Abstract
Women in physically and psychologically abusive relationships face numerous decisions related to their safety: decisions that historically have been viewed by researchers and human service practitioners as related to individual or interpersonal factors, such as how they feel about their partner, what they (or those they are close to) think is best for their children, or whether they have a safe place to go to. Social and structural factors, such as poverty, sexism, and barriers related to disability, are either left out or viewed at their individual-level consequence, such as a woman’s employment status. Using interview data and case studies from a larger study on housing instability, partner violence, and health, the authors apply ecological and macro-level theoretical models that go beyond the individual level to the stories of women who struggled with partner violence, arguing that it is critical to examine the large social and structural forces that impact women’s lives if we are to understand the decisions women make when facing a violent partner.
Women in physically and psychologically abusive relationships face numerous decisions related to their safety: decisions about how they should respond to the violence; about whether, how, or when to leave the relationship; about whether to disclose their experiences to someone and if so, to whom; and about where to go if they do leave. Particularly in North America and other “Westernized” cultures, we tend to view these decisions as being about individual choices. We ask “why does a victim stay” or “what enabled her to leave?” and even when our gaze shifts from her personal characteristics (e.g., her self-esteem or her beliefs about marriage or her history of abuse) to factors that may be beyond her control (e.g., the shelter is full and she has no place to go), we still focus on the individual-level outcome (staying or leaving) rather than on the larger structural forces that shape these decisions.
This individualistic perspective is reinforced by the scientific literature addressing battered women’s decisions to seek help, stay in, leave, or remain separated from a relationship, which tends to focus heavily on individual-level influences, such as her psychological characteristics or past experiences, or on interpersonal factors related to her relationship with her partner, her children, or others in her life (Alhalal, Ford-Gilboe, Kerr, & Davies, 2012; D. Anderson & Saunders, 2004; Enander & Holmberg, 2008; Lacey, 2010; Lerner & Kennedy, 2000). Although variables such as employment status, income, or housing are often cited as reasons that prevent women from leaving violence, they are almost always measured or discussed as individual-level variables rather than as indicative of larger social or structural mechanisms at work (D. Anderson & Saunders, 2004; Kim & Gray, 2008).
In this article, we argue that limiting our perspective on intimate partner violence (IPV) victims’ decision making to the individual and/or interpersonal levels is shortsighted. Not only does it often result in victim blaming by placing the bulk of the responsibility on the abused woman for her situation, but by ignoring the social context in which the woman, her partner, and their relationship exists, individual-focused analyses tell an incomplete story. By applying theoretical models that go beyond the individual level to the stories of women who struggled with partner violence, we argue that it is critical to examine the large social and structural forces that affect women’s lives if we are to understand the decisions women make when facing a violent partner. First, we highlight the ways in which multiple and complex factors beyond individual characteristics work both alone and synergistically to constrain women’s choices and influence their strategies for keeping themselves and their children safe when a partner is violent. Next, we take this a step farther by illustrating the importance of looking at women’s decision making through the critical lens of overlapping social inequities that drive and reinforce violence. Finally, we show how experiences related to inequities often compound across women’s lives to affect the choices they are able to make when they are in a violent partnership. By using these ecological and structural perspectives to examine the decisions women made about their relationships, we show how these systematic disparities act to not only limit women’s resources (a common argument) but often function to reinforce the violence and control of the abusive partner.
Frameworks for Understanding Abused Women’s Responses to Violence
Within the last three decades, scholars across disciplinary fields have evoked various theoretical frameworks to explain how, when, and why abused women make decisions about their relationships. Examples of prominent theories include the transtheoretical/readiness to change model (Burke, Mahoney, Gielen, McDonnell, & O’Campo, 2009; Rhatigan, Street, & Axsom, 2006; Shorey, Tirone, Nathanson, Handsel, & Rhatigan, 2013; Shurman & Rodriguez, 2006), the investment model (Edwards, Gidycz, & Murphy, 2011; Rhatigan & Street, 2005; Rhatigan et al., 2006), learned helplessness (Rhatigan et al., 2006), and reasoned action/planned behavior models (Byrne & Arias, 2004; Choice & Lamke, 1997), among others. Although many of these perspectives include an examination of influences on behavior outside of the individual (e.g., the theory of reasoned action/planned behavior suggests that the opinions of others may influence an individual’s behaviors; Glanz, Rimer, & Lewis, 2002), the bulk of the theoretical focus surrounding IPV remains on an examination of control at the individual level.
Over the years, a handful of researchers have used a social-ecological framework to look beyond individual psychological factors to examine and explain behaviors related to partner abuse (Alaggia, Regehr, & Jenney, 2012; Bliss, Cook, & Kaslow, 2006; Carlson, 1984; Heise, 1998). In this model, individual behaviors and decisions are influenced by a multitude of forces that span several dimensions, including individual characteristics (age, knowledge, mental health status), interpersonal influences (family, friends, colleagues), community and organizational factors (workplace or school policies, the availability of social services in a neighborhood), and social/political/cultural arenas (laws, cultural expectations, and normative values). For example, a woman’s response to physical violence likely depends on a combination of factors, including her self-confidence or beliefs about relationship commitment (individual), the presence of children and her perception of the impact of violence on them (interpersonal), the accessibility of social services such as shelters and housing assistance (community), and the policies surrounding social programs such as welfare assistance (social; Alaggia et al., 2012; Carlson, 1984).
One of the strengths of this framework is its acknowledgment that “causal and maintaining factors on different levels . . . can operate simultaneously, either independently or interactively” (Carlson, 1984, p. 570). In other words, it accounts for the complexity inherent in relationships. Often, however, the importance of this interaction is downplayed in favor of looking at the list of factors themselves. In addition, despite general agreement that circumstances such as being unemployed or dependent on a partner for income can increase the risks and severity of violence and/or make it more difficult for abused women to leave a violent relationship, when larger structural constructs such as poverty, racism, or sexism are mentioned in this literature, they are frequently reduced to individual-level characteristics. For example, poverty becomes actualized as the individual-level factor “unemployment,” which hinders a woman’s ability to leave a relationship because without a job she cannot take care of her children. A more comprehensive approach would examine how structural forces such as the systemic underfunding of programs intended to alleviate poverty limit access to housing, health care, child care, and parental leave for women who may want or need to leave a violent relationship (Paterson, 2009).
Rooted in Black feminist and critical race theory, intersectionality is a theoretical framework that addresses these limitations by highlighting numerous ways in which multiple structural forces interact with one another to shape individual choices and opportunities (Crenshaw, 1989). Core tenets of this framework include the notion that social identities are multiple and intersecting (a woman can be Black, female, and poor); that these identities are valued differently by society and often result in individuals achieving different levels of power based on these characteristics; that no single category is more important or dominant (just because two women have gender in common does not make their lives similar); that we must consider what happens at the intersection of these structures of discrimination and oppression; and that this is important if we are to be effective in understanding and changing power relations between dominant and oppressed populations (Bowleg, 2012; Crenshaw, 1989, 1991).
When applied in conjunction with a social-ecological model, intersectionality illustrates the necessity of examining how factors across the ecological framework overlap with each other, often reinforcing the constraints (or expanses) these influences have on the ability of an individual to make choices or take action (such as leaving an abusive relationship). Specifically, intersectionality reminds us of the power that structural and systemic forces have across the social-ecological framework, and speaks to the importance of including those forces in our research framework. Poverty, for example, interacts with violence on women’s lives in ways that go beyond many of our common perceptions, such as through the increased stress financial insecurity places on individual relationships. At a larger level, poverty and gender inequity intersect and structurally reinforce partner violence by constraining women’s choices through society’s failure to provide adequate options for necessities such as affordable housing or subsidized child care for low-income families (Crenshaw, 1991; Paterson, 2009; Purvin, 2007). When other factors such as immigration status are included in the mix, another layer of constraint is added: Poor, undocumented women face not only these consequences but additional threats of incarceration and deportation.
Finally, a life-course theoretical orientation (Elder, Johnson, & Crosnoe, 2003; Halfon & Hochstein, 2002; Halfon, Larson, Lu, Tullis, & Russ, 2014) contributes to this analysis by illustrating how the quantity, timing, and cumulative effects of events (particularly adversity and trauma that occur during key developmental points during childhood and in combination with various forms of inequity) is critically important in shaping the impact these events have. Much like the social-ecological framework, a life-course perspective informs our analysis by suggesting that trajectories of health and wellness are influenced by social, economic, and environmental exposures or experiences in addition to individual characteristics, and that these forces interact with one another, further influencing their effects.
Although life-course perspectives on health developed from a biologically deterministic framework, many current approaches argue that reducing a life-course analysis to examining “simple linear relationships” between biological or psychosocial events is insufficient to understand health; rather, multiple aspects of development “dynamically interact to shape the health development process” (Halfon et al., 2014, p. 352). In this way, aspects of life-course theory and intersectionality work well together, emphasizing that while some factors help to prevent negative outcomes, others may increase our risk for them, and it is the cumulative effects of positive and negative factors that contribute to shaping direction our lives take. In addition, life-course theory adds a further dimension, proposing that the timing of these events as well as other experiences (such as abuse or witnessing violence) within a developmental perspective is also important; some exposures may be more damaging when experienced early in life rather than later, and that some periods (such as early childhood, preconception, or pregnancy) can be times of increased vulnerability and sensitivity to these exposures (Minh et al., 2013; Elder, 1998; Elder et al., 2003). This perspective has been used in the context of IPV to explain multiple victimization, the relationship between child abuse and later victimization, as well as health outcomes (Carbone-Lopez, Rennison, & Macmillan, 2012; Davies et al., 2015; Roustit et al., 2009; Williams, 2003).
For the analysis that follows, we selected these theoretical perspectives to serve as lenses through which to understand women’s choices as they relate to managing a violent relationship. Each addresses complementary aspects of women’s lives that are inclusive of but not limited to individual characteristics that impact decision making. In applying these frameworks, we are not building new theory, per se, but aim to illustrate the importance of adopting comprehensive perspectives that acknowledge the complexities surrounding decision making, especially within the context of relationship violence. We also show how these theories inform one another, affirming the importance of applying multiple approaches to both research and intervention design and analysis (see Figure 1).

Relationship between intersectionality, the social-ecological framework, and life-course theory as they relate to abused women’s decision making.
Each of these theories sheds light on a different aspect of women’s decision making in the face of partner violence. By illustrating how structural factors such as poverty, sexism, disability, and housing instability intersect and accumulate across women’s lives to increase their vulnerability to violence and circumscribe their responses to it, we will highlight the need for intervention and prevention strategies that go beyond individual-level approaches and address the comprehensive needs of many women and their families.
Method
This study is based on data obtained from a larger study in collaboration with other community partners (Daoud, Smylie, Urquia, Allan, & O’Campo, 2013). Briefly, the study aims to better understand the interrelationships between IPV, housing stability, and health among women who experienced IPV. The study was conducted in five urban and nonurban regions of Ontario, Canada, in 2010-2011, after ethics approval by the appropriate oversight board.
Purposeful sampling method was used for recruitment of women and it was facilitated by local women’s organizations and networks such as women’s centers, social housing, and social services institutions. Flyers describing the study were distributed to potential participants by the staff of these organizations. Interested women were directed to call the study telephone line and were screened for eligibility to participate. Eligibility criteria included age 25 to 60 years, experienced domestic violence within 5 years prior to enrollment, English speaking, and residing in one of three housing types (social, transitional, or market housing). A total of 69 women called into the study team. Of those, 28 were not eligible, and 41 were eligible and agreed to participate. Most participants resided in social housing designated for low-income women who are victims of violence.
Women were interviewed by trained female interviewers using a semistructured questionnaire after signing an informed consent form. Interviews were conducted in women’s homes in a private area for confidentiality and security. Each interview lasted between 60 and 90 min and was audio recorded and transcribed verbatim. Participants received a US$50 honorarium.
Our semistructured questionnaire focused on women’s experiences of housing instability; partner violence, and physical and mental health during periods of housing instability and stability; coping strategies; and utilization of health care and social services. We did not provide women with definitions of any of these concepts; rather, we were interested in hearing women’s own perceptions of how their experiences related to these terms. Women often shared their stories about IPV, housing instability, and health in chronological order in relation to the periods of residing with their abusive partner (before leaving, during the time living with the abusive partner, and after leaving) (Daoud et al., 2016; O’Campo, Daoud, Hamilton-Wright, & Dunn, 2016).
Analysis
Our analysis was conducted concurrently with the data collection. First, we performed open coding by which each research team members separately reviewed selected transcripts, labeled passages, and highlighted themes relevant to the main questions of the study. These codes and the reasoning surrounding their selection were discussed among team members, and general agreement about the types of information captured by specific codes was reached. Because of the large number of interviews and amount of data collected, all team members did not review the subsequent transcripts in their entirety, and this initial process formed the basis on which the remainder of transcripts were analyzed and new codes identified. Using NVivo 9.2, we compiled representative quotations to describe emergent themes and patterns for more focused or in-depth analysis for specific research questions of interest. For example, we explored housing instability and health, IPV and housing instability, IPV and health, and childhood experiences and IPV (Daoud et al., 2016; Minh et al., 2013; O’Campo et al., 2016). Emergent codes and themes formed the basis of a discussion about data quality.
Using the existing coding structure, the lead author for this analysis (AV) reviewed transcript sections that were identified with codes that related to women’s decision-making processes, paying particular attention to examples with multiple or overlapping codes related to this concept (e.g., she looked closely at passages coded as “safety,” “decision to leave/not leave,” “coping,” etc.). As additional themes and patterns related to decision making emerged, the authors discussed the findings and agreed on the theoretical framework for this article.
Among the 41 participants’ transcripts, 13 women spoke in detail about decisions they made related to violent partners, as evidenced by the number of times their transcripts contained coded passages related to decision making and by the depth of the description they provided surrounding to these events. Entire transcripts for each of these participants were carefully reviewed and reanalyzed, and additional key events and experiences were noted. Because the interviews were conducted using a semistructured questionnaire, our ability to construct in-depth ethnographic descriptions of women’s relationship and life trajectories was limited by the amount of detail women chose to share when describing their experiences as they related to specific questions about violence, housing, and health. The case studies presented below were chosen because of the depth of information they provided, and because they most clearly illustrated the relationships between the most common structural inequities mentioned across the interviews (poverty, a lack of safe or stable housing, ill health, and disability), the violence in women’s lives, and how those forces impacted the decisions they made about their relationships.
Findings
Case Studies
Noreen
Noreen was in her 30s when we spoke. Her first experience with domestic violence was when she was 21, after moving in with her first husband, who she admitted had drug and alcohol addictions. He was physically and sexually abusive to her and threw her out of the house when she was diagnosed with cervical cancer. After her subsequent (more positive) relationship ended because of bad timing, she met her next partner through an advertisement looking for a singer for his band. At first, they seemed to be a good pair, but once they had children, the relationship deteriorated.
This partner exerted control over her in a number of ways: He disciplined the kids in a manner she found inappropriate, he refused to let her talk with old friends, he put her down, and he prevented her from receiving medical attention. At times, Noreen was unable to care for their children but received little assistance from her partner; he refused to visit her in the hospital after complications left her hospitalized following her second child’s birth. She blamed her in-laws (who were also his employers) for some of his behaviors, saying they “encouraged the abuse” by threatening to fire him if he took time off. After a major car accident, one of their daughters required surgery and Noreen was badly burned, unable to even pick up their other child much less take care of her; when her husband stayed home the following day to help, she said he was nearly fired.
Noreen was often in pain or tired from illness or as a result of her injuries, and felt trapped in the relationship, but told herself to “suck it up.”
I felt like, you know, here I am, so many medical problems. I can’t seem to get any relief. At least sometimes, when he is here, he might change a diaper or he might watch the kids for 20 minutes so that I can take a bath . . . it’s better than nothing.
She finally reached the end after stepping in front of him when he was upset at their daughter and he threw her across the kitchen. After that, she said, she was done. She was so frightened that something would happen that she contacted the local child protection agency to enlist them in her safety plan. “It’s not that I don’t want my kids and wouldn’t want my kids, but if I’m sick in [the] hospital and can’t take care of them, I’d rather a foster family have them than him, you know?”
At first, she took the kids to stay with her parents, but left when she realized that none of them were safe there. She was denied an emergency protection order because the judge did not find the violence serious enough, and with no court order preventing him from having contact with the kids, she felt very vulnerable. He hacked into her computer and accessed her appointment schedule and address book, stalking her and threatening to take the children. Eventually, she took her daughters and hid in a battered women’s shelter.
Lori-Anne
At the time of the interview, Lori-Anne was in her late 20s with a high school education. She grew up in a small town in Ontario, and made ends meet by qualifying for provincially subsidized housing units until her income exceeded the maximum allowed to receive aid. Unable to afford housing on her own, she moved in with her common-law partner, and soon after, relocated to a nearby city, where he said he had a job. This was Lori-Anne’s first time living in an urban area and she felt isolated from the people she knew. Moreover, his “job” turned out to be dealing drugs, which put her in harm’s way. Although she had previously been able to ignore the warning signs that the relationship was not healthy, “things started getting really ugly,” after they moved. “I found drugs in the house,” she said, and “people would show up at all hours of the night . . . We would argue a lot and it would get carried away. It would go too far, and obviously I did not feel that I was safe.” When they fought, he would restrain her or try to keep her from leaving, and she admits that there were times when she responded with violence, as well. He put her down, harassed her at work, and threw her more than once.
She thought about leaving, and although she found work, it was not enough to live on alone.
It’s the middle of the night and I’m like, “So, if I work at this grocery store, minimum wage, and I work seven days a week and never get sick for the whole year, I can pay for rent, but I won’t be able to eat.” And I’m like, “That’s really not going to be working.”
She ran through several ideas in her head, but I just didn’t find an escape route. I thought, Okay, I will need to do what it is that needs to [be] done to have a roof over my head and try to think that I’m safe, even though it’s not the life to be living.
She told herself that she might be overreacting: “without black eyes or missing teeth, it’s like, well is it really abuse?” She was also concerned about the safety of her cat, who she described as her “best buddy,” a companion who had been with her through it all and she was unwilling to leave it behind with her boyfriend.
She knew she needed to move out, but she was scared. After her mother found a veterinarian who boarded the pets of women who were leaving violent relationships, she packed up and moved to a shelter. While there, she contacted the police “to let [her partner] know I was serious,” but she was terrified that he would retaliate if she pursued any formal charges against him. The police already knew about her boyfriend, and some of the information they shared reinforced her decision to leave. Even then, however, her ordeal was not over. An altercation caused her to lose her job in a grocery store, and every time she went to the bus station behind her house, he was waiting. She began cutting herself, afraid that if she did not hurt herself, she would hurt someone else or end up in trouble with the police. “I think one of the main reasons that I was so nervous about leaving situations like that is that once you make that decision to escape, it’s not over from there.”
Claudia
Claudia was in her late 40s, living with her teenage daughter in a town along the Canadian–U.S. border. She had a history of physical and emotional health problems (related in part to a series of illnesses and automobile accidents she suffered in the past decade), and was unable to work, although previously she had been a lab technician. Separated from her husband of over 20 years, she described starting out as a middle-class couple; a nice house in a suburban area with two incomes, until they faced financial setbacks, especially once she became ill. They spent their savings, claimed bankruptcy, lost their house, and even moved in with her mother for a time. She had to undergo a series of surgeries, including a botched operation that left her addicted to pain killers.
Her husband was both verbally and physically violent, so much so that he had three assault charges pending against him. Rather than incarcerate him, however, the judge issued a 1-year restraining order after Claudia begged the court to allow him to remain free because he had a good job and she needed the support.
So the judge said to my husband, he said, “you know, you must have wife who loves you because any other wife would have just let you [go] to jail and rot because that’s what I was going to do to you.” But I needed him to work ’cause I needed money, you know?
Looking back, she wishes that she had ended the relationship then. But they got back together, and the abuse continued for several years. Finally, she said, “I snapped.” She went to a shelter and told the worker that either she or her husband was going to end up dead, and she wanted to know her options. They told her to pack her bags and come with her daughter to the shelter. She refused, not wanting to disrupt her daughter’s school and relocate her to a shelter. The shelter employee helped her complete the paperwork for subsidized housing and other services, but she returned home, remaining silent about her plans and biding her time. She declined the first apartment she was offered, which was not in a part of town where she felt she and her daughter could be safe. Finally, a unit became available in a better area and she took it.
I sorta had to tell [my husband] that I was just moving here for the interim . . . because I was afraid that if he woulda known that I was really coming here to stay and live here, I don’t think I would have made it out of the house alive.
Although they were not living together, she and her husband were still married when she was interviewed, and she had not decided what to do.
Jessica
In her 50s when she was interviewed, Jessica grew up “very poor” in a small city in central Ontario. Her father was an alcoholic who abused her mother, and at age 7, she was sent to live in an orphanage and foster homes when her parents separated and her mother was unable to care for her. She returned to her mother when she was 12, and by 14, she dropped out of school, leaving home and taking a job that paid barely more than room and board. She became involved with a man in his 20s, and although her mother threatened to have him arrested for statutory rape, she said he provided her with more stability than anyone else ever had. When her boss tried to rape her, this boyfriend protected her, and she lived with him until she was 17; in hindsight, she realizes that she broke up with him because she found it difficult to believe that she was loved.
For a short time, she lived with her father, but she was haunted by memories of being sexually abused as a child and she left, moving from place to place and job to job. She thinks that her childhood exposure to her parent’s violence left her with posttraumatic stress disorder, and after several years of stability (she had a child, a partner, and job), she slipped into drug addiction. She did not discuss the decade during which she was an addict, but said that after 2 years of sobriety, she became involved with a man who was also in recovery.
In spite of feeling healthier than she had for most of her life, she found herself in an abusive relationship, which she swore would never happen to her.
“I was always telling [her girlfriends] “don’t let them do that to you, don’t let him treat you like that, that’s not right.” . . . but all of a sudden there I am and it’s a whole different ball game, you know?
She remained in the relationship for almost 3 years.
I didn’t want to tell my friends, you know what I mean? I am in recovery, I’ve got 2 years clean, I mean, I am healthier and my life is better than it’s ever been and then all of a sudden there I am, where my mom and dad were way back when. Meanwhile, I went through 10 years of drug using, I never got hit by a guy, never. Now I am clean and sober and I am being a nicer person than I have ever been in my life and this is what I get, you know. So, to tell you the truth, I almost ended up going back out [and using] again, eh? Like giving it all up, throwing in the towel and whatever.
Eventually, she talked with her sponsor, who had her call a domestic violence hotline. The third time her partner assaulted her, she contacted the police, and he was incarcerated for several weeks. In the end, she did not pursue charges; neither, however, did she return (despite the fact that she “really loved him” and he clearly wanted her to return). She wanted to get on with her life and her recovery, but it was a difficult time, and she described the onset of panic attacks and anxiety that remained with her for over a year.
Looking Beyond the Individual: Using a Social-Ecological Framework
Across this study, the women we spoke with described how the decisions they made about their relationships—including whether to enter them, whether to stay, or whether to leave—were influenced by a variety of socioecological factors. Certainly, each woman mentioned a variety of individual-level characteristics that impacted their behaviors across the span of their relationships, such as choosing to get married, believing they needed to remain in their relationships because it was the right thing to do, falling ill and not being able to care for their children, and realizing that their partner’s behaviors were becoming lethal and they needed to leave. Yet, as central as these factors were, the women also spoke about the critical role that elements across the larger socioecological spectrum played in their lives. Friends and family, service providers, community organizations, and even social policies influenced the decisions they made. For example, Lori-Anne believed her partner when he lied about having a good job in the city; she not only moved in with him, but she left her small town, friends, and family; later, the loss of that support left her feeling “segregated” and alone, until her mother stepped in and helped her to plan her escape. She also found support from organizations in the larger community that enabled her to act on that decision, from the vet who took in her cat to the police who validated her decision to move out and the shelter that took her in.
Noreen, as well, found support from the child welfare agency workers and from the system itself when she decided that foster care would be a better alternative for her children than living with their father if she became unable to care for them; this is particularly interesting in that the child protection system is not designed as a battered women’s service, and has frequently been at odds with domestic violence advocates and victims (Findlater & Kelly, 1999). Yet because of her situation, Noreen saw few alternatives, and had the skills and ability to negotiate with the organization to obtain the support she needed. In addition, many of the women we interviewed across the larger study benefitted from a provincial-level social policy that moves victims of domestic violence to the top of the waiting list for public or social housing assistance, so rather than having to wait years for subsidized apartment to become available—a period during which time many women return to abusive partners (O’Campo et al., 2016) —they were able to access affordable housing more quickly (SPP Research Task Force Report, 2012).
Multiple factors across the social-ecological spectrum worked against women, as well, actively curtailing their options and increasing their vulnerability. Some women had little social support, were unable to access necessary services, and found that the social safety required to live safely was just not there. For example, a judge denied Noreen’s request for a protection order, leaving her and her children susceptible to her husband’s actions. Likewise, Lori-Anne moved in with her partner only after learning that she earned too much to maintain her rent subsidy (yet too little to rent an apartment on her own).
Expanding Our Perspective: Applying Intersectional and Life-Course Theoretical Orientations
As useful as the social-ecological framework is for illustrating how multiple types of factors shape women’s decision making, the application of larger, systemic-level theories such as intersectionality and life-course models suggests that we must look beyond the specific factors themselves and examine how they interact with one another and across women’s lives in systematic and structural ways. For many of the women we spoke with, inequities such as poverty, gender, unaffordable housing, disability, and other cultural and systematic conditions acted in concert with one another and across ecological levels. These forces increased women’s vulnerability to abuse, reinforced the violence and control exerted by their partners, and limited their options for coping with it.
The stories that women shared illustrated the impact that a lifetime of unstable housing, family violence and trauma, physical and mental health problems, and poverty had on the choices they faced later (Minh et al., 2013). Jessica’s story, which echoes that of other women in our study, illustrates the importance of looking at the timing and frequency of traumatic or stressful events, and at how multiple social inequities overlap and combine to influence women’s trajectories. As a child, Jessica witnessed her father beating her mother, and she reflected on the impact that had on her behaviors (such as the relationship between cleaning her mother’s blood to her subsequent bad “attitude” in later foster homes). Before she was 18, she was sexually abused, institutionalized and placed in foster care, nearly raped on multiple occasions, and dropped out of school; without question, these events impacted her developmental pathway. Commenting on why she broke up with the boyfriend who had been one of the few “secure” influences in her life, she admitted that, because of her past, she had a hard time knowing when to believe someone loved her (Minh et al., 2013). Although her life had some stable periods as well, her early experiences with multiple and overlapping incidents of trauma, poverty, violence, and instability may have contributed to her subsequent substance abuse and to her experience of betrayal when the man she related to and trusted because they were both in recovery began to hurt her.
The Intersection Between Inequities and Partner Violence
In addition to the influences of childhood and early life experiences, women described how the intersection of multiple structural inequities acted to circumscribe their choices across the trajectory of their violent relationships. These forces often had tremendous impact at three key time points: just after becoming involved with a partner, when women were living with a partner and realized the relationship was violent, and when women began to wrestle with whether to stay or leave. At each of these phases, structural-level inequities such as poverty, illness, or disability, and a lack of systemic support or safety net for women struggling with violent relationships often drove the decisions women made, frequently resulting in women becoming more dependent on the partner and reinforcing the violence and control he exerted over her. When these inequities intersected with other events that increased women’s vulnerability, the consequences were often greater. Theories that explicitly acknowledge these forces enable us to better understand both the problem and its solutions.
When becoming involved with a partner
Several women described how their decision to move in with a partner was directly related to poverty and housing instability, and how this situation then increased their vulnerability to abuse and coercion. When Lori-Anne lost her housing subsidy and moved in with her boyfriend, she ended up far away from friends and family and suddenly realized she had no place to go when she needed it. Another participant described inviting her partner (who was a carpenter) to move in so he could help pay the mortgage and renovate her kitchen—in the end, he ripped out her kitchen and became abusive, leaving her unable to force him to leave until he finished it. In both cases, unstable housing and few options reinforced the violence that was happening in women’s lives.
When living with an abusive partner
Although popular images of battered women often involve partners who threaten their lives if they leave, a more common experience among women in our study was the sense of feeling stuck and unable to change their situation because their options were limited by larger social and structural forces: having no affordable place to go, making too little to live on if they leave, and having few—if any—options for public assistance in a timely manner. For Lori-Anne, what followed after moving in with (and relocating for) her partner was a sense of feeling trapped; while she wanted to leave, she could not fathom how she could live on her income as a grocery store clerk. The timing of these events is critical to consider as well. Noreen, as a mother of two young daughters as well as someone struggling a significant level of disability and a limited income, was particularly vulnerable to the combined influences of controlling violence and poverty.
You know, I kinda think to myself, “Well, alright. I’m not allowed to have friends. I’m not allowed to go out, um, expected to look a certain way. I’m expected to cook, clean all the time. I’m expected, I’m expected, I’m expected . . . ” And, I guess, in some way, I kinda rationalized it. “Well, my kids are benefitting, so that makes it OK.” . . . And I’m throwing up as I’m cooking, and you know I think “I really don’t want to do this” and I would get bummed out, but I try to think, “Well, you know, my kids will get fed tonight. There’s kids out there that don’t, at least they’re getting fed, so [I should] suck it up.”
For a large number of women in our study, poor physical and mental health, disabilities, and poverty overlapped with one another to reinforce abusive situations that were already limited in terms of options. Both Noreen and Claudia faced physical limitations that were exacerbated by violence and low incomes—which, in turn, increased their financial vulnerability and dependence. Claudia’s situation, however, became potentially direr when she found herself begging a judge to not incarcerate her husband after his third violent attack because if he went to jail, she would be unable to pay the bills. Rather than finding an alternative approach that could keep her safe and allow her to feed her child, the judge granted her request, issuing a protection order while telling her husband how lucky he was to have a wife who loved him as much as she did. Clearly, not only was the lack of a safety net for abused women an issue, but a systemic shortcoming in how the court understood Claudia’s needs as a battered woman (as expressed through the judge’s assumption that love and not necessity drove her to make her request) illustrates the types of class and gender biases that can increase the level of danger for victims.
When leaving an abusive partner
The literature substantiates the increased danger many women face just after deciding to leave a violent relationship (J. C. Campbell et al., 2003; Sev’er, 1997). For some of the women in our study, this danger was increased several fold due to the additional limitations that go with being poor, unhealthy, and the victim of partner—rather than stranger—violence. For example, Claudia was forced to choose between staying with an abusive partner or moving with her daughter to an apartment that felt even less safe, given the level of community deprivation and crime in the neighborhood (she chose to remain until a safer option opened up). After Noreen left her husband, her options for safe housing were made even more limited when her request for a protection order was denied in spite of her husband’s documented history of violence: “I guess around here,” she was told by a court official, “unless he’s raping you guys or actually breaking bones . . . the judge won’t grant anything.”
Discussion
Within Westernized communities, women are often judged harshly by those around them for not removing themselves and their children from violent partners, even when the only options available to them may not seem much better: shelters located in unsafe neighborhoods, unemployment, and/or the possible removal of their children from their care. Although the popular emphasis is beginning to shift from asking the rhetorical question “why doesn’t she just leave?” to instead identifying the forces that may inhibit women’s actions, when analyzing or discussing women’s decisions about violent relationships, researchers, policy makers, and service providers continue to approach the topic from the perspective of individual women facing individual circumstances that make it harder or easier for them to leave a relationship. Although there has been a substantive amount of research done on select non-individual-level forces, such as the importance of social support (Coker et al., 2002; Goodman, Dutton, Vankos, & Weinfurt, 2005; Mitchell et al., 2006) and the necessity of accessible battered women’s services (Mitchell et al., 2006), few researchers have engaged in a comprehensive, ecological analysis of abused women’s actions (Carlson, 1984), and fewer still have considered the problem of women’s decision-making processes through a truly structural lens that incorporates the multitude of ecological forces that overlap and impact behaviors as well as the periods of a woman’s life in which events occur.
It should be noted that the women in this study generally described the factors that influenced their relationship decisions as being at the individual level (e.g., as having no place to go or being isolated from family or services), and these factors very much reflected the types of factors that appear across the literature (J. C. Campbell, Rose, Kub, & Nedd, 1998). But ecological, intersectional, and life-course theoretical frameworks suggest that to fully understand these factors, we need to look at a higher level, toward the social and structural forces that shape the environments in which women live and act. For example, not having a place to stay, being dependent on a partner to pay the rent, and living far away from social support because it was a cheaper location all point to larger factors such as insufficient affordable housing and failed public/social housing policies that do not meet the needs of women who experience abuse and their children (Purvin, 2007).
The experiences of the women whose case studies we shared—and many others whose stories we could not include in full—show the importance of taking this type of comprehensive approach. Women’s decisions are clearly influenced by factors outside of their individual being: They are shaped by the behaviors of their friends and family, by the community and organizational responses, and by social and political forces, both historical and current. These factors, in turn, are directly impacted by structural influences such as the ones we described here (including poverty and barriers related to disability and poor health) as well as others that were less prominent within our sample but which have been documented elsewhere, including racism and other inequitable distributions of power and influence. And finally, women’s experiences with these factors and influences early in their lives (such as Jessica’s experience with institutionalization and abuse) or at certain vulnerable times (e.g., when children are young or when women are struggling with illness) directly shape the trajectory of their lives later, such as whether they finish school or can earn a living on their own.
These results correspond with findings that show relationships between cumulative experiences with violence across the life span and negative outcomes (R. Campbell, Greeson, Bybee, & Raja, 2008; Davies et al., 2015; Roustit et al., 2009). For example, in a similar population of women who had recently left their abusers, Davies et al. (2015) found that women who reported severe and long-term abuse profiles (including child both child and partner abuse) had the lowest levels of resources and significantly worse health profiles compared with women who experienced some IPV but little to no child abuse. Likewise, Campbell and colleagues describe high levels of co-occurring child sexual abuse, adult sexual abuse, and adult IPV among African American women both in their own work with military veterans and that of others, suggesting that while African American female military veterans appear to experience higher levels of violence compared with their Caucasian counterparts, “their experiences are not markedly different from other low-income African American women” (R. Campbell et al., 2008, pp. 201-202).
By applying these theories to explore women’s decision making in the face of violence, we are going beyond highlighting factors that might influence women’s behaviors and decisions, but instead aim to identify the larger trajectories and mechanisms that drive women’s behaviors. For example, it would be easy to read Lori-Anne’s story as one in which her decision to move in with a drug-dealing boyfriend, her inability to see what was happening as “domestic violence,” her refusal to abandon her pet, and her low-wage skill set led her to feel stuck in an unsafe and unhealthy relationship. What this would miss, however, is any consideration of why and how those choices became ones that she faced. Before she moved in with her boyfriend, she received governmental assistance with her housing costs; yet when her employment status improved, she found herself making too much money to qualify for any assistance, yet not enough to actually live independently, and her situation became even worse after moving to a larger city and into a dangerous living situation. Intersectional theory suggests that the overlap of structural factors including inadequate social policies, a lack of safe housing and employment options, as well as the violence happening in both her relationship and community directly influenced Lori-Anne’s choices and thus her decision to remain in the relationship for as long as she did.
Limitations
Two of the limitations of this type of research appear antithetical to one another at first glance. On one hand, we run the risk of ecological fallacy, or inferring attributes about our individual participants based on the findings from population-level research (Neuman, 2003); for example, just because population-level findings suggest an association between experiences from childhood and later victimization does not mean we can claim that these experiences caused the women in our study to behave in a certain manner. Likewise, we must be careful not to engage in reductionism, or assume that the stories told by our sample of women are reflective of and generalizable to all women who are victims of partner violence (Neuman, 2003). To address these potential biases, we are not making direct claims of causality, but rather suggesting that by using these theories that focus on structural-level influences as a lens through which to understand women’s decision-making processes, we create a more complete understanding of the context in which women are living and making choices about their welfare and that of their children.
We also must acknowledge that our data are self-reported, and that because the focus of the original research was not specifically on women’s decisions about staying or leaving, the stories that women shared were ones that emerged naturally from our conversations about their experiences with housing instability, health, and violence, and we do not have data about these experiences from all participants. We also recognize that our small sample size further limits the generalizability of our conclusions. However, the descriptions that these women provided about their histories and decision-making processes echo what has been reported by others who have interviewed and recorded the experiences of battered women (M. A. Anderson et al., 2003; J. C. Campbell et al., 1998; Purvin, 2007).
Given these limitations, we view these findings as illustrative of the direction that future research on partner violence needs to go, particularly if we hope to expand on our understanding of the mechanisms that propel behavior related to violence. By applying these structural theories to our analysis of women’s decision making, we show the influence that systematic inequities have on women’s actions and reactions to violence. By using these and similar theories to guide both our research questions and design, we can move toward a deeper understanding of the how these inequities drive violence and violent behaviors and ultimately lead us in the direction of more effective intervention and prevention approaches.
Finally, we argue that in this political and economic climate of shrinking levels of support for social welfare programs and ongoing vilification of women who do not make the decisions that popular opinion would have them make (i.e., they do not leave abusive partners when we think they should), it is particularly important to pay close attention to the numerous ways in which structural forces shape the options women have. Previous papers from this study have examined the trajectories between housing instability, violence, and women’s health, as well as those that link early life experiences to criminality and IPV; in this article, we look specifically at how these factors overlap to impact women’s options for managing the safety of themselves and their children. By recognizing the synergistic effect of these forces on women’s safety, this article speaks to the need for service providers, program developers, and policy makers to implement comprehensive intervention and prevention strategies that acknowledge and address these complex interactions.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by an Operating Grant from the Canadian Institutes of Health Research, #97951.
