Domestic violence is a major social and health problem. Estimated prevalence rates vary depending on survey methodology, with crime surveys yielding the lowest rates and so-called ‘conflict’ studies yielding the highest. In the United States, the National Crime Victimization Survey reports that less than 1% of males and about 5% of females are physically assaulted or raped by an intimate partner (Rennison 2003). Approximately 4% of Canadian women and men are physically assaulted by their partners each year, according to the General Social Survey (LaRoche 2005). Such crime surveys frame partner violence in the context of criminal behavior, and because many victims (especially men) do not regard such violence as a crime, reported victimization rates are inhibited overall and more so for male victims (Straus 1999). However, surveys framing domestic violence in terms of interpersonal conflict and inviting fuller disclosure, have found much higher prevalence rates and no significant sex differences. For instance, based on information provided by nearly 3,000 women from the 1985 National Family Violence Survey, Straus (1993) found annual perpetration rates by women to be 12.4 per 100 couples and by men, 12.2 per 100 couples. A comprehensive meta-analytic review by Archer (2000), which examined 82 studies for a combined data sample of 64,487 respondents, also found comparable perpetration rates across men and women.

Additional research demonstrates that women initiate physical aggression as often, or more often than men, rarely in self-defense, and motivated for similar reasons as men, typically for the purposes of expressing frustration, to communicate or to control, or out of a desire to retaliate (for a review, see Hamel 2007a). Furthermore, with the exception of sexual coercion, women are as psychologically abusive as men and engage in comparable levels of non-physical control (Graham-Kevan 2007). Female victims, due to their smaller size, are more often physically injured and the impact of physical aggression on women’s emotional health also may be greater (Anderson 2002; Mirrlees-Black 1999; Straus and Gelles 1990; Williams and Frieze 2005). However, men suffer a substantial minority of physical injuries, between 25 and 43% (Archer 2000; Mirrlees-Black 1999; Straus 2004; Tjaden and Thoennes 2000), and are as affected as women by psychological abuse and controlling behaviors (Pimlott-Kubiak and Cortina 2003; Vivian and Langhinrichsen-Rohling 1994).

Given the immediate and continued negative impact of domestic violence on all victims, as well as their children (Coker et al. 2002; Plichta 2004; Wolak and Finkelhor 1998), there is a need to identify effective strategies for reducing domestic violence. A restraining order may be one such strategy. However, research suggests that judicial responses to restraining order requests may be subject to genderFootnote 1 bias (Basile 2004, 2005) and that, specifically, male victims of domestic violence may not be receiving the protection afforded to them by law. This paper reports on the findings of a study exploring equal protection issues in family law by evaluating gender and violence profiles of restraining order filings in a California district court recognized for its coordinated response to domestic violence. We begin with a review of evidence for paradigmatic thinking in the domestic violence field to provide a context within which to understand sex-differentiated responses to TRO requests.

The Patriarchal Paradigm

Public policy has focused, with few exceptions, on male-perpetrated domestic violence and the needs of female victims and their children. Women account for a small number of overall arrests, less than 20% in California (California Department of Justice 2002), and an ever smaller number of individuals who are court-mandated to treatment, about 10% overall (Price and Rosenbaum 2007). Services for male victims are scarce, as evidenced by male victims who have called the Domestic Abuse Helpline for Men and Women (Hines et al. 2007). For instance, out of nearly 2,000 shelters in the United States, only a handful offer beds to battered men and their children, and outreach programs targeting male victims are essentially nonexistent (Hines 2009). This is not surprising, given that the primary source of federal funding for victim services nationwide has come from the Violence Against Women Act, which until very recently denied funding for male victims of abuse (General Printing Office 2006).

Accounting for the discrepancy between the empirical data and current public policy has been the gender paradigm (Dutton and Nicholls 2005), also known as the patriarchal paradigm (Hamel 2007b), a set of assumptions and beliefs about domestic violence that has shaped domestic violence policy on arrest, treatment and victim services at all levels for the past several decades. A product of feminist sociopolitical theory, the paradigm posits that the causes of domestic violence can be found in patriarchy and male dominance, and ignores or minimizes the importance of other social factors as well as established findings on personality, child development, family systems, and relationship dynamics. Despite data that are inconsistent with the feminist perspective (see, for example: Dutton 2005a; Dutton and Corvo 2006; Dutton and Nicholls 2005; Felson 2002; Hamel and Nicholls 2007; Mills 2003; Straus 2006), it remains a dominant influence, supporting victim services and outreach/prevention programs for women and children and an assortment of legal remedies for male batterers, including mandatory arrest, criminal and civil restraining orders, and court-mandated interventions combining cognitive-behavioral (e.g., building interpersonal skills) and psychoeducational techniques (e.g., ‘re-educating’ abusive men in overcoming their presumed sexist, misogynistic attitudes, Pence and Paymar 1993).

Evidence of bias attributable to the patriarchal paradigm also has been found among a variety of mental health professionals (Follingstad et al. 2004; Hamel et al. 2007). Such bias is evident within the American Bar Association Commission on Domestic Violence, which perpetuates the myth that 90 to 95% of domestic violence is perpetrated by men (American Bar Association 2006; for a critique see Dutton et al. 2009), and also has been demonstrated among domestic violence and child custody researchers (Dutton 2005b; Straus 2006), as well as family court mediators, evaluators, attorneys, and judges (Hamel et al. 2009). The literature on restraining orders in particular reflects a lack of focus on male victims (Buzawa and Buzawa 2003; Etter and Birzer 2007; Logan et al. 2005).

Domestic Violence Restraining Orders

Every state in the United States now authorizes its courts to issue civil orders of protection against domestic violence. Typically, a temporary domestic violence restraining order (TRO) is issued ex parte at the request of any plaintiff who expresses an “objectively reasonable subjective fear of being injured” (Miller 2005, p. 74), without the respondent (i.e., the alleged perpetrator) having to be present in court. TROs are granted for two- to four-week periods, at which point a hearing is held to determine if a permanent order is warranted, valid in most states for a period of one to four years. In California, as of June 6th 2003, there were 227,941 active restraining orders (including temporary and permanent) issued against adults, almost all of them for domestic violence. Of the domestic violence orders, approximately 72% restrained a man from a protected woman, 19% restrained a same-sex partner, and 9% restrained a woman from a protected man (Sorenson and Shen 2005). Of particular significance to family court cases, the protected parent almost automatically obtains custody of the children, without a custody hearing or a custody decision being made (Kanuha and Ross 2004; Sorenson and Shen 2005). Many states, including California, have laws providing for mandatory arrest for anyone violating such an order, per PC 836 (c) (1) (Hirschel et al. 2007).

There is no doubt that restraining orders are a legitimate tool with which society can combat domestic violence and protect victims and their children from further abuse. Individuals who file restraining orders generally have reason to fear their abusers. Some studies have found more serious criminal histories among respondents subject to family court restraining orders than those criminally convicted (Buzawa et al. 1999). In a study conducted in Kentucky, 81% of female plaintiffs reported to have suffered severe physical abuse by the respondent (including 68% who suffered injuries) and 93% reported serious threats of physical harm (Logan et al. 2007). Male and female plaintiffs in Hawaii also reported high rates of violence and threats: 65% having been pushed, grabbed, or shoved; 35% kicked, bit, or punched; 14% chocked; 35% threatened with physical harm; and 41% who were issued death threats (Kanuha and Ross 2004). Furthermore, when children are witnesses to such abuse, they are at risk for suffering significant emotional distress and to act out with aggression, both in childhood and adolescence, and later against adult partners and their own children, regardless of the perpetrator’s sex (English et al. 2003; Fergusson and Horwood 1998; Johnston and Roseby 1997; Langhinrichsen-Rohling et al. 1995; Mahoney et al. 2003; Sommer 1996; Straus 1992; Wolak and Finkelhor 1998).

Ideally, restraining orders should be available to victims without them having to overcome unnecessary obstacles or putting themselves at risk of further harm, yet not be so freely granted that they can be manipulated by a vindictive partner. They also should be effective in protecting all genuine victims. As with most well intended laws and policy decisions, however, practice deviates considerably from the ideal. A report by the California Attorney General’s Task Force on the Local Criminal Justice Response to Domestic Violence (Seave 2006), for example, demonstrated an inconsistent pattern of enforcement throughout the state and serious shortcomings, such as many orders not containing firearm restrictions. Even when properly enforced, restraining orders are effective primarily with low-risk individuals, those with a ‘risk in conformity’, who are gainfully employed and have no previous criminal histories (Hotaling and Buzawa 2003; Mills 2003).

Partly due to lobbying by advocates who sought to overcome some of the problems listed above, restraining orders are now more liberally granted (Heleniak 2005). In many states, there is no requirement that an actual domestic violence assault have been committed, only that the plaintiff have an “objectively reasonable subjective fear of being injured” (Miller 2005, p. 74), and the process of obtaining a restraining order has been streamlined considerably (Superior Court of California 2007). Consequently, a number of judges have adopted a ‘better safe than sorry’ mindset, and anecdotal data indicate that restraining orders can be manipulated by some plaintiffs for reasons other than self-protection—to retaliate, control, or gain an advantage in family court (Kasper 2005).

“Many TRO’s and PO’s [protection orders],” concludes a Hawaiian task force on restraining orders, “are obtained by one party to a dispute to try to gain advantage over another party in future or ongoing divorce proceedings or a custody dispute” (Murdoch 2005, p. 17). In California, the Family Law section of the state bar expressed concern that domestic violence restraining orders “are increasingly being used in family law cases to help one side jockey for an advantage in child custody and/or property litigation and in cases involving the right to receive spousal support” (Robe and Ross 2005, p. 26). A retired Massachusetts judge revealed to the press that, in his experience, one-third of restraining orders are strategic ploys used for leverage in divorce cases (“Retiring Judge” 2001). Attorneys Sheara Friend and Dorothy Wright, the latter also a former board member of a battered women’s shelter, estimate that 40 to 50% of restraining orders are used to manipulate the system (Young 1999). In some cases, mothers secure custody despite a history of abuse against the father or the children (Cook 1997; Pearson 1997).

The data regarding unsubstantiated allegations of adult and child abuse are mixed, but at least one recent study reports equivalent rates for fathers and mothers (Johnston et al. 2005). Clearly, men are capable of manipulating the courts to their advantage, and this includes some who have a documented history of domestic violence (Kernic et al. 2005; Morrill et al. 2005). It could be that reports of high levels of female plaintiffs illegitimately obtaining restraining orders are a function not of any pervasive bias by judges but rather the fact that women far more often than men seek these orders in the first place. If women seek out these orders at higher rates, they are more likely to have them granted—and to possibly misuse them in some cases. This raises questions regarding why men represent such a small number of petitioners and the likelihood that their requests will be granted.

The Present Study

The present study sought to answer this second question. Holding the level of alleged violence constant, the answer may have serious public policy implications in terms of judicial conduct and the goal of maintaining public confidence in the courts. To date, we know of only one researcher who has addressed this question. Basile (2004, 2005) conducted a two-part examination of sex-differentiated court responses, examining the allegations listed in a sample of abuse prevention cases (i.e., requests for domestic violence protective orders) filed by both male and female plaintiffs in the Gardner District Civil Court in Massachusetts. We briefly review the findings of these two studies below.

Basile’s first study (2004) examined the level and type of domestic violence allegations listed in plaintiff requests for abuse prevention orders. Of the total population of cases, 73% were from women seeking protection against men, 14% from men seeking protection from women and about 13% of the cases involved same-sex litigants. Basile found no significant in the degree of alleged violence in comparing the files of male and female litigants in the aggregate. However, he noted sex differentiation in the manifestation of violence and/or the expression of prohibited conduct. For example, female defendants were much more likely to use a dangerous weapon (i.e., apparently to compensate for physical disparity with men) in attacking their victims and were more likely to scratch or gouge their victims. Female defendants also made harassing phone calls more often than male defendants. In contrast, male defendants were more likely to choke or slam their victims against a wall.

Basile’s second study (2005) examined court responses to requests for protection as a function of plaintiff sex. Results demonstrated that for a range of multiple potential court responses affecting the study population, similarly situated male victims were not afforded the same protection as female victims. For example, Basile found that women’s abuse protection requests were granted 91% of the time, compared to only 66% for men. Conversely, men were twice as likely to have their protection requests denied compared to women; 360% more likely to have their issues deferred compared to women; 110% more likely to be evicted if the litigants had a child together; and 29% more likely to be evicted if they did not share a child. Not surprisingly, these findings created significant controversy. In fact, following publication of these results, legislation was passed limiting access to the court files of protective order requests in Massachusetts (see Charalambous 2005).

Similar to the Gardner District Civil Court in Massachusetts sampled by Basile, the present study sampled court records from the Sacramento, California district court, a national model recognized for its coordinated response to domestic violence. Our goal was to determine whether there are sex-differentiated court responses to plaintiff requests for TROs. Each file was coded for level of alleged violence affording the examination of court responses relative to violence levels across plaintiff sex. In doing so, our specific objectives were to: (1) determine the gender profile (i.e., frequency of male vs. female plaintiffs) for the sample of TRO requests; (2) establish the degree and type of alleged violence (i.e., violence profile) for ex parte TRO requests; (3) compare the violence profile attributed to male versus female defendants; and (4) examine sex-differentiated case outcomes relative to level of alleged violence to determine.

Overall, it was hypothesized that plaintiff sex would predict whether judges granted or denied requests for TROs, as well as permanent restraining orders (PROs), with female plaintiffs generally receiving more favorable treatment compared to male plaintiffs. We also anticipated that level of violence would predict case outcomes such that TROs would be more likely to be granted in cases alleging more severe violence compared with cases alleging lower levels of violence. Finally, a plaintiff sex by level of violence interaction effect on case outcomes was expected, such that TROs would be most likely to be granted for cases alleging high levels of violence perpetrated against a female plaintiff and least likely to be grant for cases alleging low levels of violence perpetrated against a male plaintiff.

Method

Court Records

The present study sampled family court records of TROs maintained by the Sacramento Superior Court, William R. Ridgeway Family Relations Courthouse, located in Sacramento, California. The case files normally contain the following documents: (1) Request for Order (DV-100), (2) Description of Abuse (DV-101), (3) Child Custody, Visitation, and Support Request (DV-105), (4) Temporary Restraining Order and Notice of Hearing (DV-110), (5) Restraining Order After Hearing (DV-130), (6) Child Custody and Visitation Order (DV-140), and (7) court docket log form containing the chronological history of each case. DV-100 through DV-105 are used to provide background information about each case. Forms DV-110 through DV-140 are completed by the plaintiff (i.e., person requesting the protective order) and may become the court’s orders upon approval of the judge. The majority of the data coded in the present study (including plaintiff and defendant sex, level of violence, whether the TRO was approved, final disposition of the case) was gleaned from the DV-100, DV-101, and DV-110 forms, as well as the court’s docket log in some cases. Notably, the only description of violence available to the court is the petitioner’s written report of the most recent abuse, including responses to questions regarding date, who was present, what was said or done to make them afraid, use or threatened use of weapons, and injuries.

The Sacramento Family Court does not file domestic violence requests separately from other categories of restraining order requests, such as civil harassment and elder abuse. All types of restraining order files are identified by the year filed and assigned a case number. The files numbers are consecutive according to the order in which they are processed. Of the estimated 3,000 TRO requests processed by the Sacramento Family Court per year, approximately 79% were for domestic violence, with 80% of those filed against men and 20% against women.

Level of Violence Rating Scheme

Level of violence alleged in each TRO request (as described in the Description of Abuse, DV-101) was coded as low, moderate, or high using the following scheme developed based on a review of the literature and within the meaning of the California Domestic Violence Prevention Act (DVPA). We provide a complete description of our coding criteria in Table 1. Fairly typical of the legal scheme used in many US states, an alleged domestic violence victim is a ‘protected person’ under the DVPA and is entitled to the benefit of a protective order provided that the abuse fits the legal definition. Abuse may comprise physical assault or non-physical conduct that places the victim in “reasonable apprehension” of imminent serious bodily injury to that person or other person (Ca. Fam. Code § 6300 et seq). Domestic violence also may involve conduct that could be enjoined pursuant to Family Code § 6203(d). Thus, in the present study, domestic violence encompassed a range of interpersonal maladjustments or aberrant behaviors, from making threats, destroying personal property, emotional abuse, disturbing the peace, financial abuse, stalking, unwanted contact, or annoying telephone calls (coded as low violence) to acts of physical and sexual violence (coded as either moderate or high violence depending on the degree of physical harm, see Table 1). Importantly, all acts considered abuse under California law would have been captured in our coding scheme and vice versus.

Table 1 Level of violence rating scheme
Full size table

Level of violence was first rated by the first author of this manuscript. There was insufficient information to rate level of violence in one case. A BA-level research assistant with a double major in criminology and psychology, blind to the first violence ratings, additionally coded a subset of approximately 15% of the TRO requests (n = 22). Results calculated using the intraclass correlation coefficient (ICC) demonstrated that interrater reliability was good for the violence ratings, ICC 1 = .74, p < .001 (see Cicchetti et al. 2006). There were no disagreements in which one rater indicated that the level of violence was low and the other rater indicated ‘high’.

Data Collection

In total, 227 Family Court TRO records were randomly selected by the Court Records Clerk from the population of 2002 and 2003 domestic violence TRO petitions.Footnote 2 Under the DVPA, protection is afforded only to certain prescribed classes of persons that are potentially susceptible to domestic violence under the DVPA, including spouses, cohabitants, persons involved in dating relationships, and co-parents. Of the sampled records, 70 represented filings by non-intimate partners. Given the present study’s focus on violence between intimates, these 70 cases were removed for a final sample of 157 petitions involving intimate partners, dating couples, and married persons.

Data Analysis Plan

To compare the frequency with which TRO requests were filed by male and female plaintiffs, we conducted Chi-square analyses and calculated effect size using phi coefficients. The same analytic approach was used to examine the distribution of degree and type of alleged violence (i.e., violence profile). We then conducted a multinomial logistic regression analysis to compare whether alleged violence (nominal criterion variable) differed as a function of plaintiff sex (dichotomous predictor variable). To examine sex-differentiated case outcomes relative to level of alleged violence, we first tested separate binary logistic regression models of plaintiff sex and level of violence predicting the granting of TROs (yes/no), followed by a forced entry multinomial logistic regression model of plaintiff sex and level of violence, as well as their interaction, predicting court decisions. A binary logistic regression model was tested to determine whether plaintiff sex predicted court decisions when controlling for level of violence. Finally, we conducted separate logistic regression analyses to examine the roles of sex and level of violence in predicting whether PROs were ultimately issued or whether the case was dismissed. All analyses were conducted using SPSS Version 15.0.

Results

We present descriptive characteristics of the TRO requests in Table 2.

Table 2 Characteristics of sampled temporary restraining order requests
Full size table

Gender Profile of TRO Requests

The first objective of this study was to compare the frequency of domestic violence TRO petitions made by female and male plaintiffs. As anticipated, female requests for protection were significantly greater in number than those for men, with the vast majority of cases involving female plaintiffs alleging domestic violence by a male intimate partner. Specifically, of the 157 petitions, 131 files (83%) named female plaintiffs and 26 files (17%) named male plaintiffs, χ 2(1, N = 157) = 70.22, p < .001, Φ = .67. This pattern of results is consistent with the distribution of all domestic violence TRO requests processed by the Sacramento Family Court per year (i.e., approximately 80% filed against men and 20% against women).

Violence Profile of TRO Requests

Our second objective was to establish the violence profile of the domestic violence TRO requests (i.e., level of alleged violence). In one case, there was insufficient information to code the level of alleged violence; thus, 156 cases were included the subsequent analyses. As may be seen in Table 2, of the requests for which information was available regarding the level of alleged violence, the vast majority reported low (45%, n = 70) or moderate (39%, n = 61) levels of violence, representing no or minor physical threat to the victim. In only 16% (n = 25) of the cases were allegations of serious violence made. Chi-square analyses confirmed that alleged violence differed beyond levels expected by chance, χ 2(2, N = 155) = 21.81, p < .001, Φ = .37, and that allegations of low and moderate levels of violence were significantly more common than allegations of high levels of violence, χ 2(1, N = 95) = 21.32, p < .001, Φ = .47, and χ 2(1, N = 86) = 15.07, p < .001, Φ = .42, respectively. Prevalence of low and moderate levels of alleged violence did not differ significantly (see Table 2).

Level of Alleged Violence as a Function of Plaintiff Sex

Our third objective was to compare the levels of alleged violence in requests made by male versus female plaintiffs. Of 130 TRO requests filed by women for which violence level was known, 15% (n = 20) were rated high violence, 40% (n = 52) were rated moderate violence and 45% (n = 58) were rated as low violence. In comparison, the level of violence alleged in the TRO requests made by men was rated as high for 19% (n = 5) of the files, 35% (n = 9) were rated moderate violence and 46% (n = 12) were rated as low violence. The multinomial logistic regression model did not fit the data, indicating that severity of alleged violence did not differ significantly as a function of plaintiff sex.

Court Responses

Here we focus on court responses regarding whether the TRO was granted or denied. We additionally examined whether or not a PRO was granted, which is determined independently of whether the TRO is granted or denied. We did not examine the myriad of court responses (i.e., protective orders) available under the California DVPA once a temporary restraining order is granted (e.g., temporary custody, temporary visitation, dwelling exclusion, effectuating orders, firearms prohibitions, stay away orders, conduct and property destruction orders, wiretap orders, restitution orders, and even orders to attend batterer’s programs).

TRO Decisions

Overall, Sacramento County Judicial Officers granted approximately 88% (n = 138) of the requested TROs in our sample (see Table 2). As hypothesized, binary logistic regression revealed differences in judicial responses to TRO requests as a function of plaintiff sex, χ 2(1, N = 156) = 21.64, p < .001, Φ = .37. In fact, plaintiff sex alone accounted for 25% of the variance in court decisions, with judges being almost 13 times more likely to grant a TRO requested by a female plaintiff compared to those requested by male plaintiffs, OR = 12.89, p < .001. Of the 25 TRO petitions filed by male plaintiffs, 44% (n = 11) were denied compared to 5% (n = 7) of petitions filed by female plaintiffs (see Table 3). With respect to level of alleged violence, again, our hypothesis was supported. Specifically, level of violence was a significant predictor of case outcomes such that TROs were more likely to be granted in cases alleging more serious violence compared with cases alleging lower levels of violence, χ 2(2, N = 155) = 9.98, p < .01: Level of violence accounted for 12% of the variance in judicial decisions, with courts granting the TRO request in 80% (n = 55) of the 70 cases alleging low levels of violence, compared with 97% (n = 59) of the 61 cases alleging moderate and 92% (n = 23) of the 25 cases alleging high levels of violence (see Table 3).

Table 3 Percentage (number) of cases in which restraining orders were issued
Full size table

To address our final objective of examining sex-differentiated court responses relative to alleged violence levels, a binary logistic regression model was tested with plaintiff sex and alleged violence, as well as their interaction, as predictors of court responses to TRO requests. The model significantly predicted case outcomes, χ 2(3, N = 155) = 33.18, p < .001, accounting for 38% of the variability in court decisions. Results, however, offered only partial support for our hypothesis: Sex differentiation was limited to cases that involved allegations of low-level violence. In other words, although male plaintiffs alleging low levels of violence perpetrated by their (female) intimate partners were significantly more likely to have their TRO requests denied than were female plaintiffs alleging low levels of violence perpetrated by their (male) intimate partners, χ 2(2, N = 69) = 26.88, p < .001, Φ = .62, no significant differences as a function of plaintiff sex were observed for cases involving allegations of moderate and high levels of violence. The proportion of variance in case outcome accounted for by plaintiff sex increased substantially with level of violence included as a covariate in the logistic regression model. Again, the model significantly predicted case outcomes, χ 2(2, N = 155) = 29.37, p < .001, accounting for 34% of the variability in court decisions to grant or deny the TRO requests. When controlling for level of violence, judges were approximately 16 times more likely to grant a TRO requested by a female plaintiff compared to those requested by male plaintiffs, OR = 16.00, p < .001.

Issuance of PROs

PROs were issued in approximately one-third of the cases for which outcome information was available (see Table 3). In contrast with prior analyses, there was no evidence for sex-differentiated granting of PROs, nor was there evidence for court responses differing as a function of alleged level of violence. Although it appears that courts were almost twice as likely to issue PROs in cases filed by female than male plaintiffs and in cases involving allegations of high versus low or moderate levels of violence (see Table 3), differences were not statistically significant, p’s > .05. Results did, however, reveal sex differentiation of court responses with regards to the issuing of PROs in cases that involved allegations of low-level violence. As may be seen in Table 3, PROs were significantly more likely to be issued in cases involving female plaintiffs alleging low levels of violence perpetrated by their (male) intimate partners compared with cases involving male plaintiffs alleging low levels of violence perpetrated by their (female) intimate partners, χ 2(1, N = 59) = 26.88, p < .05, Φ = .67. No significant differences in issuance of PROs as a function of plaintiff sex were observed for cases involving allegations of moderate or high levels of violence.

Case Dismissal

Additional analyses were conducted to explore whether case dismissal (i.e., either plaintiff requested dismissal or request was dropped as a result of plaintiff’s failure to appear in court) differed as a function of plaintiff sex and alleged level of violence. It is reasonable to expect, for example, that male plaintiffs may be more likely to drop a request given the decreased likelihood that they will report their abuse (e.g., Felson et al. 1999; Straus 1993; Tjaden and Thoennes 2000). Similarly, we may anticipate differences in case dismissal as a function of level of violence, such that increased severity is associated with increased fear of the perpetrator and decreased likelihood of failure to follow through with the request. However, as with analyses of PRO issuance, there was no evidence for sex-differentiated case dismissal, nor was there evidence for differences as a function of alleged level of violence. Overall, requests were dismissed in more than half of the cases for which outcome information was available (57%, n = 79). Although review of Table 4 suggests that case dismissal was slightly more common for petitions filed by men (48%, n = 12) than women (59%, n = 67) and those involving allegations of moderate (71%, n = 39) than low (51%, n = 30) or high (42%, n = 10) levels of violence, differences were not statistically significant, p’s > .05. We found no differences in case dismissal as a function of plaintiff sex within violence levels.

Table 4 Percentage (number) of cases in which requests were dismissed
Full size table

Discussion

Despite important advances in research, policy, and practice over the last 30 years, domestic violence remains a significant problem in our society. With gender-exclusive policies and practices the norm, the field has typically been disregarding the detrimental effects of female perpetration of domestic violence and male victimization, essentially precluding the delivery of effective treatment intervention to high-risk groups (Dutton and Corvo 2006; Dutton and Nicholls 2005). Though all preventative measures, including civil orders of protection against domestic violence, should be available to all victims, prior research (Basile 2004, 2005) suggests that there may be sex-based bias in judicial responses to domestic violence restraining order requests. To extend this research, this paper reported findings of a study which evaluated the impact of gender and violence profiles of a random sample of 157 temporary restraining order petitions involving intimate partners, dating couples, and married persons on judicial decision making in a California district court. Specifically, we sought to determine whether Family Court Judges are applying a double standard in the adjudication of temporary and permanent restraining orders.

As hypothesized, results suggest that plaintiff sex plays a primary role in predicting whether judges grant or deny requests for restraining orders in certain cases (i.e., those alleging low levels of violence). Thus, the answer to our question of how likely men are to have orders granted appears to be ‘not very’ in these cases. Although the findings and research parameters are not identical, the present study replicates some of the significant results of the Basile (2004, 2005) research. Results of both the Basile and present research demonstrated that: (1) there is a sex-differentiation of court responses, and (2) that Court protective orders, as viewed in the aggregate, are preferential to women plaintiffs. Taken together, findings suggest that male victims of domestic violence are not afforded the same protection as female victims.

Interestingly, however, we found evidence for preferential treatment of women in ex parte requests for protective orders alleging low levels of violence, but found no evidence of discrimination in moderate to high violence cases. In other words, there was a shift from a situation in which facts mostly predict the Court’s response (moderate/high violence) to a situation in which plaintiff sex largely predicts the response. This suggests that judges rely on heuristics and extra-evidential information for low-level violence cases in which adjudicative fact is relatively sparse, but make decisions on legitimate evidence in more severe restraining order requests in which adjudicative fact is typically more abundant. One explanation for this finding is that judges may be undecided regarding whether the temporary restraining order is needed in cases alleging low-level violence. As a result, judges may increase reliance on heuristics and extra-evidential information, such as plaintiff sex, to arrive at a resolution not available in the petition itself.

Results of this study also demonstrated that women made significantly more requests than did men. This finding speaks to another important question we posed earlier in this paper, namely, why men represent such a small number of petitioners. One possible answer is that men do not often seek out restraining orders because they do not need them. Research indicates that male victims are less likely than female victims to express fear of their partners, at rates of approximately 3:1 (Follingstad et al. 1991; Morse 1995). Furthermore, research from both Canada and the US indicate that women are at higher risk relative to men of being killed following a relationship separation (Statistics Canada 2001; Wilson and Daly 1992). It is also true, however, that spousal abuse homicides are relatively rare in comparison to other spousal abuse and that overall post-separation assault rates are comparable across plaintiff sex. One large Canadian survey, for example, found that 32% of men and 40% of women who had previously been victimized in their intimate relationships were re-victimized after separation (Statistics Canada 2001). Men are less often physically stalked following separation than are women, but are harassed in other ways (e.g., by phone, malicious gossip; Williams et al. 2007). Although they make more threats of physical harm, men are less likely than women to actually carry them out (Meloy and Boyd 2003). Indeed, our comparisons of alleged level of violence as a function of plaintiff sex showed no differences between men and women’s requests.

Further, the observation that male intimate partners reported domestic conflict less often than female intimate partners is not necessarily an indication that men initiate domestic violence more often. Research demonstrates that men are generally far more reluctant to report domestic violence than women, even in the most severe of cases. Due to prevailing norms regarding masculinity, men may be reluctant to express fear or to call the police even when they have every reason to do so (Cook 1997; Hines et al. 2007; Tjaden and Thoennes 2000). Men may be less inclined than women to file domestic violence restraining orders because of the unique obstacles they face. In their study on female restraining order petitioners in Kansas, Logan et al. (2005) found that the biggest obstacle for sampled women in obtaining and enforcing a restraining order was the lack of resources to leave their perpetrator. Support from battered women’s advocates seems to make a difference (Kanuha and Ross 2004). As previously mentioned, however, victim advocacy work is conducted almost exclusively on behalf of female victims. For instance, the restraining order class offered in Sacramento County is conducted by an organization, Women Escaping a Violent Environment, whose very name might be off-putting to male victims (cf. Superior Court of California 2007). There is also evidence that law enforcement officers are significantly less likely to give male victims information about available services, including restraining orders, compared to female victims (Buzawa and Hotaling 2006).

The predominance of allegations of low and moderate levels of violence in the context of California law also merits some discussion. Most domestic violence restraining orders are obtained on very short notice. These ex parte orders are of limited duration and last until the next ‘order to show case’ hearing, which must be held 21 days from the issuance of the temporary order. Ex parte orders are granted to the plaintiff without ‘formal’ notice to the other party or opportunity to be heard in opposition. Under Rule 379 (Ca Rules of Ct.), absent a showing of exceptional circumstances, the applicant must notify all parties the day before the court appearance. However, not all family courts abide by this rule, and in those courts the minimum notice period may be different. Because due process requires reasonable notice and opportunity to be heard, ex parte protective orders for domestic violence are supposed to be issued with caution and under extraordinary circumstances. In fact, no restraining order governed by the California Family Code § 240 et seq may be granted on an ex parte basis unless from a showing of the facts that “great or irreparable injury would result to the applicant” (Ca Fam. law § 241). Appropriately, our findings showed that ex parte temporary restraining order requests were more likely to be granted in cases alleging moderate or high levels of violence compared with those alleging low levels. However, some may debate the appropriateness of ex parte granting in our ‘moderate’ violence cases. Despite the conservative language of the California legislation, the practical reality is that most family law courts grant the vast majority of ex parte restraining orders, and overall, the frequency of allegations of serious violence in this sample was quite low.

Limitations

This study is not without limitations which bear on the interpretation and generalizability of the results. First, some caution must be taken in extending the research findings beyond Sacramento County. Although our findings are consistent with those of Basile (2004, 2005), replication in additional jurisdictions is needed. Second, there was very little information in the files regarding the nature of the intimate relationships, in particular, regarding sexual orientation. Of the petitions sampled, only a few relationships were explicitly identified as homosexual. For this reason, we did not compare court responses regarding heterosexual and homosexual relationships. However, judicial treatment of domestic violence in gay and lesbian intimate relationships warrants further study. Third, the violence ratings reflect our coding of the cases and not the judges’ perceptions of the level of violence. As a result, interpretation of the impact of level of violence on judicial decision-making is speculative. Further, although some of the petitions were identified as consolidated cases (i.e., cases in which the restrained party is involved in other family court cases, such as divorce or custody maters), the sample size of this sub-group (n = 18) was insufficient to afford comparisons between these and standalone petitions regarding judicial treatment and disposition. Finally, we did not examine the entire protective order profile, only the court’s granting of ex parte restraining orders; thus, results do not speak to whether there is sex-based discrimination in all court responses.

Conclusions and Future Directions

In theory, all American citizens are entitled to equal protection under the laws (Haney 1991). Central to the notion of equal protection is the prevention of official conduct discriminating against any distinct class of citizen. Typically a distinct class of citizen is defined as a distinct group of persons with common and immutable characteristics; thus, persons of similar race, age, or sex would be considered a distinct class for the purpose of equal protection. Although judicial decisions constitute state action, judges are immune from discrimination claims that derive from normal judicial functions. Nevertheless, equal treatment of similarly situated persons should be a goal of the judicial system. A pattern of judicial discrimination in the granting of restraining order requests, if proven, would erode confidence in the courts and raises concerns of judicial misconduct.

Additionally, the present study, as well as the earlier Basile (2004, 2005) research, was based on protective order filings in domestic violence courts. More research is needed to establish the generalizability of these findings to the equity court justice systems of both the United States and Canada, as decision-making in different courts will likely reflect the political attitudes of their region. Nevertheless, domestic violence crosses traditional political boundaries and the influence of the patriarchal paradigm on the criminal justice system has been effective, broad, and deep (Dutton and Nicholls 2005; Hamel 2007b). Indeed, in a sample of 4,178 misdemeanor and felony domestic violence defendants (21% female) in Shelby County, Tennessee, Henning and Feder (2005) found that female defendants compared with male defendants were more likely to be released, less likely to be prosecuted, less likely to plead guilty or to be found guilty, and less likely to incarcerated when found guilty, even after controlling for other demographic factors and offense characteristics. Research examining perceptions of violence risk further supports the generalizability of our findings: Studies of risk assessment demonstrate that clinicians frequently underestimate female psychiatric patients’ violence risk, even when patients were admitted following violent incidents (e.g., Coontz et al. 1994; Elbogen et al. 2001; Skeem et al. 2005)

As previously noted, the issue of sex-based discrimination in the courts, as suggested by the present study’s findings, has clear policy implications. Results suggest that Sacramento Family Court judges may share beliefs consistent with the patriarchal paradigm and, thus, are responding to requests in a manner consistent with this one-sided and largely stereotypical viewpoint about the nature of domestic violence (i.e., that only women are victims of domestic violence). Because the bias was found only in low-level violence cases, judges may be expressing beliefs regarding the potential for future harm and escalation based on victim sex. Another explanation for the sex-differentiated responses is that judges may be engaging in risk avoidance for the serious, and potentially politically embarrassing, situation in which a woman is injured or murdered by her intimate partner following denial of her restraining order request. The preferential treatment toward women’s requests for civil protection in cases involving low-level violence also may reflect the political power of women’s organizations and their advocacy on behalf of female domestic violence victims. Although further research is needed to elucidate the reasons for the observed bias, judicial education regarding the prevalence, nature, and consequences of female aggression, and female-perpetrated domestic violence in particular, may be one means for reducing the effects of plaintiff sex on the granting of domestic violence protective orders.