score was created by summing the 11 items for that scale; internal consistencies were α = .80 for avoidance coping, α = .93 for support seeking coping, and α = .83 for problem solving coping.
Self-defense as a motive for using violence was measured with two items from the Motives Scale (Swan & Gill, 1998): “How often do you use violence to defend yourself from your partner?” and “How often do you use violence to get him to stop hitting or otherwise hurting you?” Responses ranged on a scale from 1 (almost never) to 4 (almost always) and were summed to create one score. Reliability for this subscale was α = .79.
Posttraumatic stress symptoms were assessed with 10 items from the Crime-Related PTSD scale for women (Saunders, Arata, & Kilpatrick, 1990). This scale has been useful in discriminating victims of rape from non-victims (Saunders, Arata, & Kilpatrick, 1990) as well as sexually harassed and non-harassed women (Schneider, Swan, & Fitzgerald, 1997). Responses ranged from 0 (no discomfort) to 4 (extreme discomfort) and were summed to created the measure of symptoms used in the model. Reliability for the posttraumatic stress symptoms scale was α = .87.
The Resource Utilization Questionnaire, a 33-item measure that assesses resources used to deal with domestic violence (Swan & Gill, 1998; RUQ), was administered. The RUQ is a self-report measure of a participants’ utilization of resources in six domains. The social support domain measures the extent to which a woman has utilized people in her support network. Example items include “talked to someone about violence in the relationship” and “stayed with family or friends to keep myself safe”. The criminal justice related domain assesses resources that a woman initiated herself (e.g., called the police, obtained a restraining order, talked to a domestic violence court counselor), as well as those resources the woman did not seek herself (e.g., someone else called the police, such as a neighbor). The Professional counseling or self-help domain documents the utilization of structured treatment such as substance abuse treatment, couples counseling, and family therapy, and also 12-step groups like Alcoholics or Narcotics Anonymous. Domestic violence specific resources include domestic violence groups, shelters, and hotlines. The children’s services domain measures, for those women who have children, the extent to which child-focused resources are utilized by the child (e.g., child counseling, help with school) or the woman (e.g., parenting skills training). Child protective services were also included in this category. The other resources category consists of help in finding housing and use of a telephone service that provides information about a variety of community resources, ranging from childcare to mental health treatment. Response options were “yes,” “no,” and “not in this relationship, but in the past.” Participants were instructed to focus their response on those resources that were used to deal with “the violence in (their) relationship, to keep (themselves) safe, or to deal with other problems in the relationship.” Nine items were found to be unreliable and were dropped, leaving 24 items which were summed to create the resource utilization scale, α = .78. The 24-item resource utilization scale used in the path models is based on the women’s use of resources within the current relationship. Responses were coded as 0 (did not use resource in current relationship) or 1 (did use resource in current relationship).
Study variables were assessed for assumptions of normality. All variables met these assumptions with the exception of women’s victimization and women’s violence, which showed mild levels of skew. Normal distribution of these variables was achieved by performing square root transformations (Tabachnick & Fidell, 1996).
A path model (shown in Figure 1) assessed the relationships among women’s victimization, resource utilization, coping, self-defense motive, posttraumatic stress symptoms, and women’s use of violence. The AMOS 4.0 statistical program (Arbuckle and Wothke, 1999) was utilized to analyze the path model, obtain maximum-likelihood estimates of model parameters, and provide goodness of fit indices. A model that provides a good fit to the data is generally considered to be one that has a non-significant chi-square value for the model, a root mean square of approximation (RMSEA) value of less than .05, with a p test for closeness of fit for RMSEA of .50 or greater, and a chi-square to degrees of freedom ratio of less than 3 (Byrne, 2001). Individual path coefficients, which can be interpreted like regression coefficients, are considered significant at the p <.05 level.
Hypothesis 1 (similar to victimized women, women who use violence against male partners will show high rates of resource utilization) was confirmed. All but seven of the 108 participants used at least one of the community resources assessed in this study. The most widely used resources were in the social support domain. Eighty-seven percent of the women talked to someone about the violence; 62% said that people in their support network helped in other ways; and 60% stayed with family or friends to keep themselves safe. The second most frequently used resource was calling the police (64%). Other resources among the top ten most frequently used services included receiving help in obtaining housing (such as government subsidized Section 8 housing; 50%;) others calling the police (such as a neighbor), 46%; talking to a domestic violence court counselor, 44%; substance abuse treatment, 42%; obtaining a protective order, 42%; and individual counseling, 41%.
Sixty-five percent of study participants had children living with them. Among the women living with their children, 60% reported that their children received services. Among the women whose children received services, the most frequent form of services the children received was from Child Protective Services (63%). Fifty-four percent of children received counseling; 39% received home visits, and 21% of the mothers of those children received parenting skills training.
Correlations, means, and standard deviations of those variables depicted in the path model are presented in Table 1. Women’s victimization and women’s violence are positively correlated. It is notable that the correlation between victimization and resource utilization was not significant, consistent with Gondolf & Fisher’s (1988) study. Resource utilization positively correlates with a self-defensive motive for violence, posttraumatic stress symptoms, and social support coping. Avoidance coping was positively correlated with victimization, self-defensive motive, and posttraumatic stress symptoms. Problem solving coping did not correlate with resource utilization, and so was dropped from the path model. Social support coping correlated only with resource utilization and problem solving coping; therefore, social support was portrayed in the model as an exogenous (i.e, no paths predicting it) predictor of resource utilization.
The path model is presented in Figure 1. Fit statistics indicate an excellent fit of the model to the data; Χ2 (10, n = 108) = 10.48, p = .40; RMSEA = 0.21, with a confidence interval from 0.00 to 0.11 and a p for test of close fit = 0.608. The model did confirm that a positive relationship between victimization and women’s violence exists (β = .31, p < .01), indicating that being victimized increases the likelihood that women will use violence, and vice-versa. This finding is consistent with other studies examining both violence and victimization (Archer, 2000; Magdol et al., 1997; Sullivan, Meese, Swan, Mazure, & Snow, 2005).
Our second hypothesis states that the relationship between victimization and use of resources will be mediated by a self-defensive motive for violence; use of avoidance coping to deal with relationship problems; and posttraumatic stress symptoms. This hypothesis was mostly confirmed by the model. Women who were highly victimized were more likely to use violence in self-defense (β = .36, p < .001), and to have symptoms of posttraumatic stress (β = .28, p < .01). The path from victimization to avoidance coping strategies was non-significant but did closely approach significance in the predicted direction (β = .18, p = .055), providing some evidence that greater victimization may increase the likelihood that women will use avoidance coping to manage the resultant stress. Hypothesis 2 also predicts that women who use violence for reasons of self-defense, and women with posttraumatic stress symptoms, will be more active in seeking resources; women using high levels of avoidance coping will use fewer resources. This prediction was largely supported. Women who used violence in self-defense1 were more likely to use community resources (β = .34, p < .001), and women who used high levels of avoidance coping2 were less likely to use resources (β = −.24, p < .05). The path from posttraumatic stress symptoms3 to resource utilization was non-significant but did very closely approach significance in the predicted direction (β = .18, p = .050), providing some evidence that posttraumatic stress symptoms may increase the likelihood that women will seek resources.
The third hypothesis predicts that women who engage in greater support seeking coping will use more resources. This hypothesis was confirmed: women who used support coping to deal with their relationship problems also sought more resources (β = .20, p < .05). The final hypothesis predicts that women who use resources will be less likely to perpetrate violence against their partners, and was confirmed. Use of resources decreased the likelihood that women would perpetrate violence against their partners (β = −.19, p < .05).
The bivariate correlations (Table 1) indicate positive correlations between posttraumatic stress symptoms and use of avoidance coping, and between self-defense motives and avoidance coping. Based on these correlations, paths between these variables were tested in the path model. Findings indicate that women experiencing posttraumatic stress symptoms are more likely to use avoidance coping to deal with relationship problems (β = .19, p < .05). The model also indicates that women who use greater levels of avoidance coping also were more likely to have a self-defensive motive for their use of violence (β = .27, p < .01).
We have argued that women who use violence against their partners resemble women traditionally thought of as domestic violence victims in many respects (Sullivan, Meese, Swan, Mazure, & Snow, 2005; Swan & Snow, 2002; 2003; Swan & Snow, in press; Swan, Gambone, Fields, Sullivan, & Snow, 2005). This study provides further evidence that this is the case. Participants entered the study because they recently used violence against their partner and yet, 94% experienced violence from their partners. Moreover, the women in this study actively sought and used a range of community resources to deal with the violence in their relationships. Ninety-four percent of the participants used at least some of the community resources assessed in this study. Two-thirds of the participants called the police, indicating further that most of the women in this study, far from being “batterers”, experienced the violence from their partners as out of control and necessitating police intervention. The findings are consistent with survivor theory, which posits that battered women use a number of different strategies to combat abuse and to try to achieve safety for themselves and their children (Gondolf & Fisher, 1988). We argue further that for some women, one strategy to combat abuse is the use of violence. This point is illustrated in a quotation from a participant in Burke et al.’s (2001) study: “He had me so scared of him that I didn’t even do nothing and I took so much of it. One day I hit him with a baseball bat and ever since then he hasn’t put his hand on me” (p. 1153).
The present study also found that use of resources was a protective factor that reduced the likelihood of women’s perpetration of violence. Women who used more community resources were less likely to use violence against their partners. This finding provides evidence that women struggling with domestic violence (as well as poverty, substance abuse, or mental health issues) who have access to resources will use these resources and will then be less likely to resort to violence to deal with their relationship problems.
This study found that the relationship between victimization and resource utilization is mediated by the motive to use violence in self-defense. Women who experienced greater levels of victimization were more likely to indicate that self-defense was the motive for their use of violence. Self-defensive motives then predict a greater likelihood of seeking resources. This finding suggests that when women fight back physically, they may realize that they are unable to manage the partner’s violence themselves and so must seek outside help. A complementary explanation is that, using a stages-of-change framework, women who fight back in self-defense and who seek resources may be at an action-oriented stage in which they are moving towards ending the relationship. This idea is supported by a study that found that women who utilized more community resources were more likely to end their relationships as compared to women who utilized fewer resources (Raghavan, Swan, Snow, & Mazure, 2005).
A finding that is very consistent with the literature is that greater victimization leads to increased posttraumatic stress symptoms. Further, the model suggests that women with greater symptoms of posttraumatic stress use more resources. We suggest that the uncomfortable and debilitating symptoms of posttraumatic stress motivate women to seek resources to reduce this discomfort. Our findings are consistent with survivor theory, and suggest that as the violence intensifies and leads to greater distress from posttraumatic stress symptoms, women will realize that they must seek help.
However, the model also tells another story that illustrates a less adaptive response to victimization. This pathway can be seen in the relationship between avoidance coping and resource utilization. The model suggests that women who are victimized are more likely to use avoidance coping strategies, such as trying to distract themselves from thinking about the stressor, and not talking to other people about the abuse. Similarly, Mitchell and Hodson (1983) also found that greater frequency of violence were associated with more avoidance coping. In the present study, women who used avoidance coping strategies were less likely to utilize community resources.
The model also suggests that posttraumatic stress symptoms may in some cases lead to avoidance coping and less use of resources. How do we understand that posttraumatic stress symptoms directly predict greater use of community resources, but they also predict avoidance coping and, indirectly, less use of resources? Perhaps the answer lies in the particular cluster of posttraumatic stress symptoms that are primary in an individual’s experience of PTSD. For example, one cluster of posttraumatic stress symptoms relates to avoidance and includes behaviors such as trying not to think about, talk about, or have feelings about the event, as well as to avoid people or places that remind the individual of the trauma (Foa, 1995). An individual whose posttraumatic stress symptoms were strongly avoidant would more likely use avoidance coping to manage the distressing emotions and cognitions resulting from posttraumatic stress. In turn, this type of avoidance coping may be related to a lower of level of resource utilization, since using a resource would probably force the individual to confront what she has been trying to avoid. In contrast, a different individual could have more difficulty with the re-experiencing and hyperarousal symptoms of PTSD, and fewer avoidance symptoms. This individual may be motivated to seek resources to reduce the severity of her symptoms and the associated distress. We were not able to examine this possibility in the present study, as the posttraumatic stress measure we used did not assess symptom clusters of PTSD. A fruitful avenue for future studies of resource utilization among women in violent relationships would be to assess if the different symptom clusters of PTSD predict different patterns of resource utilization. It would also be helpful to examine of other distress symptoms, such as depression, anxiety, and substance use, have an impact on women’s use of resources.
The relationship between avoidance coping and use of violence in self defense demonstrates an additional complexity in the model. While greater use of avoidance coping directly predicts less use of resources, avoidance coping also predicts a self-defensive motive for violence. Using violence in self-defense, in turn, predicts more use of resources. The complex relationships shown in the model may be a reflection of the dynamic nature of the coping process. Perhaps when women rely on avoidance coping to deny and minimize their partner’s violence, they do not develop other resources or strategies to deal with it. Therefore, lacking other strategies, when faced with violence they are more likely to have no other options but violence in self-defense. Use of violence in self-defense, in turn, may be a “wake up call” that indicates to the woman that the violence is out of control and she must seek help.
Consistent with other studies, social support was the most frequently used resource (Goodman et al., 2003; Thompson et al., 2000). Social support also played an important role in the model as a predictor of resource utilization. Consistent with theories of social support (Aspinwall & Taylor, 1997; Mitchell & Hodson, 1986; Waldrop & Resick, 2004), the more women used social support as a coping strategy, the more they tended to seek resources. Perhaps family and friends in their support network aided women in seeking resources by telling them about the sources of help that were available or by encouraging them to seek other forms of help. For example, Lewis et al. (2005) found that receiving informal help and support increased women’s readiness to engage in formal help seeking.
A limitation of the current study is the smaller sample size (N=108). These findings need to be replicated with larger samples before the results can be considered more than suggestive. Future work examining other types of abuse, such as coercive control and emotional abuse, would also be an important and informative development.
These findings have several implications for service providers who work with women in situations of domestic violence, including women who use violence and those who do not. This study found that almost all of the women who used violence were also victims of violence, suggesting that victimization needs to be examined when conducting an assessment of a woman who is violent towards her partner. The study also found that these violent women used a variety of resources to deal with both their partners’ and their own violence. The study suggests that posttraumatic stress and the types of coping the woman is using to deal with her violent relationship are important to assess. Women suffering from symptoms of posttraumatic stress clearly need to have those symptoms addressed. Women who are relying primarily on avoidance coping may be less likely to use resources, and/or may be at an earlier stage of change (e.g., pre-contemplative; Prochaska, DiClemente, & Norcross, 1992) with respect to making changes in the relationship.
Future research should take a broad view of the resources women use, including not just domestic violence-specific resources and the criminal justice system, but also social support networks; substance abuse treatment, counseling and self-help groups; help obtaining housing; and resources obtained for children. The study indicates that our usual approaches to combating domestic violence – that is, domestic violence services and the criminal justice system – are necessary but not sufficient to end domestic violence. Many women who participated in this study had needs that could not be met by domestic violence providers (e.g., poverty, housing, substance abuse treatment). This finding highlights the need for linkages between domestic violence providers and other community services (Hanson, Hesselbrock, Tworkowski, & Swan, 2002). People who provide all kinds of services, not just domestic violence providers, report that many, if not most, of their female clients are dealing with domestic violence. Service providers in all areas would benefit from domestic violence training, linkages with the domestic violence providers in their communities, and systems for cross-referral.
This research was supported in part by a grant from the Ethel F. Donaghue Women’s Health Investigator Program at Yale.
1A formal test of whether the self-defensive motive for violence mediates the relationship between victimization and resource utilization was conducted using a method based on Baron & Kenny (1986). To be a mediator, variable A must have a direct and significant path to variable C, in a model that includes only those 2 variables; and variable B (the mediator) must render this path nonsignificant in a second model that includes the direct path from A to C as well as paths from A to B and B to C. This test confirmed that self-defense does fully mediate the relationship between victimization and resource utilization. The path from victimization to resource utilization, with no other variables in the model, was found to be significant (β = .18, p < .05). However, when the indirect paths were included, the direct path from victimization to resource utilization became nonsignificant (β = .05, p > .05). The path from victimization to self-defense was significant (β = .43, p < .05), as was the path from self-defense to resource utilization (β = .27, p < .05).
2Avoidance coping is not a full mediator of the relationship between victimization and resource utilization by definition, because the path from victimization to avoidance coping failed to reach significance.
3Posttraumatic stress symptoms are not a full mediator of the relationship between victimization and resource utilization by definition, because the path from posttraumatic stress symptoms to resource utilization failed to reach significance.
4The resource utilization scale contains two social support items, “talked to someone about violence in the relationship” and “stayed with family or friends to keep myself safe”. We examined the possibility that the path in the model between social support coping and resource utilization is inflated because of these two items. We removed these two items from the resource utilization scale and re-ran the model shown in Figure 1. The path coefficient from support coping to resource utilization remained the same as in the original model. Other path coefficients and model fit were also very similar to the original model.
Suzanne C. Swan, University of South Carolina.
Tami P. Sullivan, Department of Psychiatry, Yale University School of Medicine.