The Use of Structural Intersectionality as a Method to Analyze How the Domestic Violence Civil Protective Order Process Replicates Inequality

Mandatory Arrest Policies

Given that the enforcement mechanism for protective orders is arrest, it is critical to understand domestic violence arrest policies in the United States. There are three types of arrest policies—mandatory arrest, proarrest, and discretionary arrest (Durfee, 2012; Goodmark, 2018; Hirschel et al., 2007). Mandatory arrest is the most restrictive and requires police officers to make an arrest in cases of domestic violence or protective order violations. Proarrest policies encourage arrest as the “preferred option” but do not require it. Discretionary arrest is the most flexible policy, where officers can make a warrantless arrest but are not encouraged or required to do so. An individual jurisdiction in a state may adopt a more restrictive, but not less restrictive, arrest policy than the state. As of 2016, 18 states and the District of Columbia (37%) had mandatory arrest, five states (10%) had proarrest, and 27 states (53%) had discretionary arrest policies.

Mandatory arrest policies for domestic violence were first implemented in Oregon in 1977 and were adopted nationwide after the results of the Minneapolis Domestic Violence Experiment (MDVE) indicated that the arrest of the abuser decreased the likelihood of further acts of abuse (Sherman & Berk, 1984). It is important to note that arrest was only compared with two alternatives: separation of the parties and “on-the-spot marriage counseling” by officers (Sherman, 1992, p. 20). Arrest was not tested against other types of responses. In addition, the benefit of the decreased likelihood of subsequent violence may indicate displacement rather than deterrence; after arrest, abusers may have simply “moved on to abuse new victims” (Sherman, 1992, p. 13). Since their implementation, mandatory arrest policies have dramatically changed police responses to domestic violence, and arrest rates have risen from 7–15% to 33–57% (Dugan, 2003; Durfee & Fetzer, 2016; Eitle, 2005; Hirschel et al., 2007; Jones & Belknap, 1999; Mignon & Holmes, 1995).

Any benefits of arrest should be weighed against the possible damage caused by police responses, mandatory arrest policies, and criminal prosecutions for violations of orders. Subsequent reanalyses of the MDVE and replications in other cities showed that arrest actually “had an escalation effect” for male offenders who were Black, unemployed, unmarried, or who had less than a high school education (Sherman, 1992). Mandatory arrest policies also disproportionally and negatively affect women and girls, especially Black women and girls (Durfee, 2012, 2016; Hirschel et al., 2007; Goodmark, 2018; Richie, 2012). Arrest rates of women have increased faster than men’s (DeLeon-Granados et al., 2006), and women arrested for domestic violence are four times more likely than men to have been dually arrested with their abusers (Henning & Renauer, 2005; see also Melton & Belknap, 2003). This is similar to other mandatory policies like school zero-tolerance policies, which disproportionally affect African American girls defending themselves from sexual violence and often result in their expulsion (J. Miller, 2008). There is also an increasing problem of “victim defendants”—victims who are mistakenly arrested instead of their abuser (Crager et al., 2003; see also S. L. Miller, 2005). According to Connie Burk (2008) of the NW Network of Bi, Trans, Lesbian and Gay Survivors of Abuse, over 1 year, 50% of survivors with whom they worked, who came into contact with the police, were arrested. In practice, neither protective orders nor mandatory arrests have been “victim-friendly” for many survivors, including multiply marginalized ones.

Historical Context: The Criminalization of Domestic Violence

Protective orders and mandatory arrest laws were implemented in a complex, contested national context from the 1970s through the 1990s. During this time, political rhetoric and public discourse about “crime” and poverty became increasingly neoconservative and neoliberal. In this context, the battered women’s movement made strategic choices to (a) destigmatize domestic violence victimization; (b) ignite a national conversation about domestic violence (which at the time was “hidden” and seen as a “private” or “family” matter); (c) leverage increased government funding for existing formal and informal resources for survivors; (d) force the police and criminal legal system to take domestic violence seriously; (e) create legal protections and rights for survivors; and (f) provide an alternative way for survivors to access legal protections without using the criminal legal system through the creation of protective orders.

At the time, the criminalization of domestic violence was hailed as a “symbolic” victory for women (Chesney-Lind, 2002) that would “force police officers to take intimate partner violence seriously” (Richie, 2012, p. 82). Criminalization and the “institution of mandatory practices, specifically arrest, was perceived as the first step in a process that would lead to a seismic cultural change by inscribing women’s empowerment, individual and collective accountability, and conceptions of equality” (Miccio, 2005, p. 265).

Notably, mandatory arrest was the first step in helping “victims” (“women”) as an essentialized group. Survivors from multiply marginalized communities, especially Black communities, were already suffering from the effects of neoconservative crime policies like the Anti-Drug Abuse Act of 1986, which led to the disproportionate incarceration of Black men and women (Albonetti, 2016). Overall, between 1970 and 2017, the incarceration rate of women increased by more than 750% (The Sentencing Project, 2019). As of 2017, Black women were incarcerated at twice the rate of White women, and Hispanic women were incarcerated at 1.3 times the rate of White women (rates for Native women were not reported). Levy-Pounds (2007) argues that the incarceration of poor Black mothers has lasting impacts on their children, who are then “prone to repeat the same cycles of poverty and incarceration of their parents” (p. 325). Unsurprisingly, the same pattern of disproportionate institutionalization (incarceration) is present in the juvenile justice system as well (The Sentencing Project, 2019). Thus, the criminalization of domestic violence was only one facet of the many disruptions and damage that the criminal legal system had, and continues to have, on the lives of multiply marginalized individuals and their communities and cannot be considered separately from them.

During this time, the battered women’s movement also moved away from calls for fundamental structural changes such as addressing employment and the wage gap, workplace discrimination, inadequate child care and social financial support, and barriers to education. Instead, it shifted its focus to reforming institutions through the professionalization and institutionalization of domestic violence “service provision,” creating criminal legal statutes, and calling for harsher criminal penalties for abuse (Bumiller, 2008; Coker, 2004; Goodmark, 2018; INCITE!, 2005; Kim, 2015; Richie, 2005, 2012; Schneider, 2000; Smith, 2005). Survivors and activists who had formed the first shelters and support groups in private homes were replaced by caseworkers, therapists, and trained staff. Domestic violence education and outreach also shifted to training for police, prosecutors, and other legal actors.

At several points during this process, advocates and activists in communities of color and members of the antiviolence movement raised concerns about these strategic choices (Richie, 2012). However, their concerns about criminalization were in many ways disregarded. Instead, there was a widespread adoption by politicians and policymakers of the “gendered crime frame,” which emphasizes how domestic violence and other acts of violence against women are “crimes” best addressed by strengthening law enforcement and criminal responses (Whittier, 2016). This discourse often precludes intersectional understandings of and solutions to domestic violence and is consistent with a “carceral feminist” approach with an expanded role for the state and the criminal legal system (Whittier, 2016). The result was a system that was based on the idea that domestic violence could be solved by working “inside the system rather than against it” by criminalizing domestic violence and relying on the police and the criminal legal system for enforcement (Richie, 2012, p. 97). But these changes “backfired . . . causing further harm to women of color, and to queer, immigrant, and native women” (INCITE!, 2005, p. 107).

This national shift to criminalization was solidified by the passage of the 1994 Violence Against Women Act (VAWA), which diverted financial support to criminal responses and away from social services, shelters, and other resources “that would render [battered women] less vulnerable such as transportation, or education and job training” (Coker, 2001, p. 804; see also Dragiewicz, 2008; Richie, 2012). The diversion of VAWA funding has increased over time. Although the total amount of federal funding allocated through VAWA between 1994 and 2013 nearly doubled, the share of federal funding for criminal legal responses grew from 62–85%. This has resulted in a reduction of the total amount of funds provided to domestic violence social service providers (Messing et al., 2015).

While VAWA and subsequent reauthorizations did result in the expansion of culturally specific services and resources for multiply marginalized survivors (especially immigrant and Native survivors), it also forced “survivors to navigate a discriminatory criminal legal system” (Law, 2014, as cited in Aizeki, 2019) where other, nonlegal or noncriminal options for safety were no longer considered (Goodmark, 2004). For example, one study in the 1990s found that officers who arrested abusers or took other formal actions were less likely to call shelters for victims (Jones & Belknap, 1999).

Through the heavy reliance on mandatory arrest policies, no-drop prosecution policies, “failure to protect” laws (Bierria & Lenz, 2019), and protective orders, surviving domestic violence in the United States has become a perilous venture, with multiply marginalized survivors increasingly facing both “collateral” and “unintended” consequences when they come into contact with the legal system or formal service providers (Chesney-Lind, 2002; Mauer & Chesney-Lind, 2003). The consequences of the criminalization of domestic violence are clear: Criminalization increases the vulnerabilities of survivors, putting them at greater risk whether they use the legal system or not.

Perhaps most salient to this analysis is how discourse about domestic violence victimization changed during this period. To destigmatize victimization, mobilize political support, and leverage governmental funding, the battered women’s movement adopted the language of “universal risk” (or the “universal woman” and “every woman” discourse), which conveyed the idea that all women are equally likely to experience domestic violence (Crenshaw, 1991; Kim, 2015; Richie, 2012). This discourse was then widely adopted by policy makers, advocates, and the general public. For example, in 1991, Senator David Cohen argued that we should care about domestic violence and rape because “it is our mothers, wives, daughters, sisters, friends, neighbors, and coworkers who are being victimized” (Crenshaw, 1991, p. 1260). While this discourse worked to erase some of the stigma of victimization, it also erased structural factors that shape many survivors’ experiences and choices, including the role of “intersections of systems of power (e.g., race, class, gender, and sexual orientation) and oppression (e.g., prejudice, class stratification, gender inequality, and heterosexist bias)” (Sokoloff & Dupont, 2005, p. 43). Controlling images of “victims” were highly feminized and centered White, middle-class, physically abused women. Subsequently, domestic violence policy and funding choices during this period reflected the experiences and needs of these idealized “victims.” Crenshaw (1991) argues that

where systems of race, gender, and class domination converge, as they do in the experiences of battered women of color, intervention strategies based solely on the experiences of women who do not share the same class or race backgrounds will be of limited help to women who because of race and class face different obstacles. (p. 1246)

To identify the privilege, oppression, barriers, and problems in current civil and criminal legal responses to domestic violence, including civil protective orders and mandatory arrest policies, I advocate for the use of structural intersectionality as a method of analysis. In the next section, I briefly describe intersectionality, structural intersectionality, and intersectionality as a method.

Intersectionality

The central premise of intersectionality is that “multiple grounds of identity” operate simultaneously to shape individual experiences and resources (Crenshaw, 1989, 1991; see also Collins, 1990; Combahee River Collective, 2014; King, 1988). Intersectionality provides a theoretical and methodological framework by which one can study how social identities, locations, and axes of identity such as race, gender, and class operate as a “matrix of domination” (Collins, 1990) or “interlocking” systems of oppression (Combahee River Collective, 2014) that shapes the lived experiences of multiply marginalized individuals. The use of intersectionality has come under critique as intersectionality is often used as a “buzzword . . . providing a catchy and convenient way of expressing the author’s normative commitments” rather than a reflective engagement with intersectionality (Davis, 2008, p. 75). It is important to recognize and cite intersectionality’s history and intellectual origins; by failing to do so, scholars participate in the “whitening” of intersectionality and the erasure of the extensive body of knowledge and the contributions of women of color, especially (but not only) Black feminists (Alexander-Floyd, 2012; Bilge, 2013, 2014; Carbado, 2013; Choo & Ferree, 2010; Crenshaw, 2019; INCITE!, 2007; May, 2015). I explicitly acknowledge my debt to these scholars and activists, yet due to space constraints, I cannot engage with all their work in this article. Instead, I refer readers to recent books and articles which provide a history of intersectionality, including Baca Zinn and Zambrana (2017), Carastathis (2019), Collins (2015), Collins and Bilge (2016), Hancock (2015), May (2015), Nash (2018), and Potter (2015), whose book disrupts traditional criminological scholarship using intersectionality and provides important examples of intersectional criminological research.

Crenshaw (1991) describes three forms of intersectionality in the context of violence against Black and immigrant women—structural, political, and representational. Although she separates these three forms in the article, they are often in practice interwoven. While in this article I primarily focus on structural intersectionality, both representational (controlling images of women of color derived from misogynistic and racist stereotypes) and political intersectionality (the lack of political representation of women of color due to the exclusion of their needs and goals by both the feminist movement and the antiracist movement) are important to understanding how institutions replicate broader inequalities.

For example, representational intersectionality is critical to analyzing how the power to classify and define actions and individuals as “violent,” “abusive,” and/or “criminal” directly shapes research, policy, law, and practice on domestic violence and reflects broader dynamics of both privilege and oppression (Berns, 2004; Comack & Balfour, 2004). Unsurprisingly, it has been a key “place of struggle” for feminists and domestic violence activists (hooks, 1989, as cited by Ferraro, 2006). In the context of protective orders and mandatory arrest, representational intersectionality is reflected in the power to create, define, and enforce controlling images as to what constitutes “domestic violence”; who is a “real,” “legitimate,” or “good victim”; who is a “perceivable perpetrator”; and how images of “victims” and “perpetrators” influence and shape each other (Goodmark, 2008; MacDowell, 2013; Ptacek, 1999). These controlling images often stem from media depictions of domestic violence (Berns, 2004; Crenshaw, 1991) and are themselves examples of representational intersectionality. They can shape how survivors view themselves, the decisions they make, where they seek help, and from whom they ask for it. They also influence how others see survivors themselves—including people in positions of power who can provide access to resources, refuse to help, or even impose sanctions for coming forward (such as referrals to ICE and CPS) (Aizeki, 2019; Berns, 2004; Coker, 2001; Goodmark, 2008; Ptacek, 1999; Richie, 1996, 2012; Weissman, 2019). MacDowell (2013) recommends examining the privilege and marginalization of both the survivor and the abuser, rather than focusing on just one person. Often, “domestic violence discourse does not necessarily work well for anyone, including white women” (MacDowell, 2013, p. 545), but the privilege of some survivors and abusers gives them greater power to resist the imposition of these controlling images and associated judgments. In these ways, representational intersectionality is interwoven into my discussion of structural intersectionality and my analysis of orders.

Structural Intersectionality

At the same time that intersectionality transcends an exclusive focus on identity or mere categorization, the lived experiences of racially marginalized women and girls are shaped by a range of social and institutional practices that produce and sustain social categories and infuse them with social meanings. (Crenshaw, 2012)

As stated previously, structural intersectionality focuses on how institutional practices and policies replicate inequality. The process by which this occurs need not be intentional, and often occurs when institutions make inappropriate assumptions about the individuals who work in or come into contact with those institutions. In general, those assumptions are informed by controlling images about what “X” people are like (representational intersectionality), as well as the political power to create, influence, or resist institutional policy (political intersectionality). They generally do not reflect the actual intersectional contexts in which multiply marginalized people live and make decisions. When policies and practices are created based on these inappropriate assumptions, they center the perspectives, interests, and needs of those in power and work to further subordinate those who are multiply marginalized. In the case of domestic violence, they can increase survivors’ vulnerability to abuse and trauma, and impede their efforts to be safe. They can also have other collateral and unintended consequences.

For example, Donnelly et al. (2005) analyzed shelter services for battered women. They interviewed shelter directors and found that although directors used “colorblind” language and described their services as race-neutral, White privilege shaped both the discourse they used and the services their shelters provided. According to one director, “When women come to the shelter, the one thing I notice is that color seems to cease to be a problem and they’re just battered women” (Donnelly et al., 2005, p. 21). Shelter directors also described domestic violence using the same “every woman” and “universal risk” language that equates the risks and needs of all survivors with those of White women. Relying on essentialist language and concepts led to shelter practices that were “one size fits all,” rather than culturally specific strategies that could work to counteract multiple oppressions faced by marginalized survivors (p. 20). By enacting practices that assume all survivors need the same services (“we give them all the same services and the same opportunities,” p. 22), these shelters privileged some White women whose needs and goals were aligned with those of idealized “victim” stereotypes and created barriers to multiply marginalized women. When the directors were asked why few women of color used their services, the directors relied on racist stereotypes of women (representational intersectionality) that suggested Black, Latina, Asian, and Native women did not want to use their services but instead wanted to be supported by their own communities. This then justified the continuation of the “colorblind” shelter practices and maintained barriers to service. Had the shelters engaged in a more thoughtful review of their practices and locations, all survivors (not just multiply marginalized women) would have benefited from subsequent changes, as centering the multiply marginalized and addressing their barriers and needs facilitates access to and improves services for everyone. Donnelly et al.’s (2005) analysis is an excellent example of the ways in which institutions create unrecognized and unaddressed barriers for multiply marginalized survivors which leave them more vulnerable to violence.

Structural Intersectionality as Method

Scholars who advocate for the use of intersectionality as a method argue that many studies that claim to be intersectional instead analyze inequality through the use of single-axis/unitary (which only examine one axis, such as race) or additive/multiple models (which consider multiple axes such as race and gender but analyze them separately, then add the effects together; Bowleg, 2008; Choo & Ferree, 2010; Davis, 2008; Hancock, 2007; May, 2015). A review of the literature suggests that to be truly intersectional, analyses must contain several key elements.

First, intersectional analyses should center members of multiply marginalized or the “most marginalized” groups and include their lived experiences in the analysis (Choo & Ferree, 2010; Richie, 2012). Centering those who are the “most marginalized” will reveal barriers and vulnerabilities experienced by, and lead to changes that will benefit, everyone. Second, intersectional analyses should move beyond traditional disciplinary boundaries and methods, avoiding ideological constructions that may not reflect the lived experiences of multiply marginalized individuals. In doing so, they should avoid using deficit-based models, highlighting agency and identifying resistance (May, 2015).

Third, structural intersectional analyses should focus on inequalities that have become naturalized and “attend to systematic harm and institutionalized oppression” (May, 2015, p. 244). In addition, as institutions rarely reproduce inequalities in isolation, but instead depend on one another to sustain hierarchical relationships, it is often beneficial to study overlapping institutions, such as the civil (protective orders) and criminal legal (mandatory arrest) systems (Choo & Ferree, 2010). This also requires that researchers contextualize current inequalities using a historical lens, rather than analyzing a single point in time. Finally, multilevel models that examine both individual experiences and institutional processes can best reveal how institutions replicate broader inequalities.

Mechanisms of Inequality in the Protective Order Process: Goals and Vulnerabilities

From an institutional perspective, the goal of an order is to provide safety for the petitioner and any other protected people listed on the order. This is accomplished through the specific provisions included in the order and the enforcement of orders by the petitioner and the criminal legal system. Most frequently, this involves the physical separation of the parties and restrictions on contact between the petitioner and respondent. As Judge #9 said, “I see them as sort of a way for a [petitioner] to create a boundary between herself or himself and the defendant . . . [a] way of protecting themselves.”

As orders are currently structured, several things must happen for orders to provide petitioners with safety. First, respondents should be deterred from engaging in further abuse by the threat of arrest. If they are not deterred by that threat, the petitioner must report violations by contacting the police. If contacted, the police must respond quickly and make an arrest if there is sufficient evidence that the order was violated. As orders are enforced by the criminal legal system, prosecutors must then bring charges for violations, and judges or juries must convict and sanction respondents for violations. However, even if all that occurs, the order will only provide further safety if the respondent is deterred by the threat of arrest and additional sanctions upon release.

In interviews, judges emphasized the importance of enforcement of orders and the critical role of police and the criminal legal system in achieving the institutional goal of safety through separation. Judges believed that orders influenced police responses more than respondent behaviors. For example, Judge #1 said that orders are just “a piece of paper . . . [orders mean] that the police act much quicker when there’s an order . . . it’s more of an effect on police behavior than it is on a [respondent’s] behavior.” This is consistent with the idea that orders are more of an “enhanced response” order allowing for greater criminal penalties and for arrests in a wider variety of situations than a truly protective order providing resources to prevent reabuse (Durfee & Goodmark, 2019). It also supports the idea that orders are entwined with arrest, and that the same structural issues identified through intersectional research on arrest are salient for protective orders.

Several judges expressed that arrest and criminal prosecution were necessary for orders to more effectively provide safety. Judge #9 said that

the right circumstances [for an order] are . . . a victim or plaintiff who is ready [to separate from their abuser] . . . [and] a criminal prosecution that is kind of going hand-in-hand because the defendant has more at stake.

For this judge, the abstract idea of criminal sanctions associated with an order did not have a sufficient deterrent effect to keep petitioners safe; respondents needed to directly experience criminal consequences before they would be willing to change their behavior. The majority of the police officers interviewed concurred. Officer #51 said that police encourage all survivors to obtain orders but also caution them that

orders are only effective if there’s some accountability . . . after the order’s been served, we need to know about [violations] . . . no matter how small the violation may be . . . so that we can hold them accountable.

In contrast, while petitioners agree that their primary (but not only) goal in filing for orders was to have safety through separation from their abusers, they did not share the view that arrest was the best way to accomplish this goal. Several petitioners said that their abusers were not deterred by the order or the threat of arrest; in fact, orders often incited new forms of abuse, including legal abuse (using the legal system as a tool to abuse, harass, and intimidate survivors by filing orders protecting the abuser against the survivor, contesting petitioners’ orders, and other legal actions; see Douglas, 2018; S. L. Miller & Smolter, 2011). Petitioner #70 said her abusive partner “has a problem with authority . . . so, in a way, it was almost a challenge to him to see if he could skate around” her order. Another petitioner (#69) said that her abuser

studies my orders and finds loopholes . . . it becomes a game to him . . . he challenges the law every single time and he challenges the prosecutor. I have five or six violations I’m sitting on that I haven’t even filed because of this process because I hate going to court. Every time I do anything, he retaliates 10 times worse and I have no protection.

After she filed for a protective order, the respondent filed for one against her, filed false police reports, claimed “parental alienation,” filed for sole custody, and went to CPS to make false allegations about abuse and neglect. She was also pressured by her parenting coordinator into dropping her order so that she and the respondent could “more effectively” co-parent. In effect, what was supposed to be a “protective” order elicited more abuse that was rarely sanctioned. Things like electronic abuse and harassment, legal abuse, and institutional abuse (using other institutions like CPS to harass and intimidate the petitioner) are often not legally recognized as abuse or violations of orders. In response to these types of actions, some petitioners stopped reporting violations to the police or dropped their orders.

While it is tempting to reveal things like the race, ethnicity, age, and citizenship status of both Petitioner #69 and the respondent and then draw conclusions based on that information about how gender, race, ethnicity, age, and citizenship status operate in the protective order process, such an analysis would be quite limited and likely inaccurate. Instead, Petitioner #69’s story requires an examination of the privilege and marginalization of both the petitioner and respondent (MacDowell, 2013). To harass the petitioner in several complex ways in response to her obtaining an order, the respondent must have knowledge of how multiple institutions operate and interact and be able to draw on representational images that support his claims that she is abusive and neglectful and that he is a caring parent. He has knowledge of the law to be able to find “loopholes” in orders. He also has the resources and time to engage in these activities and appears secure that these activities will not jeopardize his legal status or result in his arrest.

In contrast, Petitioner #69 has agency and engages in resistance, but is also marginalized in several ways. Although she feels unsupported and without resources, she is able to successfully file for and receive an order; she was also able to help another survivor file for an order. She can call the police and feel safe when they respond, even if she feels disregarded by them. During her interview, because of her privilege, she does not describe being in fear of arrest, incarceration, or deportation as a consequence of her contact with the legal system. At the same time, she is unable to stop the legal and institutional abuse the respondent engages in and is unable to resist the respondent framing her as an abusive, neglectful parent. She also drops her order under pressure.

Given the patterns of privilege, agency, resistance, and marginalization observed in this brief analysis of Petitioner #69’s story, how can we use structural intersectionality to trace these patterns back to constellations of social identities or axes of inequality? The important questions to ask are who shares these privileges? Who is able to exert similar agency? Who can resist, and whose efforts at resistance are relatively ineffective? Who can leverage the power of the civil or criminal legal system to achieve their goals? Who is marginalized in these types of situations, and faces consequences such as deportation, the denial of an order, the loss of custody of children, or being arrested for domestic violence or other crimes? In short, the processes observed and described here are key, and structural intersectionality is the best way to analyze them.

While these interviews may not be representative of petitioners and respondents due to the small sample size and nonrandom sampling strategies, the processes that these petitioners and respondents face are representative of many multiply marginalized individuals. Thus, an analysis of the similarities and differences between their goals and institutional goals, as well as the ways in which their decision-making is enabled or constrained by institutional policies and procedures, can be representative of how all survivors engage with and are affected by the protection order process. This requires discovering patterns in interviews, identifying mechanisms of inequality, and drawing connections to previous qualitative and quantitative research on things like divorce, child custody and child protection services, the ability to access legal representation, police decision-making, police violence, immigration and citizenship policies for domestic violence survivors and criminal offenders, and so on, and linking them to the patterns observed in these interviews. By approaching and analyzing the interviews in this way, I avoid essentializing petitioners and respondents into the categories of demographic information that I have collected for them (by predetermining that their gender, race, ethnicity, age, relationship, and citizenship status are important) and allow for intracategorical complexity which can be fruitful for drawing connections, building coalitions, creating new responses to domestic violence, and providing better resources for survivors.

What about the second institutional assumption that petitioners will call the police to report violations? While some did call the police, not all petitioners were willing to call. One reason was having prior negative interactions with police, including the petitioner being arrested when seeking help:

I called 911 . . . I was in pajamas. . . . My husband was already ready to go to work. He works at a law firm with shirt, tie, coat, you know. Just cool, calm, and collected. And I’m hysterical. They separated us, asked us some questions . . . [I said] I had broken a picture frame of us in a moment of anger . . . next thing I know, I’m being arrested. And they took me in my pajamas [to jail] . . . I was outside the house in their car in handcuffs in the neighborhood . . . eight hours, I was in custody . . . the whole experience was horrible. And I probably will not call the city, even if I’m in another situation similar or worse . . . the experience was awful. (Petitioner #66)

Note here the importance of intersectional representations of “victims” and “abusers” in the interpretation of incidents and officer decision-making. Petitioner #66 was angry and admitted to breaking a picture frame. As discussed previously, this is inconsistent with stereotypes of “victims,” who are passive and do not engage in violence. In contrast, the officers’ perceptions of her abusive partner were shaped by his display of economic and educational privilege (he was dressed in a suit and worked for a law firm), as well as his calm demeanor which conflicts with controlling images of “abusers” as violent and dangerous. In addition, this incident occurred in a mandatory arrest state, and mandatory arrest policies require the police to make an arrest because the petitioner broke a frame. Without a “primary aggressor statute,” which allows officers to consider the broader context of the relationship, the officers must arrest the person who is in reality a survivor. Although this particular petitioner and respondent may not be representative of those who share their social identities or are affected by the same axes of structural inequality, linking the processes by which this petitioner and respondent both invoked and resisted representational frames, and the eventual outcome (the arrest of the petitioner) to previous research on officer decision-making can be useful in conducting a structural intersectional analysis focusing on how policies and processes replicate broader inequalities.

As for the third institutional assumption, that police will respond quickly when called and arrest respondents for violations, petitioners described a wide range of responses from police when they called to report a violation. While some petitioners reported that the police responded quickly, treated them well, and arrested respondents for violations, other petitioners said that the police took too long to arrive and did not make arrests even when petitioners had documentation of violations. For example, Petitioner #70 said that her abuser repeatedly violated her order, that she had documentation of those violations, and that “he had even admitted to [violating] it.” However, when she called the police, they “would do nothing.” She said this lack of response from the police “taught me . . . that they were not a resource for me anymore.” Their failure to take action made her believe that orders “aren’t worth the paper that they’re written on . . . I think they’re just a prosecutor tool to lay on charges on somebody.” Given that orders only “protect” through police enforcement, the petitioner was left without other resources for safety. Another unfortunate consequence of the lack of response by police or judges is that petitioners felt that their abusers were emboldened by not being sanctioned for violating orders. Again, tracing the mechanisms and processes back to previous research would be a useful way to reveal the intersectional ways in which the mismatch of institutional assumptions about and actual goals can result in inequality despite seemingly neutral practices, policies, and laws.

Finally, while the institutional assumption is that petitioners want respondents arrested and criminally sanctioned for violating orders, many petitioners did not want respondents arrested or charged for violating orders. Some petitioners exercised their agency by not calling the police despite multiple pressures to do so. For example, one petitioner (#71) who had been severely physically abused and whose order was violated several times did not call the police because

I did not wanna see him go to jail. I just wanted to get away from him. . . . I was relying on that order or protection to stay safe, to get my kids safe, to get things calmed down. . . . It wasn’t to get people arrested and in trouble.

Petitioner #73 did call the police and had a positive experience. Yet she regretted that her calls to police resulted in the respondent’s arrest when she just wanted the police to make him stay away from her.

I wish there were just different levels. It seemed like there was just one option . . . I would have liked order of protection lite, just a civil stay-away basically . . . I think that would have been effective enough in my situation to get the message to him that it is unacceptable for him to continue to harass me, to come into my house. And I think it would be smart for the courts, too, to have different levels of that . . . why do they make even more work for themselves with this heavy-duty protective order, when there might be other things that would make the process similar . . . and still have a good impact, and still enable the police to do their job?

She felt that the other steps she had taken—changing the locks on her home, hiding the respondent’s weapons, and so on—provided her with greater safety than criminal responses to violations. She would have preferred the respondent receive “psychiatric care” rather than be arrested. It is clear that she was seeking civil solutions by filing for a civil order but was instead pushed into using criminal ones. Again, the mismatch between institutional and survivor solutions can be further analyzed through the lens of structural intersectionality.

Other petitioners felt conflicted as to whether they wanted arrest, and some who wanted arrest later regretted their decision to enforce their orders. Petitioner #66 reported an assault by her abuser to police and said that the police pushed her into getting an order that she did not want.

It had already gone too far to say no [to a protective order]. And yet, that’s really not what I wanted. Because I wasn’t feeling as, or maybe I don’t know. I’m thinking back. That maybe the police made me . . . I know I didn’t go in there asking for [a] protection [order]. I wanted him arrested. I wanted him arrested for what he had done. But I guess I didn’t want to go to that extreme of not being able to see each other. I don’t know. Am I contradicting myself? Because I know I wanted protection. I was scared when I was with the police department. So maybe, this is what you can do, get an order of protection, and I’m probably saying okay. And then here it’s happening and maybe that wasn’t what I really wanted but I had no choice but to say yes . . . I guess they’re so cut and dry. What is it really protecting?

It is clear from the interview transcript that this petitioner wanted to be able to choose options for safety not currently offered by the civil and criminal legal systems. Unfortunately, these options are even more limited by the shift of governmental funding through VAWA from nonprofits, shelters, and other social services to civil and criminal responses to domestic violence. While petitioners could choose not to file for orders, they faced enormous pressures from numerous institutions—including police, prosecutors, ICE, CPS caseworkers, and other social services—to file for one. As Petitioner #72 said, “I did what I could within the system to protect them, protect us [her family].” Faced with limited choices for safety, petitioners used their agency to pursue what they felt were the best options, given their situation, and some petitioners resisted by filing for orders and, if they had an order, deliberately choosing whether and when to report violations according to their situational needs and goals. Again, tracing those limitations and agency through a structural intersectional perspective can provide insights as to the ways in which the mismatch between institutional assumptions about and the actual goals of multiply marginalized survivors replicates broader social inequalities.