Victim Service Programs in the STOP Formula Grants Program

Services Offered and Interactions with Other Community Agencies

Read complete document: PDF


PrintPrint this page
Share:
Share on Facebook Share on Twitter Share on LinkedIn Share on Digg Share on Reddit
| Email this pageE-mail
Document date: October 01, 2000
Released online: October 01, 2000

The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.

This report was prepared for the U.S. Department of Justice, National Institute of Justice, under grant number 99-WT-VX-0010. The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders, or to the U.S. Department of Justice.

This report is available in its entirety in the Portable Document Format (PDF), which many find convenient when printing.


TABLE OF CONTENTS

HIGHLIGHTS

Purpose
Who, What, Where, and When?
Why This Study Is Important
How Was the Information for This Report Collected?
Key Findings

   STOP-Funded VS Programs and Their Agencies
   VS Program Interactions with Other Community Agencies
   Impact of STOP on Service Provision

Implications for Research
Implications for Practice
Implications for Policy

CHAPTER 1: INTRODUCTION

Introduction
Victim Service Programs
The STOP Program
The Design of this Project
First-Year Activities

CHAPTER 2: METHODS

Introduction
Subgrant Award and Performance Reports
Choosing the Eight Impact States
Victim Services Program Survey
     Sample
     Measures

CHAPTER 3: VS PROGRAMS IN THE SAMPLE AND IN STOP OVERALL

Introduction
Sample Criteria
Varieties of Victim Service
     Intended Direct Beneficiaries
     Crime Focus
     Geographic Focus
     Additional Sources of VS Program Funding
Conclusion

CHAPTER 4: STOP-FUNDED VS PROGRAMS AND THEIR AGENCIES

Introduction
Overview of Host Agencies
     Host Agency Services
     Host Agency Staffing, Paid and Volunteer
The STOP-Funded VS Program
     VS Program Activities
     STOP Funding for VS Programs
VS Program Functioning
     Service Intensity
     Accessing the VS Program
     Making Referrals to Other Agencies
     Outreach
     Victims Served
Conclusion

CHAPTER 5: VS PROGRAM INTERACTIONS WITH OTHER COMMUNITY AGENCIES

Introduction
VAW Services in VS Program Communities
     Partner Agencies for VS Programs
VS Program Interactions with Law Enforcement
VS Program Interactions with Prosecution
VS Program Interactions with Other VS Agencies
VS Program Interactions with their Primary Agencies
     Staff Involvement and Changed Interactions with Primary Agencies
Other than Law Enforcement, Prosecution, and Victim Services
     Communication, Coordination, and Collaboration Activities of VS Programs and Their Primary Agencies
The Role of Task Forces
Conclusion

CHAPTER 6: IMPACT AT THE COMMUNITY LEVEL

Introduction
Independent and Dependent Measures within the Conceptual Framework
     Box 1: Level of STOP Funds and Other Resources
     Box 2: Pre-STOP Level of Community Services
     Box 3: State STOP Program Support for Collaboration
     Box 4: Level of Coordination in Community Response
     Box 5: Post-STOP VS Program Services
     Box 6: Post-STOP Legal System Response to Victims
Descriptive Analysis
     Box 1 with Boxes 4, 5, and 6
     Boxes 2, 3, and 4
     Boxes 2, 3, and 4 with Box 5
     Boxes 2, 3, and 4 with Box 6
Impact Analysis
     Predictive Models: Boxes 2 and 3 Predicting Box 4
     Predictive Models: Boxes 2, 3, and 4 Predicting Box 5
     Predictive Models: Boxes 2, 3, and 4 Predicting Box 6
     The Conceptual Framework Revisited
Conclusions
     Other Evidence of STOP's Impact

REFERENCES


HIGHLIGHTS

Purpose

The purpose of this evaluation is to assess whether STOP's financial support for direct victim services offered through private nonprofit victim service (VS) agencies helps victims of domestic violence and sexual assault improve their safety and well-being and work successfully with legal system and other relevant agencies. We carry out this purpose by

  1. Describing the variety of VS programs funded by STOP;

  2. Understanding the community and state context in which these VS programs operate;

  3. Assessing the degree to which receipt of STOP funding for VS programs has led to improved program services and community coordination; and

  4. Examining how VS program services and the community context in which they are offered affect victim outcomes.

This report covers results of the first year of evaluation activities. It describes what we have learned with respect to the first three goals of the overall evaluation project, namely describing VS agencies, their state and community context, their interactions with other relevant agencies and organizations in their communities, and the impact of local and state activities on VS program and legal system outcomes.

Who, What, Where, and When?

In 1999, the National Institute of Justice funded the Urban Institute to conduct an evaluation to assess outcomes resulting from direct victim services offered through private nonprofit victim service agencies.1 This evaluation uses a variety of research methods to understand how VS programs help victims. Specifically, it looks at

  1. How STOP funding changes VS program and legal system activities;

  2. How VS program activities make a difference for clients, community members, and community agencies;

  3. Whether communities with greater degrees of coordinated response to violence against women are able to help victims more and in better ways; and

  4. Whether state STOP agencies are able to increase the number of communities providing a coordinated response through agencies' requirements for funding and supports for potential applicants and funded programs.

This report is the first one produced by the evaluation. It includes information submitted on standardized federal reporting forms by all STOP-funded VS programs, and information reported to us by representatives of a sample of STOP-funded VS programs during telephone interviews and follow-up contacts. Future reports will present findings on women's experiences with the service networks in their communities (to be gathered through victim interviews scheduled for 2001), and an integrated analysis detailing the roles of state and community context and VS program offerings in improving women's outcomes after domestic and/or sexual violence.

Why This Study Is Important

The STOP Violence Against Women Formula Grants Program is a major federal avenue for stimulating the growth of programs serving women victims of violence. The program's long-term goal is to promote institutionalized system change, such that women encounter a supportive and effective response from the criminal and civil legal systems and from community agencies offering services and supports. The program is authorized by Chapter 2 of the Safe Streets Act, which in turn is part of the Violence Against Women Act (VAWA), Title IV of the Violent Crime Control and Law Enforcement Act of 1994 (P.L. 103-322). It is administered by the Violence Against Women Office (VAWO) in the Department of Justice's Office of Justice Programs.

A great deal of federal money has been used to support violence-against-women services funded through the STOP program. Federal funding for the STOP program for fiscal years 1995 through 1999, the focal period of this evaluation, totaled $540.6 million. These federal funds are supplemented by a significant amount of state and local support through the match required of projects in law enforcement, prosecution, and other public agencies. States have reported on approximately 6,500 subgrants awarded as of November 15, 1999. Many STOP programs got additional STOP subgrants in the years following their initial funding, so the 6,500 subgrants translate into about 4,700 distinct projects, of which 1,200 are VS programs.

This evaluation is designed to assess the impact of STOP-funded VS programs on the clients and communities they serve. Little is known about how VS program activities influence outcomes for women and how agencies hosting VS programs interact with the legal system and other agencies to assist women victims of violence. Past research examining domestic violence and sexual assault has three limitations: (1) few studies examine the impact of a coordinated community response to violence against women; (2) most studies examine only criminal justice system outcomes (e.g., rearrests)—few studies examine outcomes for women that reflect their well-being or safety; and (3) most available studies had small samples and examined only one or two service modalities from one or two programs. This study is explicitly designed to go beyond past research efforts to cover these missing elements, and to do so on a sample of programs and women victims of violence drawn from around the nation, from communities of different types, and from communities organized in different ways to address the problem of violence against women. Findings from this study will begin to fill many gaps in our knowledge and lead to the design of more and better approaches to helping women.

How Was the Information for This Report Collected?

All programs funded by STOP are required to submit a description of their program to the Violence Against Women Office in the U.S. Department of Justice shortly after they receive funding. These descriptions come in on a federal form called a Subgrant Award and Performance Report (SAPR). The first step in this evaluation was to select and analyze these SAPRs for all STOP-funded projects that went to private nonprofit VS agencies for the delivery of direct services to women victims of domestic violence or sexual assault.

Based on this analysis, we selected a sample of 200 VS programs to participate in a telephone survey. The VS programs were sampled from the universe of about 1,200 SAPRs for VS programs according to a number of criteria. First, VS programs were sampled, rather than individual subgrant reports because many VS programs are re-funded over a number of years. Second, only private nonprofit victim service agencies were included. Third, VS programs had to have been funded for at least two years, to provide direct services to victims, and to have (or have had) STOP subgrants of at least $10,000. In addition, a subset of VS programs were sampled such that at least 10 interviews were completed within eight focus states.2 Extensive analysis after data were collected showed that the sample of programs included in the VS Program Survey strongly resembles the universe of STOP-funded VS programs on every dimension available for comparison using the SAPR database.

We collected data from the VS programs in our sample using a telephone interview and a faxed questionnaire. The faxed questionnaire covered topics such as budgets, funding, employees, and number of victims served. The phone interview covered topics such as the nature of the STOP-funded program, experiences with state STOP agencies, changes in the legal system since STOP funding became available, outreach strategies, the ability of the community to meet the needs of women victims of violence, and the extent to which the STOP-funded VS program works with other agencies in its community to address violence against women.

After interviews were completed, two trained interviewers rated each VS program on the extent to which it communicates, coordinates, and collaborates with other agencies in its community, and rated whether or not the community's service structure constituted a coordinated community response to violence against women.

Key Findings

STOP-Funded VS Programs and Their Agencies

  • One of the ways that STOP funding helped most was to increase the number of locations and/or mechanisms through which women could access victim services. Most host agencies offered services (STOP-funded and otherwise) in both disclosed (e.g., courthouses, health care facilities, and welfare offices) and undisclosed service sites (e.g., shelters).


  • One-third of STOP-funded VS projects reported focusing on both domestic violence and sexual assault issues. Of the rest, 17 percent focused exclusively on sexual assault, and half focused exclusively on domestic violence.


  • Although most STOP-funded VS projects had primary focuses on domestic violence or sexual assault, many of their host agencies reported working on both issues. Both employees and volunteers were involved in providing direct services and outreach/education activities around domestic violence and sexual assault.


  • Most VS programs used a portion of their STOP funds to support employee salaries.


  • Many VS programs reported that STOP funds have allowed their agency to provide new services to its current victim population (62 percent), that STOP funds have allowed their host agency to bring existing services to more women (72 percent), and that STOP funds helped them tap into an entirely new victim population (70 percent).


  • Victim service agencies undertook a variety of direct service activities with STOP funds, including legal/court advocacy, comprehensive safety planning, counseling, answering hotline calls, individual advocacy, medical advocacy, first response, and shelter.


  • Some types of service stand out as either particularly likely or particularly unlikely to be supported by STOP funds:


  • Court advocacy and participation in a multidisciplinary first response team were most likely to either be STOP-funded or not exist in an agency. Very few agencies supported these activities without using STOP as a funding source. This is a particularly important finding, for two reasons. First, these types of cross-agency projects are exactly what Congress intended to promote when it created the STOP program. And second, they are difficult to create and take time and energy to maintain, so they are unlikely to exist without the support of an innovative program such as STOP.


  • STOP funds were used to support major portions of projects focusing on collaboration, training, and policy/protocol development activities. STOP funding allowed these activities to proceed at a more extensive level than had been possible before STOP. Again, the fact that STOP is being used for projects such as these indicates the overall program's success in fulfilling legislative intent.


  • Host agencies were relatively unlikely to use a STOP subgrant to support shelters, offer legal representation, or answer a hotline, although many host agencies offered these services. As these are some of the oldest and best-established services for women victims of violence, they presumably have alternative sources of funding, allowing host agencies to choose to do something new with STOP support.
  • STOP funds accounted for less than half the annual budget of most host agencies.


  • Results suggest that STOP is increasing the number of women who receive needed services related to their experiences of domestic violence or sexual assault. However, it appears to be relatively difficult for many VS programs to provide statistics on the number of women they serve from year to year, so this conclusion must remain tentative.

VS Program Interactions with Other Community Agencies

  • All VS programs reported interacting with at least one law enforcement agency in their community, and most reported interacting with at least one prosecution agency (97 percent) and at least one other VS agency (94 percent) in their community.


  • VS programs identified the agencies with which they have the most or most meaningful contact, which we call "primary partner" agencies. Of all VS programs,
  • 65 percent reported law enforcement agencies;

  • 42 percent reported prosecution agencies; and

  • 25 percent reported social service agencies.

  • One-quarter of VS programs named both law enforcement and prosecution agencies as those with whom they partner the most to help women victims of violence.


  • Most VS programs reported involvement of every level of employee (frontline staff, middle management, and organizational leaders) in interactions with their primary partner agencies (law enforcement, prosecution, other VS agencies, and other types of agencies).


  • One-half of VS programs had formal policies or procedures to work with law enforcement, one-third had the same with prosecution, and one-quarter had the same with other VS agencies.


  • VS programs reported increases in five types of interaction with other agencies (law enforcement, prosecution, other VS agencies, and other types of agencies) since STOP funding. Over half reported their belief that these changes were due to their STOP-funded VS program, and between 11 and 31 percent reported changes were due to other STOP projects in their community. One-third attributed changes to both their own and another STOP subgrant, indicating that a considerable number of communities are using STOP to support activities in two or more agencies that bring those agencies into closer interaction to serve women better.


  • Most VS programs communicate in many ways with their primary partner agencies. They share general information about violence-against-women issues, have frequent phone contact, have informal meetings, and refer clients back and forth.


  • Most VS programs coordinate their activities with their primary partner agencies. Most help one another on an as-needed basis with specific cases, and facilitate referrals.


  • VS programs are more likely to provide training to law enforcement than to prosecution or other types of agencies. VS programs are more likely to receive training from other VS agencies than from law enforcement or prosecution.


  • VS programs collaborate in a variety of ways with their primary partner agencies. Most participate on task forces with partners and strategize about how to reach women victims of violence. Fewer VS programs, although still over half, influence one another's agency protocols, provide integrated services to victims, or have a regular feedback mechanism regarding their collaborative work that helps them fix problems and shape new directions.


  • Of those who named law enforcement as a primary partner, 36 percent participated on a first-response team with them.


  • Of those who named prosecution as a primary partner, 26 percent reported interacting with them on a first-response team.


  • Three-quarters of VS programs participated in some form of violence-against-women task force in their community. Every collaborative activity or arrangement was more likely to occur when the VS program and its two primary partners participated together on a task force.


  • There are levels of joint work that go well beyond task force membership. VS programs in communities that the researchers rated as providing a coordinated community response were more likely than those in communities without this level of coordination to report each collaborative activity or arrangement, even when all agencies participated on a task force together.


  • Task forces can be useful forums for agencies to work together, particularly in those communities where a coordinated community response exists. However, the existence of a task force does not guarantee joint work or collaborative activities in communities. Likewise, some communities without task forces still participate in collaborative activities.

Impact of STOP on Service Provision

  • The more communities were already addressing violence-against-women issues and were engaged in developing the ability to meet the needs of victims before STOP, the higher VS programs rated their community on its ability to meet victim needs after STOP funding. However, the greater the level of activity in communities before STOP, the less change VS programs reported when it came to addressing the needs of victims.


  • The more agencies worked together in communities, including law enforcement and prosecution agencies working with VS programs, the more likely services were to improve for both VS programs and the legal system.


  • State STOP agency support for collaboration was related to more communication among agencies and more coordinated community responses to violence against women. However, state STOP agency support for collaboration, at least as we were able to measure it, was not related to VS program or legal system outcomes.


  • Although we found that measures of the level of STOP funding to VS programs were not directly related to VS program outcomes or to changes in how legal-system agencies treat women victims, it is important to remember that every VS program in our sample did have STOP funds. The effect of receiving or not receiving a STOP grant therefore could not be assessed, but would almost certainly have revealed significant differences in community services had we been able to do so. Without being able to make this comparison, it impossible to assess the full impact of STOP funding on communities.


  • VS program representatives who attributed changes in interaction between their VS program and law enforcement, prosecution, and/or other VS agencies to STOP funding also reported greater coordination in community responses and more positive VS program and legal system outcomes.


  • Using STOP to fund certain types of activity (in particular, multidisciplinary response teams, victim witness services, and policy/protocol development activities) is associated with reports of greater coordination in community responses and more positive VS program and legal system outcomes.

Implications for Research

  • Include non-STOP funded VS programs and non-STOP funded communities in evaluation designs to compare the effect of any STOP funding versus no STOP funding on the level of coordination in communities, improved VS program outcomes, and improved legal system outcomes.


  • Include non-STOP funded communities to further illuminate the effects of state STOP agency support on the level of community coordination among agencies and on VS program and legal system outcomes.


  • Anticipate that many VS programs will have a difficult time identifying the number of victims they have served in recent years (since STOP), and an even more difficult time for previous years (especially before STOP).


  • Include interviews with women victims of violence regarding their experiences with community agencies, as the current evaluation will do next year, in order to reflect their views and perceptions in addition to those of VS program employees.


  • Include detailed behavioral questions in surveys as measures of communication, coordination, and collaboration activities. Respondents interpret the three concepts differently, and researchers will only muddy the waters if they limit themselves to questions containing only these three terms.


  • Define "institutionalized commitment to work together" for respondents, because this concept is also interpreted differently across respondents. In our usage, "institutionalized commitment to work together" entails formal and/or routine practices agencies conduct together; involvement of all levels of the agencies, from frontline workers to organization leaders, in the joint activities; and commitment of leaders to the joint work.


  • Recognize the complexity of the joint work that occurs with other agencies in local communities and structure research instruments accordingly. Include a series of questions through which respondents can report about various types of activities with several types of agencies, or with different agencies within types.


Implications for Practice

  • VS programs and legal system agencies should work together to address violence-against-women issues. The joint work should include collaborative activities, not just communication or coordination activities.


  • Task forces are not the only way communities can work toward collaborative approaches to violence-against-women issues. Some communities without task forces were working collaboratively and some communities with task forces were not working collaboratively. Community agencies, such as VS programs, law enforcement, prosecution, and the medical community, should focus on working together on particular tasks that lead to more collaborative work, with the goal of approaching or creating a coordinated community response. Such tasks include strategizing about how to address violence against women in the community, developing policies and protocols for different agencies as a joint endeavor, providing integrated services, creating feedback mechanisms about their joint work, and developing first-response teams.


  • State STOP agencies should continue to support local collaborative efforts in communities through technical assistance, training, and other subgrantee support activities.

Implications for Policy

  • State STOP agencies should continue to support local collaborative efforts in communities through funding priorities. Funding policies could be created requiring joint work as demonstrated by clear evidence of collaboration (e.g., detailed work plans, site visits by agency staff, a history of collaboration). This type of support increases coordinated responses in communities, which, in turn, are related to positive VS program and legal system outcomes.


  • VAWO should encourage states to invest in the purpose area of the recently reauthorized Violence Against Women Act that highlights collaborative efforts in local communities.


CHAPTER 1
Introduction

HIGHLIGHTS
  • In 1999, the National Institute of Justice funded the Urban Institute to conduct an evaluation to assess whether STOP's financial support for direct victim services offered through private nonprofit victim service agencies resulted in improved well-being of women victims of violence. This report describes results from the project's first year.


  • This evaluation seeks to
    1. Describe the variety of private nonprofit victim service programs funded by the STOP program;

    2. Understand the community and state context in which these programs operate;

    3. Assess the degree to which receipt of STOP funding for victim service (VS) programs has led to improved program services and community coordination; and

    4. Understand how victim service program offerings affect victim outcomes.


  • To meet these goals, the evaluation uses a variety of research methods to examine how VS programs help victims, looking specifically for the effects of


    1. STOP funding on VS program (and legal system) activities;
    2. VS program activities on clients, community members, and community agencies;
    3. Levels of coordinated community response on VS programs' ability to help their clients; and
    4. State STOP agency activities on levels of coordinated community response.

Introduction

The STOP Violence Against Women Formula Grants Program is a major federal avenue for stimulating the growth of programs serving women victims of violence. The program's long-term goal is to promote institutionalized system change, such that women encounter a supportive and effective response from the criminal and civil legal systems and from community agencies offering services and supports. The program is authorized by Chapter 2 of the Safe Streets Act, which in turn is part of the Violence Against Women Act (VAWA), Title IV of the Violent Crime Control and Law Enforcement Act of 1994 (P.L. 103-322). It is administered by the Violence Against Women Office (VAWO) in the Department of Justice's Office of Justice Programs.

In 1999, the National Institute of Justice (NIJ) funded the Urban Institute to conduct an evaluation to assess whether STOP's financial support for direct victim services offered through private nonprofit victim service agencies resulted in improved well-being of women victims of violence.1 Thus the present project is the only full-scale evaluation funded by NIJ to focus on the impact of STOP-funded victim services offered by victim service programs. It is, in addition, the only large STOP evaluation project that will include evidence gathered directly from women victims themselves. The goals of this evaluation are to (1) describe the variety of private nonprofit victim service programs (hereafter, VS programs) funded by the STOP Program, (2) understand the community and state context in which these programs operate, (3) assess the degree to which receipt of STOP funding for VS programs has led to improved program services and community coordination, and (4) understand how victim service program offerings affect victim outcomes. Further goals are to understand how STOP funding, VS program development, community and state context, and community change interact to affect victim outcomes, and to present relevant findings in ways that will best aid public policy to improve victim services and help women.

This report is the first in a series of annual reports on project findings. It covers information pertaining to the first, second, and part of the third goals of the overall evaluation project, namely describing VS agencies, their state and community context, and their interactions with other relevant agencies and organizations in their communities. Its chief sources of data are a telephone survey conducted with directors of 200 randomly selected STOP-funded VS programs offering direct services to victims, and Subgrant Award and Performance Reports (SAPRs) submitted each year by states to describe the projects they have funded. Chapter 2 describes these data sources in detail. Future reports will present findings on women's experiences with the service networks in their communities (to be gathered through victim interviews scheduled for 2001), and an integrated analysis detailing the roles of state and community context and VS program offerings in improving women's outcomes after experiencing domestic violence and/or sexual assault.

Victim Service Programs

Throughout the 1970s and 1980s, violent crimes against women gained public attention and efforts grew to improve services to meet victim needs. Grassroots advocacy played a pivotal role in this process, demanding expanded legal protections and offering direct services to women.

Public attention to the needs of domestic violence victims did not increase dramatically until the 1980s. At the federal level, the 1984 Report of the Attorney General's Task Force on Family Violence recommended coordinated community responses (CCR) to domestic violence and specific reforms in laws and operations of the justice system (Department of Justice, 1984), and the Bureau of Justice Assistance funded 11 Family Violence Demonstration programs to establish interagency coordinating committees (Bureau of Justice Assistance, 1993; Harrell, Roehl, and Kapsak, 1988). The Family Violence Prevention and Services Act of 1984 supported grants for domestic violence prevention programs, shelters, victim assistance services, and training and technical assistance for law enforcement officers and others in 20 states (Newmark, Harrell and Adams, 1995).

Estimates show that there were about 1,800 programs for women experiencing domestic violence in the late 1990s, of which 1,200 were shelters (Garner and Fagan, 1997). Coordinated community approaches to violence against women continue to be a major goal, but were not widely available before STOP (Clark, Burt, Schulte, and McGuire, 1996). Coordinated approaches to helping victims of domestic violence include a number of players, such as law enforcement, prosecution, the courts, health care, and social service agencies. Among the services in these networks are hotlines, shelters, support groups, individual and group therapy, legal advocacy, social service referral and advocacy, services for children exposed to domestic violence, transitional housing, job training, and more.

Major changes addressing the needs of women victims of sexual assault began even earlier than those for domestic violence. The first rape crisis centers opened in 1972. In an assessment of the first 10 years of rape crisis center development, Gornick, Burt and Pittman (1985) concluded that these VS programs, although widespread, had already been reduced from a high of about 1,000 to about 600, and in many instances had become part of established public (e.g., community mental health centers) and private (e.g., YWCAs, Family Services) agencies. In this transition, many developed linkages with police, prosecutors, and medical personnel. Site visits for the National STOP Evaluation in 1996 through 1999 suggest that in many places sexual violence programs are having a hard time surviving, do not have the same political clout enjoyed by domestic violence programs, and, pre-STOP, had often seen breakdowns in their system linkages (Burt et al., 2000).

Despite growth in the number of VS programs, little evaluation research addresses their "usual" offerings' impact on women (Garner and Fagan, 1997; Koss, 1993). Expanded victim services occurred before research addressing what works best, say Garner and Fagan (1997), who argue that evaluating the impact of VS programs is crucial if we are to learn how best to serve these women and whether any current practices have unintended negative consequences.

A few studies have examined the efficacy of particular psychological treatments for sexual assault victims (Koss and Harvey, 1991; Resick and Nishith, 1997), but were limited to victims of that crime, were mostly focused on women with extreme reactions to sexual assault, and did not examine "normal" VS programs. Results therefore do not generalize to other populations or services. Further, very few women victims seek the types of treatment that have been evaluated. One study shows that only 4 percent of 1,895 eligible women victims of domestic violence sought counseling services (Gondolf, 1998).

Nor does research tell us much about the impact of a coordinated community response to domestic violence or sexual assault, and what little it does offer uses criminal justice system outcomes (such as rearrest of offenders) as the measure of impact rather than outcomes reflecting victim well-being or safety. Tolman and Weisz (1995) document reduced recidivism of batterers when law enforcement officers follow protocols developed in coordination with other agencies, while Weisz, Tolman, and Bennett (1998) report a greater likelihood of a court case or an arrest when women receive both domestic violence services and at least one protective order instead of only one of these service types.

Even fewer studies examine outcomes for women other than criminal justice outcomes. In a review of 12 studies, Gordon (1996) reports that women victims most commonly sought help from the criminal justice system, then social service agencies, medical services, crisis counseling, psychological services, clergy, support groups, and women's shelters. Of the few studies that examined women's reactions to VS programs, Gordon reports that women do not necessarily find all of these services helpful—women found crisis lines, women's groups, social workers, psychotherapists, and physicians to be helpful for all types of abuse, while police officers, lawyers, and clergy were not helpful for most types of abuse.

Sullivan and colleagues (Sullivan and Bybee, 1999; Sullivan, Campbell, Angelique, Eby, and Davidson, 1994; Sullivan, Tan, Basta, Rumptz, and Davidson, 1992; Sullivan and Davidson, 1991) examined the relationship between an advocacy program for battered women and outcomes related to the program using an experimental design. Initially, women who received assistance from advocates after leaving shelters had more positive outcomes in terms of social support, effective use of resources, and quality of life than women in the control group. By the six-month follow-up, differences between groups only existed for overall quality of life and satisfaction, with women who received advocacy having better outcomes than the control group. But by the two-year follow-up, positive differences were apparent in social support, quality of life, and reabuse.

Little else is known about victim service programs and how they affect women's lives. Most available studies had small samples and examined only one or two service modalities from one or two programs. These limitations of the existing literature signal a clear need to assess the impact of VS programs on women's outcomes. As women may sometimes feel that criminal justice outcomes are in conflict with those of personal safety and well-being, more attention needs to be paid to the latter. In addition, it is important to understand how coordinated community responses may affect women's outcomes.

The STOP Program

Funding for the STOP program for fiscal years 1995 through 1999, the focal period of this evaluation, totaled $540.6 million. These funds were distributed through grants to the 50 states, the District of Columbia, and five territories (hereafter, states). The states in turn have reported on awards made through approximately 6,500 subgrants awarded as of November 15, 1999. Many STOP programs got additional STOP subgrants in the years following their initial funding, so the 6,500 subgrants translate into about 4,700 distinct projects.

From the beginning of the STOP program, VAWO and the NIJ have supported evaluations to examine the implementation and impacts of STOP funding. These include the National STOP Evaluation conducted by the Urban Institute (Burt et al., 1996, 1997, 1998, 1999, 2000); purpose area evaluations conducted by the Institute for Law and Justice, the National Center for State Courts, the University of Arizona Tribal Law Program (Luna, 1999), and the American Bar Association; and several evaluations of single programs. Of these, the only one to focus specifically on victim impact was the evaluation project conducted by the American Bar Association, but that was limited to victim services based in criminal justice agencies and thus does not address issues related to nonprofit VS programs, this evaluation's focus.

Within the STOP program, states have rarely funded projects to conduct evaluations, so the national evaluation projects are the main source of information about how these programs are making a difference. Most evaluation results pertain to the implementation of projects and their progress in changing the behavior of their own and other agencies in their communities. These changes have been documented in various reports over the years (summarized in Burt et al., 1996, 1997a, 1998, 1999, 2000; Luna, 1999). Some additional impact information on the efforts of law enforcement and prosecution agencies is being collected as part of the National STOP Evaluation, but little of this focus has been on victim services or outcomes for victims.

The Design of This Project

This evaluation is designed to assess the impact of STOP-funded VS programs on the clients and communities they serve. Of the approximately 4,700 distinct STOP-funded projects noted above, about 1,200 offer direct victim services through VS programs. These 1,200 programs comprise the universe of VS programs covered in this report.

We structured our approach to this evaluation as a multilevel, multimethod analytical design. A design seeking to understand how VS programs help victims must be able to assess the influence of many factors, including the effect (1) of STOP funding on VS program (and legal system) activities; (2) of VS program activities on clients, community members, and community agencies; (3) of the degree of coordinated community response (hereafter, CCR), possibly stimulated by VS or other STOP-funded activities, on VS programs' ability to help their clients; and (4) of state-level STOP activities on CCR levels.

The conceptual framework or logic model guiding our research design is depicted in figure 1.1. This framework incorporates all of the essential design elements to assess the effects just mentioned. It incorporates four descriptive levels of the design (victims, VS programs, communities, and states), and also shows the causal linkages expected among design elements.

To the far right in figure 1.1 are the two ultimate dependent variables or outcomes of the framework—victim and community outcomes. Victim outcomes to be examined are all important short-term effects of attempts to use services, and they include the ability to obtain needed resources; safety; and perceptions of respectful treatment, fairness or justice being done, services helping victims to achieve immediate goals, and satisfaction with various aspects of interactions and program/service contacts. Service use pattern is also a victim outcome, but is shown conceptually as a prior box (Box 6) in figure 1.1. Community outcomes include knowledge of services and attitudes toward using services.

To the far left of figure 1.1 are two factors, STOP funds and other resources, and the state STOP agency's support for collaboration, that are expected to affect elements to their right, but not to be affected by other elements in the model. In the middle positions are factors reflecting improved services within individual agency types (VS programs and legal system agencies), and levels of CCR that are assumed to be affected by elements 1 through 4 and in turn to affect the ultimate outcomes in Boxes 6, 7, and 9. Box 6, service use patterns, is both an outcome to be reported through victim interviews and a predictor of other victim outcomes (Box 7). Box 8, victim characteristics and nature of victimization, is included because these elements are likely to affect victim outcomes independent of the community's service system.

Each box in figure 1.1 represents an element that we will describe with our findings. This report covers findings for Boxes 1 through 5, based on our first year's methods and data sources described in chapter 2. These are findings about impact, but involve impact of STOP funding on VS programs and their communities. Data to assess the ultimate impact of STOP funding on women victims of violence will be collected in the coming year, directly from the women themselves. These interviews will supply information pertinent to Boxes 6 through 9 and permit assessment of personal or victim impact. Boxes 6 through 9 are shaded in figure 1 to differentiate them from Boxes 1 through 5 and to indicate that they will not be part of the analysis presented in this report. It is important, however, for the reader to have an idea of the full scope of this project's design, in order to place the findings on program and community impact into their proper context.

Each arrow in figure 1.1 represents an expected relationship, effect, or causal impact. Data collection strategies for the evaluation are designed to obtain information that describes the situation for each element in figure 1.1, and also permits assessment of the causal linkages among the elements. For example, data for Box 3 will describe both the current and pre-STOP status of VS program services, activities, and clientele, and data for Box 4 will describe the current and pre-STOP status of legal system response to women victims of violence. Analyses relating information from Boxes 1 and 2 to Boxes 3 and 4 will reveal the degree to which level of STOP and other funding, and state STOP administrator policies, affect the level of improvement in VS program and legal system services. Likewise, analyses relating information from Box 2 (state STOP policies) to Box 5 (level of CCR) will indicate how much the presence of a CCR can be attributed to the state STOP administrator's commitment to help communities move in the direction of greater coordination and interdependence.

The arrows in figure 1.1 depict some very important influences, or effects, of elements on the left on elements on their right. These arrows constitute, in effect, hypotheses about how things work. We expect the nature of state STOP administrator activities to affect VS and legal system services, and all three to affect the level of CCR. We expect improved VS and legal system services to affect the likelihood that a woman victim of violence will use services at all and to affect the pattern of services she uses (VS program services only, legal system services only, both, or neither). Effects are also expected of the level of CCR on service use patterns. Finally, we expect program and community characteristics to affect both victim outcomes and knowledge and attitudes among women in the VS program's community.

First-Year Activities

The results of our activities during the first year of this project are the subject of this report. During our first year we conducted analyses of the SAPR database containing reports of all subgrants for which awards were reported to VAWO as of November 15, 1999. From this database we identified every program meeting the definition of a VS program offering direct services to victims, and operated by a private nonprofit victim service agency. We further restricted our definition of the programs of interest by requiring that they had received at least $10,000 in STOP funds, and that they had been in operation long enough to make the expectation reasonable that they had accomplished something. As a practical matter, "long enough" was defined as operating for at least two years (although ultimately we had to accept some deviation from this criterion in our final sample). From this universe of STOP-funded VS projects, we randomly selected 250 programs to reach a target sample size of 200 completed interviews.3 Chapter 3 of this report describes our VS project sample and compares the sampled projects to all VS programs in the SAPR database. These descriptions are limited to variables that appear on the SAPR form, but they give the reader a good idea of the ways in which our VS project sample is similar to, or different from, all VS projects funded through the STOP program.

During our first year we also developed a telephone interview protocol and conducted the 200 interviews with directors of VS programs necessary to reach our target sample size. Chapters 4 and 5 provide descriptive results of this survey. In chapter 4, we examine the STOP-funded activity and describe its major dimensions and accomplishments. We also analyze the relationship of the STOP-funded program to its host VS agency, including the size of each, the types of services each offers, the types of women each serves, and other relevant matters. Chapter 5 looks at the STOP-supported VS program in relation to its larger community. It describes the interactions among the STOP program and law enforcement, prosecution, other VS agencies, and other programs with which the STOP program has a great deal of interaction. Finally, chapter 6 presents analyses linking together information from Boxes 1 through 5 using regression analysis to assess the effects on each box of the program, community, and state characteristics antecedent to it.


Chapter 2
Methods

HIGHLIGHTS
  • Representatives from 200 VS programs participated in an in-depth telephone interview. VS programs were sampled from the Subgrant Award and Performance Report database according to a number of criteria:


    1. VS programs were sampled rather than individual subgrant reports because many VS programs are re-funded over a number of years;

    2. Only private nonprofit victim service agencies were included;

    3. VS programs had to have been funded for at least two years, provide direct services to victims, and have/had STOP subgrants of at least $10,000; and

    4. A subset of VS programs were sampled such that at least 10 interviews were completed within eight focus states (CO, IL, MA, PA, TX, VT, WA, and WV).

  • We collected data from the VS programs in our sample using a telephone interview and a faxed questionnaire. The faxed questionnaire covered topics such as budgets, funding, employees, and number of victims served. The phone interview covered topics such as the nature of the STOP-funded program, experiences with state STOP agencies, changes in the legal system since STOP funding became available, outreach strategies, the ability of the community to meet the needs of women victims of violence, and the extent to which the STOP-funded VS program works with other agencies in its community to address violence against women.


  • After interviews were completed, two trained interviewers rated each VS program on the extent to which it communicates, coordinates, and collaborates with other agencies in the community, and rated whether or not the agencies in the community worked together to provide a coordinated community response.

Introduction

This chapter covers the data and methods used to describe VS programs, their host agencies, and their communities. The results reported are based on three data sources: SAPRs, the Victim Services Program Survey, and ratings of state efforts to promote coordinated community responses to violence against women.

Subgrant Award and Performance Reports

VAWO requires states and subgrantees to document information about STOP subgrant awards and the performance of supported projects. This information is recorded on a SAPR by either the subgrant recipient or the state STOP agency and submitted annually (by October 15th) to VAWO. The SAPR includes two sections. Part 1 is the award section, completed when an award is made and describing the nature of the program, its crime focus, and its goals. Part 2 is the performance section, completed annually or at the conclusion of the subgrant and documenting some aspects of project accomplishments, plus the number of victims served if the project did direct victim service.

The SAPR database available for analysis at any given time has a number of limitations of which the reader should be aware, as they affect an understanding of any results that use SAPR data. First, states vary in the promptness and completeness of their reporting of both the award and performance sections of the SAPR. The database we used to select the VS program sample was last updated in November 1999. It was missing virtually all awards made with FY 1999 appropriations, as few had yet occurred. It was also missing about half of awards made with FY 1998 funding, due to the timing of state awards and of state reporting to VAWO. This is not as serious as it sounds for VS programs, because most of them receive awards in several years, but it is something to keep in mind. In addition, the performance section has not been submitted for most subgrant awards, and those that have been submitted are often incomplete. Second, sometimes SAPRs are completed by subgrantees and sometimes SAPRs are completed by state STOP agencies. It is not clear that the form would be completed uniformly by each group, and it is likely that subgrantees would report different information than state agencies as they know more about their programs.

Third, SAPR reports do not always accurately reflect the activities or characteristics of a subgrantee, even if they are filled out consistently. For example, SAPRs frequently indicate a special focus on reaching underserved populations when further inquiry through telephone screening reveals this not to be the case. Some SAPRs do not accurately report such obvious things as urban or rural geographic focus. Fourth, the SAPR information depends on state STOP agency conventions regarding how subgrants are categorized. Some states categorize subgrants as law enforcement, prosecution, or victim services by the recipient agency regardless of the target population, while others categorize subgrants by the intended beneficiaries regardless of recipient agency. Also, some states insist that all STOP funding go through public entities, so funds for victim services are included in subgrants to prosecution or law enforcement agencies. Finally, some identical victim service agency activities (e.g., participation in a first response team, court advocacy) may be covered by either a subgrant to a victim service agency or a subgrant to a law enforcement or prosecution agency, which then hires a victim service program staff person to do the work.

Despite these limitations, the SAPR database provides the only information available for describing the universe of STOP-funded projects, so we used it as the basis for sampling in this study. The database used for this study includes SAPRs submitted from the beginning of STOP through November 15, 1999. The data include awards made from states' FY 1995, FY 1996, FY 1997, FY 1998, and FY 1999 STOP subgrants. The data set includes 6,527 subgrant awards totaling $298,844,684 of funding. Thirty-five percent of the funding was awarded to victim services programs for a total 2,788 reported subgrants, representing about 1,200 VS programs.

Choosing the Eight Impact States

We focus on eight states for this study, representing varying approaches of state STOP agencies to promoting coordinated community approaches to violence against women. State agencies were rated on three activities that promote coordinated community efforts in local communities. Ratings were recorded on 5-point scales where 1 represented the lowest level of agency promotion and 5 represented the highest level. First, states were rated based on the extent to which subgrantee funding is made contingent upon evidence of a coordinated community response to violence against women. A rating of 1 for this activity would include not requiring subgrantees to provide any documentation of interaction with other agencies in the community or requiring letters of support from other agencies that only include an agency director's signature without significant evidence of or plans for collaboration. A rating of 5 would include requiring communities to submit proposals as teams for coordinated service provision efforts that simultaneously fund multiple agencies within one community. Intermediate ratings were given for funding approaches that encouraged, but did not require, collaboration.

Second, states were rated based on the extent to which their STOP agencies offer technical assistance to subgrantees with the goal of helping localities build coordinated community responses to violence against women. This type of assistance includes providing direct technical assistance to local communities to help build teams, holding meetings at which attendance is based on teams, and giving guidance to develop proposals evidencing teamwork and collaboration. A rating of 1 for this activity would indicate the agency does not provide any of this type of assistance and a rating of 5 would indicate that much of their assistance involves promoting collaboration and coordination within communities.

Third, states were rated based on the extent to which they provide other types of assistance to subgrantees to promote collaboration with other agencies in the community. For example, agencies may help existing subgrantees obtain other supplemental (non-STOP) funding to either continue their VS program or to fill in gaps in services in the community; may facilitate contact between similar programs so that they can learn from one another, particularly when it comes to overcoming barriers to collaboration; and may assist subgrantees in involving other community systems in their programs. A rating of 1 for this activity would indicate the agency did not provide any assistance such as this and a rating of 5 would indicate they provide a great deal of this type of assistance.

Sixteen states were visited as part of the National Evaluation of the STOP Formula Grant Program (California, Colorado, Kansas, Kentucky, Illinois, Indiana, Massachusetts, Minnesota, North Carolina, Nevada, New York, Pennsylvania, Texas, Utah, Vermont, and West Virginia). Each of these states were rated and, because we knew a great deal about it based on phone calls with the state STOP administrator and statewide coalitions, examination of implementation plans, and consultations with others, Washington State was also rated. Two evaluation staff who conducted the site visit to a given state rated the state independently and then compared ratings. Consensus was achieved and states were given one rating for each of the three criteria. The three ratings were also averaged to represent an overall rating. The ratings were verified with knowledgeable staff from the STOP Technical Assistance Project, who reviewed the four ratings given to each state and confirmed their accuracy.

The eight states chosen as impact states (Colorado, Illinois, Massachusetts, Pennsylvania, Texas, Vermont, Washington, and West Virginia) were selected because they represent high (4.7-4.8), medium-high (4.0), medium (3.2-3.7), and low (1.0-1.3) levels of state agency promotion of coordinated community efforts, according to our ratings. In addition, they represent a balance of states across the nation, including the New England, Mid-Atlantic, South Atlantic, West-South-Central, East-North-Central, Mountain, and Pacific regions of the country.

Victim Services Program Survey

The Victim Services Program Survey involved in-depth telephone interviews with 200 subgrantees. The purpose of the survey was to gather descriptive information about the universe of STOP-funded victim service programs, and also to gather information about the amount of communication, coordination, collaboration and coordinated community responses taking place in the communities they serve. Survey information will also be used to select the 40 communities from within the eight impact states for victim interviews and site visits planned for the second year of this project.

Sample

VS programs for the survey were sampled from the SAPR database according to a number of criteria. First, VS programs were sampled rather than subgrants because many VS programs are re-funded over a number of years. The SAPR database was restricted such that the universe of programs that could be sampled were VS programs in private nonprofit victim service agencies that have been funded for at least two years, provide direct services to victims, and have/had STOP subgrants of at least $10,000. A subset of VS programs were sampled such that at least 10 interviews were completed within the eight impact states. To do this, all nonprofit victim service programs meeting the three criteria were sampled from those eight states.4 For the remaining respondents, we wanted programs to include equal numbers of domestic violence only, sexual assault only, and combined domestic violence and sexual assault programs, to provide enough sample members to support analyses within these categories. The SAPR database, minus programs in the eight impact states, was divided into the appropriate strata by crime focus, after which programs were randomly selected from each stratum.

The sampling strategy resulted in 200 completed interviews in 32 states. Ninety programs were interviewed in the impact states: 13 in Colorado, 10 in Illinois, 10 in West Virginia, 12 in Pennsylvania, 15 in Massachusetts, 10 in Texas, 10 in Vermont, and 10 in Washington. In order to generalize the results of the survey to the universe of VS programs, cases (programs) were weighted to reflect the relevant programs in the SAPR database as a whole.

Measures

We collected data from the VS programs in our sample using a telephone interview and a faxed questionnaire. The faxed questionnaire covered topics for which a respondent might have to examine files or records, or check with other staff, and thus do not lend themselves to telephone interviews. Fax questionnaire topics included information about budgets and funding sources; employee, volunteer, and victim numbers, characteristics and assignments to various activities related to violence against women; the relative contribution of STOP funding in comparison to other funding sources; and statistics on victims served in various years, by the host agency and the STOP-funded VS program. The fax form also collected information important to future phases of this research project, including the boundaries of the VS program's catchment area and contact information for law enforcement and prosecution agencies in its community.

The phone interview included questions about the host agency that received STOP funding and its activities, funding, crime focus, and employees. VS programs were asked about the nature of their STOP-funded program, experiences with state STOP agencies, changes in the legal system since STOP funding became available, outreach strategies, and the impact of the VS program on the number of victims served and the ability of the community to meet the needs of women victims of violence.

The major portion of the interview, however, focused on the extent to which the STOP-funded VS program works with other agencies in its community to address violence against women. All respondents were asked specifically about their interactions with law enforcement, prosecution, and other victim services agencies. They reported the personnel levels of the agencies involved in the interaction, the extent to which the interaction is institutionalized, and the nature of the interaction. Respondents were asked to identify which agency in the community was the weakest link in the network of services for women victims of violence. In addition, they were asked to identify the two primary agencies with whom they have the most or most meaningful contact to serve women victims of violence. They reported the personnel levels involved, the extent to which the interaction is institutionalized, and the nature of the interaction with these primary agencies. They also reported the extent to which they communicate, coordinate, and collaborate with their primary agencies, how successful the partnerships have been, how necessary the partnerships are, and any barriers to working with the agencies.

In addition, respondents were asked to map the service network in their community. To do this, they first identified every agency in the community that provides services to women victims of violence. Next they were asked the extent to which each agency identified interacts with others to serve such women in the community. Agencies were connected on the map if they had institutionalized or formal commitments to integrate services for women victims.

After the interviews were completed, the research team rated each VS program on its level of communication, coordination, and collaboration with other agencies in their local area, and whether or not agencies in the community are organized into a CCR. Each VS program was rated by two trained interviewers. Interviewers reviewed the interactions reported by VS programs with law enforcement, prosecution, victim services, and their two primary agencies. Interviewers also reviewed responses to specific items about communication (e.g., frequent or regular phone contact about agency services or violence-against-women issues, referring clients to one another's agencies), coordination (e.g., providing training to one another's staff, facilitating referrals to one another's agencies), and collaboration (e.g., sharing funding or mission statements, integrating services), as well as open-ended questions regarding the nature of the work with primary agencies, the service network map, and the interviewer synopsis (in which interviewers noted their perceptions of the extent to which the community interacts). Only positive interactions with other agencies were included in the ratings; negative interactions between agencies were not coded. Ratings were compared and discrepancies were resolved, resulting in one rating for each construct (communication, coordination, collaboration, and CCR).

The communication rating has four levels: 1 = little or no communication with other agencies; 2 = some communication with other agencies, but not high levels of communication; 3 = good communication with some, but not most agencies; and 4 = good communication with most other agencies or all other agencies in the community. The coordination and collaboration ratings have three levels: 1 = little or no coordination/collaboration with other agencies; 2 = good coordination/collaboration with some, but not most agencies; and 3 = good coordination/ collaboration with most other agencies or all other agencies in the community. The CCR rating has two levels: 0 = not a CCR and 1 = CCR. Communities were rated as CCRs if the victim service, law enforcement, and prosecution agencies had institutionalized commitments to work with one another and provided integrated services to address domestic violence, or were rated as CCRs if the victim service, law enforcement, and prosecution agencies, and the medical community had institutionalized commitments to work with one another and provided integrated services to address sexual assault.

Other specific questions will be described as results for them are presented throughout the report. Also, other scales, combinations of questions, or ratings will be described as they are discussed.


Chapter 3
VS Programs In The Sample And In STOP Overall

HIGHLIGHTS
  • All VS programs included in our sample of 200 programs offered direct services to victims. Within the STOP program as a whole, most, but not all, VS programs offer direct services.


  • Eighty-eight percent of the sampled VS programs had been funded for two years or more. However, according to information recorded on the Subgrant Award and Performance Report (SAPR) form, only 17 percent of these same programs were funded for two or more years, and the same was true for only 4 percent of all VS programs. Plans are underway to eliminate this inability to detect multiyear funding on the SAPRs in future years.


  • The proportions of VS programs with grants greater than $10,000, making special efforts to reach underserved communities and to enhance coordination and communication between agencies within the community, were similar for our sample of 200 VS programs and for VS programs defined by funding category or by recipient agency.


  • Victims were the number one beneficiary of STOP funds for sampled VS programs and all STOP-funded VS programs defined by funding category or by recipient agency.


  • The crime focus of STOP projects and the geographic focus of STOP projects were similar for VS programs studied and all STOP-funded VS programs in the SAPR database.


  • Our sample selection process succeeded in producing a sample of VS programs that strongly resembles the universe of STOP-funded VS programs on every dimension available for comparison on the SAPRs.

Introduction

This chapter provides a basic description of the 200 victim service programs interviewed for this project and compares these programs to all VS programs for which SAPR information was available at the time of sampling. Descriptive characteristics for which information is available on the SAPRs include types of service offered, length of time the STOP program has been funded, amount of subgrant, crime focus, intent of the project, geographic focus, types of additional sources of funding, and the top six intended beneficiaries of the STOP funding.

The database used to generate the statistics reported in this chapter contains 6,527 SAPRs submitted by states as of November 15, 1999. Of these, 2,788 were defined as VS programs using the funding category (law enforcement, prosecution, victim services, or discretionary) designated by the state STOP coordinator. If one uses, instead, the nature of the recipient agency as the criterion, focusing only on subgrants to private nonprofit victim service agencies, 3,321 SAPRs qualify as funding VS programs. We report information based on both definitions, and compare both to the characteristics of the VS programs in our sample.

Three criteria were initially specified to select VS programs for the sample for this project. We wanted a VS program to be using STOP funds to offer direct services to women victims of violence, to have been funded by STOP for at least two years, and to have received at least $10,000 in STOP funds. In reality, we were not able to use the SAPR database to establish the number of years a VS program had been funded, so we had to drop that criterion (although the vast majority of the programs we actually interviewed had, in fact, received at least two years of STOP funding). In addition, in some states we had to relax the $10,000 criterion in order to get as many VS programs as we needed from that state (see chapter 2 for a description of the full sampling strategy for this project). As the information presented in table 3.1 makes clear, the programs in our sample do not differ in any particular from the universe of STOP-funded VS programs, however defined. So our goal of selecting a representative, as well as a random, sample was achieved.

Sample Criteria

All VS programs included in our sample of 200 programs offered direct services to victims (table 3.1). Within the STOP program as a whole, most VS programs offer direct services, whether defined by funding category (86 percent) or by recipient agency (97 percent). Direct service to victims included both services to a woman and advocacy on her behalf with other agencies. (General system advocacy, as opposed to individual case advocacy, was not considered a direct service.)

The value of a STOP program can really only be assessed when it has been given a sufficient amount of time to have an effect on a community. For this reason, we wanted the sampled VS programs to have been operating for two or more years. Unfortunately, we could not derive this information from the SAPRs, and so had to drop it as a criterion. However, the agencies, during interviews indicated that 88 percent of the sampled VS programs had indeed been funded for two years or more. What is interesting is that according to information obtained from the SAPR database, only 17 percent of these same programs were funded for two or more years and only 4 percent for VS programs defined by funding category or by recipient agency. Plans are underway to make it possible to link all awards to a given program, regardless of the year in which the award was made.

The criterion that our sampled VS programs have subgrants of $10,000 or higher was not difficult to meet because the vast majority of all VS programs have subgrants this large. Of all VS programs defined by recipient agency, 93 percent had grants greater than $10,000, as was true for 89 percent of VS programs defined by funding category, and 94 percent of the sampled VS programs. For the remaining 6 percent, the requirement was waived because we needed the programs from particular states in our sample.

Table 3.1
Sample VS Programs Compared to the Universe of STOP-Funded VS Programs (weighted percentages)
Criterion/Description Sampled VS Programs (N = 200) All VS Programs, Defined by Funding Category (N = 2,788) All VS Programs, Defined by Recipient Agency (N = 3,321)
Offers direct service 100 86 97
Intent, on SAPR, to:
provide direct VS or advocacy 90 83 85
make special efforts, underserved 62 56 63
enhance coordination / communication within the community 37 32 38
Intended DIRECT beneficiaries:
Victims 93 86 97
Law enforcement personnel 41 29 35
Nonprofit victim service providers 32 29 34
Prosecution personnel 31 23 28
General public 31 21 25
Children 25 29 27
STOP project operating at least two years
Information from SAPR 17 4 4
Information from interview 88 NA NA
Grant of at least $10,000 94 89 93
Crime focus:
Domestic violence 81 77 83
Sexual assault 51 43 50
Stalking 9 12 14
Geographic focus:
Local 10 7 11
Regional 30 26 28
Statewide 1 7 10
Additional sources of funding for project:
Other federal 18 16 15
State government 11 12 12
Local government beyond match 12 8 7
Private/other 12 9 0
Source: Urban Institute analysis of SAPR database as of November 15, 1999.
Note: "Funding category" means the state STOP administrator designated the subgrant as being funded from the "victim services" category. "Recipient agency" means the subgrant went to a private nonprofit victim service agency. NA = Not Applicable.

Varieties of Victim Service

Many agencies indicating that they do direct service were not always consistent in their SAPR reporting, as can be seen in table 3.1. For example, all sampled victim service agencies indicated they do direct service as part of their STOP program, but only 90 percent responded that the specific intent of their project was to provide direct service or advocacy. The same problem exists with the VS programs defined by funding category. Eighty-three percent said that their intent was to provide direct service or advocacy, while 86 percent of the same agencies responded that they actually expected to have victims as direct beneficiaries (i.e., as service recipients). The biggest difference occurred in the VS programs defined by recipient agency, for which only 85 percent indicated that their intent was to provide direct services, yet 97 percent indicated that they expected to serve victims as part of their STOP program.

Providing direct victim service or advocacy was not the only alternative for the intent of the STOP program. Nearly two-thirds (62 percent) of the VS programs sampled included a special effort to reach underserved communities, and 37 percent intended their STOP program to enhance coordination and communication between agencies within the community. The percentages for VS programs defined by funding category or by recipient agency were basically similar.

Intended Direct Beneficiaries

Although the majority of STOP programs are intended to benefit victims, there are many other direct beneficiaries that the funds serve. For sampled VS programs and all STOP-funded VS programs by either definition, victims were the number one beneficiary. Law enforcement personnel were the second highest intended beneficiaries, and nonprofit victim service providers were the third highest.

Crime Focus

The majority of STOP programs expected to focus on domestic violence. While approximately half (51 percent) of the sampled agencies listed sexual assault as a crime focus, nearly 81 percent listed domestic violence. Only 9 percent mentioned stalking. These numbers nearly match those of the VS programs defined by recipient agency: 83 percent for domestic violence, 50 percent for sexual assault, and 14 percent for stalking. The percentages for VS programs defined by funding category are slightly lower, but not significantly different: 77 percent for a focus of domestic violence, 43 percent for sexual assault, and 12 percent for stalking. It is important to note that agencies were not limited to choosing one of these categories. These percentages reflect agencies that focus on domestic violence only, sexual assault only, stalking only, or any combination of the three, therefore the total percentages add up to more than 100 percent.

Geographic Focus

Of those reporting the geographical scope for their project (many did not), most STOP programs operated, geographically, on a regional level. Ten percent of sampled VS programs operated only on a local level, while 30 percent focused their program regionally, and only 1 percent attempted to focus their STOP program on a statewide level. The universe of VS programs, however defined, did not differ from these figures.

Additional Sources of VS Program Funding

Many of the VS programs combined their STOP funding with money from additional sources. The four main sources were other federal funds (for example, ( Victims of Crime Act (VOCA) funds, other VAWA funds, Family Vioilence Prevention Services Act (FVPSA) funds, Community Oriented Policing Services (COPS) grants, etc.), state government funds, local government funds, and/or private funds. Approximately 18 percent of the sampled VS programs combined their STOP funds with federal funds, and a little over one-tenth of the agencies also used state government (11 percent), local government (12 percent), and private (12 percent) funding. Again, many SAPRs did not report this information and again, those that did so looked essentially similar to the VS programs in our sample.

Conclusion

Our sample selection process succeeded in producing a sample of VS programs that strongly resembles the universe of STOP-funded VS programs on every dimension available for comparison using SAPR information. These results give us confidence that the analysis reported in the following chapters will, as intended, be representative of all VS programs funded through the STOP program.


Chapter 4
STOP-Funded VS Programs And Their Agencies

HIGHLIGHTS
  • One way that STOP funding helped most was to increase the number of locations and/or mechanisms through which women could access victim services. Most host agencies offered services (STOP-funded and otherwise) in both disclosed (e.g., courthouses, health care facilities, and welfare offices) or undisclosed service sites (e.g., shelters).


  • Most host agencies specialized in either domestic violence or sexual assault services, but many agencies actually reported working on both issues. Both employees and volunteers were involved in providing direct services and outreach/education activities around domestic violence and sexual assault. In addition, one-third reported that their STOP project focused on both domestic violence and sexual assault issues.


  • Most VS programs used a portion of their STOP funds to support employee salaries. The resulting increased staff capacity allowed these agencies to provide new services to their current victim population. STOP funds have also allowed host agencies to bring existing services to more women, and have helped them tap into an entirely new victim population.


  • Victim service agencies undertook a variety of direct service activities with STOP funds including legal/court advocacy, comprehensive safety planning, counseling, answering hotline calls, individual advocacy, medical advocacy, first response, and shelter.


  • Some services stand out as particularly likely or particularly unlikely to be supported by STOP funds:

  • Court advocacy and participation in a multidisciplinary first-response team were most likely to be STOP-funded or to not exist in an agency. Very few agencies supported these activities without using STOP as a funding source.

  • STOP funds were used to support major portions of projects focusing on collaboration, training, and policy/protocol development activities. STOP funding allowed these activities to proceed at a more extensive level than had been possible before STOP.

  • Host agencies were relatively unlikely to use a STOP subgrant to support shelters, offer legal representation, or answer a hotline.

  • STOP funds accounted for less than half the annual budget of most host agencies. Most agencies earmarked STOP funds for specific projects.


  • Results suggest that STOP is affecting the number of women who receive needed services related to their experiences of domestic violence or sexual assault. However, it appears to be relatively difficult for many VS programs to provide information on the number of women served from year to year.

Introduction

This chapter and the next one examine the information we collected about STOP-funded VS programs, their host agencies, and their communities. In this chapter we stay within the walls of the VS program and its host agency. We look first at the characteristics of the private nonprofit victim service agencies that host the VS programs.5 Then we look at the specific activities that make up the VS program, and how they fit into the larger host agency. Chapter 5 goes beyond the boundaries of the VS program and its host agency to identify the variety of services available in the VS program's community for women victims of violence and reviews how the program relates to the agencies that provide those services.

Overview of Host Agencies

Though all agencies that hosted VS programs provided victim services, they differed from each other on a number of dimensions. Differences in agency type, history, service focus, staffing, and location distinguish agencies in the sample and reveal the variation existing across the universe of STOP programs.

Host Agency Services

All host agencies provided direct services to victims. However, the type of agency varied somewhat. Nonprofit, non-governmental victim service agencies comprised the majority (90 percent) of VS program host agencies. The remaining 10 percent of VS programs were hosted by a variety of other nonprofit agencies, including those offering legal services (4 percent) and social services (2 percent).6

Host agencies also varied in the number of years they had been in operation. Ten percent had been open for less than 10 years, 36 percent for 10 to 19 years, and 54 percent for 20 years or more. Only 1 percent of all host agencies had begun with the receipt of STOP funds.

Most host agencies specialized in either domestic violence or sexual assault services, with 86 percent focusing primarily on domestic violence and 13 percent focusing primarily on sexual assault. Seventy-nine percent of all host agencies reported that more than half of their whole agency's activities focused on domestic violence, while 24 percent noted this level of concentration with respect to sexual assault, and 8 percent did so for stalking. Likewise, 80 percent of all agencies reported that more than half of their whole agency's annual budget was used for domestic violence services, while 24 percent noted this level of financial commitment for sexual assault services, and 14 percent did so for services related to stalking.

Host agency services were available to victims in many different types of location, with most host agencies offering more than one access point, or an access mechanism such as a hotline. Most host agencies offered services (STOP-funded and otherwise) in both disclosed or undisclosed service sites. Eighty-one percent provided services at locations known to the general public (disclosed sites), and 71 percent provided services at undisclosed sites (e.g., shelters whose location was kept a secret). Some agencies also provided services at other locations, such as courthouses (25 percent), health care facilities (8 percent), welfare offices (5 percent), and minority or special population service sites (3 percent). Though these latter sites were occasionally an agency's only service location, they typically functioned as satellite offices for a host agency. One of the ways that STOP funding helped most was to increase the number of locations and/or mechanisms through which women could access victim services.

Host Agency Staffing, Paid and Volunteer

Host agencies for VS programs employed staff in a variety of service and managerial positions. All host agencies employed "frontline" workers to interact directly with victims. Frontline workers often served as victim advocates, legal advocates, case managers, outreach coordinators, and therapists. Seventy-two percent of all host agencies employed "middle management" staff to fill roles as coordinators, program directors, supervisors, and the like. Approximately three-quarters of host agencies with middle management staff reported that these staff also provided at least some direct services to victims. All but one host agency had an agency head or leader (most often, executive directors), 59 percent of whom also provided some direct services to victims.

Host agencies for VS programs ranged dramatically in size, for both paid staff and volunteers. Host agencies reported having from 1 to 62 paid staff involved in direct services to women victims of violence, with half reporting 10 or more such staff, and half reporting fewer than 10 paid staff. They also reported having from 0 to 48 staff involved in outreach or education on violence-against-women issues, with half reporting between 2 and 5 such staff.

In addition to paid staff, victim service agencies have long relied on the assistance of volunteers to provide both direct service and outreach/education. Indeed, many victim service agencies have more volunteers than paid staff, and many of these volunteers devote a considerable amount of time to their volunteer activities. Host agencies for VS programs reported working with from 0 to 608 volunteers engaged in direct services to women victims of violence, with half reporting a volunteer corps of 17 or more, including 10 percent with 65 or more volunteers doing direct service. They also had volunteers working on outreach/education activities, reporting from 0 to 97 such volunteers (half reported working with 7 or more such volunteers, including 10 percent that had 27 or more).

From information about numbers of staff and volunteers, combined with the type of work they did related to violence against women, we calculated proportions of paid staff and volunteers engaged in either direct victim services or outreach/education work. We used information about numbers of staff involved in each type of work who focused exclusively on domestic violence, exclusively on sexual assault, and worked on both issues to calculate other proportions. The results appear in table 4.1. All respondents, regardless of their agency's concentration on domestic violence or sexual assault, were asked to describe their staff's involvement in these issues, so each column of table 4.1 includes responses from the entire sample. Once we examine the overall patterns, we will look at staffing patterns separately for agencies that focus exclusively on domestic violence, exclusively on sexual assault, or address their efforts to victims of both crimes.

Looking first at staff involvement in direct services to women victims of violence (first panel of table 4.1, first cell), we can see that for 51 percent of host agencies, between three-quarters and all of their staff are involved in direct services of all varieties related to violence against women. Another 29 percent of host agencies have half to three-quarters of their staff involved in such direct services, while very few (9 percent) have less than one-quarter of their staff involved in such services. The staff of host agencies are thus heavily oriented toward providing direct services related to issues of violence against women.

Table 4.1
Host Agency Staff and Volunteers Engaged in Violence-against-women Work
(weighted percentages)
    Of Staff/Volunteers Doing Any VAW Work, those Doing:
  Any VAW Work Work Relating Only to DV Work Relating Only to SA Work Relating to Both DV and SA
Staff involvement in direct VAW services
25% or less 9 56 91 46
26 to 50% 11 3 1 6
51 to 75% 29 9 0 4
76 to 100% 51 32 8 44
Volunteer involvement in direct VAW services
25% or less 17 66 85 43
26 to 50% 24 1 4 1
51 to 75% 18 4 1 1
76 to 100% 42 29 9 55
Staff involvement in VAW outreach/education
0 to 5% 17 64 87 41
6 to 15% 31 0 1 0
16 to 25% 32 2 0 0
26 to 100% 19 34 11 59
Volunteer involvement in outreach/education
0 to 5% 33 70 91 35
6 to 25% 25 0 1 1
26 to 50% 29 1 0 1
51 to 100% 13 28 8 63
Source: Urban Institute analysis of VS program telephone interview responses; n = 200.

The remaining three cells in this first panel of table 4.1 show, for staff involved in direct services, what proportion focused on domestic violence, sexual assault, or both. Forty-four percent of respondents indicated that three-quarters or more of their direct service staff were involved in serving victims of both domestic violence and sexual assault, while another 32 percent of respondents indicated that three-quarters or more of their direct service staff worked exclusively with victims of domestic violence. Far fewer programs (8 percent) reported this level of concentration on exclusive work with victims of sexual assault. The very high proportion—91 percent—of programs reporting that 25 percent or less of their staff work exclusively with sexual assault victims reflects the general dearth of VS programs with this exclusive focus (only 17 percent).

The Effects of Crime Focus. To see whether the crime focus of STOP projects and their host agencies were basically the same, or whether some host agencies requested STOP funds to do something different from (either broader or more concentrated) the work of the host, we divided the sample into STOP projects focusing only on domestic violence (87 projects), those focusing only on sexual assault (35 projects), and those with a dual focus (66 projects). Then, for each group, we looked at the types of crime that formed the focus for the staff of the host agency. The answer seems to be that the majority of STOP projects mirror the work of their host agencies. Thus three out of five STOP projects focusing exclusively on domestic violence were in host agencies where three-quarters or more of the staff focused exclusively on domestic violence. Almost half (47 percent) of STOP projects focusing exclusively on sexual assault were in host agencies where three-quarters or more of the staff focused exclusively on sexual assault. And 7 out of 10 STOP projects focusing on both domestic violence and sexual assault were in host agencies where three-quarters or more of the staff also worked with victims of both crimes.

Only a few agencies went in the direction of broadening their crime focus with their STOP funding. Only 3 percent of host agencies used their STOP funding to develop a dual focus when they had previously had an exclusive focus. However, about 3 out of 10 STOP projects with an exclusive crime focus, either domestic violence or sexual assault, were in agencies with very high proportions of staff devoted to a dual focus. These agencies appear to have decided to apply their STOP-supported efforts to one type of crime. In doing so, they undoubtedly had one or more reasons, which may have included their experiences of which victims needed the most help, which activities they were most likely to be successful in performing, which activities would be most likely to stimulate cooperation from other agencies in the community, or which activities they could fund in other ways.

Volunteers Doing Direct Services. The second row of table 4.1 reports the involvement of volunteers in direct services. The volunteer corps of host agencies were less heavily devoted to direct services than were agency staff, but 42 percent of agencies said that three-quarters or more of their volunteer corps engaged in direct victim services. Within this group of volunteers, the distribution of those who focused exclusively on either domestic violence or sexual assault, or who worked on both, was fundamentally the same as the distribution for paid staff. Analyses conducted separately by the crime focus of the STOP project also mirrored the results for paid staff.

Outreach and Community Education Activities. The last two rows of table 4.1 show results for staff and volunteer involvement in outreach and community education activities related to violence against women. Note that the percentage categories for outreach/education are very different than for direct service, and they are also different for staff compared to volunteers. Relatively few host agencies devoted much of their paid staff time to outreach and education, with 81 percent reporting that one-fourth or less of their staff engaged in these activities. For 58 percent of host agencies, the involvement of the volunteer corps was this low, but the remainder used many more of their volunteers for outreach and education activities. Staff and volunteer likelihood of devoting themselves exclusively to domestic violence or sexual assault, or to both, was essentially similar. About three out of five programs reported that 25 percent or more of their volunteer corps doing outreach/education covered both domestic violence and sexual assault; while another three in ten reported this level of volunteer effort going to outreach and education related only to domestic violence. At most, one in ten agencies devoted a significant proportion of volunteer time to an exclusive focus on sexual assault outreach and education. Staff and volunteer commitment to doing outreach/education with respect to each crime followed directly from the crime focus of the STOP project.

The STOP-Funded VS Program

STOP-funded VS programs filled different niches in their host agencies. In some cases VS programs supported services entirely new to a host agency, and in other cases they allowed the agency to expand existing services to be more comprehensive or to reach more women. In many cases the VS program served several functions. Sixty-two percent of VS programs reported that STOP funds have allowed their agency to provide new services to its current victim population (figure 4.1). In other words, the same victims were being served through STOP funding, but could either receive more services (e.g., more counseling hours, longer stays at a shelter) or new types of service as a result of STOP. Likewise, 72 percent of VS programs reported that STOP funds have allowed their host agency to bring existing services to more women, especially women who would not otherwise have been able to receive services. In addition, 70 percent of the VS programs have used STOP funds to tap into an entirely new victim population, such as women living in a neighboring county, women with language/cultural service barriers, or women who go to court for protective orders but who would not otherwise seek VS program services.

Thus STOP-funded VS programs have heightened the service capacity of their host agencies. More than half (55 percent) of all victim service agencies credit STOP with bringing new, first-of-their-kind services to their communities, while 45 percent of all agencies have used STOP funding to supplement previously existing work/programs.

VS Program Activities

Victim service agencies undertook a variety of activities with STOP funds (second column of table 4.2). Direct service components such as legal/court advocacy (77 percent), comprehensive safety planning7 (58 percent), counseling (53 percent), answering hotline calls (50 percent), individual advocacy (49 percent), medical advocacy (48 percent), first-response (34 percent), and shelter (30 percent) ranked among the most common VS program activities.

STOP-funded VS programs were less likely than host agencies to offer any given type of direct victim service, but some types of service stand out as either particularly likely or particularly unlikely to be supported by STOP funds. Of prime importance, considering some of the emphases in the VAWA legislation on promoting multiagency responses to violence against women, is that STOP is being used in a major way to foster this type of response. Two activities that entail serious interagency involvement—court advocacy and participation in a multidisciplinary first-response team—were most likely to be STOP-funded or not to exist, as depicted by the narrow black band at the top of their bars in figure 4.2. Ninety-four percent of the agencies offering court advocacy did so with STOP funds, while 82 percent of the agencies participating in a first-response team supported that effort with STOP funding. Multidisciplinary first-response teams appear to be rather rare (only 22 percent of host agencies are involved in them), but when they occur, STOP funding apparently plays a major role in their existence. Conversely, STOP funding was not used extensively to fund basic shelter services.

Three activities indicating a serious commitment to interagency work are ones that indirectly benefit women victims of violence by helping to change the system of response (as opposed to offering direct service). They were also very likely to be supported by STOP funds. For example, 91 percent of all VS programs reported engaging in interagency collaboration, 95 percent said they offered domestic violence and/or sexual assault training to other agencies, and 61 percent participated in policy or protocol development with other agencies. STOP funds were used to support collaboration, training, and policy/protocol development activities in 73, 66, and 61 percent, respectively, of the instances where they occurred. STOP funding allowed these activities to proceed at a more extensive or intensive level than had been possible before STOP.

Conversely, host agencies were relatively unlikely to use a STOP subgrant to support shelters, offer legal representation, or answer a hotline (only 39, 46, and 57 percent, respectively, of host agencies offering these services did so). Some of this pattern may be due to having other sources of funding for these activities, as they tend to be among the oldest and best established of services for women victims of violence. Thus, VS programs seem to use their STOP funds to supplement their activities in the direction of developing a greater capacity to serve women throughout the various systems in their community and to develop more extensive working relationships with other agencies to do so.

Table 4.2
Activities of STOP-Funded VS Programs and Their Host Agencies
(weighted percentages)
Type of Activity Activity Conducted by Host Agency Activity Conducted by STOP-Funded VS Program Of Host Agencies Doing Activity, Proportion Supporting It with STOP Funds
Direct victim services
Court advocacy 82 77 94
Comprehensive safety planning 77 58 75
Counseling 81 53 65
Answering hotline calls 87 50 57
Case advocacy (helping individual women get public benefits, housing, employment, etc.) 81 49 60
Medical advocacy 73 48 66
First-response 48 34 71
Shelter 76 30 39
Multidisciplinary first-response team 22 18 82
Legal representation 26 12 46
Activities indirectly benefiting victims
Interagency collaboration 91 66 73
Providing DV/SA training or education 95 63 66
Policy/protocol development 61 37 61
Batterer intervention programs 23 5 22
Source: Urban Institute analysis of weighted VS program telephone interview responses; n = 200.

STOP Funding for VS Programs

No host agency relied solely on STOP for its funding. In fact, STOP accounted for less than half the annual budget of over 70 percent of host agencies. Host agencies usually earmarked STOP funds for specific program components. However, 13 percent reported that rather than earmarking funds, they instead blend their STOP funds into their general operating budget to be used for general program support. Interview responses indicated that host agencies received STOP funding over the course of five years (from FY 1995 to FY 1999), though most received it for the first time in either FY 1996 (32 percent) or FY 1997 (45 percent) (table 4.3). All agencies received STOP funds for two or more years. By FY 1997, 96 percent of host agencies had procured STOP funding at least once. Ninety-eight percent of agencies had their funding renewed for FY 1999. Additionally, STOP funding levels increased over time (table 4.3).

Table 4.3
Years and Amounts of STOP Funding for VS Programs
(weighted percentages)
Subgrant Size Percent of VS Programs:
FY 1995 FY 1996 FY 1997 FY 1998 FY 1999
No subgrant 87 62 25 14 18
Under $10,000 5 8 4 3 2
$10,000 to $19,999 4 10 17 15 15
$20,000 to $49,999 2 12 31 44 41
$50,000 to $99,999 1 8 20 22 21
$100,000 or more 0 3 4 3 4
Source: Urban Institute analysis of weighted VS program telephone interview responses; n = 200.

VS Program Functioning

Nearly all VS programs were currently underway when we conducted the interviews for this study (Spring 2000). However, implementation status varied. Most VS programs (81 percent) were fully up and running at the time of their phone interview, and 4 percent were beginning a second program component in addition to their first one. Only 3 percent of programs had just begun, and only 4 percent reported that they had made some progress but still had a way to go to achieve full operating status.

Service Intensity

Most VS programs (99 percent) used a portion of their STOP funds to support employee salaries, which often allowed these staff to devote considerable time to individual women. Women received differing amounts of staff time depending on the nature of the services being provided. In most VS programs (55 percent), staff spent an average of 10 hours or less serving an individual woman. Twenty-one percent of VS programs estimated spending between 10 and 25 hours of staff time per victim; 8 percent estimated spending between 25 and 50 hours; 8 percent estimated spending between 50 and 100 hours; and 7 percent estimated spending 100 hours or more of staff time per victim. Obviously these more intensive programs are likely to be shelter or long-term transitional or ameliorative programs involving more than one staff person per woman served. There may also be some mis-estimation occurring, as programs may have reported the number of hours women spend with them (e.g., 100 hours would translate into two and a half weeks in a shelter) rather than actual direct staff contact with an individual woman. Please note, also, that we do not have any estimates of the time that VS programs spent with women before STOP, so we cannot tell whether the figures just reported represent an increase over pre-STOP service intensity or not.

Accessing the VS Program

Survey respondents reported that women accessed their VS program services through a variety of means. Sometimes they found out about services through word of mouth or agency outreach activities, and other times they were referred to the VS program by other community agencies (table 4.4). Most STOP programs (88 percent) received referrals from law enforcement agencies. Many also received referrals from health agencies (60 percent), government social service (57 percent), prosecution (52 percent), and other private nonprofit victim service agencies (37 percent).8

Table 4.4
Referral Sources and Resources for VS Programs
(weighted percentages)
Agency Type VS Program Receives Referrals from Agency VS Program Makes Referrals to Agency
Law enforcement 88 57
Health agency (e.g., emergency room, clinic) 60 57
Governmental social service agency (e.g., welfare, child welfare, housing) 57 76
Prosecution agency 52 49
Another private nonprofit victim service agency 37 50
Mental health/substance abuse agency 33 65
Court 32 29
Religious organizations/clergy 17 9
Legal services 9 45
Government victim service agency 9 11
Community service agency (nongovernmental) 6 14
Professional association 6 7
Judges 4 3
Source: Urban Institute analysis of weighted VS program telephone interview responses; n = 200.

Once women contacted a VS program, they could usually receive some type of service immediately, such as counseling, referrals to other services, or information about community systems. Only 7 percent of VS programs reported a waiting period before any type of service was provided. Some services required advance scheduling (e.g., court advocacy), could only occur on the timetable of another agency, or required waiting until space was available (e.g., the shelter was full, or a counseling group was not starting for a few weeks). Over half of all VS programs with waiting periods mentioned a 4- to 10-day wait, and over one-third reported a wait of 11 days or more, usually for shelter or specialized services.

Making Referrals to Other Agencies

When VS programs could not provide needed services in-house, they referred clients to other providers for assistance (table 4.4). For example, over three-quarters of victim service agencies referred clients to government social service agencies (e.g., welfare, child welfare, housing, and similar agencies), and 65 percent referred clients to mental health and/or substance abuse treatment agencies. Nearly 60 percent of all victim service agencies referred clients to law enforcement and health agencies, and roughly half referred clients to prosecution agencies, agencies offering civil legal assistance, and/or other private nonprofit victim service agencies.

Outreach

In keeping with one of the aims of the STOP program, to reach more women who previously did not receive the benefits of victim service programs, STOP-funded VS programs undertook activities with a specific outreach intent—to bring their services to a larger portion of their community (table 4.5). In examining this information, please note that many other agencies did the same activities but did not name them as part of their outreach effort, or did not think of them as part of their STOP program. For example, 87 percent of agencies had hotlines, but only 32 percent name them as an outreach strategy within their STOP program.

Agencies employed different strategies to make women aware of and encourage women to use their STOP programs. Community education programs were by far the most heavily used outreach strategy, with 84 percent of VS programs employing the strategy. Use of flyers (74 percent), public service announcements (66 percent), newspaper advertisements/articles (48 percent), and posters (47 percent) to spread information about services was also common among STOP programs. Interestingly, approximately 40 percent of programs viewed their collaboration and referral systems with other agencies as important outreach strategies. This finding offers direct support that the collaborative emphases of the VAWA are having an effect for local VS programs.

Table 4.5
Outreach Strategies of VS Programs
(weighted percentages)
Strategy Percent of VS Programs
Using the Strategy
Community education programs 84
Flyers 74
Public service announcements (radio and TV) 66
Newspapers 48
Posters 47
Collaborating/referral system with other agencies serving victim populations 42
Community events (e.g., health fairs) 42
Word of mouth among women 40
Victim service information cards distributed by law enforcement 34
Hotline 32
Materials/services in native languages 20
Source: Urban Institute analysis of weighted VS program telephone interview responses; n = 200.

Though most VS programs used several outreach strategies, they were often able to identify a single strategy as their most successful one (only 15 percent did not do so). VS programs frequently cited community education programs as their most successful strategy (26 percent), though other strategies such as collaboration/referral systems (13 percent) and flyers (8 percent) were also mentioned. Overall program quality might also be said to be an outreach strategy of sorts, as 14 percent of programs relied on word of mouth among women as their most successful outreach strategy.

We found no differences in the probability that one strategy would be preferred over another based on type of host agency, but one thing stood out with respect to the VS program's involvement with other community agencies. VS programs that were not on a task force with their two primary agencies were more likely to report that word of mouth was their most successful outreach strategy than were VS programs that shared task force membership with their primary agencies. The same was true for VS programs whose community was not rated as a CCR, compared to those in a CCR community. It would seem that when systems do not work, word of mouth is the only thing left to rely on.

Victims Served

If STOP funding did nothing else, we would expect it to increase the number of women receiving services from VS programs. We might also anticipate that having STOP funding might enable VS programs to offer women different types of service, or get more of their needs met. Unfortunately, it appears to be relatively difficult for many VS programs to provide the number of women they served from year to year. This difficulty is compounded by the fact that the activities undertaken by many STOP-funded VS programs did not exist in their host agencies before the VS program began, so there are no "before" data in that sense. It is also sometimes difficult for many agencies to sort out which women should be counted as clients of the host agency and which should be counted as clients of the STOP-funded VS program, when people such as researchers try to make them choose. Add to this the confounding influence of federal and state fiscal years, calendar years, and program years counted from the day the agency actually received permission to begin spending STOP funds, and calculating a before-and-after statistic can be close to impossible. When we tried, we ended up with so much missing data that we had before-and-after information for only 40 percent of the VS programs in our sample, and we were not sure we could trust even that.

Our solution, therefore, for purposes of this report, is to present the full range of responses to various questions about service levels of the host agency. Also, because host agencies differed so much in size, we also think it is important to convey an idea of victims served by small, medium, and large programs. To do this, we represent small programs by showing the size (in number of victims served) of the program at the 25th percentile (that is, one-fourth of all host agencies are smaller than this one, and three fourths of all host agencies are bigger). For medium-sized programs we show the number of victims served by the program "in the middle," for which half the host agencies serve more victims and half serve fewer victims. Finally, to represent large agencies we show the number of victims served by the agency at the 75th percentile (only one-fourth of host agencies are bigger). All of this information is displayed in table 4.6, which also shows the range in the number of victims receiving particular types of service from host agencies in particular years. This information is tempered by knowing how many agencies supplied data for a particular type of service, as indicated in the first column of table 4.6.

To walk through one example from the table, 76 VS programs reported the number of domestic violence victims their host agency served in 1995. This number ranged from 26 women up to 19,985 women. One-fourth of these host agencies served 175 or fewer domestic violence victims (that is, the agency at the 25th percentile served 175), half served less and half served more than 480, and one-fourth served 1300 or more women. Obviously there are some very large agencies in this sample; the reader may want to know that figures for the 75th percentile barely changed when the few largest agencies were dropped from the calculations (and the 25th and 50th percentiles did not change at all).

VS programs were asked for statistics with respect to the following types of services and recipients for their host agency, for each calendar year from 1990 through 1999 (we report only 1995, the year before STOP funding became available, through 1999):

Table 4.6
Victims Served by the Host Agencies of VS Programs, FY 1995 through FY 1999
(weighted percentages)
Victims Served/Services Received from VS Program's Host Agency Number of Programs Answering Range Number of Women Served by Host Agency at the ____ Percentile:
25th 50th 75th
Domestic violence victims served in calendar year:
1995 76 26 - 19,985 175 480 1,300
1996 108 7 - 22,005 209 528 1,448
1997 121 2 - 23,390 226 459 1,331
1998 139 15 - 22,545 255 534 1,354
1999 141 7 - 28,558 215 598 1,413
Percentage change, (1999-1995)/1995: + 23% + 25% + 9%
Domestic violence victims receiving assistance with protective/restraining orders in calendar year:
1995 55 0 - 19,253 26 86 156
1996 80 0 - 19,385 35 95 171
1997 102 0 - 13,982 43 91 265
1998 117 1 - 15,221 59 113 292
1999 127 0 - 11,349 47 117 250
Percentage change, (1999-1995)/1995: + 81% + 38% + 64%
Domestic violence victims receiving legal system advocacy services in calendar year:
1995 33 0 - 19,253 44 128 272
1996 63 0 - 19,385 47 107 235
1997 83 0 - 13,982 41 128 296
1998 94 1 - 15,221 65 154 390
1999 106 0 - 11,349 45 187 422
Percentage change, (1999-1995)/1995: + 0% + 46% + 55%
Sexual assault victims served in calendar year:
1995 52 0 - 7,972 8 62 192
1996 79 0 - 7,380 5 63 235
1997 94 0 - 10,906 11 62 313
1998 110 0 - 12,745 13 73 287
1999 115 0 - 12,020 15 73 303
Percentage change, (1999-1995)/1995: ** + 18% + 58%
Sexual assault victims receiving legal system advocacy services in calendar year:
1995 27 0 - 1,495 1 11 72
1996 52 0 - 1,460 1 7 40
1997 61 0 - 2,400 1 24 60
1998 78 0 - 2,400 2 24 100
1999 86 0 - 2,916 3 25 91
Percentage change, (1999-1995)/1995: ** ** + 26%
Source: Urban Institute analysis of VS program telephone interview responses; n = 200.
** Too few for reliable change calculations.

  • Number of domestic violence victims—
  • Served by the host agency for each calendar year,
  • Assisted by the host agency to obtain protective/restraining orders,
  • Receiving individualized advocacy services related to the criminal justice system from the host agency.
  • Number of sexual assault victims—
  • Served by the host agency for each calendar year,
  • Receiving individualized advocacy services related to the criminal justice system from the host agency.

One can infer9 from the information in table 4.6 that an increasing number of programs received STOP funding for each type of service between 1995 and 1998, at which time things seem to have stabilized or the increase in programs slowed. One can also infer that fewer programs were funded to deliver services to sexual assault victims than to domestic violence victims and that these programs were generally considerably smaller (median program size for all domestic violence services is around 500 to 600, while median program size for all sexual assault services is around 60 to 70).

The information on the range of victims served indicates that in every year, STOP-funded VS programs varied tremendously in how many women were served. However, information on the different percentiles indicates that three-quarters of the programs in every category of service were far smaller than the few very large programs that comprised the top of the range. In addition, it is clear that significantly fewer women received legal advocacy services (services that help women negotiate legal systems such as civil courts for protection orders, issues of child visitation, child support, etc., and criminal courts to pursue the case against a batterer or sexual assailant) than received the more generalized victim services available through shelters, other domestic violence programs, and sexual assault programs. While half of VS programs served up to 500 to 600 women a year through all services, only one-quarter to one-fifth of these women appear to have received advocacy services with respect to cases pending against perpetrators with law enforcement or prosecution agencies or criminal courts, or with respect to civil court procedures such as protection/restraining orders.

Examining the question of whether STOP funds have supported services to more women victims of violence, it is hazardous to try to infer an answer from the information in table 4.6. First, it is important to know that of the VS programs that supplied information on victims served, only 45 percent drew this information from actual counts or records. The remainder estimated the figures they submitted, probably using a variety of dissimilar approaches to do so.

Second, there definitely are more programs being included in the statistics on victims served, from which one might be tempted to infer that more women would be benefitting because more programs were on board, even if the average number of women per program receiving services did not change. But this inference is a dangerous one to draw, because the obvious increases from year to year in programs reporting in table 4.6 may reflect only a propensity to begin keeping statistics in the year that STOP funding was first received. Therefore missing programs do not necessarily mean fewer services. Recall that only 1 percent of the host agencies of VS programs began with their STOP subgrant; the others were functioning for some years earlier, and should they have been able to supply data on service levels if their record keeping systems predated STOP funding.

However, the 25th, 50th, and 75th percentile levels do contain some indications that more women were being served, as agencies at these levels appear to have increased the number of women they served between 1995 and 1999. Looking only at the median (50th percentile), the data suggest that perhaps between 20 percent and 46 percent more women, depending on the specific service examined, were receiving help from VS programs' host agencies in 1999 than in 1995. The increases suggest that STOP is, indeed, affecting the number of women who receive needed services related to their experiences of domestic violence or sexual assault. These increases would be smaller if one began with 1996 statistics rather than those from 1995. This conservative approach might be warranted on the grounds that substantially more programs reported data for 1996 than for 1995.10

Conclusion

This chapter has examined the nature of STOP-funded VS programs and their integration into their own host agency. VS programs have fit into their host agencies in a number of ways, usually providing increased and often unique services to help women victims of violence. For the most part, they have received STOP support for two or more years and are firmly established in their communities, albeit almost certainly dependent on continued STOP funding to maintain their level of activities. They have paid particular attention to increasing access to services, including provision of services for the first time to many women who never before would use or could access victim services. Consequently, the evidence we have suggests that more women are using these services. In the next chapter, we explore the ways in which these VS programs interact with the legal system and other agencies in their larger community.


Chapter 5
VS Program Interactions with Other Community Agencies

HIGHLIGHTS
  • All VS programs reported interacting with at least one law enforcement agency, and most reported interacting with at least one prosecution agency and at least one other VS agency in the community.


  • VS programs reported primary partner agencies with whom they have the most or most meaningful contact:

  • 65 percent reported law enforcement agencies;

  • 42 percent reported prosecution agencies;

  • 25 percent reported social service agencies.

  • One-quarter of programs named both law enforcement and prosecution agencies as those with whom they partner the most to help women victims of violence.


  • Most VS programs reported involvement of every level of employee (frontline staff, middle management, and organizational leaders) in interactions with other agencies (law enforcement, prosecution, other VS agencies, and other types of agencies).


  • One-half of VS programs had formal policies or procedures to work with law enforcement, one-third had the same with prosecution, and one-quarter had the same with other VS agencies.


  • VS programs reported increases in five types of interaction with other agencies (law enforcement, prosecution, other VS agencies, and other types of agencies) since STOP funding. Over half reported these changes were due to their STOP-funded VS program, and between 11 and 31 percent reported changes were due to other STOP projects in their community.


  • Most VS programs communicate with their primary partner agencies in several ways. They share general information about violence-against-women issues, have frequent phone contact, have informal meetings, and refer clients back and forth.


  • Most VS programs also coordinate with their primary partner agencies in several ways. Most help one another on an as-needed basis with specific cases and facilitate referrals.


  • VS programs are more likely to provide training to law enforcement than to prosecution or other types of agencies. VS programs are more likely to receive training from other VS agencies than from law enforcement or prosecution.


  • VS programs collaborate with their primary partner agencies at several levels. Most participate on task forces with partners and strategize about how to reach women victims of violence. Fewer VS programs, although still over half, influence one another's agency protocols, provide integrated services to victims, or have a regular feedback mechanism regarding their collaborative work that helps them fix problems and shape new developments.


  • Of those who named law enforcement as a primary partner, 36 percent participated on a first-response team with them.


  • Of those who named prosecution as a primary partner, 26 percent reported interacting with them on a first-response team.


  • Three-quarters of VS programs participated in some form of violence-aga inst-women task force in their community. Every collaborative activity or arrangement was more likely to occur when the VS program and its two primary partners participated together on a task force.


  • VS programs in communities rated as CCRs were more likely to report each collaborative activity or arrangement than those not in CCRs but still participating on a task force with both primary agencies.


  • Task forces can be useful forums for agencies to work together, particularly in those communities where a CCR exists. However, the existence of a task force does not guarantee joint work or collaborative activities in communities. Likewise, some communities without task forces still participate in collaborative activities.

Introduction

The communities served by private nonprofit VS programs may contain many other agencies that also pursue activities directed toward women victims of violence. We include legal system agencies such as police, prosecution, and the courts among these, even though they do not offer "services" but rather respond to crimes. We were very interested in learning what STOP-funded VS programs know about the array of services for victims in their communities. We also wanted to know whether VS programs interacted with these different agencies, and the extent to which STOP funding had improved or expanded the network in which VS programs operate. This chapter examines the variety of services in VS program communities, and the extent to which VS programs interact with them.

VAW Services in VS Program Communities

Every VS program reported the presence of law enforcement agencies in its community. We asked every VS program about its interactions with law enforcement agencies, and every VS program said it interacted with at least one law enforcement agency in some way. We describe the nature of these interactions in a later section of this chapter.

Ninety-nine percent of VS programs also reported having prosecution agencies in their community. We asked every VS program about its interactions with prosecution agencies, and learned that all but 3 percent worked with at least one of them in some way. We describe these interactions below, in the section on VS program relationships with prosecution agencies.

The final type of agency about which we asked every VS program in our sample was "other victim service agencies." Virtually all (94 percent) worked with at least one other victim service agency. Seventy percent reported the presence in the community of at least one other private nonprofit victim service agency, and 15 percent reported having at least one governmental VS agency. In addition, 40 percent reported having one or more agencies with a special focus on serving racial, ethnic, cultural, language, or other minority populations and which offered a victim service component as part of its service array. As with law enforcement and prosecution, we describe the interactions of the sampled VS program with other victim service agencies in a separate section below.

Several other types of agencies offering services to women victims of violence were available in most of the communities served by STOP-funded VS programs. These included:

  • health care facilities such as emergency rooms and clinics (reported by 89 percent of VS programs);

  • government social service agencies such as cash assistance, child welfare, or housing agencies (reported by 88 percent of VS programs);

  • mental health and substance abuse treatment agencies (reported by 81 percent of VS programs); and

  • legal aid agencies (reported by 73 percent of VS programs).

Other types of services for women victims of violence were reported by significantly fewer VS programs. About half (52 percent) said that courts in their community had some arrangements for women victims of violence, 34 percent identified services through probation and/or parole agencies, and 22 percent each cited arrangements by or with judges and services through community service agencies.

Partner Agencies for VS Programs

Despite the presence of a great array of services and agencies offering assistance to women victims of violence in the communities served by VS programs, not all of them were equally involved in coordinated work. In addition to inquiring directly about interactions with law enforcement, prosecution, and other victim service agencies, we asked VS programs to identify the two agencies in their community, of any type, with which they had the most or the most meaningful interactions. Law enforcement and prosecution agencies were named most frequently, indicating that VS programs often enjoyed productive relationships with these essential agencies. These were followed by governmental social service agencies and other victim service agencies. The proportion of VS programs naming each type of agency as a working partner were:

  • 65 percent named a law enforcement agency,

  • 42 percent named a prosecution agency,

  • 25 percent named a governmental service agency such as a child welfare, cash assistance, or housing assistance agency,

  • 22 percent named another private nonprofit victim service agency,

  • 16 percent named courts or judges,

  • 11 percent named a legal aid program,

  • 8 percent named a health agency such as an emergency room or a clinic, and

  • less than 4 percent named various other types of agency, including community service agencies, mental health or substance abuse agencies, probation/parole agencies, or governmental victim service agencies.

For the most part, the crime focus of the VS program did not affect which agencies they named as primary partners. Programs focusing exclusively on domestic violence or on sexual assault, or working with victims of both, were equally likely to name law enforcement and prosecution as primary agencies, and these were the agencies they named most often. There were, however, a few exceptions for other agency types. VS programs focused exclusively on sexual assault were more likely to name a health agency as a primary partner than were programs with an exclusive domestic violence focus or programs focusing on both crimes. This makes a good deal of sense given the importance of forensic medical examinations as part of evidence gathering in sexual assault cases. VS programs working with domestic violence victims, either exclusively or along with sexual assault victims, were more likely than those with an exclusive sexual assault focus to name a social services agency or a court as a primary partner. These choices also relate directly to the needs of women facing issues of domestic violence for dealing with courts to get protection orders, and with social service agencies to obtain cash assistance, housing options, and other benefits.

In the remainder of this chapter, we describe how VS programs interact with the various agencies in their community. We look first at interactions with law enforcement, then with prosecution, then with other VS agencies, and then with other types of agency. These topics are followed with an examination of the ways that VS programs communicate, coordinate, and collaborate with the agencies they identified as their primary community partners. The chapter ends with a discussion of the role of task forces in developing services for women victims of violence.

VS Program Interactions with Law Enforcement

As already noted, all VS programs said they interacted with law enforcement agencies, and 65 percent named them as one of their two primary agencies. The interview asked them to describe the interactions between the two agencies, including which members of each agency interacted with each other, whether the agencies had formal policies or procedures for working together, whether interactions had changed since the VS program began receiving STOP funds, and whether the VS program attributed any changes to the effects of the STOP-funded VS program. Table 5.1 shows the responses to these questions. The VS program's crime focus did not make any difference for any of the responses reported in this table.

Table 5.1
Interactions of VS Programs with Other Community Agencies
(weighted percentages)
  Percent of VS Programs Interacting with:
Law Enforcement Agencies
(n = 200)
Prosecution Agencies
(n = 200)
Other VS Agencies
(n = 200)
Other Types of Agency
(n = 200)
Working relationships
VS program works with agencies of this typea 100 97 95 Not applicable
VS program names an agency of this type as a primary partner agencyb 65 42 22 72
n's for remaining answers: (n = 200) (n = 193) (n = 182) (n = 143)
Staff involvement —percent with interactions between VS and other agency staff (if agency has that level of staff) among:
Frontline staff 95 86 97 98
Middle management 89 85 92 83
Agency leadership 83 78 90 80
Policies —percent with formal policy or procedure guiding the work the two agencies do together 51 31 27 Not asked
Increased interactions —percent reporting increased interactions of the following types since receiving STOP funding:
1. Contact of any type 87 88 81 95
2. VS program helps women deal with agency 89 89 84 94
3. Agency refers to VS program 88 90 85 88
4. The two agencies coordinate their work 86 87 85 83
5. The two agencies do joint planning, funding, and/or have institutionalized level of commitment 81 83 82 60
Summary of increases in:
None (no changes reported) 7 7 13 1
One or two types of interaction 10 8 4 5
Three types 15 16 6 8
Four types 28 32 28 33
All five types of interaction 41 37 51 54
Percent saying:
Changes were due to VS STOP project 80 76 57 64
Changes were due to other STOP project 31 29 20 11
These changes were more true for relationship with some agencies of this type than with others 67 34 37 Not asked
Source: Urban Institute analysis of VS program telephone interview responses. aFor law enforcement, prosecution, and other VS agencies, responses are to interview questions 16, 17, and 18. bFor "Other Types of Agency," the agencies are those that respondents identified as a primary agency; responses are to interview question 20.

Very large proportions of VS programs reported the involvement of every level of employee in interactions between their program and law enforcement. We asked whether frontline workers of the two agencies interacted with each other, whether middle management did so, and whether the agency heads or leaders did so. (Some agencies did not have a middle management level; the percentages in table 5.1 reflect only those that did have this level of staff.) Ninety-five percent of VS programs reported cross-agency interactions among frontline staff, 89 percent reported such interactions among middle management staff, and 83 percent reported them between the leaders of the VS program and law enforcement agencies. Half (51 percent) of VS programs had a formal policy or procedure in place that specified how they would work together with law enforcement. This is a pretty remarkable level of commitment and was significantly more common than the level of formality achieved by VS programs with either prosecution agencies or other victim service agencies. However, these reports also indicate that almost one in five VS programs (17 percent) operate without connections between the leaders of VS programs and law enforcement agencies.

When asked whether specific types of interaction had increased between the two agencies since the advent of STOP funding for the VS program, most respondents said that this had happened (4 percent could not say). Eighty-one percent said they had increased their amount of joint planning, joint funding, and/or an institutionalized level of commitment to work together (figure 5.1). Equally high proportions of VS programs reported that contact of any type, advocacy work for individual women (as opposed to system advocacy), referrals back and forth, and coordination of the two agencies' actions with respect to victims had increased since STOP funding. Most (69 percent) reported that four or all five types of interaction had increased.

On the other hand, while these increases are desirable, they do not necessarily mean that all problems have been solved. They could be increases over very low levels of interaction to start with. Seven percent of VS programs reported that there had been no increases with respect to law enforcement in any of the types of interaction we examined, and 17 percent said that the leaders of law enforcement agencies were not involved with their own leaders in setting policy or procedures.

As might also be expected, not all law enforcement agencies were alike. Most VS programs operated in communities with more than one law enforcement agency. Two-thirds (67 percent) of VS programs reporting some changed interactions with law enforcement said that these changes were more true for their program's relationship with some law enforcement agencies than with others. This proportion is higher than prevails for VS program relations with prosecution or other victim service agencies, but that difference probably is due to the greater likelihood of having several law enforcement agencies compared to the probability of having several prosecution or other victim service agencies in the same community as the STOP-funded VS program.11

Of VS programs indicating some increased interactions between their program and law enforcement, 80 percent attributed these changes to the STOP funding that supports the VS program, 9 percent felt the changes were not due to STOP, and 11 percent declined to say (including the 7 percent who felt that no changes had occurred). In addition, 31 percent attributed the changes to STOP funding going to another project in their community. Thus most respondents felt both that some changes had occurred and that the changes should be attributed, at least in part, to STOP funding.

Of the VS programs that answered these questions and felt that change had occurred, 33 percent felt that the changes they reported in their interactions with law enforcement were due to the joint effects of their own STOP funding acting in conjunction with activities supported by at least one other STOP subgrant operating in their community (figure 5.2). On the other hand, 59 percent felt the changes were due solely to their own STOP subgrant, and only 1 percent felt the changes should be attributed completely to someone else's subgrant (7 percent did not think the changes were due to STOP at all).

VS Program Interactions with Prosecution

We noted above that 97 percent of VS programs reported interacting with prosecution agencies, and 42 percent named them as one of their two primary agencies. As with law enforcement, most VS programs reported that all levels of employee were involved in cross-agency interactions (table 5.1, second column). Eighty-six percent of VS programs reported interactions among the frontline staff of their own and prosecution agencies, 85 percent reported interactions among middle management staff (if these existed in the two agencies), and 78 percent reported interactions between the leaders of the VS program and prosecution agencies. About one-third (31 percent) of VS programs had a formal policy or procedure in place that specified how the two agencies would work together, which is fewer than for law enforcement but still impressive. However, it is also clear from these reports that more than one in five VS programs (22 percent) operated without connections between the VS program leader and the leaders of prosecution agencies.

When asked whether specific types of interaction had increased between the two agencies since the advent of STOP funding for the VS program, most respondents said that they had. Eighty-three percent said the two agencies had increased their amount of joint planning, joint funding, and/or an institutionalized level of commitment to work together. As many or more VS programs reported that contact of any type, individual advocacy work, referrals back and forth, and coordination of the two agencies' actions with respect to victims had increased since STOP funding. Most (69 percent) reported that four or all five types of interaction had increased.

However, 8 percent reported increased interactions with prosecution agencies of only one or two types, and 7 percent said that no types of interaction with prosecution agencies had increased (10 percent felt they could not say). And without the commitment of agency leadership, changing interactions or frontline and even middle management staff can only go so far to change a whole system of response. And, as with law enforcement, not all prosecution agencies were alike. Many VS programs were in communities with more than one prosecution agency, usually split by city and county, misdemeanor and felony, and sometimes other separations. One-third (34 percent) of VS programs reporting some changed interactions with prosecution agencies said that these changes were more true for their program's relationship with some agencies than with others.

Of VS programs indicating some increased interactions between their program and prosecution, 76 percent attributed these changes to the STOP funding that supported the VS program while 6 percent said that the changes were not due to STOP and 17 percent could not say (including the 7 percent who did not report any changes). In addition, 29 percent attributed the changes to STOP funding that went to another project in their community.

As with changed interactions with law enforcement agencies, most VS programs felt both that some changes had occurred and that the changes should be attributed, at least in part, to STOP funding. Of the VS programs that answered these questions and felt that change had occurred between themselves and prosecution, 33 percent (the same as for law enforcement) felt that the changes they reported in their interactions were due to the effects of their own STOP funding acting in conjunction with activities supported by at least one other STOP subgrant. On the other hand, 60 percent felt the changes were due solely to their own STOP subgrant, and only 2 percent felt the changes should be attributed completely to someone else's subgrant. Five percent did not think the changes were due to STOP.

VS Program Interactions with Other VS Agencies

Virtually all VS programs (95 percent) interacted with another VS agency in their community. However, only 22 percent named them as one of their two primary agencies (table 5.1). VS programs reported that every level of both agencies was involved in their interaction, but only about one in four (27 percent) had a formal policy or procedure in place for working together with the other VS agency. Involvement of VS agencies at the leadership level may be somewhat higher than that between VS programs and either law enforcement or prosecution, with 9 out of 10 VS programs reporting connections between the agency leaders, but the difference is not statistically significant.

Most respondents said that specific types of interaction had increased between the two agencies since the advent of STOP funding for the VS program. Eighty-two percent said they did more joint planning, had more joint funding, and/or had a higher institutionalized level of commitment to work together. As many or more VS programs reported that contact of any type, individual advocacy work, referrals back and forth, and coordination of the two agencies' actions with respect to victims had increased since STOP funding. Most (79 percent) reported that four or all five types of interaction had increased, and only 1 percent said there had been no change in the level of interaction between the VS program and other victim service agencies. Also, 37 percent of those who reported changes with other victim service agencies said that the changes were more true of interactions with some agencies than with others.

Fifty-seven percent of VS programs attributed increased interactions with other victim service agencies to the STOP funding that supports the VS program (lower than for changes with either law enforcement or prosecution agencies), 15 percent felt that STOP funding and the STOP project were not responsible for the changes, and 28 percent could not say (including the 13 percent who did not report any changes). In addition, 20 percent attributed the changes to STOP funding going to another project in their community.

As with changes between their agency and law enforcement and prosecution, most respondents felt both that some changes had occurred in their interactions with other victim service agencies and that the changes should be attributed, at least in part, to STOP funding. Of the VS programs that answered these questions and felt that change had occurred, one in four (26 percent) felt that the changes they reported in their interactions with other victim service agencies were due to the joint effects of their own and other STOP funding. On the other hand, 55 percent felt the changes were due solely to their own STOP subgrant, and only 3 percent felt the changes should be attributed completely to someone else's subgrant (17 percent did not think the changes were due to STOP).

To summarize attributions of changed interactions to the influence of STOP, 31 percent of VS programs reporting these changes with law enforcement, prosecution, or other victim service agencies attributed them to the joint effects of their own and other STOP subgrants. This finding reflects what we think is a fairly high degree of synergy operating in communities, reflecting the inclination of state STOP agencies to fund either joint projects or several related projects in a community. It probably also reflects the inspiration that one STOP project in a community can give to other local agencies to get their own subgrants and work together. Even without this joint impetus to change, however, change can and does occur as a consequence of a single STOP subgrant to a VS program. These changes are reflected in the 58 percent of VS programs reporting their belief that their changed interactions with law enforcement, prosecution, and other victim service agencies occurred under the influence of their own STOP subgrant.

VS Program Interactions with Their Primary Agencies

Interviewers asked VS program directors to identify "the two primary agencies you work with to serve women victims of violence. By primary I mean those with whom you have the most or the most meaningful contact to provide services."

Sixty-five percent of VS programs named a law enforcement agency as one of their two primary agencies, 42 percent named a prosecution agency, 22 percent named another victim service agency, and 72 percent named other types of agency. Less than 2 percent of VS programs failed to name two primary agencies. As described earlier in this chapter, these other agencies included governmental and nongovernmental social service agencies such as welfare, child welfare, housing assistance, community action, and family service agencies; courts, judges, or corrections agencies; legal aid; and health, mental health, and substance abuse agencies. The issues we explored about primary agencies focused on communication, coordination, and collaboration with STOP-funded VS programs. However, if a VS program identified as a primary agency one that it had not already described in response to questions about its interactions with law enforcement, prosecution, and other victim service agencies, we also asked questions about staff interactions by level of staff, and changed interactions as a consequence of the VS program's STOP subgrant. Responses to these questions complete table 5.1, where the fourth column reports results for "other types of agency."

Staff Involvement and Changed Interactions with Primary Agencies other than Law Enforcement, Prosecution, and Victim Services

Descriptions of staff involvement for other types of agency, by level of staff, did not differ from findings for VS program interactions with law enforcement, prosecution, and other victim service agencies, even though the latter categories in table 5.1 include many agencies that VS programs did not name as a primary agency. However, there are some very interesting ways in which these "other" agencies named as primary agencies by VS programs do differ from the interactions reported in the first three columns of table 5.1.

First, the proportion of VS programs (87 percent) reporting increases in four or all five of the types of interaction asked about was higher than for law enforcement agencies (at 69 percent) and prosecution agencies (at 69 percent), but not for other victim service agencies (at 79 percent). This was true even though fewer VS programs (60 percent) reported increases in joint planning, funding, and/or institutionalized commitment with these primary agencies than they did with law enforcement (81 percent) or prosecution (83 percent).12 Attribution of these changed interactions to receipt of STOP funding did not differ for interactions between VS programs and their primary agencies of "other types." On the other hand, fewer VS programs attributed changes with their primary "other type of agency" to a different (non-VS) STOP subgrant than was true for changes with law enforcement agencies.

Communication, Coordination, and Collaboration Activities of VS Programs and Their Primary Agencies

For the two agencies named as primary, we wanted to know the nature of their interactions with VS programs with respect to communication, coordination, and collaboration. We asked VS program respondents to describe these interactions with respect to a series of specific behaviors relating to each level of interaction. We expected to see some degree of hierarchy in their responses, such that more agencies would report communication activities than would report coordination activities, which in turn would be more common than collaboration activities. For the most part this is what we found, but there are some exceptions. Table 5.2 reports the results, separately for law enforcement, prosecution, other victim service, and other types of agency named as a primary agency by the VS programs in our sample.13

Communication. Very high proportions of VS programs, mostly above 90 percent, reported participating in each of the four communication activities with their primary agencies. No differences were found depending on the type of agency named as primary. Of the four communication activities we asked about (i.e., share general information about violence against women, have frequent or regular telephone contact about services, have informal meetings to share general information, and cross-refer clients), most VS programs said they did all four with each type of primary agency. This was true for 82 percent of VS programs with respect to a law enforcement as primary, 87 percent with respect to a prosecution agency as primary, 86 percent with respect to another victim service agency as primary, and 79 percent with respect to another type of agency as primary.

Table 5.2
Communication, Coordination, and Collaboration Activities
Between VS Programs and their Primary Agencies
(weighted percentages)
Activities that VS Programs Report as Part of Their Interactions with Agencies Named as a Primary Agency Primary Agency Named Was a:
Law Enforcement Agency
(n = 127)
Prosecution Agency
(n = 82)
Other VS Agency
(n = 45)
Other Type of Agency
(n = 143)
Communication activities (percent reporting):
1. Share general VAW information 98 98 98 98
2. Have frequent/regular telephone contact about the services each agency provides 91 96 95 92
3. Have informal meetings to share general information (NOT case conferences) 86 87 93 86
4. Refer clients 100 100 97 91
Summary: percent reporting all 4 communication activities 82 87 86 79
Coordination activities (percent reporting):
1. Help one another on an as-needed basis for specific cases by sharing information 97 98 99 94
2. Facilitate referrals by contacting one another to coordinate service provision for specific victims 96 97 97 95
3. VS program provides training to this agency 95 73 76 72
4. Provide coordinated community awareness/education activities 84 72 79 56
5. VS program receives training from this agency 61 56 78 53
6. Have regularly scheduled meetings to discuss specific cases 46 57 51 46
Summary: percent reporting 0, 1, or 2 4 4 1 16
3 6 19 10 15
4 23 24 20 25
5 39 29 45 23
6 coordination activities 28 25 23 21
Collaboration activities (percent reporting):
1. Participate on a task force together 83 80 91 72
2. Strategize together about how to reach VAW victims 81 78 83 68
3. Influence one another's agency protocols 72 66 77 56
4. Routinely provide integrated services to victims 53 66 64 56
5. Have a regular feedback mechanism between agencies to ensure that collaboration is working 59 59 64 47
6. Participate together on a first-response team 36 26 8 11
7. Share funding 18 26 21 16
8. Share a joint mission statement 15 20 19 11
Summary: percent reporting 0, 1, or 2 17 16 14 33
3 18 17 16 14
4 23 22 15 21
5 17 13 37 18
6 14 7 13 10
7 8 11 5 3
8 collaboration activities 3 1 0 1
Type of commitment between VS program and agency:
  • Major organizational commitment to work together
  • 50 38 62 46
  • Organizational commitment, but workers are mostly left to build relationships on their ow
  • 45 52 38 49
  • All or almost all relationships are personal; little organizational commitment
  • 5 7 0 5
    Source: Urban Institute analysis of VS program telephone interview responses to interview questions 20b5, 20c5, 21, 22, and 23.
    Agencies are those that respondents identified as a primary agency, meaning the agencies with which they have the most, or the most meaningful, contact.

    Coordination. For coordination activities, we asked whether the VS program and its primary agency (presented in order or decreasing frequency of endorsement):

    1. Help one another on an as-needed basis for specific cases by sharing information;

    2. Facilitate referrals by contacting one another to coordinate service provision for specific victims;

    3. Provide training (VS to primary);

    4. Provide coordinated community awareness/education activities;

    5. Participate in training (primary to VS); and

    6. Have regularly scheduled meetings to discuss cases, such as a multiagency team.

    Virtually all VS programs reported sharing the first and second coordination activities with their primary agencies, regardless of type (proportions ranged from 94 to 99 percent). The last type of coordination activity, serving on a multiagency team or having regularly scheduled case conferences, was also equally likely to occur regardless of the type of primary agency, but was less common (proportions ranged from 46 to 57 percent).

    For training going from the VS program to the primary agency, law enforcement agencies stand out as more likely to have received this training than either prosecution or other types of agency. For training going to the VS program from the primary agency, VS programs were more likely to have received training from other victim service agencies than from other types of agency (the differences with law enforcement and prosecution agencies were not significant). Finally, similar proportions of VS programs reported participating with law enforcement (84 percent), prosecution (72 percent) and other victim service agencies (79 percent) in joint community awareness or education activities. The likelihood of joint community education/awareness activities with other types of primary agency was lower (56 percent) than that for law enforcement agencies, but did not differ from that for prosecution or other victim service agencies.

    Between one-fifth and one-fourth of all VS programs reported participating in all six coordination activities with a primary agency, with no significant differences across agency type. However, when one looks at the proportion participating in five or six activities, and the proportion with low participation (in none, one, or two activities), other types of agencies stand out as less involved. VS programs were less likely to report sharing none, one, or two of the six coordination activities with law enforcement or prosecution (4 percent) or other victim service (1 percent) primary agencies than they were to report this for other types of agency they named as primary (16 percent). Conversely, only 44 percent of VS programs said they shared five or six coordination activities with their primary agency of an "other" type. This proportion was lower than for law enforcement primary agencies (67 percent) but did not differ from the situation when another victim service primary agency (68 percent) or a prosecution agency was named as a primary agency (54 percent).14 These patterns probably reflect the fact that "other" agencies tended to be generic health, welfare, housing, or social service agencies from which VS program clients needed help to obtain particular benefits, but which were not involved in a "first" or "primary" response to victims.

    Collaboration. For collaboration activities, we asked whether the VS program and its primary agency (presented in order of decreasing frequency of endorsement):

    1. Participate on a task force together;

    2. Strategize together about how to reach women victims of violence;

    3. Influence one another's agency protocols;

    4. Routinely provide integrated services to victims;

    5. Have a regular feedback mechanism between agencies to ensure that collaboration is working;

    6. Participate together on a first-response team;

    7. Share funding; and

    8. Share a joint mission statement.

    There were very few significant differences between types of primary agency in the degree to which VS programs participated with them in collaborative arrangements. The modal number of collaboration activities shared by VS programs and their primary agencies was four, and there was essential similarity across primary agency types in the proportion of VS programs reporting each aspect of collaborative work.

    Four out of five VS programs participated on a task force with their law enforcement and prosecution primary agencies. If VS program respondents named another victim service agency as primary, then the vast majority (91 percent) also shared task force membership with that agency. The least likely collaborative arrangements, regardless of primary agency type, were sharing a joint mission statement and sharing funding. Only slightly more common was joint participation in a first-response team. For this activity, VS programs were significantly more likely to interact with law enforcement than with another victim service agency or another type of agency. This finding may in part reflect the nature of "first-response," which is associated with law enforcement activities. However, one in four of the VS programs that named a prosecution agency as primary (10 percent of all VS programs) indicated that it was part of a first-response team, and one might expect that hospitals would be part of a first-response team for sexual assault.

    Given that responses in table 5.2 all relate to agencies that VS programs named as primary (those with which they had the most, or the most meaningful, interaction), it may not be surprising to learn that substantial proportions were involved with VS programs in major organizational commitments to work together. The nature of these commitments can be seen in the level of reporting for the first five collaborative arrangements listed in table 5.2. We expect task force participation as a sine qua non of joint work. But too often communities have a task force and little else. Among VS programs and their primary agencies, considerably more real joint work appears to be happening, including mutual influences on agency protocols and routinely providing integrated services. By integrated services, we mean serious efforts of two agencies to work together regularly, for most women, to provide them with the services they need from each agency, and to do this without getting in each other's way or making the woman feel that she is in a revolving door and that no one knows what is going on.

    Barriers to Collaboration. We have reported significant levels of communication and coordination among VS programs and their primary partners, as well as some but not pervasive collaborative arrangements. This is not surprising, as collaboration is harder to achieve than communication and coordination. As part of other evaluation work on the STOP program (Burt et al., 1997, 1998, 1999, 2000), we have written about efforts to reach collaborative systems and barriers encountered. What we learned from the VS programs we interviewed for this project is similar to previous information, so we present only a brief summary here.

    We asked VS program representatives to describe the most important barriers they had encountered to developing and maintaining collaborative interactions with the agencies in their community that need to be involved if women victims of violence are to experience prompt, respectful, and effective responses to their victimization. No barrier was named by more than 30 percent of respondents. Four barriers were named by more than 20 percent of VS programs. These were:

    • Attitude problems on the part of staff in other agencies;

    • History, old antagonisms, lack of sensitivity to racial, cultural, language, and other issues of different groups of women;

    • Territoriality, turf issues, disputes about which agency should provide services, resistance to hearing feedback about one's agency from people outside the agency; and

    • Specific difficulties related to working with law enforcement agencies.

    In addition, between 15 and 20 percent of VS programs mentioned staff burnout because too few people had too much work to do; difficulties in getting people interested in working together—inertia; and differences in approach, ideology, professional training and professional languages, and the specific missions of different agencies. Anyone familiar with efforts to develop collaboration on any issue, in any community, at any time, will recognize these barriers as important, and as completely independent of the issue at hand. When the issue is violence against women, however, the usual barriers to collaboration are further strengthened by cultural resistance to taking these crimes seriously, believing women, and combating cultural myths about blame and responsibility for victimization. Furthermore, establishing and maintaining collaborative arrangements absolutely requires the devotion of a coordinator as well as the administrative support of a data collector as well as regular office activities. Funders are extraordinarily reluctant to pay for this work, presumably because they think of it as "administration." Yet without the continued attention of a coordinator, whose time is paid to do this work, collaborative arrangements will collapse, as has happened in community after community when funding is withdrawn from this vital function.

    The Role of Task Forces

    Almost three-quarters (72 percent) of VS programs participated in some form of violence-against-women task force in their community. It is not uncommon, when people are talking about the extent of cooperation among different agencies in town, for them to mention having a task force. The implication is that the task force is a major form of cooperation; sometimes the implication is that no more needs to be said. However, task forces can also be a good way to do nothing more, as countless politicians have demonstrated. On site visits for the national evaluation of the STOP program (Burt et al., 1998, 1999, 2000), people in many communities expressed their perception that although they had had a task force, and amicable relationships among agencies had prevailed before STOP, the advent of STOP funding and pressure to work more closely together galvanized the community to much greater levels of collaboration. Alternatively, some communities are so small and close-knit that a formal task force may be superfluous.

    We can use the information from our VS program interviews to examine the relationship between having a task force, having the right people on the task force, and various indicators of collaborative activity. We do this using VS program information about their interactions with their primary agencies coupled with the ratings of communication, coordination, collaboration, and CCR given to each VS program community by Urban Institute researchers on the basis of all the information we collected about the program. The first four columns of table 5.3 show the relevant information. The first column includes the 16 percent of VS programs who did not participate on a task force with either of their two primary agencies (this may not mean that no task force exists in the community, but it is a strong indicator that the VS program is not a part of such a mechanism with the two agencies it says it works with most closely). The second column includes the 12 percent of VS programs who did participate on a task force, but only one of each program's two primary agencies was also on the task force. The third column contains the remaining 72 percent of VS programs, whose task force participation was shared with their two primary agencies. The fourth column of table 5.3 shows the 15 percent of VS programs/communities that Urban Institute researchers rated as CCRs (all of which are also included in the previous, i.e., third, column).

    Two conclusions are obvious from these four columns of table 5.3. First, every activity or arrangement was more likely to occur when all three agencies (the VS program and its two primary agencies) participated together on a task force than when only one, or neither, of the two primary agencies served on a task force with the VS program. The second is that VS programs in communities rated as providing a CCR to women victims of violence were even more likely to report each activity or arrangement than the entire group of agencies participating on a task force with both of their primary agencies. Thus our CCR rating reflects a substantially higher level of interagency interaction than simple participation together on a task force.

    Table 5.3
    Implications of Task Force Participation
    (weighted percentages)
    Activity/Arrangement Of Those that ............ , Proportion Doing the Activity/Arrangement: Of Those Doing the Activity/arrangement, Proportion on Task Force with Agencies 1 AND 2
    Do Not Participate on a Task Force with Either Agency 1 or 2
    (16% of sample)
    Participate on a Task Force with Agency 1 OR 2 but Not Both
    (12% of total)
    Participate on a Task Force with Agencies 1 AND 2
    (72% of total)
    Are Rated as a CCR
    (15% of total)
    Have the highest level of communication (4) 54 33 70 100 80
    Have the highest level of coordination (3) 14 19 64 100 91
    Have the highest level of collaboration (3) 0 7 23 100 95
    Rated as a CCR 0 0 21 100 100
    Strategize about VAW issues with agency 1 and 2 25 14 85 100 92
    Influence agency protocol in agency 1 and 2 24 34 65 91 86
    Provide integrated services with agency 1 and 2 21 29 56 77 86
    Have feedback mechanism with agency 1 and 2 23 16 56 80 88
    Have first-response teams with agency 1 and 2 0 4 20 42 96
    Share funding with agency 1 and 2 0 0.4 15 29 100
    Have joint mission statements with agency 1 and 2 5 2 18 33 93
    Task force agencies participate in joint projects Not applicable 89 83 87 85
    Source: Urban Institute analysis of weighted VS program telephone interview responses.
    Note: "Agency 1" and "Agency 2" are the primary agencies, of any type, named by respondents.

    However, it is also important to observe that more than half of the VS programs that do not serve on a task force with either of their primary agencies still received the highest rating of communication ("good communication with most or all other agencies in the community"), and about one in seven (14 percent) received the highest rating for coordination ("good coordination with most or all other agencies in the community"). In addition, between 21 and 25 percent engaged in various collaborative activities with their two primary agencies, including strategizing about how to address issues of violence against women in their community, influencing each other's protocols, providing integrated services to women, and having a feedback mechanism to assess the appropriate functioning of coordination mechanisms. These are not trivial accomplishments, and obviously can be achieved in some communities without benefit of a task force. It is true that none of these communities received the highest rating for collaboration, and none was rated as being a CCR. But it does not follow that having a task force would, of itself, have promoted greater levels of collaboration.

    Table 5.3 contains one final column, which is also of interest. Instead of showing the proportion of a particular group, defined by task force status, that participated in each activity or arrangement, it shows the proportion of VS programs participating in each activity/arrangement who were on a task force with their two primary agencies. These proportions are uniformly high, ranging from 80 to 100 percent. Task force participation, one can conclude, is an extremely common mechanism which is associated with, and probably both promotes and develops along with, joint activities and arrangements for helping women victims of violence. It is certainly possible to get along without a task force, and having a task force is not a guarantee that collaboration occurs or will develop in the future. But the absence of a task force, especially in communities with relatively more complex service structures, probably indicates a relatively low level of movement toward the goal of helping women who experience domestic violence and sexual assault in a coherent and integrated manner.

    Conclusion

    The communities in which STOP-funded VS programs operate included several other programs or agencies that could be part of a network of services to support women victims of violence. For the most part, VS programs worked at some level with essential parts of the legal system such as law enforcement and prosecution. They also often reported working with other victim service agencies and other types of agency that either offered explicit victim services or services/benefits needed by victims (e.g., cash assistance, housing assistance, or mental health services). Many VS programs reported working very closely with law enforcement and prosecution, while others had less involvement with these agencies.

    Even with agencies named as primary community partners (each VS program could name two), levels of communication, and especially coordination and collaboration, varied considerably. About half of VS programs had institutionalized commitments to work with their primary agencies, and half did not. Further, many changes in levels of interaction among agencies in their community were attributed by VS programs to the impact of STOP funding. Finally, task forces were shown to be important, but neither necessary nor sufficient, to produce high levels of collaboration or a coordinated community response to women victims of violence.


    Chapter 6
    Impact at the Community Level

    HIGHLIGHTS
    • The more communities were already addressing violence-against-women issues and were engaged in developing the ability to meet the needs of victims before STOP, the higher VS programs rate their community on its ability to meet victim needs after STOP funding. However, the greater the level of activity in communities before STOP, the less change VS programs report when it comes to addressing the needs of victims.


    • The more agencies work together in communities, including law enforcement and prosecution agencies working with VS programs, the more likely services are to improve for both VS programs and the legal system.


    • State STOP agency support for collaboration was related to more communication among agencies and more coordinated community responses to violence against women. However, state STOP agency support for collaboration, at least as we were able to measure it, was not related to VS program or legal system outcomes.


    • Although we found that measures of the level of STOP funding to VS programs were not directly related to VS program outcomes or to changes in how legal system agencies treat women victims, it is important to remember that every VS program in our sample did have STOP funds. The effect of receiving or not receiving a STOP grant therefore could not be assessed, but would almost certainly reveal significant differences in community services had we been able to do so. Without being able to make this comparison, it impossible to assess the full impact of STOP funding on communities.


    • VS program representatives who attributed changes in interaction between their VS program and law enforcement, prosecution, and/or other VS agencies to STOP funding also reported greater coordination in community responses and more positive VS program and legal system outcomes.


    • Using STOP to fund certain types of activity (in particular, multidisciplinary response teams, victim witness services, and policy/protocol development activities) is associated with reports of more coordination in community responses, and more positive VS program and legal system outcomes.

    Introduction

    In this chapter we estimate a revised version of the relationships depicted in our conceptual framework for program and community impact (figure 1.1). The original framework hypothesized direct and indirect relationships among aspects of states and communities that may affect the level of community response to violence against women. As analysis progressed, we revised the framework to reflect additional relationships among concepts as depicted in figure 6.1. Boxes 3 and 4 of the original framework specified "improved" services, entailing the difference between a condition before STOP and a condition after it. Instead of combining these concepts, we have separated measures of post-STOP VS program services and of post-STOP legal system responses to victims from measures of the pre-STOP level of community services. Boxes 5 and 6 of the new framework now include only post-STOP measures of VS programs and of legal system responses, and a new box (Box 2) includes the pre-STOP measures. In addition, we moved level of coordination in community response (now Box 4) to become a predictor, rather than a consequence, of post-STOP outcomes.

    The new framework reflects our original hypotheses as well as some new ones. First, we hypothesize that the level of STOP funds and other resources (Box 1) is directly and positively related to the level of coordination in community response (Box 4), post-STOP VS program services (Box 5), and post-STOP legal system responses to victims (Box 6). The reader should note that only communities with STOP-funded VS programs were included in this analysis and our prediction about STOP-funding should be interpreted as such. We expect that the presence or absence of STOP funding would be related to outcomes in Boxes 4, 5, and 6 if we could compare communities in which one or more STOP subgrants are operating with communities that have never had a STOP subgrant. However, we cannot test these particular associations in the present study, but can only test if differences in the levels of STOP funding received by VS programs, all of which did receive STOP funding, are related to outcomes in Boxes 4, 5, and 6. Therefore, although we theorize that relationships between Box 1 and Boxes 4, 5, and 6 exist, we may not be able to observe them using the data we have.

    Second, we hypothesize that pre-STOP level of community services (Box 2) and state STOP program support for collaboration (Box 3) are directly and positively related to level of coordination in community responses (Box 4), post-STOP VS program services (Box 5), and post-STOP legal system responses to victims (Box 6). Third, we hypothesize that level of coordination in community responses (Box 4) is directly and positively related to post-STOP VS program services (Box 5) and post-STOP legal system responses to victims (Box 6). Finally, we hypothesize that pre-STOP level of community services (Box 2) and state STOP program support for collaboration (Box 3) also have indirect effects on post-STOP VS program services (Box 5) and post-STOP legal system responses to victims (Box 6) mediated through the level of coordination in community response (Box 4).

    Independent and Dependent Measures within the Conceptual Framework

    To the extent possible, each box in the conceptual framework is represented by more than one measure. Figure 6.2 lists the measures included in each box.

    Box 1: Level of STOP Funds and Other Resources

    Box 1 includes four measures of the level of STOP funding and other resources that support the VS program. The first measure is the total STOP dollars received by the VS program, combining subgrants from different fiscal years as appropriate. Responses ranged from $3,000 to $805,136 in total STOP funds. Eighteen percent of VS programs received a total of $50,000 or less, 30 percent received between $50,000 and $100,000, and 52 percent received $100,000 or more.

    The second measure in Box 1 is the total number of STOP subgrants the VS program has received, ranging from 1 to 5. Three percent received only one STOP subgrant, 17 percent received two, 45 percent received three, 22 percent received four, and 13 percent received five subgrants. The third measure is the proportion of the VS program activity funded with STOP dollars. This is not a measure of funding for total agency activity, but rather a measure of the extent to which host agencies supplement their STOP subgrant with other funds to support the VS program activities. For 43 percent of VS programs, STOP funds support 25 percent or less of total program activity, while for 23 percent, STOP funds support 75 percent or more of total program activity. The remaining 34 percent of VS programs support between 26 and 74 percent of their program activity with STOP funds.

    The fourth measure in Box 1 is the number of other grants (Community Oriented Police Services funds, Victims of Crime Act funds, Byrne funds, other federal funds, other local funds, and private funds) the host agency combines with STOP funds to support VS program activities. This measure has a range from 0 to 6. Less than 1 percent of participants reported not having any other funding sources, 53 percent reported one or two other funding sources, 37 percent reported three or four other funding sources, and 10 percent reported five or six other funding sources being combined with STOP dollars.

    Box 2: Pre-STOP Level of Community Services

    Box 2 contains two measures of pre-STOP level of community services: one for domestic violence (DV) and one for sexual assault (SA). The measures are based on three questions related to DV and three related to SA. The questions asked respondents to rate their perceptions of the: (1) level at which community agencies were working together before STOP funding around issues of DV or SA; (2) ability for the community to meet the needs of DV or SA victims before STOP funding; and (3) quality of the response of the legal system toward victims before STOP funding. The response scale ranges from 1 (the lowest level) to 5 (the highest level). The means of these variables range from 2.1 to 2.5 (table 6.1).

    Table 6.1
    Perceptions of the Level of Community Services Before STOP Funding
    (weighted percentages)
    Measure Mean 1
    (lowest level)
    2 3 4 5
    (highest level)
    Pre-STOP ratings for DV:
    Level of community agencies working together to address domestic violence before STOP 2.4 15.9 40.2 33.3 7.9 2.7
    Ability of community to meet the needs of DV victims before STOP 2.5 10.1 42.0 40.2 6.9 0.8
    Quality of the response from the legal system toward DV victims before STOP 2.1 20.0 47.3 30.9 1.7 0
    Pre-STOP ratings for SA:
    Level of community agencies working together to address sexual assault before STOP 2.1 32.8 35.6 21.9 7.8 1.9
    Ability of community to meet the needs of SA victims before STOP 2.2 27.5 39.2 25.3 5.5 2.5
    Quality of the response from the legal system toward SA victims before STOP 2.1 28.7 43.3 21.5 6.5 0
    Source: Urban Institute analysis of weighted VS program telephone interview responses.

    Respondents perceived higher levels of community responses to DV than to SA before STOP funding. Overall chi-square estimates15 comparing level of community agencies working together around DV versus around SA indicated significant differences in the distributions of respondents' ratings for DV and SA (p < .05). Overall chi-square estimates were also significantly different (p < .05) when comparing the community's ability to meet victims' needs related to DV versus their ability to meet victims' needs related to SA, and when comparing the quality of the legal system's response to DV victims versus SA victims (p < .05).

    In general, more respondents rated the pre-STOP community response to SA at the lowest level (between 29 percent and 33 percent) than did so for the response to DV (between 10 percent and 20 percent). Greater proportions of respondents rated activities around DV at the midpoint of the scale (between 31 and 40 percent) than SA (between 22 and 25 percent). However, participants reported responses to DV and SA at similar rates at the higher end of the scale (levels four and five).

    The pre-STOP measures for DV are significantly correlated with one another (r's range from .27 to .55). Therefore, the three measures for DV were averaged to create one pre-STOP rating for DV (M = 2.3, s.d. = 0. 7). Likewise the pre-STOP measures for SA are significantly correlated (r's range from .43 to .62), and the three measures for SA were averaged to create one pre-STOP rating for SA (M = 2.1, s.d. = 0. 9). The average pre-STOP rating for DV is statistically higher than the average pre-STOP rating for SA (t = 2.95, p < .05).

    Box 3: State STOP Program Support for Collaboration

    Box 3 contains one measure of state STOP agency support for collaboration. The measure is based on respondent reports of activities conducted by state STOP agencies (SSAs) to assist subgrantees. Respondents were asked about 12 activities their agencies may or may not do to support subgrantees, 6 of which directly relate to supporting coordinated efforts in local communities. Examples include: "Does the SSA provide technical assistance with project implementation?" "Does the SSA provide guidance or training to help develop a coordinated community response including your agency and other agencies in the community?" "Does the SSA require that your agency work with other agencies in the community in order to be eligible for STOP funds?" and "Does the SSA provide guidance or training on team-building, trust-building, or other issues necessary to collaboration?" Responses were summed to create a total number of SSA activities ranging from 0 to 12. We created a second summary score, including only the six collaboration-oriented questions, but it did not perform as well in analyses as the score based on all 12 questions. Thirty-four percent of respondents reported 1 to 4 activities, 46 percent reported 5 to 9 activities, and 30 percent reported 10 to 12 activities (M = 6.9, s.d. = 3.1).

    Box 4: Level of Coordination in Community Response

    Box 4 contains five measures of coordination in community responses. It is important to note that these measures are of behavioral practices of agencies, albeit as reported by VS programs, rather than perceptions of the extent to which agencies work together. Three of the measures are ratings we created to capture communication, coordination, and collaboration between VS programs and other agencies in the community to address violence against women, and a fourth measure rates whether or not a community seemed to be organized into a CCR. The definitions of these four variables and the process by which we rated them were presented in chapter 2. No community was rated as a 1 for communication and 63 percent of communities were rated at the highest level of communication (4), meaning that positive communication existed with most or all other agencies in the community. The majority of communities (51 percent) were also rated at the highest level of coordination (3), meaning agencies were coordinating their activities with most or all other agencies in the community. Only 6 percent of communities were rated at the lowest level of coordination.

    Conversely, the majority of communities were rated as either the lowest level of collaboration (36 percent) or the mid-level of collaboration (46 percent). Only 18 percent of communities were rated at the highest level of collaboration (3), at which collaborative activities occurred with most or all other agencies in the community. Similarly, only 15 percent of communities were rated as having a CCR, meaning that at least law enforcement, prosecution, and victim services participated in collaborative activities to address domestic violence, or at least law enforcement, prosecution, victim services, and the medical community participated in collaborative activities to address sexual assault.

    The fifth measure of community coordination in Box 4 is a measure of the extent to which legal system agencies are the primary agencies with which VS programs interact. VS programs reported the two agencies with which they had the most or most meaningful interaction. Twenty-six percent named both law enforcement and prosecution as their two primary agencies. Another 53 percent reported that either law enforcement or prosecution was one of their primary agencies but did not name both. The remaining 21 percent did not name either law enforcement or prosecution as a primary agency. Conceptually, this measure indexes the degree to which a STOP-funded VS program has substantial, regular, and important interactions with the two legal system agencies most important for women victims of violence.

    Box 5: Post-STOP VS Program Services

    Box 5 includes seven outcome measures that capture post-STOP VS program services. The first measure captures the number of changes in interactions between the STOP-funded VS program and other victim service agencies that have occurred since STOP funding (see chapter 5 for a description). We asked respondents about their interactions with other victim service agencies before and after STOP for five activities: contact, advocacy, referral, coordinating activities, and joint planning and/or institutionalized commitment. The number of activities for which change was reported ranged from 0 to 5, with 0 representing no changes and 5 representing positive changes on all five types of interaction with other victim service agencies. The majority of respondents reported changes on all five (51 percent). Additionally, 28 percent reported four types of changed interaction, 6 percent reported three types, 2 percent reported two types, 2 percent reported one type, and 13 percent reported no changed interactions.

    The second measure in Box 5 is the number of STOP-funded activities the host agency conducts (see chapter 4 for a list of activities). The scale ranges from 0 to 17, with 8 activities being the most frequent number of activities reported by agencies (M = 7.2, s.d. = 3.6). The third measure captures the extent to which STOP enabled VS programs to initiate new types of work. It categorizes programs into two types: those for which STOP money has either started programs that are the first of their kind in a community (coded as 1), or those for which STOP money allowed agencies to supplement work that was already underway (coded as 2). Forty-five percent of agencies reported that STOP funding allowed them to supplement prior work.

    The remaining four measures used to describe post-STOP VS program services are based on respondents' perceptions of their community's ability to meet the needs of victims. The first two measures correspond to those included in the pre-STOP ratings for DV and SA. Respondents rated the level at which their community has been able to meet the needs of victims since STOP funding, using a scale of 1 to 5 with 1 being the "needs of victims are not met at all" and 5 being "the needs of victims are completely met" (table 6.2). Few respondents perceived that victims were able to have their needs completely met—12 percent for DV victims and 8 percent for SA victims. In addition, few respondents perceived that victims were not able to have any of their needs met—0 percent for DV and 3 percent for SA. For both DV and SA, the most common response was that most needs were being met—62 percent for DV and 40 percent for SA.

    The last two measures in Box 5 are also based on respondents' perceptions of the ability of the community to meet the needs of victims. The measures reflect the extent to which communities have changed from before STOP to after STOP on ratings of meeting victim needs. A change score was calculated for both DV and SA by subtracting the before-STOP ratings shown in table 6.1 from the after-STOP ratings shown in table 6.2. The resulting ratings were then categorized into three levels of change: the first level represents a negative incremental change or no change, the second level represents a change from one category to the next one (e.g., from 2 to 3), and the third level represents a change of two or more steps (e.g., from 2 to 4). We combined reports of two or more steps because respondents who were at the midpoint to begin with could not move more than two steps, and we did not want to create a skewed measure (table 6.3).

    Table 6.2
    Perceptions of the Community's Ability to Meet the Needs of Victims Since STOP Funding
    (weighted percentages)
    Measure Mean 1
    (The needs of victims are not being met at all)
    2
    (Some needs are being met)
    3
    (A moderate amount are met)
    4
    (Most needs are being met)
    5
    (The needs of victims are being completely met)
    Ability of community to meet the needs of DV victims since STOP funding 3.9 0 0.8 25.2 61.6 12.4
    Ability of community to meet the needs of SA victims since STOP funding 3.4 2.8 15.7 32.9 40.2 8.4
    Source: Urban Institute analysis of weighted VS program telephone interview responses.

    Table 6.3
    Levels of Change in Perception Before and After STOP in the
    Community's Ability to Meet the Needs of Victims
    (weighted percentages)
    Measure Mean -1, 0 1 2 or more
    Change in the ability of community to meet the needs of DV victims 1.3 4.6 57.2 38.2
    Change in the ability of community to meet the needs of SA victims 1.1 20.3 45.4 34.3
    Source: Urban Institute analysis of weighted VS program telephone interview responses.

    The majority of respondents reported a change of one unit in the ability of their community to meet the needs of victims from before to after STOP funding — 57 percent for DV and 45 percent for SA. Fewer respondents reported negative or no change. However, the proportion was four times higher for SA victims (20 percent) than for DV victims (5 percent), indicating less positive change around sexual assault than around domestic violence issues (x2 = 41.4, p < .05).

    Box 6: Post-STOP Legal System Response to Victims

    Box 6 includes 11 measures of post-STOP legal system responses to victims. Three categories are included: changes in VS program interaction with the legal system, respondent perceptions of changes in legal system responses to victims, and respondent perceptions of changes in behaviors of legal system agencies.

    Box 6 includes two measures of change in VS program agency interactions with the legal system—one focusing on law enforcement and the other on prosecution. Similar to the measure in Box 5 regarding other VS programs in the community, these measures capture the number of changed interactions between the VS program of interest and law enforcement or prosecution that have occurred since STOP funding (chapter 5, table 5.1).

    As with Box 5, Box 6 also includes four measures that capture respondent perceptions of the legal system's post-STOP response to victims. The first two measures correspond to those included in the pre-STOP ratings for DV and SA. Respondents rated the level at which the legal system responds to the needs of women victims of violence on a scale of 1 to 5, with 1 being the "the legal system failed to respond to the need of women victims of violence" and 5 being "the legal system did an excellent job responding to the needs of women victims of violence" (table 6.4). Few respondents perceived that the legal system did an excellent job meeting the needs of victims — 7 percent for DV victims and 6 percent for SA victims. In addition, few respondents perceived that the legal system failed to respond to victims needs — 0 percent for DV and 6 percent for SA. For DV the most common response was level 4, with 55 percent of respondents rating the legal system at this level at the time of the survey, after STOP funding. For SA, the most common response was level 3—an average response by the legal system, with 41 percent of respondents rating the legal system at this level since STOP funding.

    The other two perception measures in Box 6 reflect the extent to which the legal system within communities has changed from before STOP to after STOP on ratings of responsiveness to victims' needs. A change score was calculated for both DV and SA by subtracting the before-STOP ratings shown in table 6.1 from the after-STOP ratings shown in table 6.4. The resulting ratings (table 6.5) were then categorized into three levels of change: the first level represents a negative incremental change or no change, the second level represents a change from one category to the next one (e.g., from 2 to 3), and the third level represents a change of two or more steps (e.g., from 2 to 4).

    Table 6.4 Perception of the Legal System Response toward Victims Since STOP Funding
    (weighted percentages)
    Measure Mean 1
    The legal system failed to respond to the needs of women victims of violence
    2 3
    The legal system did an average job responding
    4 5
    The legal system did an excellent job responding to the needs of women victims of violence
    Quality of the response from the legal system toward DV victims since STOP 3.7 0 3.8 34.3 54.5 7.4
    Quality of the response from the legal system toward SA victims since STOP 3.2 6.1 13.3 40.9 33.6 6.1
    Source: Urban Institute analysis of weighted VS program telephone interview responses.

    Table 6.5
    Levels of Change in Perception Before and After STOP in the Legal System Response toward Victims
    (weighted percentages)
    Measure Mean -1,0 1 2 or more
    Change in quality of the response from the legal system toward DV victims 1.4 5.3 48.0 46.7
    Change in quality of the response from the legal system toward SA victims 1.1 21.5 46.1 32.3
    Source: Urban Institute analysis of weighted VS program telephone interview responses.

    The majority of VS programs reported a change of one unit in the legal system response to victim needs from before to after STOP funding — 48 percent for DV and 46 percent for SA. Fewer respondents reported negative or no change. However, the proportion is four times higher for SA victims (22 percent) than for DV victims (5 percent) indicating less positive change around sexual assault than domestic violence issues (x2 = 40.5, p < .05).

    The remaining five measures in Box 6 are respondent reports of changes in behaviors of legal system agencies since STOP funding became available in 1996. Four measures reflect changes in behavior of law enforcement, prosecution, and judges, and behavior around protective orders. The fifth measure reflects a total score for all legal system changes combined, ranging from 1 to 16 (M = 10.5, s.d. = 3.5).16

    The law enforcement measure ranges from 0 to 6 and captures changes such as collecting better evidence in DV and SA cases and arresting more perpetrators. Thirty-five percent of the respondents reported all six changes in their community, 21 percent reported five changes, 21 percent reported four changes, 14 percent reported three changes, 8 percent reported two changes, 1 percent reported one change, and 1 percent reported no change (M = 4.6, s.d. = 1.4).

    The prosecution measure ranges from 0 to 7 and captures such changes as charging perpetrators with more offenses, trying more DV and SA cases in court, and getting more convictions in DV and SA cases. Only 14 percent of respondents reported all seven changes. Ten percent reported six changes, 17 percent reported five changes, 21 percent reported four changes, 14 percent reported three changes, 7 percent reported two changes, 10 percent reported one change, and 7 percent reported no change (M = 3.9, s.d. = 2.0).

    The judges measure ranges from 0 to 2 and captures sentencing offenders to stricter punishments and violating offenders who do not comply with probation, parole, or conditions of protective orders. Forty-three percent of respondents reported both changes, 37 percent reported one change, and 21 percent reported no change (M = 1.2, s.d. = 0. 8). The measure of behaviors around protective orders ranges from 0 to 3 and is based on questions about changes in such behaviors since STOP funding became available. It includes if easier processes for women to obtain protective orders exist, if police enforces orders by arrest or other means, and if judges violate offenders who do not comply with conditions of protective orders. Fifty-seven percent of respondents reported all three changes, 26 percent reported two changes, 17 percent reported one change, and less than 1 percent reported no changes in behaviors around protective orders (M = 2.4, s.d. = 0. 8).

    Descriptive Analysis

    Before examining the interrelationships hypothesized in our conceptual framework, we looked at the simple associations between the variables in one box and those in another. Bivariate correlations among independent and dependent variables are discussed below in relation to the arrows in the conceptual framework (figure 6.2).17 Each arrow is described separately.

    Box 1 with Boxes 4, 5, and 6

    Box 1 (level of STOP funds and other resources) was hypothesized to relate to outcomes in Box 4 (level of coordination in community response), Box 5 (post-STOP VS program services), and Box 6 (post-STOP legal system responses to victims). Few relationships exist among Box 1 independent variables and outcomes in Boxes 4, 5, and 6, and those that do exist are only moderate correlations. Total STOP funds and the percent of the VS program funded with STOP money are not significantly correlated with any of the outcomes in Boxes 4, 5, or 6. The number of STOP subgrants and the number of other sources supplementing STOP funding are significantly correlated with some, but not all, outcomes in Boxes 5 and 6 and are not significantly related to any outcomes in Box 4.

    More specifically, the total number of STOP subgrants that agencies received relates to three items in Box 5: less change in interactions with other VS agencies (r = -.18, p < .05), more STOP-funded activities (r = .17, p < .05), and more activity supplementing prior work than work that is the first of its kind (r = .15, p < .05). The number of other funding sources for the VS program, in addition to STOP subgrants is correlated with one outcome in Box 5 and three outcomes in Box 6: more STOP-funded activities (r = .27, p < .05), more changes in interactions between law enforcement and VS programs (r = .18, p < .05), higher levels of legal system response to DV victims' needs post-STOP (r = .22, p < .05) and higher levels of legal system response to SA victims' needs post-STOP (r = .23, p < .05). It is important to note that 25 to 36 percent of the sample are missing data for questions regarding other funding sources that supplement STOP funding and therefore the sample included is smaller.

    Based on descriptive statistics, the hypothesized relationships between Box 1 predictors and outcomes in Boxes 4, 5, and 6 are not strong. These few correlations do not provide compelling evidence that the level of STOP funds and other resources are related to coordination in community responses, post-STOP VS program services, and post-STOP legal system responses to victims. However, we may have failed to find a relationship between STOP funding levels and level of coordination in community responses and between STOP funding levels and VS program and legal system outcomes because this survey included only STOP-funded programs. We cannot compare outcomes based on whether or not communities have STOP-funded programs at all. If we could, we would be likely to see significantly greater effects. Instead, we have examined if, given some STOP funding, variation in the total amount of STOP funds makes a difference. Total STOP funding may merely be a proxy for size of agency or size of community in which the STOP-funded program operates. If this is the case, as is quite likely, we would not expect more post-STOP changes in communities with more money.

    In addition, communities that combine STOP funds with other funding sources seem to have greater levels of change, but a consistent pattern of relationships between other funding sources and outcomes in Boxes 5 and 6 does not exist. Again, more funding sources may imply larger agencies or communities. We cannot conclude from the analyses possible with our survey data that communities whose VS programs received only STOP funding have more or fewer positive outcomes than communities whose VS programs combine other funds with STOP funds.

    Boxes 2, 3, and 4

    We hypothesized that Boxes 2 (Pre-STOP level of community services) and 3 (state STOP program support for collaboration) would relate directly to Box 4 (level of coordination in community response). In addition, we expected that the effects of independent variables in Boxes 2 and 3 on Boxes 5 and 6 would be mediated by variables in Box 4. Since variables from Boxes 2, 3, and 4 were included together in models, correlations among the predictors are presented in table 6.6. Pre-STOP ratings for DV and pre-STOP ratings for SA are highly correlated (r = .63, p < .05). The pre-STOP rating for DV is also significantly related to collaboration and CCR ratings, and the pre-STOP rating for SA is significantly associated with coordination and collaboration ratings. Neither of the pre-STOP ratings (for DV or SA) are related to the measure of SSA activities. However, SSA activities are positively related to communication, collaboration, and CCR ratings. These relationships indicate that the more communities worked together to address violence against women before STOP, the more likely agencies were to interact post-STOP (the level of interaction also changed, as described below). In addition, the more the SSA supported subgrantee efforts around collaboration the more likely agencies were to interact.

    As one would expect, communication, coordination, collaboration, and CCR ratings are all correlated with one another. The closer the ratings are in concept, the more highly they are correlated. Specifically, communication is more highly correlated with coordination (r = .48, p < .05) than it is with collaboration (r = .39, p < .05) or being a CCR (r = .27, p < .05). Coordination is more highly correlated with collaboration (r = .65, p < .05) than with being a CCR (r = .35 p < .05) and collaboration is highly correlated with being a CCR (r = .62, p < .05).

    Table 6.6
    Correlations Among the Variables in Boxes 2, 3, and 4
      1 2 3 4 5 6 7 8
    Box 2: Pre-STOP level of community services
    1. Pre-STOP rating of community for DV .63* .07 .14+ .10 .21* .15* .06
    2. Pre-STOP rating of community for SA   .03 .08 .16* .17* .11 .12+
    Box 3: State STOP program support for Collaboration
    3. SSA activities     .20* .06 .16* .23* .10
    Box 4: Level of coordination in community response
    4. Communication rating       .48* .39* .27* .20*
    5. Coordination rating         .65* .35* .17*
    6. Collaboration rating           .62* .20*
    7. CCR rating             .17*
    8. Primary agencies              
    Source: Urban Institute analysis of weighted VS program telephone interview responses.
    Note: * = p < .05  + = p < .10.

    Agencies that are part of a CCR are likely to be high on collaboration, agencies that are high on collaboration are likely to be high on coordination, and agencies that are high on coordination are likely to be high on communication. Additionally, when the primary agencies named by VS programs are the important legal system agencies of law enforcement and prosecution, ratings for communication, coordination, collaboration, and CCRs are higher.

    Based on these descriptive statistics, the hypothesized relationships between Box 2 and Box 4, and between Box 3 and Box 4, are evident. Pre-STOP ratings of community services and SSA activities are related to measures of coordination in community response when these relationships are looked at one at a time.

    Boxes 2, 3, and 4 with Box 5

    Boxes 2 (Pre-STOP level of community services), 3 (state STOP program support for collaboration), and 4 (level of coordination in community response) were hypothesized to relate directly to outcomes in Box 5 (post-STOP VS program services). Table 6.7 presents the correlations of independent variables in Boxes 2, 3, and 4 with dependent variables in Box 5.

    For Box 2, higher pre-STOP ratings for DV and SA are related to a greater likelihood that VS programs are using STOP to supplement previous work rather than starting something entirely new, and to post-STOP perceptions that communities are more able to meet the needs of victims. The more communities worked together and addressed the needs of victims before STOP, the more the communities could supplement work that was already underway and the more the communities were able to meet victims' needs after STOP (in the view of the VS program).

    Table 6.7
    Correlations Among Independent Variables in Boxes 2, 3, and 4 with Dependent Variables in Box 5
    Box 5: Post-STOP VS Program Services Box 2 Box 3 Box 4
    Pre-STOP Rating DV Pre-STOP Rating SA SSA Activities Communication Rating Coordination Rating Collaboration Rating CCR Rating Primary Agencies
    Changes in interactions with victim services -.08 -.02 .07 .14+ .14+ .18* .16* -.03
    Index number of STOP-funded activities -.04 -.04 .18* .08 -.10 .05 .08 .03
    STOP program supplements previous work .30* .27* -.05 -.06 .03 .03 .01 -.05
    Post-STOP perception that community can meet the needs of DV victims .42* .20* .20* .17* .09 .24* .24* .12+
    Post -STOP perception that community can meet the needs of SA victims .39* .57* .14+ .19* .23* .31* .22* .25*
    Change in perception from before to after STOP that community can meet the needs of DV victims -.54* -.41* .08 .04 .01 .05 .07 .01
    Change in perception from before to after STOP that community can meet the needs of SA victims -.18* -.31* .08 .07 .01 .11 .14+ .11
    Source: Urban Institute analysis of weighted VS program telephone interview responses.
    Note: * = p < .05   + = p < .10.

    Pre-STOP ratings for DV and SA were also negatively related to the amount of change that occurred regarding meeting victims' needs from before to after STOP funding. The less communities worked together around violence-against-women issues and the less they addressed the needs of victims before STOP, the greater the positive change they reported from before to after STOP funding. So, building on pre-existing relationships puts a community at a higher level of organization to meet victims' needs post-STOP, but the most dramatic change occurs in communities that started with no work around domestic violence and sexual assault issues or close to it.

    For Box 3, SSA activities are positively related to the number of STOP-funded activities that VS programs are able to conduct. SSA activities are also positively related to post-STOP perceptions that communities are able to meet the needs of DV victims and are marginally related to post-STOP perceptions that communities are able to meet the needs of SA victims.

    For Box 4, collaboration and CCR ratings are related to changes in more types of interaction with other victim service agencies. Additionally, greater levels of communication, coordination, collaboration, and CCR are related to perceptions that communities are better able to meet the needs of DV and SA victims post-STOP. Further, having law enforcement and prosecution as primary agencies for VS programs is related to perceptions that communities are more able to meet the needs of SA victims post-STOP and is marginally related to these perceptions for DV victims. In sum, communities in which agencies work together to address violence-against-women issues make a difference for victims when it comes to getting their needs met, in the view of VS program representatives. VS program involvement with legal system agencies is also important for improving service provision for victims.

    Based on these descriptive statistics, relationships between Box 2 and Box 5 are strong, with pre-STOP ratings of community services being associated with most outcomes in Box 5. Relationships between Box 3 and Box 5 are weaker, with only two outcomes in Box 5 related to SSA activities. Relationships between Box 4 and Box 5 are somewhat mixed, with some measures of coordination levels related to three measures of post-STOP VS program services.

    Boxes 2, 3, and 4 with Box 6

    Boxes 2, 3, and 4 (pre-STOP level of community services, state STOP program support for collaboration, and level of coordination in community response) were also hypothesized to relate directly to outcomes in Box 6 (post-STOP legal system response to victims). Table 6.8 presents the correlations between independent variables in Boxes 2, 3, and 4 with dependent variables in Box 6.

    As with relationships between Box 2 and Box 5, pre-STOP ratings for DV and SA are also related to outcomes in Box 6. Pre-STOP ratings for DV and SA are positively related to post-STOP perceptions about the legal system's response to the needs of DV and SA victims. The more communities worked together and addressed the needs of victims before STOP, the more the legal system responded to needs of victims after STOP funding. The pre-STOP ratings are also negatively related to the amount of change that occurred in the legal system to address victims' needs from before to after STOP funding. The less communities worked together and the less they addressed the needs of victims before STOP, the greater the perceived changes in the actions of legal system agencies in addressing victim needs from before to after STOP funding.

    Table 6.8
    Correlations Among Independent Variables in Boxes 2, 3, and 4 with Dependent Variables in Box 6
    Box 6: Post-STOP Justice System Response to Victims Box 2 Box 3 Box 4
    Pre-STOP Rating DV Pre-STOP Rating SA SSA Activities Communication Rating Coordination Rating Collaboration Rating CCR Rating Primary Agencies
    Changes in interactions with law enforcement -.04 -.06 .15* .25* .17* .33* .29* .13+
    Changes in Interactions with prosecution -.02 .04 .03 .26* .24* .41* .29* .18*
    Post-STOP perception about the legal system's response to DV victims .32* .22* .11 .01 -.10 .16* .21* .03
    Post-STOP perception about the legal system's response to SA victims .32* .55* .04 .09 .12 .24* .13+ .24*
    Changes in perception from before to after STOP about the legal system response to DV victims -.50* -.36* -.06 -.10 -.11 -.07 -.00 -.02
    Changes in perception from before to after STOP about the legal system response to SA victims -.20* -.28* -.02 .17* .13+ .18* .09 .16*
    Overall legal system changes .18* .16* .12 .25* .15* .28* .22* .18*
    Changes in law enforcement .12 .12 .09 .19* .20* .23* .17* .08
    Changes in prosecution .19* .13+ .09 .27* .14+ .31* .22* .22*
    Changes in judges .18* .30* .04 .11 .07 .20* .11 .15*
    Changes in behaviors around protective orders .15* .18* .12 .04 .03 .18* .13+ .12+
    Source: Urban Institute analysis of weighted VS program telephone interview responses.
    Note: * = p < .05   + = p < .10.

    For Box 3, SSA activities was related to one legal system outcome — changes in interactions with law enforcement. The greater the number of SSA activities that support subgrantees and promote collaboration, the greater the number of interactions for which change occurred between VS programs and law enforcement agencies since STOP funding.

    Predictors in Box 4 are related to many of the legal system outcomes in Box 6. Higher ratings on communication, coordination, collaboration, and CCR, and having law enforcement and prosecution as primary agencies, are all related to change since STOP funding in more types of interaction between law enforcement and VS programs and between prosecution and VS programs. Higher ratings on collaboration and CCR are related to more complete, or more satisfactory, legal system responses to victims needs. Naming both law enforcement and prosecution as primary agencies is positively related only to post-STOP legal system responses to victims' needs for SA, and not for DV. Similarly, greater collaboration and having law enforcement and prosecution as primary agencies is related to more changes in legal system responses to SA victim needs from before to after STOP funding.

    Higher ratings on communication, coordination, collaboration, and CCR, and naming both law enforcement and prosecution as primary agencies, are related to more legal system behavior changes. Measures from Box 4 are related to overall legal system changes, as well as to changes specific to behaviors of law enforcement and of prosecution. In addition, greater levels of collaboration and having law enforcement and prosecution as primary agencies are related to more reported changes in judges' behaviors and behaviors around protective orders.

    Based on descriptive statistics, hypothesized relationships between Box 2 and Box 6 are evident, with pre-STOP ratings of community services related to 8 of the 11 measures of legal system response to victims. The relationship between Box 3 and Box 6 is not evident, with only one measure of legal system response to victims (changes in the number of interactions between law enforcement and VS programs) related to the number of SSA activities. Relationships between Box 4 and Box 6 are strong. All 11 legal system outcomes are related to some measure of coordination in community response.

    However, associations between two measures that appear strong by themselves sometimes change, or even disappear, when one controls for other related variables using statistical models. Therefore, we need to assess these relationships while at the same time taking into account the effects of other factors. Such analyses will also let us calculate how much of the effects of Boxes 2 and 3 on outcomes in Boxes 5 and 6 "goes through" or is accounted for by the level of coordination the community has achieved (Box 4).

    Impact Analysis

    We tested models of the relationships within our conceptual framework to further illustrate the impact of communities and states on VS program and legal system outcomes. We conducted multistage (hierarchical) regression models reflecting the stages of the conceptual framework.18 Because few relationships existed between independent variables in Box 1 and outcomes in Boxes 4, 5, and 6, measures from Box 1 were excluded from the analysis.

    The first stage of the analysis involved examining the relationships between variables in Boxes 2 and 4 while also looking at the relationships between Boxes 3 and 4 at the same time. To do this we first included pre-STOP ratings of DV and pre-STOP ratings of SA in models to predict measures of community interaction in Box 4. Next, we included pre-STOP ratings of DV and SA in models along with SSA activities as independent variables predicting communication, coordination, collaboration, CCRs, and primary agencies.

    The second stage of the analysis assessed the relationship between Box 2 and the individual outcomes in Boxes 5 and 6. Because pre-STOP ratings for DV and SA are highly correlated (r = .63), only one independent variable from Box 2 was included in each model. Pre-STOP ratings for SA were included in models when the outcomes were specific to SA victims, otherwise pre-STOP ratings for DV were included in models. The third stage of the analysis examined the relationship between Box 3 and individual outcomes in Boxes 5 and 6, after taking into account the relationships between Box 2 and the outcomes. SSA activities were included in models along with one pre-STOP rating of DV or SA.

    The fourth stage assessed the relationships between Box 4 and individual outcomes in Boxes 5 and 6, after accounting for the relationships between Boxes 2 and 3 and the outcomes. Because communication and coordination are highly correlated (r = .48), as well as coordination and collaboration (r = .65), the coordination rating was not included in the models. Similarly, the CCR rating was not included in analysis because it is also highly correlated with collaboration (r = .62). Therefore, the communication rating, the collaboration rating, and the primary agencies measure were included in models along with one pre-STOP rating of DV or SA and the variable indexing SSA activities.

    Predictive Models: Boxes 2 and 3 Predicting Box 4

    Table 6.9 presents the results predicting Box 4 (level of coordination in community response) by Boxes 2 and 3 (pre-STOP level of community services and state STOP program support for collaboration). Despite the fact that bivariate correlations exist, when predictors from Boxes 2 and 3 are combined to predict outcomes in Box 4, few relationships remain significant. Communication ratings are significantly predicted by SSA activities, collaboration ratings are significantly predicted by pre-STOP ratings for DV, CCR ratings are significantly predicted by SSA activities, and having law enforcement and prosecution as one's two primary agencies is not predicted by either the pre-STOP ratings of DV or SA or SSA activities. It is also clear from the estimates of variance explained that the measures of Box 2 and Box 3 are quite independent of each other, accounting for different portions of the variance in Box 4 variables.

    Predictive Models: Boxes 2, 3, and 4 Predicting Box 5

    Table 6.10 presents the results predicting Box 5 (post-STOP VS program services) from variables in Boxes 2, 3, and 4 (pre-STOP level of community services, state STOP program support for collaboration, and level of coordination in community response). Pre-STOP ratings of DV or SA significantly predict five of the seven outcomes in Box 5 while accounting for effects of other independent variables. Estimates for pre-STOP ratings of DV or SA remain as strong or stronger with the addition of independent variables in Box 4, and only decrease, but not to non-significance, for one outcome in Box 5 (post-STOP perceptions that the community is able to meet the needs of DV victims). This indicates that only a part of the relationship between Boxes 2 and 5 runs through Box 4, and that only for one outcome. The effects of Box 2 on the remaining outcomes in Box 5 do not run through Box 4.

    Table 6.9
    Predictive Models Boxes 2 and 3 Predicting Box 4
    Box 4: Level of Coordination in Community Response Box 2 Box 3 Model R2
    Pre-STOP
    Rating DV
    Pre-STOP
    Rating SA
    SSA Activities
    Communication rating .09 .01   .01
    Communication rating .06 .02 .04* .05*
    Coordination rating .02 .10   .03
    Coordination rating -.00 .11 .01 .03
    Collaboration rating .18+ .05   .05*
    Collaboration rating .22* .07 .03+ .09*
    CCR rating .09+ .00   .03+
    CCR rating .09+ .01 .03* .09*
    Primary agencies -.03 .11   .02
    Primary agencies -.00 .07 .01 .01
    Source: Urban Institute analysis of weighted VS program telephone interview responses.


    Table 6.10
    Predictive Models Boxes 2, 3, and 4 Predicting Box 5
    Box 5: Post-STOP VS Program Services Box 2 Box 3 Box 4 Model R2
    Pre-STOP
    Rating DV
    Pre-STOP
    Rating SA
    SSA Activities Communication
    Rating
    Collaboration
    Rating
    Primary
    Agencies
    Changes in interactions with victim services -.20         .01
    Changes in interactions with victim services -.09 .04       .01
    Changes in interactions with victim services -.25 .02 .28 .48* -.23 .06+
    Index number of STOP-funded activities -.21         .00
    Index number of STOP-funded activities .10 .21*       .03+
    Index number of STOP-funded activities .09 .20* .39 -.13 .05 .04
    STOP program supplements previous work .21*         .09*
    STOP Program supplements previous work .23* -.01       .09*
    STOP program supplements previous work .24* -.01 -.07 .01 -.06 .11*
    Post-STOP perception that community can meet the needs of DV victims .37*         .17*
    Post-STOP perception that community can meet the needs of DV victims .36* .03*       .18*
    Post-STOP perception that community can meet the needs of DV victims .32* .03* .04 .10 .05 .20*
    Post -STOP perception that community can meet the needs of SA victims .63*         .33*
    Post -STOP perception that community can meet the needs of SA victims .63* .03+       .31*
    Post -STOP Perception that community can meet the needs of SA victims .57* .02 .08 .24* .19* .38*
    Change in perception from before to after STOP that community can meet the needs of DV victims -.43*         .29*
    Change in perception from before to after STOP that community can meet the needs of DV victims -.41* .02+       .25*
    Change in perception from before to after STOP that community can meet the needs of DV victims -.44* .02 .06 .09 .02 .27*
    Change in perception from before to after STOP that community can meet the needs of SA victims -.26*         .10*
    Change in perception from before to after STOP that community can meet the needs of SA victims -.24* .02       .08*
    Change in perception from before to after STOP that community can meet the needs of SA victims -.28* .01 .02 .13 .18* .13*
    Source: Urban Institute analysis of weighted VS program telephone interview responses.
    Note: * = p < .05   + = p < .10 .

    The effect of SSA activities on outcomes in Box 5 is also partially mediated through Box 4. For two outcomes (post-STOP perceptions that communities are able to meet the needs of SA victims and the change in perception before and after STOP that communities are able to meet the needs of DV victims), SSA activities was marginally significant in the reduced model and then reduced to non-significance in the full model with the addition of Box 4 independent variables. However, for two other measures (total STOP-funded activities and supplementing prior work), SSA activities is related to the outcomes in both the reduced and full models.

    In full models (those that include all the variables), independent variables in Box 4 are only somewhat related to outcomes in Box 5, since one is also accounting for the effects of the independent variables in Boxes 2 and 3. The communication rating is not related to any of the outcomes. Collaboration is related to the number of changed interactions between VS programs of interest and other VS programs in the community since STOP funding. Both collaboration and primary agencies predict post-STOP perceptions that communities are able to meet the needs of SA victims. Only the probability that VS programs are involved with legal system programs as primary agencies predicts changes in perception from before to after STOP funding that communities are able to meet the needs of SA victims.

    In sum, pre-STOP ratings of DV and SA are directly related to five of seven outcomes measuring post-STOP VS program services. These effects are also partially mediated by the level of coordination in community response. SSA activities are directly related to two of seven outcomes measuring post-STOP VS program services and are indirectly related to two other outcomes through the level of coordination in community response. Finally, levels of coordination in community response are related to three of seven outcomes measuring post-STOP VS program services: two about SA and one about activities. Coordination in community response was not related to DV-specific measures.

    Independent variables in models predicting post-STOP VS program services account for different amounts of the variance depending on the outcome measure (R2's range from .01 to .38). Independent variables explain little of VS program characteristics (number of interactions changed with other victims services, index of STOP-funded activities, and whether the STOP program supplements previous work), accounting for only 1 to 11 percent of the variance in models. Conversely, independent variables explain large portions of variance in outcomes of respondent perceptions of post-STOP level of community ability to meet victim needs and of changes in ability to meet victim needs from before to after STOP, accounting for 8 to 38 percent of the variance. Independent variables explain greater amounts of variance for SA than for DV with respect to post-STOP level of community ability to meet victim needs. However, they explain greater amounts of variance in change in perception of community ability to meet victim needs from before to after STOP for DV than for SA. Further, the addition of independent variables from Box 4 in models predicting respondent perception does not seem to increase greatly the overall amount of variance explained in prior models by pre-STOP ratings of DV or SA.

    Predictive Models: Boxes 2, 3, and 4 Predicting Box 6

    Table 6.11 presents the results of the hierarchical models predicting Box 6 (post-STOP legal system response to victims) by Boxes 2, 3, and 4 (pre-STOP level of community services, state STOP program support for collaboration, and level of coordination in community response). Pre-STOP ratings of DV and SA predict 6 of the 11 outcomes in Box 6 while taking into account the effects of Boxes 3 and 4. Some effects were strengthened with the addition of predictors in Box 4 and others were reduced, some to non-significance, indicating that the effects of Box 2 on Box 6 are partially mediated by measures in Box 4.

    Pre-STOP ratings of DV and SA are negatively related to the number of changed interactions between VS programs and law enforcement and between VS programs and prosecution. The negative relationships indicate that the less communities worked together to address violence against women and the needs of victims before STOP, the greater the number of interactions that changed with STOP funding. Pre-STOP ratings of DV and SA are also negatively related to changes in perceptions of the legal system response to needs of DV victims and SA victims. The negative relationships indicate that the less communities worked together to address violence against women and the needs of victims before STOP, the more the legal system changed since STOP to address victims' needs.

    Table 6.11 Predictive Models Boxes 2, 3, and 4 Predicting Box 6
    Box 6: Post-STOP Justice System Response to Victims Box 2 Box 3 Box 4 Model R2
    Pre-STOP Rating DV Pre-STOP Rating SA SSA Activities Communication Rating Collaboration Rating Primary Agencies
    Changes in interactions with law enforcement -.08         .00
    Changes in interactions with law enforcement -.19 .07+       .03
    Changes in interactions with law enforcement -.37* .04 .19 .64* .07 .14*
    Changes in interactions with prosecution -.05         .00
    Changes in interactions with prosecution -.13 .01       .00
    Changes in interactions with prosecution -.38* -.03 .26 .78* .19 .21*
    Post-STOP perception about the legal system's response to DV .31*         .10*
    Post-STOP perception about the legal system's response to DV .30* .02       .10*
    Post-STOP perception about the legal system's response to DV .28* .02 -.09 .10 -.02 .11*
    Post-STOP perception about the legal system's response to SA .64*         .31*
    Post-STOP perception about the legal system's response to SA .66* .01       .29*
    Post-STOP perception about the legal system's response to SA .62* .01 .02 .11 .25* .33*
    Changes in perception from before to after STOP about the legal system's response to DV -.42*         .25*
    Changes in perception from before to after STOP about the legal system's response to DV -.42* -.00       .22*
    Changes in perception from before to after STOP about the legal system's response to DV -.42* -.00 -.04 .03 .00 .22*
    Changes in perception from before to after STOP about the legal system's response to SA -.25*         .08*
    Changes in perception from before to after STOP about the legal system's response to SA -.24* -.00       .06*
    Changes in perception from before to after STOP about the legal system's response to SA -.29* -.02 .12 .21* .18* .17*
    Overall legal system changes .91*         .03*
    Overall legal system changes .73+ .10       .03+
    Overall legal system changes .35 .02 1.00* 1.00* .46 .14*
    Changes in law enforcement .23         .01
    Changes in law enforcement .19 .03       .01
    Changes in law enforcement .07 .01 .22 .40* -.03 .07*
    Changes in prosecution .55*         .01
    Changes in prosecution .42+ .05       .03
    Changes in prosecution .18 -.00 .65* .65* .49* .18*
    Changes in judges .20*         .03*
    Changes in judges .22* .01       .03*
    Changes in judges .17+ -.00 .03 .17+ .11 .08*
    Changes in behaviors around protective orders .16*         .02*
    Changes in behaviors around protective orders .18* .02       .03*
    Changes in behaviors around protective orders .14 .02 -.09 .19* .03 .06+
    Source: Urban Institute analysis of weighted VS program telephone interview responses.
    Note: * = p < .05  + = p < .10.

    SSA activities are not related to any of the 11 outcomes in Box 6, in either reduced or full models. Only one outcome (the number of changed law enforcement interactions with VS programs) was marginally related to SSA activities, but this was reduced to non-significance with the introduction of measures in Box 4 into the model. The results indicate no relationship between state support for collaboration and legal system outcomes.

    Finally, the level of coordination in community responses affects post-STOP legal system responses to victims, while accounting for pre-STOP levels of community services and state STOP program support for collaboration. Eight of the 11 outcomes in Box 6 were significantly related to at least one measure of community interaction from Box 4. Communication ratings significantly predicted overall levels of legal system behavior changes and changes in prosecution behavior. Collaboration ratings significantly predicted the number of changed interactions between VS programs and law enforcement and VS programs and prosecution, changes in perceptions before and after STOP about legal system responses to SA victims, overall legal system behavior changes, and changes in law enforcement behavior, prosecution behavior, and behaviors around protective orders. Having law enforcement and prosecution as primary agencies predicted post-STOP perceptions about legal system responses to SA victims, changes in perception before and after STOP about legal system responses to SA victims, and changes in prosecution behavior.

    In sum, pre-STOP ratings of DV or SA are directly related to six of eleven outcomes measuring post-STOP legal system response to victims. Pre-STOP ratings are alsoindirectly related to four other outcomes in Box 6. The relationships are mediated through Box 4 as they were reduced to non-significance with the addition of measures of coordination in community response into the model. Also, measures of coordination in community response are directly related to 8 of 11 measures of post-STOP legal system response to victims.

    Independent variables predicting post-STOP legal system response to victims explain different degrees of variance depending on the outcome of interest (R2 's range from .00 to .33). Variables in Boxes 2 and 3 explain little to no variance in outcomes related to number of changed interactions with law enforcement and prosecution, but with the addition of variables in Box 4 the variance explained increases to 14 and 21 percent respectively. The same pattern is evident when predicting legal system behavior changes. Variables in Boxes 2 and 3 explain little to no variance in legal system behavior change outcomes. However, the addition of variables in Box 4 increases the variance explained to between 6 and 18 percent.

    An opposite pattern is evident when examining the variance explained in models predicting respondent perceptions. As with outcomes in Box 5, the addition of independent variables in Box 4 does not increase the variance explained much beyond what is explained in models that included pre-STOP ratings of DV and SA, with the exception of changes in perception from before to after STOP about the legal system response to SA victim needs. Independent variables explain large portions of variance in outcomes of respondent perceptions of post-STOP level of legal system response to victim needs and of changes in the legal system response to victim needs from before to after STOP, accounting for 10 to 33 percent of the variance. Independent variables explain greater amounts of variance for the post-STOP level of legal system response to victim needs for SA than for DV. However, they explain greater amounts of variance in change in perception of legal system response to victim needs from before to after STOP for DV than SA.

    The Conceptual Framework Revisited

    The results of the predictive models indicate that some of the hypothesized relationships in our conceptual framework were supported, while others were not. Based on these results, the conceptual framework is revised in figure 6.3 to reflect which relationships were supported and which were not. The new framework reflects the lack of relationship between Box 1 (level of STOP funds and other resources) with Boxes 4, 5, and 6 (level of coordination in community response, post-STOP VS program services, and post-STOP legal system responses to victims). Arrows between Boxes 2 and 3 (pre-STOP level of community services and state STOP program support for collaboration) with Box 4 (level of coordination in community response) have become dashed to indicate the limited relationships among these independent and dependent variables. The same is true for the arrow between Box 4 (level of coordination in community response) and Box 5 (post-STOP VS program services). The arrow between Box 3 (state STOP program support for collaboration) and Box 6 (post-STOP legal system response to victims) has been eliminated entirely.

    The remaining arrows from Boxes 2 and 3 to Box 5 and Boxes 2 and 4 to Box 6 remain bolded because the predictive models indicate strong and consistent relationships between independent variables in these boxes with outcomes in Boxes 5 and 6.

    Conclusions

    Four important findings emerge from the current analysis. First, the greater the pre-STOP activity in communities addressing violence against women and developing the ability to meet the needs of victims, the higher communities are rated on their ability to meet victim needs after STOP funding. However, the greater the level of activity in communities before STOP, the less change they experience when it comes to addressing the needs of victims. From the VS program perspective, STOP funding has allowed communities already working together to continue to work on such issues and to achieve incremental improvements. In comparison, communities starting at the beginning had not gotten as far by the time of our survey, but the proportional amount of change was greater.

    Second, levels of coordination among community agencies are particularly important for system behavior changes. The more agencies work together in communities, including law enforcement and prosecution agencies working with VS programs, the more likely services are to improve for both VS programs and the legal system. Interactive activities in communities are particularly important in stimulating changes in law enforcement and prosecution behavior, as well as behavior around protective orders. In the current analysis, measures of coordination among community agencies were related to outcomes such as more arrests, better evidence collection, more convictions, and more adequate enforcement of protective orders. The relationship between coordination and changes in VS program services may be weaker than the relationship between coordination and changes in the legal system, because VS programs may be able to improve access to services, increase their ability to meet the needs of victims, and develop new/enhance existing program activities without coordinating with other agencies in the community.

    Third, contrary to our prediction, state STOP agency support for collaboration was not related to VS program outcomes or legal system outcomes. The level of this support may not be what matters for outcomes, but rather if this type of support exists at all. Because only STOP-funded VS programs are included in the current analysis, we may not be able to discern if the lack of this type of agency contact, such as prevails in communities without STOP funds, would make a difference as compared to communities with this type of agency contact. State STOP agency support, however, was related to more communication and more CCRs in communities. Findings from site visits conducted as part of the national evaluation of the STOP Formula Grant Program (Burt et al., 2000) and the work of the STOP TA project would support the finding that when a state STOP agency helped subgrantees to build collaborative relationships in communities, more communities in the state moved toward greater coordination and joint work.

    Finally, whether or not communities receive any STOP funding may result in differences in VS program and legal system outcomes. Although we found that measures of the level of STOP funding to VS programs was not directly related to VS program outcomes or changes in the legal system treatment of women victims, it is important to note that the current study only includes communities with STOP-funded VS programs. If we were to compare non-STOP communities to the communities included here, we might see much greater effects of STOP funding on VS program and legal system outcomes. Additionally, we might see that greater levels of coordination existed between VS programs and legal system agencies in STOP-funded communities compared to those not funded through STOP.

    In the current analysis we have examined if, given some STOP funding, what matters for outcomes is the total amount of STOP funds supporting the VS program. Total STOP funding may merely be a proxy for size of agency or size of community in which the STOP-funded program operates. Indeed, total STOP funding is significantly and positively related to the number of employees providing direct services to victims in agencies (r=.19, p < .05) and the number of volunteers providing direct services to victims in agencies (r=.22, p < .05), and is marginally related to the likelihood that programs serve urban geographic areas (r=.13, p < .10). Also, the number of STOP grants VS programs received is significantly related to the likelihood that programs serve urban geographic areas (r=.17, p < .05).

    Other Evidence of STOP's Impact

    Because we believe STOP funding has made a difference in communities across the nation based on numerous reports during site visits and telephone surveys, we have conducted additional analysis to examine how STOP funding is related to outcomes of interest. First, we examined if outcomes were related to respondent perceptions that changes in types of interaction with law enforcement, prosecution, and other VS programs were due to STOP funding. Second, we examined if outcomes were related to the types of activity VS programs were supporting with STOP funds. The results of these analyses indicate that they are.

    In general, agencies attributing changes in types of interaction with law enforcement to their STOP subgrant also reported greater coordination among community agencies and more positive VS program and legal system outcomes. More specifically, agencies reporting changes in types of interaction with law enforcement due to their own STOP project had lower ratings for communication (r=-.16, p < .05), but had higher ratings for collaboration (r=.20, p < .05), reported greater change in the community's ability to meet the needs of DV and SA victims from before to after STOP (r=.21 and r=.25, p < .05 respectively), reported changes in more types of interaction with law enforcement since STOP (r=.16, p < .05), reported greater levels of legal system response to DV and SA victim needs since STOP (r=.19 and r=.17, p < .05, respectively), reported greater change in the levels of legal system response to DV and SA victim needs from before to after STOP (r=.25 and r=.24, p < .05, respectively), and reported more law enforcement changes (r=.17, p < .05). Agencies attributing changes in types of interaction with law enforcement to the effects of other STOP projects in their community were more likely to be rated as CCRs (r=.21, p < .05), reported higher levels of community ability to meet the needs of SA victims since STOP (r=.24, p < .05), reported less change in the community's ability to meet the needs of SA victims from before to after STOP (r=-.18, p < .05), and reported less change in the legal system response to DV and SA victim needs (r=-.16 and r=-.18, p < .05, respectively).

    In general, attributing changes in more types of interaction with prosecution to the effects of STOP also reported greater coordination among community agencies and more positive legal system outcomes. More specifically, those who felt their own STOP program had affected these changes had lower ratings for communication (r=-.16, p < .05), reported more changed types of interaction with prosecution since STOP (r=.23, p < .05), reported a greater number of legal system changes since STOP (r=.18, p < .05), and reported a greater number of law enforcement and prosecution changes since STOP (r=.18 and r=.19, p < .05, respectively). Agencies attributing changes in types of interaction with prosecution to the effects of another STOP project had higher ratings for communication (r=.19, p < .05), were more likely to be rated as CCRs (r=.22, p < .05), reported more changed types of interaction with prosecution since STOP (r=.21, p < .05), and reported a greater number of legal system changes (r=.17, p < .05).

    In general, agencies reporting changes in types of interaction with other victim services due to STOP also reported greater coordination among community agencies and more positive VS program and legal system outcomes. Agencies reporting changes in types of interaction with other VS programs due to their own STOP project had higher ratings for collaboration (r=.34, p < .05), were more likely to be rated as a CCR (r=.21, p < .05), and reported more types of interaction with other victim services changed since STOP (r=.40, p < .05). Agencies reporting changes in types of interaction with other VS programs due to other STOP projects in their community were more likely to be rated as a CCR (r=.23, p < .05), reported more types of interaction with other victim services changed since STOP (r=.19, p < .05), had greater numbers of STOP-funded activities (r=.20, p < .05), reported more changed types of interaction with prosecution since STOP (r=.20, p < .05), reported higher levels of legal system response to SA victim needs since STOP (r=.24, p < .05), and reported less change in the legal system response to DV victim needs from before to after STOP (r=-.25, p < .05).

    In general, spending STOP funds on multidisciplinary first-response teams, victim witness services, and policy/protocol development resulted in the greatest numbers of associations with measures of coordination in community responses, and VS program and legal system outcomes. More specifically, spending STOP funds on multidisciplinary first-response teams was significantly related to higher ratings of communication, coordination, collaboration, and CCR (r=.17, .28, .32, and .21, p < .05, respectively), to greater numbers of STOP-funded activities (r=.28, p < .05), to the community's greater ability to meet the needs of SA victims since STOP (r=.16, p < .05), and to more changed types of interaction with law enforcement and prosecution since STOP (r=.18 and r=.15, p < .05, respectively). Spending STOP funds on victim witness services was significantly related to higher ratings of CCR (r=.17, p < .05), to greater numbers of STOP funded activities (r=.37, p < .05), to the community's greater ability to meet the needs of DV and SA victims since STOP (r=.17 and r=.20, p < .05, respectively), to more changed types of interaction with law enforcement and prosecution since STOP (r=.17 and r=.23, p < .05, respectively), to improvements in the legal system's response to DV and SA needs victims since STOP (r=.25 and r=.26, p < .05, respectively), and to greater behavior changes around protective orders (r=.17, p < .05). Spending STOP funds on policy and protocol development was significantly related to higher ratings of coordination, collaboration, and CCR (r=.18, r=.28, and r=.21, p < .05, respectively), to more changed types of interaction with law enforcement and other victim services since STOP (r=.17 and r=.17, p < .05, respectively), to greater numbers of STOP-funded activities (r=.42, p < .05), the community's improved ability to meet the needs of SA victims since STOP (r=.16, p < .05), the change in the community's improved ability to meet the needs of SA victims from before to after STOP (r=.17, p < .05), and the change in the legal system response to SA victims needs from before to after STOP (r=.17, p < .05).

    In sum, respondents reported the impact of STOP funding on their community in a number of different ways. The coordination of community agency activities and VS program and legal system outcomes achieved in communities were often attributed to their own STOP projects or other STOP projects in their community. Outcomes were also attributed to the type of activity for which STOP funding was used, with multidisciplinary first-response teams, victim witness services, and policy/protocol development activities associated with more outcomes than other STOP-funded activities.


    REFERENCES

    Bureau of Justice Assistance. 1993. Family Violence: Interventions for the Justice System. Program Brief. Washington, D.C.

    Burt, M.R., Zweig, J., Schlichter, K., Kamya, S., Katz, B., Miller, N., Keilitz, S., and Harrell, A.V. 2000 Report: Evaluation of the STOP Block Grants to Combat Violence Against Women Under The Violence Against Women Act of 1994. Prepared for the National Institute of Justice, forwarded to Congress. Washington, DC: The Urban Institute, 1999.

    Burt, M.R., Harrell, A.V., Raymond, L.J., Iwen, B., Schlichter, K., Katz, B., Bennett, L., and Thompson, K. 1999. 1999 Report: Evaluation of the STOP Block Grants to Combat Violence Against Women Under The Violence Against Women Act of 1994. Prepared for the National Institute of Justice, forwarded to Congress. Washington, DC: The Urban Institute.

    Burt, M.R., Newmark, L., Jacobs, L., and Harrell, A.V. 1998. 1998 Report: Evaluation of the STOP Block Grants to Combat Violence Against Women Under The Violence Against Women Act of 1994. Prepared for the National Institute of Justice, forwarded to Congress. Washington, DC: The Urban Institute.

    Burt, M.R., Newmark, L., Olson, K., Aron, L.Y., and Harrell, A.V. 1997. 1997 Report: Evaluation of the STOP Block Grants to Combat Violence Against Women Under The Violence Against Women Act of 1994. Prepared for the National Institute of Justice, forwarded to Congress. Washington, DC: The Urban Institute.

    Burt, M.R., Newmark, L., Norris, M., Dyer, D., and Harrell, A.V. 1996. The Violence Against Women Act of 1994: Evaluation of the STOP Block Grants to Combat Violence Against Women. Prepared for the National Institute of Justice, forwarded to Congress. Washington, DC: The Urban Institute.

    Clark, S. J., Burt, M. R., Schulte, M. M., and Maguire, K. 1996. Coordinated Community Responses to Domestic Violence in Six Communities: Beyond the Justice System. Prepared for the U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. Washington, DC: The Urban Institute.

    Department of Justice. 1984. Report of the Attorney General's Task Force on Family Violence. Final Report, September. Washington, D.C.

    Garner, J. and Fagan, J. 1997. Victims of domestic violence. In Davis, R. C., Lurigio, A. J., and Skogan, W. G. (Eds.), Victims of Crime, Second Edition (pps. 53-85). Thousand Oaks, CA: Sage Publications.

    Gondolf, E. W. 1998. Service contact and delivery of a shelter outreach project. Journal of Family Violence, 13 (2): 131-145.

    Gordon, J. S. 1996. Community services for abused women: A review of perceived usefulness and efficacy. Journal of Family Violence, 11 (4): 315-328.

    Gornick, J., Burt, M.R., and Pittman, K. 1985. The structure and activities of rape crisis centers in the early 1980s. Crime and Delinquency, 31(2): 247-268.

    Harrell, A.V., Roehl, J.A., and Kapsak, K.A. 1988. Family Violence InterventionDemonstration Programs Evaluation: Volume II, Case Studies. Report to the Bureau ofJustice Analysis. Washington, D.C.: Institute for Social Analysis.

    Koss, M. P. 1993. Rape: Scope, impact, interventions, and public policy responses. American Psychologist, 48 (10): 1062-1069.

    Koss, M. P. and Harvey, M. R. 1991. The Rape Victim: Clinical and Community Interventions. Newbury Park, CA: Sage Publications.

    Luna, E. M. 1999. Impact Evaluation of STOP Grant Programs for Reducing Violence Against Indian Women. Project report submitted to the National Institute of Justice. University of Arizona Tribal Law and Policy Program.

    Newmark, L., Harrell, A.V., and Adams, B. 1995. Evaluation of Police Training Conducted Under the Family Violence Prevention and Services Act. Project report submitted to the National Institute of Justice. Washington, D.C.: The Urban Institute.

    Resick, P. A. and Nishith, P. 1997. Sexual assault. In Davis, R. C., Lurigio, A. J., and Skogan, W. G. (Eds.), Victims of Crime, Second Edition (pps. 27-52). Thousand Oaks, CA: Sage Publications.

    Sullivan, C. M., and Bybee, D. 1999. Promising findings of a community-based advocacy project for women with abusive partners. Journal of Consulting and Clinical Psychology, 67(1): 43-53.

    Sullivan, C. M., Campbell, R., Angelique, H., Eby, K. K., and Davidson, W. S. 1994. An advocacy intervention program for women with abusive partners: Six-month follow-up. Journal of Community Psychology, 22 (1): 101-122.

    Sullivan, C. M., Tan, C., Basta, J., Rumptz, M., and Davidson, W. S. 1992. An advocacy intervention program for women with abusive partners: Initial evaluation. American Journal of Community Psychology, 20 (3): 309-332.

    Sullivan, C. M. and Davidson, W. S. 1991. The provision of advocacy services to women leaving abuse partners: An examination of short-term effects. American Journal of Community Psychology, 19 (6): 953-968.

    Tolman, R. M. and Weisz, A. 1995. Coordinated community intervention for domestic violence: The effects of arrest and prosecution on recidivism of women abuse perpetrators. Crime & Delinquency, 41 (4): 481-495.

    Weisz, A. N., Tolman, R. M., and Bennett, L. 1998. An ecological study of nonresidential services for battered women within a comprehensive community protocol for domestic violence. Journal of Family Violence, 13 (4): 395-415.


    Notes

    1. This project is supported by Grant No. 99-WT-VX-0010, awarded by the National Institute of Justice, U.S. Department of Justice. Points of view in this document are those of the authors and do not necessarily represent the official position or policies of the U.S. Department of Justice or of other staff members, officers, trustees, advisory groups, or funders of the Urban Institute.

    2. This structure was necessary as a prelude to set up the next phase of the project, in which we will interview women who have used services, and also women in the community. The eight states were Colorado, Illinois, Massachusetts, Pennsylvania, Texas, Vermont, Washington, and West Virginia.

    3. Chapter 2 explains in more detail how programs were selected for the sample.

    4. In order to interview 10 programs in Vermont, we had to take one program that had not received funding for two years or more and had a subgrant less than $10,000. In several other states, we had to call the state Stop administrator to learn about VS programs that had not yet been included in the SAPR database, or that did not appear to have had at least two years of funding when in reality they had.

    5. Throughout this report, we refer to the activities supported with Stop funds as the "Stop program," and the larger agency of which they are a part as the "host agency." The reader should remember, also, that the data on which this chapter is based have been weighted to make the results reported here representative of the approximately 1,200 Stop-funded programs offering direct services to victims that are operating within private nonprofit victim service agencies and have received at least $10,000 in Stop funds.

    6. Due to different funding arrangements or requirements in various states, not every host agency was a private nonprofit victim service agency, as initially required by our selection criteria. Of those that were not, all used or contracted for the services of such victim service agency staff. In each case, interviewers made the judgment that the VS program in question was substantially similar in spirit and actual functioning to a program hosted by a private nonprofit victim service agency to be included in the sample.

    7. "Comprehensive safety planning" was defined for respondents as "addressing safety in multiple contexts"—home, work, school, etc.; addressing implications of safety strategies—legal alternatives, emergency housing, etc.

    8. We also asked respondents to report numbers of referrals to and from their VS program and their different referral partners, in order to assess whether referrals in both directions had increased as a result of STOP. However, so much data were missing that no reliable assessment of change could be calculated.

    9. It is not clear why more VS programs could not report activities for their host agency from the years before receipt of Stop funding, but the changes over the years 1995 to 1999 in the number of agencies reporting the number of victims served suggest that they did not.

    10. The ideal statistic to report would be the per-program change from the year before the program received Stop funding to the year after. Unfortunately, as noted earlier, this statistic is quite unreliable as we lose 60 to 90 percent of our sample, depending on the specific type of service involved, when we try to make this calculation. Therefore, we do not report it.

    11. When the text refers to two percentages as being different, that difference is statistically significant at p < .05 or better. The tests of significance employed have a conservative bias, taking into account the departures from simple random sampling in the sample design and the non-independence of responses across columns (answers from the same respondent, describing program interactions with various different agencies, may appear in every column of table 5.1, in two columns of table 5.2, and in two or three columns of table 5.3). Conversely, statements in the text that one percentage did not differ from another percentage mean that the difference is not statistically significant at p < .05. A statement that a difference is marginal means that p is greater than .05 but less than .10.

    12. It is possible that, as primary agencies, their level of joint planning, funding, and/or institutionalized commitment was already high, and therefore did not need to change.

    13. There is no indication that STOP-funded programs specializing in serving women victims of sexual assault are any more included in or excluded from interactions with other community agencies than programs serving domestic violence victims. We analyzed the relationships in table 5.2 separately for VS programs focusing exclusively on domestic violence or sexual assault, or serving both groups of women. Patterns of response were essentially similar for communication and coordination activities, regardless of crime focus. Some differences may exist with respect to collaboration activities, but they are not strong or consistent. We do not report them because of all the various comparisons we made, only one reaches statistical significance, and this could have happened by chance. The very small cell sizes in a number of cases do not give us much confidence that the data really reflect the reality of program interactions rather than idiosyncracies of a few programs.

    14. One difference may be statistically significant while another that appears similar may not, due to differences in the number of VS programs that named each type of agency as primary.

    15. Chi-square estimates are statistical tests of the differences between the expected proportion of the sample with particular responses compared to the actual proportion of the sample with particular responses.

    16. Some indicators appear in more than one measure, as appropriate. For example, law enforcement behavior around protective orders is counted as a law enforcement behavior change as well as change in behaviors around protective orders.

    17. Bivariate correlations are simple associations between only two variables, such as communication ratings and changes in interaction with law enforcement, they are created without taking into account other factors that may affect the association, such as pre-STOP ratings for legal system response to victim needs. The correlation is indicated by an r in the text, along with the level of significance of the association between the two variables. The level of significance is the extent to which we can be confident the association of interest did not result due to chance. The lower the number, the more confidence in the association.

    18. Hierarchical regression models are statistical techniques that allow a researcher to isolate the effects of any particular variable by holding constant the effects of other variables. Because so many factors are associated with VS program outcomes and legal system outcomes, but are also strongly associated with each other, analyses of the outcomes of VS programs and the legal system must use techniques such as regression to take these associations into account simultaneously. This approach allows researchers to sort out what factors are really important and what factors are not. The term "hierarchical" means that the researcher looks at variables in stages, first using one (small) set of variables, then adding another set, then another, etc., looking after each addition at how much more has been explained.



    Topics/Tags: | Crime/Justice | Families and Parenting


    Usage and reprints: Most publications may be downloaded free of charge from the web site and may be used and copies made for research, academic, policy or other non-commercial purposes. Proper attribution is required. Posting UI research papers on other websites is permitted subject to prior approval from the Urban Institute—contact publicaffairs@urban.org.

    If you are unable to access or print the PDF document please contact us or call the Publications Office at (202) 261-5687.

    Disclaimer: The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders. Copyright of the written materials contained within the Urban Institute website is owned or controlled by the Urban Institute.

    Source: The Urban Institute, © 2012 | http://www.urban.org