Abstract
Objectives. We compared types and frequencies of intimate partner violence experienced by women before and after receipt of a 2-year protection order.
Methods. Participants were 150 urban English- and Spanish-speaking Black, Hispanic, and White women who qualified for a 2-year protection order against an intimate partner.
Results. One woman committed suicide 6 weeks into the study. The remaining 149 women completed all interviews. Results showed significant reductions in threats of assault, physical assault, stalking, and worksite harassment over time among all women, regardless of receipt or nonreceipt of a protection order.
Conclusions. Abused women who apply and qualify for a 2-year protection order, irrespective of whether or not they are granted the order, report significantly lower levels of violence during the subsequent 18 months.
Abused women use a variety of methods in seeking assistance to halt violence inflicted upon them, including court orders of protection. Such orders restrict the access of 1 person (e.g., a male abuser) to another person (e.g., an abused woman) for a specified time. (The synonym “restraining order” is used in some jurisdictions.) Protection orders, both temporary and permanent, represent public documentation that abuse has occurred, and if the order is violated, the assailant is subject to prosecution. A protection order offers the victim legal action when the victim does not want the abuser charged criminally or jailed for an offense. However, choice of this action does not preclude other civil or criminal action.
Results of research on the effectiveness of protection orders are inconsistent. We identified 8 recent longitudinal studies that measured additional intimate partner violence committed against women after a protection order had been filed. Six of the studies reported positive results,1–6 meaning that the respondents felt the protection order helped to end or reduce the violence. The remaining 2 studies reported high reassault rates after filing of the protection order.7,8 In the case of most of these studies, low response rates, short follow-up periods, and lack of comparison groups do not allow generalizations to be made.
In addition, we did not identify any studies that included non–English-speaking women or measures of worksite harassment. To test the effectiveness of protection orders, we entered into a partnership with a local district attorney’s office in a large urban city in an attempt to determine whether women who are granted a 2-year protection order experience lower levels of violence than women who apply and qualify for such an order but are not granted one.
METHODS
Our study was conducted from January 2001 to June 2002 at a special family violence unit of the Houston, Tex, district attorney’s office that serves an ethnically diverse population of 3 million citizens. The primary service of the family violence unit is processing of protection orders. During the 12 months preceding this study, 2932 women applied to the unit for a protection order; 1980 (68%) met qualifying criteria, and 962 (49%) were granted the protection order. Qualification criteria for protection orders are set by state law and include applicants providing evidence (i.e., police or witness report, visible injury) that the respondent (e.g., abuser) has been violent with them and is likely to continue this violence toward them. In addition, the applicant must have previously lived with the abuser in the same household, or they must be the biological parents of the same child.9
If the applicant’s case is accepted, the attorneys file the case with the family law court and ask for a court date to be set for a hearing. After the case has been filed, the court issues a temporary protection order. A copy of this order is sent to the applicant by mail, and a copy is served to the abuser in person. The temporary protection order is similar to the final 2-year protection order in that it informs the abuser that he or she must stay 200 ft (60 m) away from the applicant’s home and workplace and prohibits the respondent from assaulting the applicant, from threatening the applicant directly or through another person, and from harassing or stalking the applicant.
However, the temporary protection order differs from the final 2-year protection order in that a violation of the temporary order cannot be charged as a criminal offense; it can be filed only as a civil contempt of court. Furthermore, the temporary protection order is valid for only 20 days. The court date is set within those 20 days, and the order expires whether or not the abuser is served or the hearing takes place. However, the temporary protection order may be extended if the abuser is not served by the hearing date.
The applicant is not responsible for any fees in association with the protection order. The order is granted for 2 years and can result in both criminal and civil penalties if violated. Applicants are informed at the time of application as to whether they do or do not qualify to receive the order. All qualifying applicants are assigned to a case worker who provides them with educational information about violence and safety planning as well as information regarding community resources (e.g., emergency shelters, counseling, legal and medical assistance). Applicants are encouraged to contact the case worker for further questions about the protection order process.
All women who presented to the special family violence unit at the district attorney’s office to apply for a protection order, completed the application process, qualified for the protection order, and met our inclusion criteria (e.g., female, 18 years or older, English or Spanish speaker) were invited into the study by 1 of the 6 investigators until 150 women agreed to participate and were entered into the study. Four women refused to participate. One woman committed suicide 6 weeks into the study. All of the remaining 149 women completed the 3-month, 6-month, 12-month, and 18-month follow-up interviews, resulting in a retention rate of 99%.
Instruments
Demographic data form.
This form was used to document information on participants’ age, education, income, self-identified race/ethnicity, employment status, relationship to the abuser, and primary language.
Severity of Violence Against Women Scales (SVAWS).
This 46-item instrument is designed to measure threats of physical violence (19 items) and physical assault (27 items).10 Examples of behaviors that threaten physical violence are threats to destroy property, do bodily harm, or harm other family members. Examples of behaviors that represent physical violence are kicking, choking, beating up, and engaging in forced sex. For each item, respondents use a 4-point scale to indicate how often the behavior occurred (1 = never, 2 = once, 3 = 2–3 times, 4 = 4 or more times). Possible score ranges were 19 to 76 for the threat of abuse dimension and 27 to 108 for the physical abuse dimension. The higher the score was, the more violence that was reported.
Internal consistency reliability estimates in studies of abused women have ranged from 0.89 to 0.91 for the threat of abuse dimension and from 0.91 to 0.94 for the physical abuse dimension.5,11,12 In the present study, reliabilities (measured with Cronbach α coefficients) were 0.91 for the threat of abuse dimension and 0.94 for the physical abuse dimension.
Stalking Victimization Survey.
This 17-item yes/no questionnaire was used to document the frequency and type of stalking engaged in by the perpetrator. The initial stalking survey instrument consisted of 7 items (e.g., being followed or spied on, being sent unsolicited letters or written correspondence, or finding the perpetrator standing outside one’s home, school, or workplace) developed by Tjaden and Thoennes13; 10 items were added from the Sheridan14 HARASS instrument to form the overall 17-item instrument used here. Examples of items added include threats by the abuser to harm the children or to commit suicide if the woman left the relationship, leaving threatening notes on the woman’s car, and threatening her family. The possible score range was 0 to 17. In this study, reliability (Cronbach α coefficient) was 0.83.
Danger Assessment Scale.
This instrument, which consists of 15 items with a yes/no response format, assists women in determining their potential risk of becoming a femicide victim.15 All of the items refer to risk factors that have been associated with murder in situations involving abuse. Examples of risk factors include the abuser’s possession of a gun, use of drugs, and violent behavior outside the home. The possible score range was 0 to 15. Scale reliability coefficients have ranged from 0.60 to 0.86 in several studies.16 In this study, the reliability (Cronbach α coefficient) was 0.67.
Worksite harassment.
Eight yes/no questions were asked about worksite harassment. Questions were derived from a congressional report17 that reviewed studies of worksite harassment of women by intimate partners. Questions focused on, for example, repeated calls/visits to the woman’s worksite and difficulties experienced by the woman in regard to going to work. The possible score range was 0 to 8. Reliability (Cronbach α coefficient) was measured as 0.76.
Procedure
Data collection began after institutional review board approval had been received and consent had been obtained from the district attorney’s office. Women meeting the study criteria were escorted to a private room in the offices of the family violence unit where the investigators provided an explanation of the study’s purpose, protocol, instruments, administration time, and follow-up schedules. Women who agreed to take part in the study signed an informed consent form, and the investigators administered the study instruments. Instruments were offered in English and Spanish according to women’s language preference. All measures focused on women’s reports of violence and health status during the preceding 3 months.
To assist in maintaining contact with each of the women, we formed a safe contact list of at least 6 persons the woman granted permission for us to contact in the event she could not be reached. This list consisted of close relatives (i.e., mother, grandmother, sister, and adult children), neighbors, friends, work colleagues, and other acquaintances. In each case, name, relationship, address (home and work), and contact telephone numbers (i.e., home, work, and cellular) were listed. When contacted, the person was told that the woman was involved in a health study and had given permission for the researcher to contact individuals who may know of her current address/telephone number(s). During all subsequent interviews with the women, both their contact information and that of each safe contact were reviewed and updated.
The safe contact list proved the best method for maintaining contact with the women over the 18-month study period. We ensured women’s safety in completing the follow-up telephone interviews by establishing a convenient, private, and safe time for these interviews. A safety protocol was used for each follow-up telephone interview. Women were reimbursed $20 for the first interview; $30 for the 3-month interview; $40 for the 6-month interview; $50 for the 12-month interview; and $60 for the 18-month interview. They were reimbursed an extra $40 for completing all of the interviews.
Data Analyses
Means, standard deviations, and frequencies were used in descriptions of the demographic characteristics of the 150 women who applied for protection orders. We conducted independent t tests to determine whether the women who were granted an order differed significantly in terms of age or years of education from the women who were not granted an order. Chi-square analyses were used to determine whether the groups of women differed significantly with respect to race/ethnicity, income, employment status, or status of relationship with abuser.
Using Cohen’s power analyses and tables,18 we calculated the a priori power of our between-groups repeated measures analysis of variance to produce a small-to-moderate multivariate effect size by conducting a multivariate analysis of variance with 4 dependent variables (i.e., the 4 score differences from intake scores). Given a significance level of .05, 150 participants, 4 dependent variables, 2 groups, and the goal of a small-to-moderate effect size, we calculated the power of the analysis as 91%. Assumptions of independent observations, normality, and homogeneity of (co)variance were examined. Results indicated that the study’s robustness, procedure, number of participants, and sample size ratio satisfied these assumptions.
Violence scores for women who were granted or not granted the protection order were subjected to repeated measures analyses. We initially considered as covariates demographic characteristics that exhibited significant between-groups differences at intake and were shown to be univariately associated with the dependent variables. However, we retained only significant covariates in the final analyses. We conducted a 1-factor repeated measures multivariate analysis of covariance on SVAWS scores to accommodate the 2 dependent subscales (threats of violence and physical violence scores). We performed 1-factor repeated measures analyses of covariance (ANCOVAs) on danger, stalking, and work harassment scores.
We calculated adjusted means, standard deviations, and multivariate effect sizes (0.02 = small, 0.10 = small to moderate, 0.15 = moderate, 0.35 = large).18 To achieve a balance between type I and type II error, we set the significance level at .025 for each SVAWS subscale. In the case of within-group (time, Group × Time interaction) contrasts, we set significance levels at .006 for subscale scores and .0125 for stalking, danger, and worksite harassment scores.
RESULTS
The women were stratified into 2 groups: those who were granted a 2-year protection order (n = 81) and those who were not granted such an order (n = 69). Reasons for nonreceipt of the protection order were as follows: the woman dropped the order (n = 40), inability to locate the abuser and serve papers to appear in court (n = 18), and dismissal of cases (n = 11).
Reasons for Not Being Granted a 2-Year Protection Order
Forty women dropped the protection order before their court date. Most did so because they returned to the relationship with the abuser or because the protection order process was “too much of a hassle” or “inconvenient.” To obtain a protection order, applicants must be willing to arrive at the district attorney’s office with proper photo identification and complete paperwork, and they are required to complete an interview with a caseworker, be photographed, and sign an affidavit. This process requires about 2 to 3 hours. Applicants must wait approximately 6 weeks for a court date and then appear in court in front of a judge, at which time the abuser may contest the protection order.
In addition, many women need to return to the district attorney’s office at a later date with additional required paperwork/witnesses to the abuse. For some women, these trips to the district attorney’s office mean work absences and loss of income. We did not ask the participants in our study who had dropped the protection order when they did so; however, at the 3-month interview, many women reported dropping the order within the first 2 weeks after application.
Eighteen women were not granted a protection order because the abuser could not be found and served papers to appear in court. Eleven women were not granted the order because their case was dismissed. Six cases were dismissed by the district attorney’s office owing to incomplete applications (e.g., required documents not being supplied). Seven cases were dismissed by the judge, 2 because the protection order was contested by the abuser and the remaining 5 because the women did not appear in court. One of these women committed suicide. As mentioned, the remaining 149 women completed the 4 follow-up interviews, for a retention rate of 99%.
Between-Groups Differences in Demographic Characteristics and Violence Scores
Frequencies, percentages, and the results of tests assessing demographic differences among women who were and were not granted a 2-year protection order are shown in Table 1 ▶. Relationship status was significantly (χ21 = 4.407, P = .036) associated with receipt of a protective order. Slightly more than half of the women who were granted a protection order were involved in relationships, as compared with 71% of women who were not granted the order. No other significant differences were found.
TABLE 1—
Demographic Characteristics and Results From χ2 Tests of Independence Assessing Differences Between Women Who Were Granted (n = 81) and Not Granted (n = 69) a 2-Year Protection Order
| Characteristic | Protection Order | No Protection Order | Total No. | Test Statistic (χ2 or t) (P) |
| Age, y, mean (SD) | 33.5 (9.2) | 31.2 (9.1) | 150 | 1.503a (.135) |
| Education, y, mean (SD) | 11.7 (3.0) | 11.9 (2.6) | 150 | 0.503a (.615) |
| Race/ethnicity, No. (%) | 3.320b (.190) | |||
| African American | 31 (38.3) | 18 (26.1) | 49 | |
| White | 22 (27.2) | 18 (26.1) | 40 | |
| Latino/Hispanic | 28 (34.6) | 33 (47.8) | 61 | |
| Relationship status, No. (%) | 4.407b (.036) | |||
| Current spouse/boyfriend | 44 (54.3) | 49 (71.0) | 93 | |
| Ex-spouse/friend | 37 (45.7) | 20 (29.0) | 57 | |
| Family income, $, No. (%) | 0.672b (.412) | |||
| ≥19 000 | 22 (32.8) | 22 (40.0) | 44 | |
| >19 000 | 45 (67.2) | 33 (60.0) | 78 | |
| English speaking, No. (%) | 0.831b (.362) | |||
| No | 15 (18.5) | 9 (13.0) | 24 | |
| Yes | 66 (81.5) | 60 (87.0) | 126 | |
| Employed, No. (%) | 2.916b (.088) | |||
| No | 8 (9.9) | 2 (2.9) | 10 | |
| Yes | 73 (90.1) | 67 (97.1) | 140 |
at test.
bχ2 test.
Adjusted means and standard deviations for violence scores at intake and at 3, 6, 12, and 18 months among women who were granted (n = 81) and not granted (n = 69) the 2-year protection order are shown in Table 2 ▶. After adjustment for age, race/ethnicity, and relationship status, results of the multivariate analysis of variance focusing on SVAWS scores yielded a significant (F8,1144 = 16.123, P < .001) multivariate main effect for time. The magnitude of this multivariate effect size was in the small-to-moderate range (0.10). Univariate tests revealed a significant main effect for both SVAWS subscales: threats (F4,572 = 19.077, P < .001) and physical abuse (F4,572 = 36.261, P < .001). The group main effect and the Group × Time interaction were not significant. The effect size between the groups was small (0.02). Examination of withinsubject contrasts showed that intake scores were significantly (P < .001) higher than subsequent scores.
TABLE 2—
Adjusted Means and Standard Deviations for Violence Scores at Intake and 3, 6, 12, and 18 Months: Women Who Were Granted (n = 81) and Not Granted (n=69) a 2-Year Protection Order
| Measure and Group | Intake, Mean (SD) | 3 Months, Mean (SD) | 6 Months, Mean (SD) | 12 Months, Mean (SD) | 18 Months, Mean (SD) |
| SVAWSa | |||||
| Threats of abusea | |||||
| No order | 44.7 (13.7) | 21.6 (7.7) | 20.7 (7.1) | 21.4 (8.6) | 21.9 (8.1) |
| Order | 47.5 (13.7) | 23.1 (7.8) | 22.9 (7.2) | 24.9 (8.7) | 22.7 (8.1) |
| Physical abuseb | |||||
| No order | 49.2 (17.0) | 27.7 (5.5) | 27.54 (4.6) | 27.2 (7.4) | 28.3 (7.6) |
| Order | 48.5 (17.0) | 29.2 (5.5) | 28.46 (4.6) | 31.1 (7.5) | 29.2 (7.6) |
| Stalkingb | |||||
| No order | 7.7 (4.0) | 2.2 (3.3) | 1.4 (2.8) | 1.9 (3.0) | 2.1 (3.0) |
| Order | 7.0 (4.0) | 3.0 (3.3) | 1.8 (2.8) | 2.4 (3.0) | 1.6 (3.0) |
| Dangerb | |||||
| No order | 7.1 (3.0) | 1.5 (2.2) | 1.2 (2.0) | 1.1 (2.1) | 1.4 (2.1) |
| Order | 7.1 (3.0) | 2.2 (2.2) | 1.7 (2.0) | 2.0 (2.1) | 1.6 (2.1) |
| Worksite harassmentc | |||||
| No order | 3.7 (1.8) | 2.1 (1.5) | 1.5 (1.1) | 1.4 (1.1) | 1.3 (0.8) |
| Order | 4.3 (1.8) | 2.1 (1.5) | 1.2 (1.1) | 1.6 (1.1) | 1.3 (0.8) |
Note. One participant committed suicide; analyses were performed on a sample of 149 participants. SVAWS = Severity of Violence Against Women Scales.
aAdjusted for age, race/ethnicity, and relationship.
bAdjusted for age and race/ethnicity.
cAdjusted for relationship status.
After adjustment for age and race/ethnicity, repeated measures ANCOVAs showed significant (F4,141 = 16.17, P < .001, and F4,141 = 18.33, P < .001, respectively) effects over time of stalking and danger scores. The time effect size for stalking was in the medium-to-large range (0.31), and the effect size for danger was moderate (0.18). There was no significant group main effect or significant Group × Time interaction. Between-groups effect sizes were zero or small (0.02). After adjustment for relationship status, the repeated measures ANCOVA of work harassment scores also showed a significant (F4,80 = 13.88, P < .001) effect over time. There was no significant group main effect or significant Group × Time interaction. The between-groups effect size was zero. Examination of within-subject contrasts for the main effect of time showed that intake scores were significantly (P < .001) higher than subsequent scores.
Protection Order Violations
Finally, women were asked, at each interview, whether a violation of the 2-year protection order had occurred since the previous interview. Among the 81 women granted a protection order, 36 (44%) reported at least 1 violation over the 18 months of the study. Violations were reported by 17 women (21%) at 3 months, 16 women (20%) at 6 months, 20 women (25%) at 12 months, and 19 women (23%) at 18 months. Four women (5%) reported a violation during each of the 4 time periods measured. Most violations involved nonadherence to the order to stay 200 ft from the woman’s home or workplace; stalking, threats of violence, and a combination of these infractions were other examples of violations. Women reporting a violation also were asked whether they had called the police. Among these 36 women, 21 (58%) had called the police at least once to report a violation.
DISCUSSION
The 149 women who took part in this study reported significantly lower levels of intimate partner violence, including worksite harassment, up to 18 months after applying for a protection order. Whether women were granted or not granted the protection order made no significant difference in terms of the amount of violence they reported at the time of application for the order or during the subsequent 3, 6, 12, or 18 months. Forty-four percent of the women granted a 2-year protection order reported at least 1 violation over the 18-month study period, and half of these women reported the violation to the police.
This study followed women after they had qualified for a protection order, irrespective of whether or not they were granted the order. Our results agree with those of others5,19 reporting significantly lower levels of violence experienced by women seeking assistance from the justice system, irrespective of the justice system outcome. One other study, to our knowledge, involved the use of victim interviews to measure levels of violence toward women granted and not granted a protection order.20 Although this study reported that violence frequency was not significantly decreased by receipt of a protection order, the study’s low response rate and short follow-up period limited the generalizability of the findings. Other researchers have focused only on women who received an order of protection against the abuser1–4,7 or have relied solely on police reports.6 Because fewer than half of abused women ever report intimate partner violence to law enforcement personnel,21 relying on police reports may severely underrepresent levels of violence experienced by women both with and without a protection order.
Our findings of significant reductions in violence scores over time among all of our participants, regardless of receipt or nonreceipt of the protection order, are consistent with abuse intervention findings reported by social and health researchers. In one study, abused women exiting a shelter and receiving home social support were compared, at 6 months, with abused women not receiving such support; women in both groups reported decreases in physical abuse.22 In 2 health clinic studies involving comparisons of abused women receiving intensive counseling and outreach support and abused women offered a wallet-sized card listing community abuse resources, women in both groups reported significantly lower levels of abuse at 6, 12, and 18 months postintervention.23,24 Although we found no other studies with which to compare our results, the economic implications of the significant decline in worksite harassment experienced by abused women after contact with the justice system merit further research.
Do these findings indicate that the justice intervention of a protection order and the health and social service interventions of counseling, support, and referrals are no more of a deterrent to future violence than an abused woman’s contact with assistance agencies? When an abused woman decides to contact a criminal justice, civil justice, health, or social service agency, information about the abuse is shared, and contact is made. Just as the privatization of domestic violence contributes to its continuation, perhaps the contact and public knowledge stemming from justice encounters can prevent reoccurrence of violence. Perhaps just as legal sanctions (e.g., requirements involving the use of helmets and seat belts) have proven effective in reducing unintentional injuries, such sanctions can reduce the occurrence of intentional intimate partner violence.
An earlier qualitative study focusing on why women seek civil orders of protection revealed a desire among women to regain some measure of control in their lives by making the abuse public.25 These women discussed using the application for a protection order as a “loudspeaker” to notify the abuser that the law knew about his behavior. They viewed the legal system as a force larger than themselves and as having power over the abuser that they themselves had lost as a result of the abuse. Moreover, they felt a need to have the legal system both approve and reinforce their decision to leave the abuser. The protection order becomes an announcement that the abused woman refuses to “take it” anymore and is acting on her own behalf. Our results appear to quantify these qualitative findings. Once a woman applied and qualified for a protection order, a rapid and significant decline in violence scores occurred and was sustained for 18 months.
Our study involved limitations that are important to the generalizability of the findings. Our sample was small and limited to women from a single urban agency who were seeking assistance. Furthermore, we relied exclusively on self-reports, possibly leading to underreporting as a result of inadequate recall or lack of voluntary disclosure. If we are to learn more about the occurrence of intimate partner violence in the absence of justice system contact, there is a need for future research with larger, representative samples of abused women that include those who are victimized but do not apply for a protection order. In addition, replication is essential in rural settings with diverse ethnic groups. Despite these limitations, our urban sample of English- and Spanish-speaking women demonstrates the important effect of justice system contact in terms of reductions in future episodes of violence.
CONCLUSIONS
Ensuring the safety of victims of intimate partner violence is of utmost importance to health care providers, justice agencies, shelter workers, and other service providers. This study clearly demonstrates that, irrespective of whether or not a 2-year protection order was granted, abused women who sought a protection order reported significantly lower levels of threats of abuse, physical abuse, stalking, work harassment, and risk factors for femicide at 3, 6, 12, and 18 months after their initial contact with the justice system.
Acknowledgments
This project was supported by grant 200-WT-VX-0020 awarded by the National Institute of Justice, Office of Justice Programs, US Department of Justice.
We wish to thank the Family Criminal Law Division of the Harris County District Attorney’s Office for its unflagging support and assistance in collecting the data for this study. We also thank the 149 women who shared their stories and maintained contact with the research team for 18 months.
Note.The points of view offered in this article are those of the authors and do not necessarily represent the official positions or policies of the US Department of Justice.
Human Participant Protection This study was approved by the Texas Woman’s University institutional review board. All participants provided informed consent according to the guidelines specified by the Texas Woman’s University institutional review board.
Contributors J. McFarlane conceived the study, supervised all aspects of its implementation, and wrote the first draft of the article. A. Malecha assisted with supervision of all aspects of the study, synthesized the analysis, and edited the drafts. J. Gist assisted with data collection and managed the study. K. Watson completed the data analyses. E. Batten translated all instruments and collected and coded data on Spanish speakers. I. Hall and S. Smith collected data, coded data, and synthesized analyses. All of the authors helped to conceptualize ideas, interpret findings, and review drafts of the article.
Peer Reviewed
References
- 1.Kaci JH. Aftermath of seeking domestic violence protection orders: the victim’s perspective. J Contemp Criminal Justice. 1994;10:201–219. [Google Scholar]
- 2.Keilitz SL, Hannaford PL, Efkeman HS. Civil Protection Orders: The Benefits and Limitations for Victims of Domestic Violence. Williamsburg, Va: National Center for State Courts; 1997.
- 3.Ptacek J. Battered Women in the Courtroom: The Power of Judicial Responses. Boston, Mass: Northeastern University Press; 1999.
- 4.Carlson JJ, Harris SD, Holden GW. Protective orders and domestic violence: risk factors for re-abuse. J Fam Violence. 1999;14:205–226. [Google Scholar]
- 5.Gist J, McFarlane J, Malecha A, et al. Protection orders and assault charges: do justice interventions reduce violence against women? Am J Fam Law. 2001;15:59–71. [Google Scholar]
- 6.Holt VL, Kernie MA, Lumley T, Wolf ME, Frederick PR. Civil protection orders and risk of subsequent police-reported violence. JAMA. 2002;288:589–594. [DOI] [PubMed] [Google Scholar]
- 7.Harrell A, Smith BE. Effects of restraining orders on domestic violence victims. In: Buzawa ES, Buzawa CG, eds. Do Arrests and Restraining Orders Work? Thousand Oaks, Calif: Sage Publications; 1996:214–242.
- 8.Klein AR. Re-abuse in a population of court-restrained male batterers: why restraining orders don’t work. In: Buzawa ES, Buzawa CG, eds. Do Arrests and Restraining Orders Work? Thousand Oaks, Calif: Sage Publications; 1996:192–213.
- 9.Texas Family Code, Title 4, ch 71 (1995).
- 10.Marshall LL. Development of the Severity of Violence Against Women Scales. J Fam Violence. 1992;7:103–121. [Google Scholar]
- 11.Wiist W, McFarlane J. Severity of spousal and intimate partner abuse to pregnant Hispanic women. J Health Care Poor Underserved. 1998;9:248–261. [DOI] [PubMed] [Google Scholar]
- 12.Wiist W, McFarlane J. Use of police by abused pregnant Hispanic women. Violence Women. 1998;4:677–693. [Google Scholar]
- 13.Tjaden P, Thoennes N. Stalking in America: Findings From the National Violence Against Women Survey. Washington, DC: US Dept of Justice; 1998. [DOI] [PubMed]
- 14.Sheridan D. Measuring Harassment of Battered Women [dissertation]. Portland, Ore: Oregon Health Sciences University; 1998.
- 15.Campbell JC. Assessment of risk of homicide for battered women. Adv Nurs Sci. 1986;8:36–51. [DOI] [PubMed] [Google Scholar]
- 16.Campbell JC. Assessing Dangerousness: Violence by Sexual Offenders, Batterers, and Child Abusers. Thousand Oaks, Calif: Sage Publications; 1995.
- 17.Domestic Violence: Prevalence and Implications for Employment Among Welfare Recipients. Washington, DC: US General Accounting Office, Health, Education, and Human Services Division; 1998. GAO/HEHS publication 99-12.
- 18.Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988.
- 19.McFarlane J, Willson P, Lemmey D, Malecha A. Women filing assault charges on an intimate partner: criminal justice outcome and future violence experienced. Violence Women. 2000;6:396–408. [Google Scholar]
- 20.Grau J, Fagan J, Wexler S. Restraining orders for battered women: issues of access and efficacy. In: Schweber C, Feinman C, eds. Criminal Justice Politics and Women: The Aftermath of Legally Mandated Change. New York, NY: Haworth Press Inc; 1985:13–28.
- 21.Tjaden P, Thoennes N. Extent, Nature, and Consequences of Intimate Partner Violence. Washington, DC: National Institute of Justice; 2000.
- 22.Sullivan C, Campbell R, Angelique H, Eby K, Davidson W. An advocacy intervention program for women with abusive partners: six-month follow-up. Am J Community Psychol. 1994;22:101–122. [DOI] [PubMed] [Google Scholar]
- 23.McFarlane J, Soeken K, Wiist W. An evaluation of interventions to decrease intimate partner violence to pregnant women. Public Health Nurs. 2000;17:443–451. [DOI] [PubMed] [Google Scholar]
- 24.Parker B, McFarlane J, Soeken K, Silva C, Reel S. Testing an intervention to prevent further abuse to pregnant women. Res Nurs Health. 1999;22:59–64. [DOI] [PubMed] [Google Scholar]
- 25.Fischer K, Rose M. When “enough is enough”: battered women’s decision making around court orders of protection. Crime Delinquency. 1995;41:414–429. [Google Scholar]