Abstract
Protection orders are widely used legal interventions intended to reduce the risk of future harm by one person considered to be a threat to another. However, there has been controversy about when and whether such orders are useful. This article is a review of empirical studies of outcomes associated with protection orders and factors associated with violations of the orders. Although protection orders are not a panacea, the results support that they can serve a useful role in threat management. We describe factors to consider before implementing a protection order.
Protection orders are considered by forensic evaluators, threat assessment specialists, and mental health clinicians when individuals are victims of domestic violence, are stalked, or are otherwise threatened or abused. Forensic evaluators may recommend the use of protection orders during the course of risk assessments for law enforcement agencies and threat-management teams. Clinicians may also consider the use of a protection order when making a Tarasoff warning or when they themselves are stalked. Although they are rare, when situations potentially warranting protection orders arise, forensic evaluators must incorporate both ethics-related and legal considerations and then make thoughtful recommendations, often in the context of limited information. Because of the high stakes for the victim in such cases, many professionals err on the side of caution by invoking a legal protection order with the goal of providing the highest level of security to the victim.
Protection orders are legal interventions designed to reduce the risk of future threat or harm by a person who is determined to pose a threat to another. However, the specifics of protection orders can vary greatly across jurisdictions.1–3 They can be issued by criminal courts to persons charged with assault or other crimes, by family courts in the context of divorce proceedings, or by civil courts after a hearing in which a petitioner presents a case of violence, stalking, or harassment and asks for a court order to stop these behaviors. Such orders can also vary by duration, depending on the jurisdiction. These variations give rise to one of the major obstacles in assessing the effectiveness of these legal interventions systematically and are thus among the major difficulties in providing a recommendation based on a firm foundation.
Protection order legislation was first implemented in the 1970s, and by 1989 all 50 states and the District of Columbia had enacted statutes providing civil remedies for battered women via protection orders.4 Protection orders have also been called restraining orders, civil protection orders, orders of protection, stay-away orders, protection from abuse orders, domestic violence restraining orders, civil harassment restraining orders, no-contact orders, and anti-harassment orders. Although these various orders may be different, in the interest of simplicity, the term protection order will be used in this review.
In general, protection orders are designed to: …prohibit the abuser from committing acts of violence; exclude the abuser from the residence shared by the petitioner and abuser; prohibit the abuser from harassing or contacting the petitioner by mail, telephone, or in person; award temporary custody of minor children; establish temporary visitation and restrain the abuser from interfering with custody; prohibit the abuser from the jurisdiction of the court; and order the abuser to participate in treatment or counseling [Ref. 5, p. 589].
In most jurisdictions, civil protection order statutes grant judges wide discretion to issue any warranted relief that is constitutionally defensible.6,7 As such, civil protection orders enable victims of violence or harassment to petition for stay-away orders, no-contact orders, vacate orders, property rights and access orders, custody orders, visitation orders, and orders for child support, monetary support, and relief for the victim, including medical and counseling bills and orders for treatment or counseling for the defendant.6 In these orders, a person is usually mandated to have no contact with another person, with specific restrictions for proximity.6 Although state statutes vary, the Interstate Stalking Punishment and Prevention Act of 1996 made a restraining order issued in one state enforceable in other states and violation of such an order across state lines a federal crime punishable by both imprisonment and a fine.8
Sorenson and Shen3 summarized the protection order statutes of the state of California as follows: “emergency orders” last for the shorter of five court days or seven week days, “temporary orders” can remain in effect for three weeks, “restraining orders” for up to three years, and “permanent orders” for the life of either party. Overall, the state maintains approximately 880 protection orders for every 100,000 adults, and between 84 and 92 percent of these orders are implemented for domestic violence. Their estimates show that 17% (roughly 147 of 880) of the orders initiated go unserved and so are not technically enforceable.
Consequences for violation of protection orders range from civil contempt penalties to misdemeanor or felony criminal charges. Punishments range from fines to imprisonment.6,7,9 In select jurisdictions, police have been given the authority to enhance the legal response to protection orders and are permitted to make warrantless arrests for misdemeanor offenses. As of 2004, 20 states and the District of Columbia had instituted mandatory arrest laws for order violations.10
While experts in fields in which protection orders are commonly considered (typically the judiciary, law enforcement, legal scholars, sociologists, and mental health clinicians) may provide opinions as to whether it is appropriate to seek such an order, limited data are available to guide the decision-making process. Most studies that have addressed this topic arise from fields that may be less familiar to the mental health clinician, existing primarily in the domestic violence, social science, or legal literature. As a consequence, the findings in these studies, while important, may be less clinically relevant in the mental health setting. The purpose of this article is to review the current literature in the area of protection orders with attention to those studies that provide data on rates of violation and characteristics associated with increased risk of violation. In light of this information, we propose a framework for decision-making when considering protection orders.
Methods
A search of PubMed and PsycINFO was conducted with the terms protective order and restraining order. References from articles retrieved from this search were then used to expand the scope of the search. LexisNexis was used to find relevant articles in the legal and law review literature. Google Scholar was searched with the same terms to identify other potentially relevant articles missing from the original searches. Many of the references from the relevant articles identified in these searches were also reviewed. We attempted to include the broad scope of legal, sociological, and mental health literature in this review.
After the review was limited to accessible articles published in peer-reviewed journals, official policy statements, and books currently in print, 15 original articles, book chapters, and several internet references were eligible for inclusion. From these sources, references focusing on rates of protection order violation and factors associated with violation and the key findings of these empirical studies were synthesized. The methodologies used in these studies included review of court documents and police reports, victim and defendant interviews, and random telephone surveys. The studies ranged from one-time interviews to longitudinal investigations with various follow-up periods. Although the diversity of research methods posed challenges to synthesizing the literature, an attempt was made to group relevant findings into a cohesive framework.
Results
Protection Orders and Violence
Table 1 provides an overview of the selected studies, noting the study design, sample size, focus (victims of intimate partner violence, protection order defendants, and stalking), study limitations, rates of protection order violations, and variables associated with protection order violation.
The most widely reported measure of protection order effectiveness is the rate of violation. Another important measure is the specific nature of the violation (e.g., verbal contact, psychological abuse, or physical violence). The reported rates of protection order violation vary widely across studies, from as low as 7.1 percent to as high as 81.3 percent.5,11–23
Many studies of protection orders have been limited by design problems such as lack of comparison groups, small sample size, and short follow-up periods. In the first controlled study comparing a group of women with and without protection orders, Grau et al.19 found a non-statistically significant trend toward reduction in further abuse (56% versus 59%) and violence (24% versus 27%). However, they found a significant reduction in a subset of the study group that had experienced less severe prior injuries.19 In another study, McFarlane et al.20 noted that abused women who applied and qualified for a two-year protection order reported significantly lower levels of violence in the subsequent 18 months, regardless of whether the order was implemented.
The most compelling evidence that protection orders are effective5,21 comes from studies that include control groups as well as large sample sizes, long follow-up periods, and more representative samples. In a study involving 2,691 women who reported an incident of intimate partner violence to police, Holt et al.5 found that having a permanent protection order in effect was associated with an 80 percent reduction in police-reported physical violence in the next year. Women with permanent protection orders were significantly less likely than those without protection orders to be physically abused (relative risk over 12 months, 0.2; 95% CI, 0.1–0.8). Similarly, in a prospective study of 448 women who had police-reported intimate partner violence, Holt et al.21 found that the odds of violation after protection order implementation were less for contact (OR = 0.4), threat (OR = 0.4), sustained psychological abuse (OR = 0.5), and physical abuse (OR = 0.3) by the perpetrator when compared with women who had reported intimate partner violence but had not obtained protection orders. These differences were even larger with longer follow-up.21
Variables Associated With Violation of Protection Orders
To inform the mental health clinician's consideration of protection orders, Table 2 summarizes variables that have been associated with violation. Overall, the studies reviewed suggest that the risk of violation of a protection order is associated with the amount of time since placement of the order; characteristics of the victim, abuser, and their relationship; and legal system factors. The following sections highlight important aspects of these associations.
Time
Time is a factor related to protection order violation. Much of the violation activity occurs within the first three months after issuance of the order.5,15–17,21,22 Holt et al.5 noted a quadrupling of psychological abuse during the period of a temporary protection order (the time shortly after an index incident, when most temporary protection orders are issued), although it should be noted that the same authors did not find evidence of increased risk of nonphysical abuse at any point during the 9 month follow-up in a subsequent study from the same setting and a similar time period.21 In their subsequent study, Holt et al.21 found “an apparent dose-response relationship according to the duration of the civil protection order” (Ref. 21, p 21), with a 70 percent reduction in physical abuse and a 60 percent reduction in psychological abuse among women who maintained their protection orders throughout the follow-up. This result is similar to Klein's finding that a lower likelihood of re-abuse is associated with more time since the incident leading to the protection order.17
Victim Characteristics
Several victim characteristics, while not consistent across all studies, have been associated with renewed abuse after placement of the initial protection order. These include socioeconomic status, presence of biological children with the abuser, race/ethnicity, and prior drug use by the victim.16,18,24 Mears24 and Carlson et al.18 found that the socioeconomic status of the victim is related to increased risk of renewed abuse after a protection order has been placed. Women of very low versus low/medium socioeconomic status experience a significantly lesser decline in reported violence after a restraining order (71%–53%, p < .05),18 suggesting that they are at overall higher risk of re-victimization.
The presence of biological children between the victim and defendant may be another important factor to consider. Carlson et al.18 found that the presence of children increases the odds of reporting re-abuse by a factor of 4.5 in relationships lasting less than five years (p < .01). Other investigators found that, although the overall abuse risk is not higher, women with children are more likely to experience violence (OR = 1.7, p < .10) and more likely to be threatened or have property damage (OR = 1.5, p < .10).16 However, the results in this evaluation did not reach accepted statistical significance and so must be considered suggestive.
Race of the victim has also been identified as a significant factor in renewed abuse. Black women are at elevated risk of renewed abuse after legal intervention (PO or arrest of partner for DV incident).18,24 More specifically, Mears reported a relative risk of re-abuse of 2.93 (p < .001) for black women when compared with white women (risk among Hispanics was not higher relative to that among whites).24 Drug use by the victim is also associated with increased re-victimization after legal intervention.24
Finally, in a study examining stalking, the authors found that men with protection orders are more likely to experience violation, although both sexes report violation in excess of 65 percent of the time.14
Perpetrator Characteristics
Several perpetrator characteristics may predict renewed abuse after initiation of a protection order. These include a history of violence or criminal acts,12,15,16,17 being male,15 youthful age,17 less than full-time employment,12 substance abuse,12,15 and other mental health contact.15 Overall, a violent or criminal history is the characteristic most frequently associated with protection order violation. In a limited study by Chaudhuri and Daly,12 men with criminal histories all violated protection orders. In addition, the offender's criminal history, including alcohol and drug convictions, has been associated with placement of protection orders22 and with continued violence afterward.12,15,17 Harrell and Smith16 concluded that more resistance by a defendant to a protection order during court proceedings predicts a higher likelihood of violation of that order, and the level of resistance at the time of issuance increases the probability of severe violence (OR = 3.57, p < .05), threats (OR = 2.87, p < .10), and psychological abuse (OR = 4.12, p < .01).
In contrast with victim characteristics, the race of the defendant was not found to be a factor related to violation of protection orders.15
Relationship Factors
Several characteristics of the abuser-victim relationship may be related to the likelihood of renewed abuse after initial protection order placement. These relate to both the nature of the relationship between the victim and the abuser and to the nature of the abuse itself.
In terms of the nature of the relationship, Harrell and Smith16 reported that cohabitation with the abuser at the time of the original event leading to a protection order reduces the odds of future violation of the protection order (OR = 0.61, p < .005). In their study, the duration of abuse before initiation of the protection order was not related to incidence or type of abuse that occurred afterward. In addition, Carlson et al.18 noted that women in relationships of more than five years are less likely to report re-abuse (OR = 0.12, p < .01) than are women in relationships of less than one year.
Regarding the nature of the abuse, Harrell and Smith found that the severity of abuse before placement of a protection order does not predict its violation. However, they found that “persistence in the pattern of violence” is related to the likelihood of renewed abuse (Ref. 16, p 232). When a violation occurs, the severity of prior abuse is significantly related to the severity of abuse in the year after the original protection order is placed. Similarly, the nature of behavior before the placement of a protection order predicts the occurrence of similar behavior afterward: violent behavior, threats of violence and property damage, and psychological abuse.16
Stalking, while often a component of intimate partner violence, may warrant special attention apart from violence when considering protection orders. Data from a large, national telephone survey investigating stalking indicated that significantly more of the women (28%) than the men (10%) surveyed obtained a protection order in response to having been stalked.14 In the same study, among those (women and men) who obtained a protection order, more than 68 percent reported violation of the order by their stalker.14 In another study evaluating women with protection orders and reported stalking, more severe violence histories, increased violation of orders, and more negative perceptions of protection order effectiveness were noted in a comparison with women who had protection orders but reported that they had not been stalked.23
Legal System Factors
The role of the legal system in protection orders may affect their implementation and effectiveness. The roles of the police, both in arresting suspects and in supporting victims, have been identified as factors related to future violation of protection orders.
Studies vary regarding the association between arrest for the incident leading to a protection order and the risk of future abuse. Harrell and Smith16 reported that the probability of post-protection order abuse is lower if the abuser is arrested at the time of the incident that leads to the protection order. Similarly, Carlson et al.18 found that arrest of the perpetrator before the protection order reduces reported re-abuse for women of low socioeconomic status. However, neither Klein17 nor Mears et al.24 found any association between such arrests and the likelihood of re-abuse.
Harrell and Smith16 reported that the perceived helpfulness of police interventions at the time of the initial abusive incident is related to a reduced probability of severe violence after a protection order is issued, and women who report that they need more protective measures in their orders are more likely to report future serious violence (OR = 1.7, p < .01) and psychological abuse (OR = 1.7, p < .10).
Many have hypothesized that the legal system's responsiveness to violations of protection orders (such as arrest at the time of a protection order violation or prosecution for such an offense) may play an important role in the overall effectiveness of the orders by reducing re-victimization. However, this has not been systematically evaluated.
Discussion
Do protection orders protect? The answer seems to depend on how one understands the question. The diverse responses to the question are reflected in the variation of expert opinions and the conclusions drawn from the data and literature cited herein. While the most widely used measure of the success of protection orders is their rate of violation, this rate varied widely in the studies reviewed.5,11–23 Similarly, in a meta-analysis by Spitzberg25 of 32 studies published before 2002, the rate of violation ranged from 3 to 79 percent with a mean of 40 percent. Although the rate of violation is a concrete measure of efficacy, it may not describe the full impact of protection orders. Other factors, including risk for violence/escalation, timing of violation, and the potential for under-reporting of violations must be considered when assessing the overall effectiveness of protection orders.
Assessing the literature on the topic of protection orders is challenging for many reasons. There is no consistent language or statutory construction for protection orders, making it difficult to identify and then compare the relevant literature. In addition, most of the relevant available literature is published by those in fields not routinely accessed by mental health clinicians and is presented with a nonclinical, sociological or policy-oriented focus.
Methodological differences also limit conclusions derived from the available data related to protection orders. Some of the common limitations include absence of a comparison group without a protection order11–13,15–18,20,23; failure to account for selection bias (i.e., characteristics associated with seeking a protection order may affect the likelihood of future violence and other abuse, limiting conclusions about whether it is the protection order that accounts for the outcomes)18–20,26; studies that rely on official records do not account for violations that may have occurred but were not reported to police5,13,15,17,18,22,24; attrition (some participants are lost to follow-up because they fail to complete study components or because they leave the jurisdiction under study or enter another one that is under study)5,13,15,21,22,24; and reporting bias (interview-based studies can be affected by inaccuracy of self-report).12,14,19–21,23 Our review identified no randomized controlled trials of protection order effectiveness.
Despite the limitations in the existing research literature related to protection orders and their violation, failure to consider the literature as a framework for decision-making would leave the clinician with only speculation or anecdotal experience as a guide.
In this review, we have identified several variables that may help in the clinician's consideration of protection orders. First, available research supports the conclusion that there is a substantial chance that a protection order will be violated, and that the risk of a violation is greatest soon after its initiation, such as during the time span of a temporary order.5,15–17,21,22 Second, the presence of stalking behavior appears to elevate further the risk for protection order violation.14,23,25,26 Third, because future violence after protection order placement can have serious consequences for the victim, the nature of previous violence should be taken into account. That the severity of violence before protection order placement predicts the severity of future violence is an important safety consideration,16 especially in light of the conclusion of Spitzberg25 that approximately 20 percent of protection orders are associated with escalation of violence against the victim and the findings of Chaudhuri and Daly12 that some of the women they interviewed believed that they were beaten in response to having obtained protection orders.
The available literature shows that the time since the initiation of the protection order; potential for escalation or further violence; characteristics of the victim, the abuser, and their relationship; and matters related to the legal system are areas worthy of attention in the consideration of a protection order. In individual cases, some of these factors may inform decision-making, both before and after the placement of a protection order. For instance, if a victim obtains a protection order but the defendant is resistant to the order in court, extra care may be taken to ensure that additional safety plans are put in place.
Protection orders are widely used in an attempt to reduce harm to those who are at risk. Although there is controversy about their effectiveness, available research supports the conclusion that they are associated with reduced risk of violence toward the victim. Despite the limitations of the existing research base, the current literature offers helpful information that can be considered, in conjunction with other risk management strategies, to guide decision-making. Nevertheless, protection orders are only one component of any effective threat-management strategy, and the decision to use these and other tools to promote safety requires thoughtful clinical judgment.
Footnotes
-
Disclosures of financial or other potential conflicts of interest: None.
- American Academy of Psychiatry and the Law