Investigating the Justice System Response to Domestic Violence in Missouri
Brandi L. Byrd1 |
M. Meghan Davidson1 |
Mary M. Beck1 |
Niels C. Beck1 |
Gregory F. Petroski1 |
Introduction
One in four women will experience domestic violence in their lifetimes.2 Each year, 5.3 million domestic violence assaults occur in the United States alone3 and domestic violence is the leading cause of injury to women.4 Yet, despite the prevalence of domestic violence, little empirical research on the justice system’s response to it exists. This paper seeks to describe a state funded project that was created to assess and compare responses to domestic violence throughout the state of Missouri.5 The project lasted for three years and was conducted by an interdisciplinary team of University of Missouri-Columbia (MU) professors and students.
History of the Project
In 1992, the MU School of Law opened the Family Violence Clinic, in which Rule 13 certified law students, supervised by a law professor, represent indigent victims of domestic violence in rural Missouri counties. Over the years, Professor Mary M. Beck, director of the Family Violence Clinic, sometimes observed divergent responses to domestic violence in the various counties, including three particularly disappointing experiences.
In one such instance, a woman was coerced into her former husband’s truck when he threatened to shoot their daughter’s boyfriend with a rifle if she did not follow his orders. She did as she was told and was taken to a remote spot where her ex-husband began raping her with tools he had in the truck. The police arrested the man for assault and for violating the civil protective order that the woman had previously obtained against him. The prosecutor declined to prosecute the ex-husband, including the violation of the ex-wife’s protective order.
Another instance occurred when a woman who had obtained a civil protective order was fired when the respondent harassed her at her workplace and threatened to disrupt the workplace if the manager did not fire the woman.6 Despite the woman’s report to the police, they refused to arrest the man for violating the protective order.
In the third instance, a woman was raped in the bathroom of a bar by a friend of her abusive husband. The woman went to the hospital and obtained a forensic rape examination that could have constituted evidence, but the police told the victim that they would not arrest the man because they did not think the prosecutor would charge him.
Frustrated by these cases, Professor Beck invited a colleague, Professor Kent Collins of the MU School of Journalism, to co-investigate. This collaboration began during the 2002-2003 academic year when students at the MU Schools of Law and Journalism worked together to investigate four Missouri counties’ responses to domestic violence. Relatively marked differences were observed in the rate of responses. This early collaboration produced both a televised investigative series and a Journal of The Missouri Bar article.7 It also precipitated grant funding from the Missouri Department of Public Safety STOP Violence Against Women Program to quantitatively and qualitatively conduct the investigation statewide.
The project consisted of a four-phased protocol involving faculty and students in the Schools of Law, Journalism, Psychology, and Medicine. In phase one, law faculty supervised law students who compiled statewide, county-by-county data on domestic violence incident reports, civil protective orders, domestic violence related prosecutions, population, numbers of police and prosecutors, average income level, and unemployment level. In phase two, social science faculty created computerized databases and applied multivariate statistical procedures to generate lists of predicted versus actual domestic violence response rates for each county. Per capita rates of domestic violence indicators were also calculated together with summary tables that listed statewide means and percentile ranks. In phase three, journalism faculty supervised journalism students traveling into counties of interest and interviewed domestic violence victims, law enforcement officers, prosecutors, judges, and shelter workers. In the fourth phase, faculty and students reported the results of their investigations in newspapers, on television, in scholarly papers, and in presentations to statewide conferences of sheriffs, police, prosecutors, and judges.
Purpose of the Project & Assumptions of Investigators
When working effectively, the justice system’s response to domestic violence typically begins when a law enforcement officer receives a call for help from an abuse victim or a witness to the abuse. Theoretically, the law enforcement officer responds to the scene, investigates, identifies the primary aggressor, and arrests that person.8 The officer offers the victim help, such as a ride to the home of a family member or to a shelter, information on counseling, and information on how to obtain a civil protective order against the abuser. The officer also collects evidence pertinent to the crime. The next aspect of response is initiated by prosecutors, who charge the offender with a criminal offense relying on evidence collected by the police.9 The victim may also petition the circuit court for a civil protective order in which a judge prohibits the abuser “from abusing, threatening to abuse, molesting, stalking, or disturbing the peace” of the victim for up to one year.10 The civil and criminal court actions for domestic violence are independent of one another.11 When the justice system works ineffectively, one or more aspects of this coordinated response go undone and offender conduct may be unchecked, leaving victims vulnerable and unprotected.
This interdisciplinary MU project was created to assess the response of judges, prosecutors, and law enforcement officers to domestic violence in each Missouri county and the City of St. Louis. The investigators selected three variables as indices of the legal system’s response to domestic violence. We chose the first variable, domestic violence incident reports (DVIRs), because the filing of the reports is the responsibility of law enforcement officers. We chose the second variable, civil protective orders filed (CPOs), because ruling on these adult and child abuse petitions is the responsibility of judges. The third variable, domestic violence assaults (DVAs), represented a compilation of crime codes pertaining to domestic violence related crimes. It was chosen because charging these crimes is the responsibility of prosecutors.
The investigators assumed that assessing the response of judges to domestic violence can be accomplished by tallying the number of civil protective order petitions filed in a county and comparing it to that same number in other counties. The assumption was that if judges uniformly apply the adult abuse statute, § 455 RSMo, in civil protective order cases, victims in each county will file a number of civil protective order petitions proportional to that county’s population.
Investigators also assumed that assessing the response of prosecutors to domestic violence could be accomplished by tallying the number of domestic violence related charges filed in a county and comparing it to charges filed in other counties. Investigators included in domestic violence related filings both domestic assaults12 and violations of civil protective orders.13 The assumption was that if prosecutors uniformly charge domestic violence crimes according to Missouri’s criminal statutes, a county’s per capita domestic violence-related charges would be proportional to that county’s population.
Lastly, investigators assumed that assessing the response of law enforcement to domestic violence can be accomplished by tallying the number of domestic violence incident reports from a county and comparing it to the numbers from other counties. The assumption was that if law enforcement uniformly responded to domestic violence calls, a representative number of victims and witnesses would seek law enforcement help and a county’s per capita domestic violence incident reports would be proportional to county population.
Methodology
MU law students gathered annual domestic violence-related data and demographics from each county and the City of St. Louis. To evaluate the police response, students collected the number of domestic violence incident reports in each county. To evaluate the prosecutorial response, students collected the number of prosecutions for domestic violence assaults and civil protective order violations. To evaluate the judicial response, students collected the number of civil protective order petitions filed. Lastly, to fairly compare the counties, students collected data relating to population, income, unemployment rate, number of law enforcement officials, and number of prosecutors in each county. These data were compiled from the Office of State Court Administrator, the Missouri Highway Patrol, United States census data, and phone calls to counties. For the purpose of this paper, we collected data for fiscal years (July through June) 2003/2004 and 2004/2005.
Next, social science students and professors conducted statistical analyses on the data. These analyses were conducted using techniques described more thoroughly below. Overall, the respective data were examined by county to generate an expected rate of response. This expected rate of response was then contrasted with the actual rate of response. This process produced lists of counties that approximated the expected rates, counties that exceeded the expected rates, and counties that fell short of the expected rates. In a parallel and alternative set of analyses, the investigators calculated the per capita rates of civil protective orders (CPOs), domestic violence incident reports (DVIRs), and domestic violence related charges (DVAs) per 100 county residents. Tables of these rates were constructed, with counties rank-ordered based on rate magnitude. Rate data was also broken down in terms of percentile ranks so that any individual county’s rates could be evaluated with reference to statewide norms.
Following the data analyses, students from the MU School of Journalism conducted interviews of police, prosecutors, judges, victims, and shelter workers in selected counties to find out why response rates in those counties were statistically higher or lower than others. These interviews were used to produce television broadcast segments, radio reports, and newspaper articles that were released in an attempt to inform the public of the study, its results, and possible reasons for those results. The professors and students also presented the results of the data and interviews to conferences of Missouri judges, prosecutors, and law enforcement officials.
Analysis of Data
In the current project, variables such as population, unemployment, income, number of law enforcement officials, and number of prosecutors served as the predictor variables while domestic violence incident reports, number of prosecutions for domestic violence assaults and civil protective order violations, and number of civil protective order petitions filed served as the predicted variables.
Because the variables of primary focus were event counts per year (the number of civil protective order petitioners filed, the number of domestic violence-related charges filed, and the number of domestic violence incidents reported), an adjustment for population was made before comparing counties. Two methods of adjusting for differing county sizes were employed.
The first was simply to express the number of events in a per capita metric, specifically the number of events per 100 of county population. Counties were then ranked ordered from lowest to highest on each of the per capita domestic violence measures. Descriptive statistics such as the mean and standard deviations, medians, and selected percentiles of each variable were also calculated.
A second method of accommodating population size was to use regression methods to generate an expected number of events in each county based on a statistical modeling procedure. In this regard, a negative binomial regression model, which is most appropriate for event count data, was used to estimate the number of each type of events per county given county population and percent of unemployment.14 The number of prosecutors and the number of law enforcement officers in a county were also considered as predictor variables, but were not included in the final regression models because they were so highly correlated with population. Given the expected number of events conditional on population and unemployment, counties were then rank-ordered on the percent difference error between their observed and expected performance. The relative percent difference for each of the outcomes was calculated as
#Expected-#Reported
___________________ x 100
#Reported
where the #Expected is derived from the regression model and the #Reported is the actual number of the given event type reported. Nine Missouri counties with populations greater than 100,000 were excluded from the regression analysis, because they are atypical of the state in general and exhibited extreme influence on the regression results.15
The results of the rate analysis are entitled the “UMC Domestic Violence Project Results” and may be found at the Missouri Department of Public Safety website at http://www.dps.mo.gov/home/dpshome.htm. Regression analysis results are available from the authors.
Summary of the Investigation
The data analyses demonstrate that a large number of Missouri counties have an overall response to domestic violence that differs substantially from the expected response. Some counties fall far below the response expected and some counties far exceed the expected response. However, it is important to interpret the data with some caution due to two factors. First, several of the most highly populated counties skewed the data (served as outliers themselves due to population size) and thus were removed from the analyses to control this statistical problem. Second, these statistical analyses do not provide information on causation; that is, the analyses only pinpoint potential outlier counties and do not explain why the county is an outlier. Nevertheless, the analyses demonstrate a great deal of variance between and among counties and suggest a need to further ascertain what is occurring in these selected counties, in terms of their response to domestic violence, that makes them so discrepant from expected values.
A number of factors may be operating to create these outlier counties. MU journalists uncovered several explanations for counties with responses to domestic violence lower than predicted values. For example, the problem may lie in a lack of resources; if a county has a small number of law enforcement officers, it may be nearly impossible to respond efficiently and effectively to each domestic violence call that is received. Also, journalists’ investigations have revealed that not all law enforcement organizations develop incident reports and/or report domestic violence incidents to the Missouri State Highway Patrol. Filing and clerical errors, plus misreports to the Office of State Court Administrator or the Highway Patrol, provide additional explanations for counties which fall below expected response rates.
Additionally, negative bias toward domestic violence cases held by law enforcement officers, prosecutors, and judges serves as a potential explanation. Some systemic issues reduce a county’s responsiveness, such as: one county permitting either prosecution of a domestic assault or the filing of a protective order against the offender, but not both; another county charging the innocent parent, rather than the abusive parent, for guardian ad litem fees if the non-abusive parent wishes to get an order of protection for her/his child; or the refusal of a court to provide the child custody or child support remedies provided by statute. These are only a sample of potential explanations for a particular county to perform below the level expected regarding its response to domestic violence.
Similarly, MU journalists uncovered explanations of why a county might demonstrate a high domestic violence response rate. For example, many of the counties that perform above predictions request and receive grants from the Missouri Department of Public Safety, which they subsequently use to train their responders and purchase equipment, and thereby create a coordinated response to domestic violence. Some of these counties have instituted team response systems that consist of multidisciplinary groups such as law enforcement officers, prosecutors, shelters, victims’ advocates, and attorneys who work collaboratively to effectively and efficiently respond to domestic violence incidents in their communities.16 To illustrate this coordinated effort, law enforcement officials in these counties respond on site to a domestic violence call and make every attempt to arrest the primary aggressor. They take photos of the victim’s injuries and any property damage, such as spray-painted cars, broken telephones, or broken windows. Officers also collect evidence such as crow bars, knives, guns, or other objects used in the commission of the abuse. Either an advocate or law enforcement officer offers to transport the victim to a safe place such as a shelter or friend’s home. Police then write a detailed report and turn the evidence and report over to the prosecutor, who then takes appropriate legal action against the abuser. Victims’ advocates are called to support the victim and discuss Victim of Crime Act benefits, testifying in a prosecution, obtaining a civil protective order, and developing a safety plan. The success of these counties in responding to domestic violence confirms the efficacy of the grant programs.
Additional explanations for counties with a response to domestic violence higher than predicted values may lie in tourism or in prosecution policies. Some counties with a large number of tourists visiting annually may perform above average and such results are more related to a population “bump” than to the justice system’s response. On the other hand, prosecutors’ offices may develop “No Drop” policies, which increase domestic violence crimes charged (with and without victim testimony). Such policies encourage law enforcement to routinely collect evidence and arrest abusers, which enables prosecutors to uniformly charge and successfully prosecute domestic assaults and civil protective order violations. Thus, No Drop policies may result in more domestic assaults prosecuted, as well as more perpetrators arrested. These are a few potential explanations for a particular county to perform above the level predicted.
MU journalists also exposed some issues that interfere with assessing the prosecutors’ and law enforcement responses. Both responders are part of an interdependent system and their conduct may reflect factors outside of their own control. For example, prosecutors are particularly sensitive to influence from judges and police, because a local judge who will not admit hearsay evidence under accepted hearsay evidence exceptions impairs the prosecutor’s ability to prosecute successfully. This may influence the prosecutor’s charging decisions. Also, prosecutors are influenced by law enforcement practices where evidence of domestic crimes is not collected (e.g., taking photographs, collecting weapons used, etc.), and this impairs the ability of the prosecutors to successfully litigate charges and ultimately influences charging decisions. Similarly, law enforcement is discouraged both from responding to domestic violence, particularly on site, and completing the required paperwork when the prosecutor does not routinely file commensurate charges.
Conclusion
Through their interdisciplinary investigative approach, MU professors and students have illustrated considerable variability among Missouri counties regarding the judicial, prosecutorial, and law enforcement responses to domestic violence. The results of this study have provided a new way to quantify the justice system’s responsiveness to a common crime. MU journalists have conducted interviews in and investigated the results of the data analysis in counties of interest, but each county’s judges, prosecutors, and law enforcement officers are best positioned to find the study’s data online at the Missouri Department of Public Safety website and self-evaluate their own county’s position in relation to its predicted response rates. Where a county’s response rate is substantially below counties of comparable demographics, local justice system officials are in the best position to determine why this is the case and to endeavor to improve the system’s responsiveness.
The stakes are high. Domestic violence is the leading cause of injuries to American women.17 The National Coalition Against Domestic Violence reports that “61% of female homicide victims were wives or intimate acquaintances of their killers.”18 Further, victimization extends to children because men who witness domestic violence as children are twice as likely to abuse their own partners and children.19 Missouri reported 39,097 incidents of adult domestic violence and 51 domestic homicides in 2004.20
The finite nature of justice system resources, the complexity of human relationships, including violent ones, and the multidisciplinary nature of these issues confound efforts to identify and quantify a reasonable response to domestic violence. Nonetheless, this investigation has used a broad brush stroke to attempt just that. The Missouri counties data may be used to respond to newspaper reporters, to implement changes in county policies, and to support grant proposals for improving or maintaining their responses to domestic violence.
Footnotes
1 Brandi L. Byrd is an associate at The Copeland Law Firm in Columbia. She is a 2003 graduate of Drury University and a 2006 graduate of the University of Missouri-Columbia School of Law.
M. Meghan Davidson, Ph.D., is an assistant counseling psychology professor at the University of Nebraska at Lincoln. She earned her doctorate in counseling psychology from the University of Missouri-Columbia.
Mary M. Beck is a professor of clinical law at the University of Missouri School of Law in Columbia, where she directs the Family Violence Clinic. She earned undergraduate and graduate nursing degrees at St. Louis University in 1970 and in 1971, respectively, and graduated from the University of Missouri School of Law in 1988.
Niels C. Beck received his Ph.D. in clinical psychology from St. Louis University in 1977. He is a professor of psychiatry at the University of Missouri-Columbia Medical School and chief psychologist at the University of Missouri Health Sciences Center.
Gregory F. Petroski, Ph.D., is a statistician with the Office of Medical Research in the University of Missouri-Columbia School of Medicine. He is also research assistant professor in the Department of Health Management and Informatics.
2 Patricia Tjaden & Nancy Thoenne, Extent, Nature, and Consequences of Intimate Partner Violence, National Institute of Justice and Centers of Disease Control and Prevention (NIJCDC 2000). Available at http://www.ncjrs.gov/pdffiles1/nij/181867.pdf.
3 National Center for Injury Prevention and Control, Costs of Intimate Partner Violence Against Women in the United States, Atlanta (GA): Centers for Disease Control and Prevention; 2003. Available at http://www.cdc.gov/ncipc/pub-res/ipv_cost/IPVBook-Final-Feb18.pdf.
4 Domestic Violence in the Workplace Statistics, American Institute on Domestic Violence, 2001. Available at http://www.aidv-usa.com/Statistics.htm.
5 Missouri has 114 counties and the City of St. Louis.
6 The clinic works to help domestic violence victims develop action plans designed to increase their independence. The victim in this case had just landed her first job in many years as part of her action plan when the respondent’s harassment resulted in her being fired after her second day at work.
7 Rebekah E. Bromberg & Brendan O. Kelley, Investigating the Response to Domestic Violence in Four Missouri Counties, 60 J. MoBar 70 (2004); Battered Justice investigative series aired in 2004.
8 Bromberg, note 6.
9 Id.
10 Section 455.045, RSMo 2006. The judge may order other remedies including exclusive possession of the home, child custody and visitation, counseling, division of property, and payment of certain expenses including maintenance, child support, court costs, and attorney’s fees.
11 Section 455.070, RSMo 2006.
12 Sections 565.063, 565.072 - 565.074, RSMo 2006.
13 Section 455.085, RSMo 2006.
14 A. Colin Cameron & Pravin K. Trivedi, Regression Analysis of Count Data (Cambridge University Press 1998).
15 Those nine counties include Boone, Clay, Greene, Jackson, Jasper, Jefferson, St. Charles, St. Louis City, and St. Louis County.
16 Boone County’s DOVE (Domestic Violence Enforcement) Unit is an interdisciplinary team that coordinates its response to domestic violence.
17 Note 3.
18 See Homicide & Domestic Violence Facts: When Men Murder Women (2004); http://www.ncadv.org/files/WhenMenMurderWomen2004.pdf. Last visited 08/21/2007.
19 http://www.ncadv.org/files/childrenandchildcustody.pdf, last visited 08/27/2007.
20 See Missouri Domestic Violence Facts (2004); http://www.ncadv.org/files/Missouri.pdf. Last visited 08/21/2007.