Domestic Violence and Sexual Assault

Data Collection Systems in the States

October 1999

Justice Research and Statistics Association

ACKNOWLEDGMENTS

This report was written by Dr. Stan Orchowsky, JRSA’s Research Director, and Dr. Candace Johnson, Consultant. JRSA Project Assistant Laura Parisi and intern Kate Wagner assisted with the research on the project.

The following SAC directors and staff members conducted the three case studies presented in this report: Dolly Reed, Connecticut Office of Policy and Management; Dr. Gerard Ramker, Teresa Hirsch and Andrea Carr, Illinois Criminal Justice Information Authority; and Richard Moore and Dr. Paul Stageberg, Iowa Criminal and Juvenile Justice Planning. We would also like to thank the following SAC Directors and staff members who shared with us information about their state data collections systems: Kim English, Colorado Division of Criminal Justice; Evelyn Nestlerode, Delaware Statistical Analysis Center; Susan Burton, Florida Department of Law Enforcement; Dr. Christine Boyle, New Jersey Division of Criminal Justice; Richard Ely, New York State Division of Criminal Justice Services; and Stephen Grohmann, Wisconsin Office of Justice Assistance.

We would like to acknowledge all of the people in the states who shared information about their data collection systems, and especially staff of the state Uniform Crime Reporting programs and state domestic violence and sexual assault coalitions, and regret they are too numerous to name individually.

Thanks also to Grant Monitor Angela Moore Parmley and to Bernie Auchter at NIJ for their support and suggestions throughout the project.

Joan C. Weiss
Executive Director

This project was supported by Grant No. 96-IJ-CX-0057 awarded by the National Institute of Justice, Office of Justice Programs, 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.

TABLE OF CONTENTS

LIST OF TABLES

EXECUTIVE SUMMARY

BACKGROUND AND METHODOLOGY

SECTION I. DATA COLLECTION SYSTEMS

OVERVIEW OF STATE DATA COLLECTION SYSTEMS

LAW ENFORCEMENT INCIDENT-BASED CRIME REPORTING SYSTEMS

SPECIALIZED DOMESTIC AND SEXUAL VIOLENCE DATA COLLECTION SYSTEMS: INCIDENT-BASED

SPECIALIZED DOMESTIC AND SEXUAL VIOLENCE DATA COLLECTION SYSTEMS: SUMMARY-BASED

SERVICE PROVIDER SYSTEMS - CLIENT-BASED

SERVICE PROVIDER SYSTEMS - SUMMARY-BASED

KENTUCKY'S CENTRAL REGISTER

SUMMARY

SECTION II. STATE CASE STUDIES

BACKGROUND AND METHODOLOGY

IOWA'S INCIDENT-BASED CRIME REPORTING SYSTEM

CONNECTICUT'S SPECIALIZED DATA COLLECTION SYSTEM

ILLINOIS' SERVICE PROVIDER DATA COLLECTION SYSTEM

SECTION III. RECOMMENDATIONS

RECOMMENDATIONS

APPENDIX A. CRITERIA FOR DOMESTIC VIOLENCE FLAGS AND CASE SELECTION

APPENDIX B. STATE DATA COLLECTION SYSTEM CONTACT AGENCIES

LIST OF TABLES

Table 1. Statewide Domestic and Sexual Violence Data Collection Systems

Table 2. Relationship Codes in NIBRS

Table 3. Statewide Incident-Based Crime Reporting Systems

Table 4. Specialized Incident-Based Reporting Systems

Table 5. Specialized Summary-Based Reporting Systems

Table 6. Service Provider Client-Based Systems

Table 7. States with Summary-Based Service Provider Data Collection Systems

Table 8. Factors Required for Determining a Domestic Violence Case

Table 9. Individuals Responsible for Identifying Domestic Violence Cases

Table 10. Law Enforcement Uses of Domestic Violence Data

Table 11. Domestic Violence Data Used by Service Providers

Table 12. Service Provider Uses fo Domestic Violence Data

Table A. Offense Criteria for Domestic Violence Flags and Case Selection

Table B. Relationship Criteria for Domestic Violence Flgas and Case Selection

EXECUTIVE SUMMARY

The Violence Against Women Act of 1994 specified that a study be conducted of "how the States may collect centralized data bases on the incidence of sexual and domestic violence offenses within a State." The National Institute of Justice (NIJ) asked the Justice Research and Statistics Association (JRSA) to undertake a study of domestic and sexual violence incident data collection by the states. The study involved convening a panel of experts and surveying state Statistical Analysis Center (SAC) directors to determine how these data were collected in their states. The findings of this study were published in July 1996 in an NIJ Research Report entitled Domestic and Sexual Violence Data Collection.

In 1997, JRSA began the second phase of the study, which was designed to expand upon the findings of the first effort. The first step in this second study was to update the survey of SAC directors which provided the information included in the 1996 report. Changes in data collection procedures were noted, and the SAC directors were asked to provide the names of contacts in their states who could provide additional information on the states’ databases. JRSA then interviewed these contacts, along with the directors of the state domestic violence and sexual assault coalitions, to obtain detailed information on the data collection systems. Copies of data collection forms and other system documentation were also obtained.

The study also examined the data collection systems in three states in greater detail by having the SACs in those states study and report on their state systems. The three state "case studies" were conducted in: (1) Iowa, which studied its National Incident-Based Reporting System (NIBRS)-compatible incident-based crime reporting system; (2) Connecticut, which examined its Family Violence Reporting Program, a specialized domestic violence data collection system; and (3) Illinois, which studied its new automated system for collecting domestic violence and sexual assault service provider data. Each SAC conducted interviews with knowledgeable individuals who provided information about how the data systems functioned, and surveyed data providers to determine how the information was collected and used at the local level.

OVERVIEW OF STATE DATA COLLECTION SYSTEMS

In our classification of state domestic violence and sexual assault data collection systems, we used as a baseline the Federal Bureau of Investigation’s (FBI’s) summary-based Uniform Crime Reporting (UCR) program. The study included only those data collection systems which provide more detailed data on domestic violence and sexual assault than can be obtained from the national summary system. Crime reporting to the FBI’s UCR program under NIBRS was included in our study.

The systems reported here are divided into two basic types, depending on the source of the data: (1) law enforcement databases, which collect data on offenses reported to or arrests by local law enforcement agencies; and (2) service provider databases, which collect data on clients served by local domestic violence and sexual assault programs. A total of 34 states have some type of law enforcement data collection system for domestic violence, and 14 have this type of system for sexual assault data. A total of 16 states collect statewide domestic violence data from service providers, while 17 states maintain statewide systems for collecting sexual assault data. The service provider systems are further subdivided into incident-based systems (6 for domestic violence and 8 for sexual assault data) and summary systems (10 for domestic violence and 9 for sexual assault data).

Law Enforcement Incident-Based Crime Reporting Systems

Twelve states were identified that currently capture either domestic or sexual violence data statewide via an incident-based crime system. Some of these 12 state systems have been developed as a result of NIBRS, while others are state-based systems that are not necessarily compatible with NIBRS.

A total of 46 of the 54 states and territories surveyed indicated that they have implemented, or are working toward or planning to meet, the NIBRS data collection standards. For purposes of this study, only those states which estimated that the vast majority of crime in the state is reported through the state IBR system are considered to have "statewide" NIBRS-compatible data collection systems. This definition resulted in seven states being classified as NIBRS states, regardless of their status with regard to the FBI’s process of certifying states to submit data to NIBRS.

NIBRS provides significant enhancements over the summary UCR system for reporting and analyzing domestic violence and sexual assault. In expanding the number of crimes for which offenses reported to the police are tracked, NIBRS includes the additional assault offense categories of simple assault and intimidation, which will facilitate the study of domestic violence, and the inclusion of the additional sex offense categories of forcible sodomy, sexual assault with an object, and forcible fondling, which will enhance the study of sexual violence.

NIBRS also includes an extensive list of codes for identifying the relationship between the victim and offender in every violent incident. These codes include relationships within the family (such as spouse, common-law spouse, and sibling), and outside the family (such as acquaintance, ex-spouse, and boy/girlfriend). These relationship codes allow for the identification of domestic violence incidents, and also permit more detailed analysis of sexual violence information (for example, specifying the extent to which perpetrators were known to victims, were friends, acquaintances or neighbors of victims, or were family members of victims).

There are, however, two potential drawbacks to the use of NIBRS to identify domestic violence cases. A comparison of NIBRS relationship codes with the possible various relationship criteria used in the states, shows that NIBRS is missing several possible relationship codes that could be relevant in domestic violence cases. For example, while NIBRS includes former spouses as a relationship, it omits other former intimate relationships, such as boyfriend/girlfriend, from its list. This omission may result in many domestic-related cases not being identified as such in NIBRS-compatible data collection systems. States can add their own additional relationship codes to their NIBRS systems, but these will not be reported when the data are examined at the national level.

NIBRS includes 11 offenses for which only arrests, and not crimes, are reported. These include several offenses which are considered by some states to be domestic in nature, including "nonviolent family offenses," "trespassing," and "disorderly conduct." Thus some domestic-related offenses which do not result in an arrest will not be included in NIBRS-compatible data collection systems.

In addition to providing the ability to identify most domestic violence offenses, the NIBRS offense codes also allow for the analysis of sexual violence information. Under NIBRS, it will be possible to determine the extent to which perpetrators were known to victims, were friends, acquaintances or neighbors of victims, or were family members of victims.

In general, IBR systems provide information on characteristics of the victim and offender, along with information on the nature of the offense, whether a weapon was involved and if so, what type, and whether the victim was injured. Other variables captured in some of these systems include: whether or not a child was present during the incident, information on protection orders, and information on referrals to service providers.

Specialized Domestic and Sexual Violence Data Collection Systems: Incident-Based

Fourteen states collect statewide information on domestic violence or sexual assault incidents using specialized incident-based data collection forms. Incident-based data collection forms and activities are of two general types. About half of these states collect detailed information for every incident involving domestic or sexual violence. In the remaining states, only minimal information is recorded for each incident.

Specialized domestic violence and sexual violence systems generally capture similar, but more detailed, information regarding these incidents than the NIBRS/IBR systems. In addition to information about victims, offenders, and offenses, these systems may capture information on the specific circumstances surrounding the incident (for example, nature of the dispute or type of behavior involved); alcohol or drug involvement; presence of children; whether or not an arrest was made; prior abuse history; and whether a protection order was in effect.

Specialized Domestic and Sexual Violence Data Collection Systems: Summary-Based

Summary-based domestic violence and sexual assault reporting systems are used in nine states to capture the frequency of domestic and sexual violence incidents in each jurisdiction. In general, summary domestic violence and sexual assault specialized data collection systems collect data in the same manner as summary UCR systems.

Summary reporting forms typically provide less information on domestic violence and sexual assault incidents than incident-based forms. In general, little more is reported than the frequency of calls for service or incidents, with specifications for offense type or relationship. While most of the systems specify the nature of the offense, only a few capture victim or offender characteristics or victim/offender relationship.

Service Provider Systems: Client-Based

Spurred by the reporting requirements of federal and state agencies which fund domestic violence and sexual assault programs, and facilitated by coordination efforts of statewide coalitions, many states are now moving toward the development of centralized statewide data systems to capture client and service information on victims of both domestic and sexual violence. Nine data collection systems were identified that maintain information at the state level on each client served. In addition, nine states indicated that they are in the process of developing statewide client-based systems.

The information maintained in client-based systems is collected by staff from clients who request services from the program. In most systems, information is collected from hotline calls as well as from programs providing face-to-face services to victims. The data for most of these systems are obtained through a client intake process utilizing standard client or intake forms. Some of the newer systems involve direct computer entry at each local program.

One of the issues in considering client-based data systems as a source of information on the incidence of domestic and/or sexual violence is being able to clearly identify the client population. Thus, data systems should have the capacity to identify primary victims (as opposed to children or significant others who are also receiving services), new clients or incidents (as opposed to multiple contacts with the service provider related to a single incident), the type of abuse (especially in those systems which collect information on both domestic violence and sexual assault), and the time period during which the incident occurred (for example, whether a rape victim is calling about an incident that occurred recently or five years ago).

It should be noted that the capability of these systems to identify clients may vary depending upon which reporting source is considered. Many of these systems employ separate reporting procedures for hotline or crisis calls, for example, which provide less information than for other types of sources. In addition, crisis data may be more incomplete than those from other sources, since programs place a priority on service provision rather than data collection in these types of cases.

Service Provider Systems: Summary-Based

Many states utilize statewide summary systems to collect information from service providers. These systems, like the client-based systems previously discussed, have been implemented for the purpose of providing information to funding sources. Thus these summary-based service provider systems tend to focus on the number of clients and services provided. Because of the limited nature and purposes of these systems, rarely does information summarized at the state level appear to provide an indicator of the frequency of incidents or offenses. This is primarily due to the lack of information available in these systems for identifying clients. Thus, while some of the data from these summary systems may provide useful estimates of the incidence of domestic violence in a state (such as the number of new victims), the available information from such systems is of limited utility.

Summary

Each of the three types of data collection systems examined here has advantages and disadvantages. The two approaches that yield the most complete data on domestic violence and sexual assault are the specialized incident-based data collection systems and the service provider incident-based systems. The former systems are based on official reports to police, and are therefore limited to the extent that domestic violence and sexual assault incidents are not reported to the police. Service provider incident-based systems provide information on all clients who receive services, regardless of whether and when an incident has occurred. In order to be useful for estimating the incidence of domestic violence, these systems must allow for the identification of a primary victim, and for individuals who receive services on more than one occasion for the same incident. Regardless of which system is implemented, it should provide detailed information on the victim, the offender, and the characteristics of the incident.

NIBRS provides the most promise for comparing incident rates across states. NIBRS has the advantage of allowing for standard definitions of domestic violence and sexual assault based on offense and relationship codes. States can also add codes to identify domestic violence cases, and codes for their unique state statutes. Since NIBRS is a general crime reporting system, however, it does not provide as much detailed information on domestic violence and sexual assault incidents as do specialized or service provider systems. Nevertheless, states which implement incident-based crime reporting systems such as NIBRS may find that it is no longer necessary nor desirable to maintain specialized data collection systems for domestic violence or sexual assault.

STATE CASE STUDIES

Iowa’s Incident-Based Crime Reporting System

Conversion from summary to incident-based reporting (IBR) in Iowa was completed January 1, 1991. Iowa was the fifth state to be accepted as a certified "reporting state" of incident-based crime data to NIBRS. Iowa incorporated its incident-based domestic violence data and hate/bias crime data as part of the new IBR system, housing all crime data in one computerized system. Including domestic violence data collection as part of the new IBR system was relatively straightforward, since the existing domestic violence data collection was already incident-based, and since the data elements included in the new IBR system were compatible with those collected in the previous incident-based domestic violence system.

One of the issues associated with the switch to incident-based reporting in Iowa was the resulting decrease in reporting on the part of local law enforcement agencies. In the final year of the summary-based system, all 225 eligible agencies in the state reported crime figures directly to the Department of Public Safety. In 1991, the first full year of reporting under the new IBR system, only 61% of proportion eligible agencies reported data. In 1996, at the end of its sixth year of operation, 185 departments, or about 80% of eligible agencies, were direct contributors to the state. Some of the current non-reporters are among the largest departments in Iowa: Cedar Rapids, the state’s second-largest city, and Council Bluffs, the sixth-largest, are among them.

The most frequently reported reasons for non-reporting to the state IBR system were lack of compatible software and lack of data entry personnel. Other reasons given for not participating included lack of compatible hardware and having no computer system appropriate for UCR participation.

Connecticut’s Family Violence Reporting Program

Connecticut’s Family Violence Reporting Program was instituted in 1986. Completion of a family violence offense report is required for each family violence incident regardless of whether or not an arrest occurs. The data form completed by law enforcement agencies collects information on the date and time of the offense; the nature of the offense; number and type of weapons involved; seriousness of injury; whether or not alcohol or drugs were involved; whether or not there was a prior court order; the victim-offender relationship; and whether children were present or involved.

Connecticut's law enforcement agencies are currently in the process of converting from summary-based crime reporting to incident based reporting (NIBRS). Approximately 30 of Connecticut's 99 law enforcement agencies are currently collecting NIBRS data. The data components of the Family Violence Reporting Program are being incorporated into the NIBRS reporting program. Connecticut will continue to collect data using the current reporting program until NIBRS becomes operational statewide.

One of the advantages of specialized data collection systems is their ability to collect more detailed information on domestic violence than can be collected under more general crime reporting systems. One of the main advantages of Connecticut’s Family Reporting Program is its ability to provide consistent data on family violence over a long period of time, allowing researchers and policymakers access to information on long-term trends in domestic violence in their state.

Illinois’ InfoNet System

The Illinois Criminal Justice Information Authority’s (ICJIA) InfoNet is a new system designed to collect victim data statewide. The InfoNet is a tool to automate the required victim and service information that is reported by ICJIA-funded service agencies. This tool also allows each service provider to easily collect a variety of case level information including the victim’s circumstances, the court proceedings, and the services provided to the victim, and to create reports for other funding agencies. All of the information recorded by the service providers is kept confidential using a unique identifier for each client.

ICJIA staff has worked closely with the state’s domestic violence and sexual assault coalitions to create data entry screens which were customized to local programs’ needs. Data entry using the InfoNet software began at local domestic violence agencies in October 1997 and at sexual assault service provider agencies in July 1998. The InfoNet will allow the coalitions to answer questions about the amount and nature of victim services provided by their member agencies.

The InfoNet database was designed to link a program’s entire structure in order to both record and calculate a variety of administrative and service information. The relational database includes information on victims and offenders, program staff, volunteers, and financial information. Data entry for the InfoNet system is completed at the reporting agency’s site. Staff from ICJIA created the manual to guide the agencies as they set up the software and security systems of the InfoNet. ICJIA staff also held user group meetings to train and pilot the InfoNet system. Throughout the pilot and training process ICJIA compiled the opinions and reactions of users. The results from individual agencies have been overwhelmingly positive despite the difficulties of learning this new and complex automated system of data collection.

RECOMMENDATIONS

Based on the assessment of current state efforts and the case studies of three different state systems, the following are recommendations for the states with regard to domestic violence and sexual assault data collections systems:

• States should implement incident-based reporting systems which use offense and relationship codes that are compatible with the National Incident-Based Crime Reporting System (NIBRS).

• States should move toward implementing incident-based service provider domestic violence and sexual assault data collection systems.

• States should develop guidance and implement training on how to identify and report cases of domestic violence and sexual assault.

• States, with assistance from the federal government, should develop initiatives to analyze and validate domestic violence and sexual assault data being collected by statewide incident-based systems.

• States, with assistance from the federal government, should begin developing linkages among the various state data systems that collect information relevant to domestic violence and sexual assault incidents.

BACKGROUND AND METHODOLOGY

The Violence Against Women Act of 1994 specified that a study be conducted on "how the States may collect centralized data bases on the incidence of sexual and domestic violence offenses within a State." The National Institute of Justice (NIJ) asked the Justice Research and Statistics Association (JRSA) to undertake a study of domestic and sexual violence incident data collection by the states. In response to this request, JRSA convened a panel of experts representing backgrounds in criminal justice statistics, law enforcement, and victim services to provide comments and suggestions regarding domestic and sexual violence data collection. In addition, JRSA surveyed state Statistical Analysis Center (SAC) directors to determine how these data were collected in their states. The findings of this study were published in July 1996 in an NIJ Research Report entitled Domestic and Sexual Violence Data Collection. The current study expands on the findings of our previous effort by conducting a more in-depth assessment of states’ efforts to collect domestic violence and sexual assault data.

The first step in the current study was to update the survey of SAC directors which provided the information included in the 1996 report. Changes in data collection procedures were noted, and the SAC directors were asked to provide the names of contacts in their states who could provide additional information on the states’ databases. The contacts provided were most often the SAC directors themselves or contacts in the state’s Uniform Crime Reporting (UCR) office.

In order to collect more detailed information about the data collection systems in the states, we interviewed by phone the contacts provided by the SACs. These systems were limited to law enforcement-based systems and specialized domestic violence and sexual assault systems. Contacts in all 50 states and four additional jurisdictions (Washington, D.C., Northern Mariana Islands, Puerto Rico, and the Virgin Islands) were interviewed to verify the status of their systems, to identify any newly developed systems, and to collect additional information on the systems identified. If the contact person could not provide the required information, that person was asked for additional names of individuals who would be knowledgeable about the systems. These individuals were then contacted. Appendix B lists the names of the agencies that were contacted during the course of the study.

Copies of data collection forms or content descriptions of automated systems were requested of the individuals who were contacted regarding the data collection systems. Forms and other system documentation were received from most of the states contacted. This material was reviewed and used in conjunction with the interview information to develop the system descriptions and classifications discussed in the study.

As information was being collected, the decision was made to expand the scope of the data collection to include service provider-based data collection systems. In some states, contacts for the other systems provided us with information regarding service provider systems. However, in order to be sure that we had obtained comprehensive information on all such systems in the country, we obtained lists of the contacts for the domestic violence and sexual assault coalitions

in each state. We then contacted each coalition in each state and asked them about the existence in their state of domestic violence or sexual assault data collection systems in which service providers collected information regarding victims receiving services. Again, contacts were asked for some basic information regarding any identified systems, and were asked to forward copies of data collection forms or system descriptions. Appendix B also lists the individuals who were contacted from the coalitions.

All of the above-referenced contacts with the states were carried out in late 1997 and early 1998. Follow-ups were conducted to obtain data collection forms where they had been promised but not received. These follow-ups were generally completed by the summer of 1998. Thus the information presented in this report is accurate for that time period, but does not reflect changes and additional systems that were put in place after the middle of 1998, with the exception of the states’ status with regard to NIBRS certification, which is accurate as of August 1999.

Early in the project, a meeting of nine SAC representatives was convened to identify and discuss the issues associated with statewide domestic violence and sexual assault data collection. The nine SAC representatives (Colorado, Connecticut, Delaware, Florida, Iowa, Massachusetts, New Jersey, New York, and Wisconsin) presented overviews of the data collection systems in their states. The group reviewed the various issues associated with the collection of data from summary and incident-based systems, including how best to collect the data and assure their accuracy. The group confirmed the utility of the classification of state data collection systems ultimately used in the study, and discussed the advantages of a fully developed incident-based reporting system such as the Federal Bureau of Investigation’s (FBI’s) National Incident-Based Reporting System (NIBRS) or a state’s own incident-based reporting (IBR) system for the collection of domestic violence and sexual assault data.

Finally, three state SACs were selected to conduct more in-depth studies of the data collection systems in their states. Each was selected to serve as an example of a specific type of data collection system: (1) Iowa (an incident-based crime reporting system); (2) Connecticut (a specialized incident-based domestic violence data collection system); and (3) Illinois (a service provider data collection system). The SACs conducted interviews and surveys of data providers and users in order to provide a history and description of the data collection system, and information regarding how the data are collected and used.

This report is divided into three major sections. The first section presents the findings of the results of our analysis of state systems for collecting and reporting domestic violence and sexual assault data. The second section presents the findings of the state Statistical Analysis Center studies of the data collection systems in their states: Iowa’s incident-based crime reporting system, Connecticut’s specialized domestic violence data collection system, and Illinois’ service provider information system. The final section of the report presents recommendations to the states for improving the collection of domestic violence and sexual assault data.

SECTION I. DATA COLLECTION SYSTEMS

OVERVIEW OF STATE DATA COLLECTION SYSTEMS

In our classification of state domestic violence and sexual assault data collection systems, we have used as a baseline the FBI’s summary-based UCR program. We have not included state systems which report summary data to UCR as part of our assessment, since these systems provide limited data on domestic violence and sexual assault. Rather, we have included only those data collection systems which provide more detailed data on domestic violence and sexual assault than can be obtained from the national summary system. Crime reporting to the FBI’s UCR program under NIBRS is included in our study.

The systems reported here are divided into two basic types, depending on the source of the data: (1) law enforcement databases, which collect data on offenses reported to or arrests by local law enforcement agencies; and (2) service provider databases, which collect data on clients served by local domestic violence and sexual assault programs. A total of 34 states have some type of law enforcement data collection system for domestic violence, and 14 have this type of system for sexual assault data. The other main category of state data collection systems examined here are those which collect data on numbers of clients served from local domestic violence and sexual assault service provider programs, including domestic violence shelters, rape crisis centers, non-residential domestic violence and sexual assault programs, and hotlines. A total of 16 states collect statewide domestic violence data from service providers, while 17 states maintain statewide systems for collecting sexual assault data from service providers. The service provider systems are further subdivided into incident-based systems (6 for domestic violence and 8 for sexual assault data) and summary systems (10 for domestic violence and 9 for sexual assault data).

Table 1 provides a summary of the states’ domestic violence and sexual assault data collection systems according to the classification scheme outlined above. States which currently operate the system under consideration are marked by the symbol "X," while states which are developing or plan to develop the system are marked by the symbol "x."

The first three columns of the table provide each state’s status with regard to NIBRS. Since NIBRS represents the possibility of national-level reporting of detailed domestic violence and sexual assault data, it is discussed extensively in the section on statewide incident-based crime reporting systems. The remaining columns of the table indicate which states have domestic violence and sexual assault data collection systems for each of the remaining categories discussed above.

Table 1. Statewide Domestic and Sexual Violence Data Collection Systems ^a
(X = Statewide system; x = System being developed or planned)

^a The information in this table is accurate as of August 1998 with the exception of the states' status with regard to NIBRS certification, which is accurate as of August 1999.
*  NIBRS certified.
**  System is no longer statewide - data from a developing NIBRS system is combined to allow for statewide reporting on domestic or sexual violence.
***  Estimate is not available.
****  NIBRS system is not statwide, but NIBRS coding of violent crimes in non-participating agencies allows for statewide domestic and sexual violence reporting.

Definitional Issues

One of the difficulties in developing estimates of the incidence of domestic violence from state and local data collection systems is the lack of standardized definitions of domestic and family violence. Most states define domestic violence in their state statutes, while others define it specifically for data collection purposes. State definitions vary according to which offenses are specified, and which relationships are included. About half of the states have specific domestic violence battery offenses, which may be misdemeanors or felonies.¹ The remaining states classify domestic violence offenses according to the nature of the offense, most commonly simple or misdemeanor-level assaults.

Table A in Appendix A illustrates variations in offense criteria that are used by the states, and which result in a case being counted a domestic violence case. The table lists 29 states which collect relatively detailed information on domestic violence offenses. The table shows the variety of offenses that are specified in the states’ statutes or database definitions of domestic violence. For example, only 7 of the 29 states specifically identify property offenses, including destruction of property and vandalism, in their definitions of domestic violence. Similarly, 9 states specify sexual offenses, such as sexual assault, as part of their criteria for identifying domestic violence cases.

Table B in Appendix A shows the relationship criteria specified by the same 29 states. Again, wide variation in the definitions is apparent. For example, while almost all states include spouses, ex-spouses, household members, and those who have a child in common as relationships that define domestic violence, only 10 states include other intimate relationships, such as boyfriend/girlfriend, in their definitions of domestic violence.

As noted above, these criteria determine which cases are identified in databases as being domestic in nature. This makes comparing domestic violence cases across states where definitions vary difficult, since a case that might be included in one state would be excluded in another. These definitional issues should be considered in the ensuing discussions regarding the various state data collection systems.

Confidentiality Issues

Another concern that underlies the development of domestic violence and sexual assault data collection systems is the issue of the confidentiality of the data collected. To maintain the safety of clients and to provide the best services possible, it is important that the identity of clients remain confidential, along with any information they may provide. This is especially relevant for service provider data collection systems, which use client referrals for service as their basic source of information. Moreover, as states and localities see the benefits of sharing information among various agencies, the issue of how to maintain clients’ confidentiality will become even more challenging.

_____________________

¹ Miller, Neal, Domestic Violence Legislation Affecting Police and Prosecutor Responsibilities in the United States: Inferences from a 50-State review of State Statutory Codes.   Institute for Law and Justice, September, 997.

LAW ENFORCEMENT INCIDENT-BASED
CRIME REPORTING SYSTEMS

Statewide incident-based crime reporting (IBR) systems involve the collection and maintenance at the state level of standardized information on each incident of crime reported by law enforcement agencies throughout the state. An incident is typically defined as a single event, independent of the number of offenders, victims, or subsequent charges. Law enforcement officers responding to an incident usually complete a standard, pre-coded incident reporting form which includes information on the victim(s), offender(s), offense(s) and charge(s), property involved, and/or arrest(s). The amount of information pre-coded on the form, as contrasted with the amount left as free text, varies from state to state.

As noted previously, 12 states were identified that currently capture either domestic violence or sexual violence data statewide via an incident-based crime system. These 12 state systems can be categorized as one of two major types: NIBRS-compatible systems and state IBR systems. The characteristics of each of these types will be discussed in turn.

NIBRS-Compatible IBR Systems

The National Incident-Based Reporting System has been proposed as the new uniform crime reporting system for the country. NIBRS is an incident-based reporting system through which data are collected on each single crime occurrence. NIBRS collects data on each single incident and arrest within 22 offense categories made up of 46 specific crimes called Group A offenses. Facts about each crime are collected for each of the offenses coming to the attention of law enforcement. In addition to the Group A offenses, there are 11 Group B offense categories for which only arrest data are reported. The FBI has published standards for the submission of statewide incident-based crime information to NIBRS. States which meet the standard are certified by the FBI to submit data regardless of the number or percentage of local law enforcement agencies reporting data to the state. Since NIBRS has the potential to provide a great deal more information regarding domestic violence and sexual assault than the summary-based UCR system, the current study collected information on the status of NIBRS implementation in the states, summarized in Table 1.

A total of 46 of the 54 states and territories surveyed indicated that they have implemented, or are working toward or planning to meet, the NIBRS data collection standards. For purposes of this study, only those states which estimated that the vast majority of crime in the state is reported through the state IBR system are considered to have "statewide" NIBRS-compatible data collection systems. This definition resulted in seven states being classified as NIBRS states: Delaware, Idaho, Iowa, Kansas, North Dakota, South Carolina, and Vermont. These states are not all currently certified NIBRS states by the FBI (Delaware and Kansas are currently being tested for certification; the other five states are certified). An additional 12 states (Colorado, Connecticut, Massachusetts, Michigan, Nebraska, Ohio, Tennessee, Texas, Utah, Virginia, West Virginia and Wisconsin) are certified NIBRS states by the FBI, but are not classified as having statewide systems in Table 1, since most of the crime in the state is not covered under the IBR system (for example, Virginia has been NIBRS-certified by the FBI, but as of August 1998, only about 19% of the crime in the state is currently covered by law enforcement agencies reporting to the NIBRS system).

As the summary data presented in Table 1 suggest, there is considerable variation among the states in terms of their progress toward development of NIBRS-compatible crime reporting systems and the degree to which those which do have such systems have been successful in getting local law enforcement agencies to report their crime data to the system. These issues are examined in greater detail in a later section of this report.

NIBRS provides significant enhancements over the summary UCR system for reporting and analyzing domestic violence and sexual assault. First, NIBRS expands the number of crimes for which offenses reported to the police are tracked. Under the current summary UCR system, offenses reported to the police are reported for only the most serious crimes (known as Part I crimes). Under NIBRS, offenses reported are tracked for all Group A offenses. Of greatest importance is the inclusion of the additional assault offense categories of simple assault and intimidation, which will facilitate the study of domestic violence, and the inclusion of the additional sex offense categories of forcible sodomy, sexual assault with an object, and forcible fondling, which will enhance the study of sexual violence.

In addition to an expanded offense list, NIBRS includes an extensive list of codes for identifying the relationship between the victim and offender in every violent incident (see Table 2). These codes include relationships within the family (such as spouse, common-law spouse, and sibling), and outside the family (such as acquaintance, ex-spouse, and boy/girlfriend). Since the current UCR is summary-based, no information regarding relationships is available (except for homicides, for which the Supplemental Homicide Reporting form collects information on each incident). The inclusion of this information for all violent offenses provides the ability to identify offenses in which the offender and victim are related, thus providing the capability for identifying domestic violence offenses. Moreover, the extensive range of relationship codes in NIBRS allows for the identification of cases based on differing definitions of domestic violence (e.g., violence between spouses or ex-spouses).

There are, however, two potential drawbacks to the use of NIBRS to identify domestic violence cases. A comparison of the relationship codes shown in Table 2 with the possible various relationship criteria used in the states, as depicted in Table B in Appendix A, shows that NIBRS is missing several possible relationship codes that could be relevant in domestic violence cases. For example, while NIBRS includes former spouses as a relationship, it omits other former intimate relationships, such as boyfriend/girlfriend, from its list. This omission may result in many domestic-related cases not being identified as such in NIBRS-compatible data collection systems. States can add their own additional relationship codes to their NIBRS systems, but these will not be reported when the data are examined at the national level.

As noted previously, NIBRS includes 11 Group B crimes for which only arrests, and not offenses, are reported. These include several offenses which, as can be seen from Table A in Appendix A, are considered by some states to be domestic in nature, including "nonviolent family offenses," "trespassing," and "disorderly conduct." Thus some domestic-related offenses which do not result in an arrest will not be included in NIBRS-compatible data collection systems.

In addition to providing the ability to identify most domestic violence offenses, the NIBRS offense codes also allow for the analysis of sexual violence information. Under NIBRS, it will be possible to determine the extent to which perpetrators were known to victims, were friends, acquaintances or neighbors of victims, or were family members of victims.

State Incident-Based Crime Reporting Systems

As Table 1 shows, five states in addition to those which are NIBRS-compatible maintain incident-based crime reporting systems. Two states (Alabama and Montana) and the District of Columbia report domestic violence and sexual assault data through their state IBR systems. Two additional states, Nebraska and Oregon, maintain special IBR systems that require further explanation. Both states have systems which consist of abbreviated information on incidents provided in the form of entries on a logging form.² Nebraska’s IBR system collects data on forcible rape only, and so cannot provide any more detailed information on sexual assault incidents than would be available from a summary-based system. The state does identify domestic-related assaults with a special code, however, so it is shown in Table 1 as an IBR system providing data on domestic violence only. Oregon’s system captures crimes reported to the police for both Part I and Part II UCR offenses, and is therefore capable of providing more sexual assault information than a summary-based system. Since Oregon’s system is not capable of identifying domestic violence offenses, it is classified in Table 1 as providing sexual assault data only.

Table 2. Relationship Codes in NIBRS

Variations in Data Collection Practices

In addition to the data systems maintained by Nebraska and Oregon, there are some other notable variations in data collection practices among the 12 states with incident-based crime reporting systems. These include the following:

• Information from Delaware’s IBR system is converted to meet both NIBRS and current UCR reporting specifications. In January of 1998, law enforcement officers began to use information on domestic violence incidents. The domestic incident report contains information concerning prior abuse-related activity, protection orders, prior system contacts, and a risk assessment.

• North Dakota’s NIBRS system covers approximately 80% of reported crime. To complete statewide data collection on domestic and sexual violence, grant funds from the S·T·O·P (Services·Training·Officers·Prosecutors) Violence Against Women Formula Grants Program are being used to complete NIBRS coding on violent offenses at non-participating agencies.

• The Vermont system is totally automated and uses no standardized forms. Officers enter relevant information directly into the system via computer terminals.

• Montana has several agencies which do not submit data in the NIBRS format. Various local automation systems are being utilized as well as some hard copy submissions. The data are converted by the state as best as possible to provide compatibility and complete the statewide system.

Characteristics of Incident-Based Crime Reporting Systems

Table 3 summarizes the characteristics of the 12 IBR systems. Some of the key components of the systems are discussed below.

Table 3. Statewide Incident-Based Crime Reporting Systems

* No enhanced sexual violence data available.
** No domestic violence data available.
*** Modified UCR code.
**** Montana is converting its existing IBR system to one that is NIBRS-compatable.    The current IBR system covers almost the entire state, and about 80% of the data are NIBRS-compatible.
X- = fewer codes available than NIBRS
X+ = more extensive codes than NIBRS

Domestic Violence Indicators and Offense Codes

There are four methods by which domestic violence cases can be identified in the 12 IBR systems: relationship and offense codes, flags, specific offense codes, and specific crime statutes.  The seven NIBRS-compatible systems can identify domestic violence cases using NIBRS relationship and offense codes. While states may vary in their individual definitions of what constitutes a domestic violence offense, comparable estimates across these seven states could be obtained using a standard definition compatible with NIBRS relationship and offense codes. Of the five non-NIBRS systems, two (Alabama and the District of Columbia) use relationship and offense codes to identify domestic violence cases.

Several of the NIBRS states (Delaware, Iowa, Kansas, and Vermont) and two of the states with state IBR systems (the District of Columbia and Nebraska) flag domestic violence cases when they are reported to the system. In most states a special line or box is included on the incident-based reporting form which the responding officer checks to indicate that the incident meets the appropriate criteria for reporting a domestic violence incident. In most states these criteria involve statutory definitions of domestic violence.³

Montana uses a specific UCR offense code for domestic assault that can be used to identify domestic violence cases. Two of the NIBRS/IBR states (Vermont and the District of Columbia) have specific statutes related to domestic violence which are coded on the data reporting form.

Information Available

Table 3 shows the information available from the systems regarding the victim, the offender, and other information about the incident. The seven NIBRS-compatible systems, along with the Alabama and D.C. systems, all collect similar information. Two of the state IBR systems (Nebraska and Oregon) appear to be primarily logging systems for type of offense and, as a result, can provide little information other than counts of incidents by offense type. Montana’s system, once it is converted to meet NIBRS specifications, will collect the same types of data as the other NIBRS-compatible systems. The characteristics shown in Table 3 for Montana refer to the reporting capabilities of the state’s IBR system in its current form.

Victim and Offender Characteristics

The age or date of birth, race/ethnicity, and gender of victims and offenders are consistently captured by the NIBRS-compatible systems and the state IBR systems in Alabama, the District of Columbia, and Montana. The only variation is the lack of ethnicity coding in the Alabama system. Oregon only captures the victim’s gender in its incident-based UCR system. The relationship between the victim and the offender is captured in all systems except those of Montana, Nebraska and Oregon. The variations in coding of relationships was noted above in the discussion of domestic violence indicators.

Other Information

Several factors related to the offense are also typically captured by statewide IBR systems:

• Type of offense involved - this is included in all systems, although the categories used vary. These variations were examined above in the discussion of the identification of domestic violence offenses.

• Type of weapon involved - this factor is captured by all NIBRS systems and by systems in Alabama, the District of Columbia, and Nebraska. Both Alabama and Nebraska, however, utilize less specific weapon codes than those used in NIBRS.

• Injury to the victim - all NIBRS systems collect information about the type of injury to the victim. The District of Columbia’s system captures injury in a similar manner, while Alabama’s system records only treatment for an injury, but not the type of injury.

• Other variables - additional factors related to domestic violence are captured in four of the statewide systems. The presence of a child during the incident is recorded in both the Iowa and Kansas systems. Whether referrals were made to service providers is captured in Delaware and Iowa’s systems. Finally, whether a victim had a protection order at the time of the incident is documented in Delaware and the District of Columbia’s systems.

Documents Produced

Ten of the 12 states with IBR systems produce documents which focus on or include information on domestic or sexual violence. Nine states produce statistics in these areas as part of their state annual crime reports. North Dakota does not include this information in an annual crime report, but produces a special report focusing on domestic and sexual violence. In addition to including this information in their annual crime reports, Delaware and Alabama produce special domestic violence reports. Vermont, whose NIBRS system only recently became statewide, intends to begin producing reports in the near future.

Discussion of Incident-Based Crime Reporting Systems

Advantages

Incident-based crime reporting systems offer the potential for more in-depth analysis of crime data than is possible with summary-based systems. Incident-based information available at the state level allows for the examination of more specific types of crime, the reorganization of data to address more specific questions at the state level and across jurisdictions within the state, and the ability to explore complex relationships between multiple offenders and victims.

Under IBR systems such as NIBRS, states are better able to study the extent and nature of both domestic violence and sexual violence. The ability to connect victims and offenders through the use of extensive relationship codes, along with an expansion of the offenses for which crimes reported to the police are captured, allows states to explore the nature of domestic and sexual violence incidents reported to the police. The ability to select incidents of this specific nature for further study is available only with an incident-based system.

In addition, the use of standardized relationship and offense codes in NIBRS creates the potential for conducting comparisons between states based on similar criteria. Although the definitions and coding practices of individual states with regard to domestic violence and sexual assault offenses may vary, cross-state comparisons may still yield valuable information regarding the nature of these crimes. The addition of state-specific codes, statutes, and flags to NIBRS-compatible systems provides for both in-state and cross-state comparability, as well as providing data to analyze differences in definitions.

Finally, the collection of incident-based data at the state level provides the potential for merging this information with data from other sources to produce a more comprehensive assessment of domestic and sexual violence in the state. Many states indicated that crime information is subsequently merged with court or corrections files to provide an even more in-depth look at specific issues of interest.

Concerns

While incident-based systems are providing a wealth of information and possibilities for analysis at the state level, they have also added a level of complexity throughout the state which many are struggling to overcome. Many states are having difficulty in bringing all agencies on line with a standard system, providing adequate training to get consistent data collection, and obtaining all information requested in a timely fashion. These difficulties have the potential to negate the benefits of these systems.

Since incident-based systems collect data on all major crimes reported to the police, they do not provide detailed information about specific types of offenses, such as domestic or sexual violence. As a result, incident-based crime reporting systems are not as useful as specialized data collection systems for examining specific questions about the circumstances surrounding domestic violence or sexual violence incidents.

While the incident-based system allows for a greater wealth of information than summary-based systems, the complexity of the system, with multiple offenses, victims, and offenders, makes analysis more complicated. For example, 30 sexual assault incidents may yield 50 victim-to-offender relationships, which makes the interpretation and explanation of the data more difficult.

As noted previously, incident-based systems that use only the NIBRS offense and relationship codes may miss some domestic violence cases that fall outside of these codes, such as relationships between former boyfriends and girlfriends. Thus, these systems may be defining domestic violence more narrowly than is usually the case.

Finally, as with any law enforcement reporting system, agencies can only report the offenses of which they are aware. These systems, therefore, suffer from the weakness of any crime reporting system in underestimating the incidence of domestic violence and sexual violence to the extent that these crimes are underreported by victims.

_____________________

² Nebraska has a certified NIBRS system which includes only part of the crime in the state, and is therefore not included as a "statewide" incident-based crime reporting system for purpose of this study.  The statewide IBR system for Nebraska that is depicted in Table 1 and discussed in this section is based on the logging forms, and is not the state's NIBRS system.

³ In Delaware, a separate incident form is completed in domestic violence cases meeting criteria indicated through special instructions from the Attorney General.  When the information from these forms is merged with remaining state data, a flag is included for domestic violence cases.

SPECIALIZED DOMESTIC AND SEXUAL VIOLENCE
DATA COLLECTION SYSTEMS: INCIDENT-BASED

As shown in Table 1, 14 states (Connecticut, Georgia, Illinois, Maryland, Michigan, Nevada, New Jersey, New York, Oklahoma, Rhode Island, Texas, West Virginia, Wisconsin, and Wyoming) were identified that collect statewide information on domestic violence incidents using specialized incident-based data collection forms. Two of these states, Rhode Island and Wisconsin, also collect incident-based information on sexual assault - Rhode Island with the same form and Wisconsin with a separate form. Table 4 summarizes the characteristics of these specialized systems.

Incident-based data collection forms and activities are of two general types. For eight of the 15 systems (Connecticut, Georgia, Nevada, New Jersey, New York, Rhode Island, Texas, and Wisconsin ⁴ ), officers complete a standardized form for every incident involving domestic violence or sexual violence, as applicable. These are typically very comprehensive forms that provide detailed information about each incident. In the remaining seven systems (Illinois, Maryland, Michigan, Oklahoma, West Virginia, Wyoming, and Wisconsin ⁵), information is logged onto a multiline form on which each line represents one incident. As would be expected, information captured by these latter systems is typically less comprehensive than that maintained by the eight states which use detailed forms to track incidents.

The data collection procedures and systems in several states represent notable variations from those typical of specialized reporting systems. These data collection practices are as follows:

• Connecticut requires local law enforcement agencies to submit domestic violence incident forms to the state system only when an arrest is made. Under Connecticut statutes, arrest is mandatory in domestic violence cases.

• Georgia’s system involves the submission of incident forms monthly. These are followed by the submission of clearance forms, as applicable.

• Approximately 50% of reported crime in Michigan is reported via a certified NIBRS system. Local law enforcement agencies not submitting NIBRS data must complete a supplemental domestic violence report. Domestic violence data from the two systems are summarized separately and added together for reporting purposes.

• Oklahoma’s domestic violence logging form captures only date, time, and type of offense.

• Nevada’s central repository accepts not only completed forms, but other paperwork with the required information highlighted. Coding is completed centrally, as necessary.

• Rhode Island’s form is used for both domestic and sexual violence cases. Originally developed for domestic violence in 1988, it was revised in 1996 to cover sexual violence incidents as well.

Characteristics of Incident-Based Specialized Domestic Violence and Sexual Assault Systems

Table 4 summarizes the key characteristics of the incident-based specialized domestic and sexual violence data collection systems.

Defining Domestic Violence

While the data collected by specialized incident-based forms are similar to those collected by the statewide incident-based crime reporting systems, law enforcement officers in states with the former systems only provide information for incidents which have been classified as involving domestic violence and/or sexual violence. The determination by the officer or other agency

personnel as to whether any given incident constitutes domestic violence or sexual assault is thus critical to a case’s inclusion or exclusion from the data collection system.

In most of the states with these data collection systems, statutory definitions of domestic violence or sex crimes are used to determine the types of cases to be included in the system. In two states (Maryland and New York), domestic violence is defined by instructions on the incident form itself, without referral to statutory language. Michigan appears to have no standard statutory definition or instructions for how cases are to be identified as domestic violence or sexual assault.

Table 4. Specialized Incident-Based Reporting Systems

* These systems provide separate reporting forms for each individual incident.
** Michigan's form is only used by jurisdictions not collecting NIBRS information.

Information Available

Specialized domestic violence and sexual violence forms generally capture more information specifically related to these offenses than the NIBRS/IBR systems. Table 4 summarizes the specific information available in each state’s system.

Victim and Offender Characteristics

As with the NIBRS/IBR systems, most specialized forms capture victim and offender characteristics. All states’ specialized incident-based forms collect information on the age and gender of the victim and offender except Oklahoma’s. Most forms also capture the race of both victim and offender.

Victim/Offender Relationship

While all forms except Oklahoma’s include information on the relationship between the victim and offender, there is considerable variation among the states in the specific categories used to characterize the relationship. Several states use categories very similar to those used in NIBRS, but include additional categories such as "child-in-common." Other states employ categories that reflect the relationships specified in their state statutes as defining domestic violence; for example, "dating relationship," "co-habitant," or "former co-habitant." Two states (New Jersey and Rhode Island) include a separate data element to reflect living arrangement (in same household, for example) independent of relationship.

Offense-Related Information

All forms recorded some offense-related information, although variations in coding schemes were again evident. Offense codes used by states include NIBRS and UCR codes, state statutes, lists of offenses reflecting the state domestic violence definition, and lists of specific abusive behaviors (for example, biting and kicking). Agencies can either fill in the appropriate offense code or check one of a list of possible items.

Most states also capture weapon and injury information. Only Oklahoma does not record weapon use and Oklahoma and West Virginia do not capture victim injury. In general, codes found for weapon and injury on these forms have fewer categories than those used in NIBRS-compatible systems.

Other Information

The eight states which use single incident forms are more likely to include additional factors related to the incident than those states which log multiple incidents on the same form. Among the factors included by those eight states were:

• specific circumstances surrounding the incident (for example, nature of the dispute or type of behavior involved);
• alcohol or drug involvement;
• presence of children;
• whether or not an arrest was made;
• prior abuse history;
• whether a protection order was in effect.

Documents Produced

Most states with specialized incident-based reporting systems summarize the data collected by the system in published reports. Eleven states include domestic violence sections in their annual state crime reports, while three produce special reports on domestic violence. Of the two states collecting sexual violence information, Rhode Island includes the information in its annual crime report while Wisconsin produces a special report.

Discussion of Specialized Incident-Based Domestic Violence and Sexual Assault Data Collection Systems

Advantages

Specialized incident-based systems provide greater analytic capability than summary-based systems. The advantages of incident-based data discussed previously for crime reporting incident-based systems like NIBRS also apply to these state systems. In addition, since these systems include only domestic violence and sexual assault incidents, they are likely to contain more data elements of direct relevance to these offenses. They therefore provide the potential for more detailed analysis of the factors related to domestic violence and sexual assault. This specificity also applies to the data elements and coding schemes used in the data collection systems. For example, the specific coding of the relationship between the victim and the offender to correspond with state definitions may be extremely useful for the state’s study of its domestic violence policies and issues. However, the same categories would not be particularly instructive in the analysis of other types of crime.

Concerns

One of the drawbacks of this type of data collection system relates to concerns about its comprehensiveness in capturing domestic violence and sexual assault incidents. The decision regarding whether or not to report a given incident to the system is left to the discretion of the local law enforcement agency. The need for interpretations of what may be vague state definitions, the use of discretion based on the nature of the offense, and the disincentive associated with completing additional paperwork are all factors which may work to exclude some incidents from being reported to the system.

The fact that these specialized systems include specific elements and coding based on state definitions suggests that attempting to compare the data across states would be problematic. The opportunity afforded by NIBRS-compatible state systems to select cases based on standardized relationship and offense codes would not be as easily available in attempting comparisons across states with specialized systems.

Finally, state incident-based crime reporting systems also provide rich data on domestic violence and sexual assault incidents. As states move toward developing NIBRS-compatible crime reporting systems, there will be fewer states that are willing to maintain specialized domestic violence and sexual assault systems as well. Thus the types of specialized systems discussed here may be phased out in years to come as states move toward incident-based crime reporting.

_____________________

⁴  Wisconsin' domestic violence form.

⁵  Wisconsin's sexual violence form.

SPECIALIZED DOMESTIC AND SEXUAL VIOLENCE
DATA COLLECTION SYSTEMS: SUMMARY-BASED

Summary-based domestic violence and sexual assault reporting systems are used in nine states (California, Florida, Maine, Missouri, Ohio, Oregon, Puerto Rico, Tennessee, and Washington) to capture the frequency of domestic violence incidents in each jurisdiction. As part of its summary UCR system, Florida also captures enhanced sexual violence information.

In general, summary domestic violence and sexual assault specialized data collection systems collect data in the same manner as summary UCR systems. In fact, in three states (Florida, Maine, and Missouri) relevant domestic violence and/or sexual violence data are captured on the same UCR form required for all crime reporting. In Washington, while the information is captured on a separate form, it is submitted to the state with the UCR forms.

Each agency is required to submit summary reports on domestic violence incidents to a central agency. Reports are typically submitted monthly, although some states have quarterly reporting. Each agency is responsible for summarizing the required information for all incidents reported in its jurisdiction. At the state level, the information is again summarized across all jurisdictions. Notable variations in the methods by which states collect summary domestic violence and sexual assault data via specialized systems are as follows:

• California’s law enforcement agencies report domestic violence-related calls for assistance. Information is also collected from other "jurisdictions," such as Parks and Recreation, school districts, and railroads.

• Florida’s UCR form captures information on four types of sexual violence as well as domestic violence, making it the only system which provides both domestic violence and enhanced sexual violence information.

Characteristics of Specialized Summary Domestic Violence and Sexual Assault Data Collection Systems

Table 5 summarizes the characteristics of the nine domestic violence and one sexual assault summary data collection systems.

Table 5. Specialized Summary-Based Reporting Systems

* Combination age/gender/relationship use, all factors incomplete.
** Estimate only.
*** Information not available.

Defining Domestic Violence

As with the prior systems discussed, state definitions of domestic violence determine which cases are reported to the state system. As Table 5 shows, most of the summary systems use statutory definitions to make this determination.

Information Available

Summary reporting forms typically provide less information on domestic violence and sexual assault incidents than incident-based forms. In general, little more is reported than the frequency of calls for service or incidents, with specifications for offense type or relationship. The Florida UCR system provides the frequency of offenses reported, clearances, adult/juvenile arrests, and weapon involvement for four sexual violence offenses.

Table 5 shows the information available under each state’s system. Unlike the incident-based systems, which can produce data to analyze relationships between factors, these systems are capable of reporting only the total number of incidents falling under each category. Only five of the nine forms capture victim or offender characteristics or victim/offender relationship. Six of the nine forms specify the type of offense involved in the domestic violence incidents.

Documents Produced

Of the nine states with summary domestic violence reporting, six include this information in their state annual crime reports. Florida also includes its enhanced sexual violence information in its annual crime report. Florida and Puerto Rico include their summary domestic violence information in a separate report on domestic violence.

Discussion of Summary-Based Specialized Reporting Systems

Advantages

Specialized summary systems on domestic and/or sexual violence provide basic information on the extent and perhaps seriousness of the domestic or sexual violence problems reported to law enforcement. Counts for different categories of offenses, and various victim and offender characteristics, are also available from many of these systems. Thus these systems do provide basic information on the incidence of domestic violence or sexual assault as reported to law enforcement.

Concerns

These summary-based systems provide little information beyond numbers of incidents. Compared with other types of systems examined here, these systems allow for no further reorganization or analysis of the data at the state level. At the same time, systems which require counts across multiple categories of multiple factors (relationship by injury, for example) are likely to be cumbersome to complete unless factors are already automated. For these more complex summary forms, the submission of incident-based information might be easier for many local agencies.

As noted with the other systems, the identification of relevant cases may be of concern with summary forms. Such concerns are compounded by the use of summary forms since the determination of which cases to include in the local summary counts may actually be made by clerks or other office staff responsible for completing the reporting forms, as opposed to officers who have responded to the incident. Not only must the officer interpret and document information relevant to determining whether to report the incident, but another individual must interpret the offense definition and apply that definition to the information presented in the incident report.

SERVICE PROVIDER SYSTEMS - CLIENT-BASED

Spurred by the reporting requirements of federal and state agencies which fund domestic violence and sexual assault programs, and by coordination activities through statewide coalitions, many states are now moving toward the development of centralized statewide data systems to capture client and service information on victims of both domestic and sexual violence. As noted previously, these data collection systems can be divided into those which collect information on each client (comparable to incident-based law enforcement systems), and those which collect summary information. Table 1 shows that nine data collection systems were identified that maintain information at the state level on each client served. In addition, nine states indicated that they are in the process of developing statewide client-based systems. The newer client-based systems are designed to capture standard statewide client and service data directly through automated systems at the programs or facilities, while providing local programs with the capability of completing the reports needed to meet the requirements of multiple granting sources. While the data collection activities in many states place emphasis on capturing information on services provided, many record information on the victims, offenders, and characteristics of incidents of domestic and sexual violence that is as extensive as that maintained by incident-based law enforcement systems.

The information maintained in client-based systems is collected by staff from clients who request services from the program. In most systems, information is collected from hotline calls as well as from programs providing face-to-face services to victims. The data for most of these systems are obtained through a client intake process utilizing standard client or intake forms. Some of the newer systems (Alabama, Illinois, and New Hampshire) involve direct computer entry at each local program. In the remaining states intake forms are either submitted to a central agency for automation, automated at each local program, or both. The central agencies are usually state agencies or domestic violence and/or sexual assault coalitions or other similar organizations.

Characteristics of Client-Based Service Provider Systems

Table 6. Service Provider Client-Based Systems

* Information only available for cases admitted.
** This system records incident-based, not client-based, information.
*** Information is currently not available.
XO = This data element is an optional field.

Table 6 summarizes the characteristics of the nine client-based service provider systems.⁶ Of the nine systems identified, one (Alabama) focuses solely on domestic abuse victims, five (Alaska, Illinois, New Hampshire, Oklahoma, and Rhode Island) include both domestic and sexual abuse, and three (the District of Columbia, Massachusetts, and Washington) focus solely on sexual assault victims. In these systems, information on self-referred victims is obtained from hotline calls, non-residential programs, domestic violence shelters and service provider programs, and rape crisis centers. Table 6 shows which sources are used for each state’s system. Notable characteristics of the various state systems include the following:

• Alabama’s system collects only basic information on victims who call the crisis line and those who receive non-residential services. More extensive data on the incident and perpetrator are collected only for those clients admitted to a shelter.

• Alaska’s system has been operating since 1985 and reporting is mandated statewide. This is Alaska’s only source for domestic violence and sexual assault information, and the system captures a great deal of information.

• Illinois’ system just became operational in 1998. Automated reporting of information from programs to the states via the Internet will begin next year.

• New Hampshire’s new automated system is primarily geared to capturing information on clients and services as needed to satisfy granting sources. Little information is available on the offender and offense.

• Oklahoma’s system is operated by the Department of Mental Health and focuses on the provision of services. The system covers a variety of problems, including domestic violence and sexual assault victimization. The nature of the specific problem is not collected unless the client is admitted to a facility. Because the system covers so many problems, little information specific to domestic violence or sexual assault is available.

• Rhode Island’s system collects information from a newly developed hotline for all victims of crime. It includes information on victims, offenders, and offenses. Rhode Island is currently developing another automated system to support its victim service programs.

• The District of Columbia’s system consists of data collected from its rape crisis center.

• Massachusetts has collected sexual abuse information since 1986 in summary format and has only recently converted to a statewide client-based system. The system is incident-based, rather than client-based.

Information Available

Identifying Clients

One of the issues in considering client-based data systems as a source of information on the incidence of domestic and/or sexual violence is being able to clearly identify the client population. Thus data systems should have the capacity to identify primary victims (as opposed to children or significant others), new clients or incidents (as opposed to multiple contacts with the service provider related to a single incident), the type of abuse (especially in those systems which collect information on both domestic violence and sexual assault), and the time period during which the incident occurred (for example, whether a rape victim is calling about an incident that occurred within the last year or five years ago).

Table 6 summarizes the capabilities of the nine state systems with regard to identifying clients. All systems except one allow for the ability to indicate whether a report is for a new client or incident or represents additional services related to an incident already reported. All but one also allow service providers to identify a primary victim along with others (such as children) who may be receiving services. Most of the systems provide some categorization of the type of abuse that occurred. In the case of systems that collect both domestic violence and sexual assault information, this may simply be an indicator of which of the two types of violence the victim suffered. In the other systems, the notation under this factor in Table 6 indicates whether specific information regarding the nature of the incident is recorded. Of the nine systems, six identify the specific type of abuse that occurred. Finally, the table shows whether the systems can identify when the incident occurred, with at least the ability to identify incidents that occurred during the last year. Six of the nine systems have this capability.

It should be noted that the capability of these systems to identify clients may vary depending upon which reporting source is considered. Many of these systems employ separate reporting procedures for hotline or crisis calls, for example, which provide less information than for other types of sources. In addition, crisis data may be more incomplete than that from other sources, since programs place a priority on service provision rather than data collection in these types of cases.

Victim, Offender and Incident Information

All systems capture some basic information about the client. Six of the nine systems (Alabama, Alaska, Rhode Island, the District of Columbia, Illinois, and Massachusetts) capture information similar to that described for the law enforcement systems, including victim and offender characteristics (age, race, gender), as well as information on the nature and/or circumstances surrounding the incident (for example, weapon use). As would be expected, however, coding schemes for this information vary from state to state.

Five of the nine systems capture information as to whether the relevant incident was reported to the police. This information could be valuable in statewide attempts to estimate the incidence of domestic or sexual violence through comparison or combination with other information systems based on police reporting.

Documents Produced

In general, client-based systems have only recently been developed and have not yet focused on publishing reports concerned with the incidence of domestic assault or sexual abuse. Most systems are currently focused on providing grantors with required information on the numbers of clients being served and the types of services provided. The exception is Alaska, which publishes the data from its service provider system in an annual report.

Discussion of Incident-Based Service Provider Systems

Service provider systems have the potential to address the failure of law enforcement systems to collect data on many domestic violence and sexual assault incidents. To the extent that victims are more likely to call crisis hotlines or seek services than to contact the police, service provider systems should be more comprehensive in scope than law enforcement systems. However, these systems suffer from a similar coverage problem: not all victims seek services, and some victims who report incidents to police do not seek services. Therefore, not all victims would necessarily be identified under a service provider system. Moreover, it is not possible to accurately assess the degree of overlap between law enforcement and service provider systems. Determining the overlap between these two systems would be necessary in order to gain a comprehensive picture of the incidence of domestic violence and sexual assault.

Since these systems report information on all clients who receive services, they do not have the same definitional problems as law enforcement systems. A similar problem for these systems, however, is the definition of a client. Overcounting in such systems (at least in relation to how law enforcement systems count cases) can occur when the same client receives services on multiple occasions for the same incident, or when a female victim and her children all receive services, and all are counted. Thus those systems which allow these circumstances to be identified are most useful for obtaining statewide estimates of the incidence of domestic violence or sexual assault.

Finally, service provider systems vary widely in the amount of information they collect. Some systems collect detailed data on only a small subset of the clients who receive services. This makes comparing information across states which use such systems a difficult task.

_____________________

⁶ The list of information available includes primarily data elements chosen for comparison with the other systems described in this report.  Client-based systems are likely to include additional data elements regarding clients and the services they receive which are not documented in the table.

SERVICE PROVIDER SYSTEMS - SUMMARY-BASED

Many states utilize statewide summary systems to collect information from service providers on hotline calls, clients receiving services from non-residential domestic violence and other victims programs, domestic violence shelters, and rape crisis centers. These systems, like the client-based systems previously discussed, have been implemented for the purpose of providing information to funding sources. Thus these summary-based service provider systems tend to focus on the number of clients and services provided.

Table 7 shows which states collect summary-based service provider data. In all of these systems, client or service information is submitted in summary form from programs or facilities to a central agency. Because of the limited nature and purposes of these systems, rarely does information summarized at the state level appear to provide an indicator of the frequency of incidents or offenses. This is primarily due to the lack of information available in these systems for identifying clients. As noted in the discussion of incident-based service provider systems, adequate incidence information would require the ability to identify clients who are primary victims, first time clients, the type of abuse, and the recency of the offense. Because of the aggregate nature of information from these summary systems, these data elements are not available. Thus, while some of the data from these summary systems may provide useful estimates of the incidence of domestic violence in a state (such as the number of new victims), the available information from such systems is of limited utility.

Table 7. States with Summary-Based Service Provider Data Collection Systems

* Domestic violence and sexual assault data are collected in a single summary reporting system.

KENTUCKY'S CENTRAL REGISTER

One additional state system is worth mentioning because of its unique nature. The state of Kentucky collects adult abuse information, including information on spouse abuse, in a system operated by its Department of Social Services (DSS). The Kentucky Adult Protective Services Central Register, implemented in 1976 and automated in 1986, was designed to provide a central statewide information system on reported cases of adult abuse, self-neglect, neglect by caretaker, exploitation, and spouse abuse. With reports submitted from multiple sources, this system combines the benefits of both the specialized incident-based law enforcement forms and victim service provider client-based forms. This is the only state system identified that combines information from multiple sources into a single case-based system.

Because reports are mandated from "any person who has reasonable cause to suspect that an incident has occurred," this central register is initiated by reports not only from law enforcement, but from victim service providers, hospitals, relatives, friends, and the victims themselves. Law enforcement officers in the state complete a special domestic violence incident form which is forwarded to the DSS. Other reports are more likely to be initiated by phone or office visits.

Information captured in the DSS database includes victim characteristics (age, race/ethnicity, gender, marital status, living arrangement, and victim/offender relationship), offender characteristics (age, gender, and race/ethnicity), reporting source, investigation findings, service needs and referrals, and legal action. While the abuse is categorized as adult abuse or spouse abuse, no additional information on the nature of the abuse is contained on the form.

As the major source for statewide information on adult and spouse abuse in Kentucky, this information is utilized for management reporting and to identify trends. Data are also available to be used for research. The information is published annually in a special report from DSS and portions are included in Kentucky’s annual crime report.

SUMMARY

Each type of data collection system examined here has advantages and disadvantages. The two approaches that yield the most complete data on domestic violence and sexual assault are the specialized incident-based data collection systems and the service provider incident-based systems. The former systems are based on official reports to police, and are therefore limited in their scope to the extent that domestic violence and sexual assault incidents are not reported to the police. Service provider incident-based systems provide information on all clients who receive services, regardless of whether and when an incident has occurred. In addition, these systems must allow for the identification of a primary victim and for individuals who receive services on more than one occasion for the same incident. Regardless of which system is implemented, it should provide detailed information on the victim, the offender and the characteristics of the incident.

NIBRS provides the most promise for comparing incident rates across states. NIBRS has the advantage of allowing for standard definitions of domestic violence and sexual assault based on offense and relationship codes. States can also add codes to identify domestic violence cases, and codes for their unique state statutes. Since NIBRS is a general crime reporting system, however, it does not provide as much detailed information on domestic violence and sexual assault incidents as do specialized or service provider systems. Moreover, NIBRS is missing at least one important relationship code, and includes several domestic-related offense categories only when an arrest occurs. Despite these limitations, states that implement incident-based crime reporting systems such as NIBRS may find that it is no longer necessary nor desirable to maintain specialized data collection systems for domestic violence or sexual assault. This may be especially true if states add additional variables and codes to the NIBRS system that will allow them to collect additional data on domestic violence offenses.

Many states have had difficulty in implementing NIBRS. States which have implemented the system have difficulty in obtaining the cooperation of local law enforcement agencies, especially large ones, in reporting data to the system. Thus the conversion to NIBRS may result in an apparent drop in the number of domestic violence and sexual assault incidents caused by fewer agencies reporting data to the system, not to an actual decrease in incidents. For this reason, obtaining national estimates of domestic violence and sexual assault based on NIBRS will remain difficult for the foreseeable future.

One issue common to all of the law enforcement data collection systems is the need for consistency in how cases are reported to the system by local law enforcement agencies. States need to provide training on an ongoing basis to ensure that all local agencies are using the same criteria to report incidents to the system.

Once domestic violence and sexual assault databases are firmly established, states should consider linking their law enforcement and service provider databases to other data collection systems, including courts, corrections, and social service agencies. Information on case dispositions and services provided in non-criminal justice settings would considerably enhance states’ ability to conduct meaningful analyses of domestic violence and sexual assault incidents.

SECTION II. STATE CASE STUDIES

BACKGROUND AND METHODOLOGY

In order to give states more detailed information about the various types of domestic violence and sexual assault data collection systems, JRSA arranged for the SACs in three states to conduct in-depth studies of the data collection systems in their states. The three state SACs and their corresponding data systems were: (1) Iowa, which studied its NIBRS-compatible incident-based crime reporting system; (2) Connecticut, which examined its Family Violence Reporting Program, a specialized domestic violence data collection system; and (3) Illinois, which studied its new automated system for collecting domestic violence and sexual assault service provider data.

Iowa was one of the first states to convert to NIBRS and become certified to submit NIBRS data to the FBI. The vast majority of the state’s localities report NIBRS data. In addition, Iowa had a domestic violence data collection system in place when it implemented NIBRS, and supplements its NIBRS data collection with additional variables related to domestic violence. These factors make the state a good choice for examining a NIBRS system.

Connecticut’s Family Violence Reporting Program is one of the oldest in the country. Connecticut has been collecting specialized domestic violence data since 1986, and the continuity of the system has produced a great deal of policy-relevant data. The Connecticut system was chosen for detailed study because of these characteristics.

Illinois’ automated service provider information system is new, and represents the "state of the art" in collecting data from service providers. Since service provider information systems are relatively new to the states, it seemed important to examine the issues related to the development and implementation of such a system. For these reasons, Illinois was selected to represent a state collecting service provider data.

Each of the three SACs used the same basic methodology to collect information on its system. Each SAC conducted interviews with knowledgeable individuals who provided information about how the data systems were developed and implemented, and how they functioned. The SACs also surveyed data providers (law enforcement agencies in Iowa and Connecticut; domestic violence and sexual assault programs in Illinois) to determine how the information was collected and used at the local level. The results of these interviews and surveys were compiled in reports which were submitted to JRSA. These reports were edited for inclusion in this document. More detailed information regarding these three systems is available from the SACs which conducted the studies.

As noted previously, states define domestic violence in different ways, and some of these definitions are more narrow than others. The data collection systems in the three states discussed here are limited by their definitions of domestic violence and sexual assault. State decisionmakers need to consider how the systems described here might fit with the definitions of domestic violence and sexual assault used in their states.

IOWA'S INCIDENT-BASED CRIME REPORTING SYSTEM

History of the Uniform Crime Reporting Program in Iowa

In 1974, the Iowa General Assembly enacted a provision of the Code of Iowa requiring law enforcement agencies to submit reports of crime and arrests to the Department of Public Safety’s (DPS) Bureau of Criminal Investigation. On January 1, 1975, the Iowa Uniform Crime Reporting program was implemented, with forms being sent to 210 local law enforcement agencies around the state. The forms used were provided by the Federal Bureau of Investigation (FBI), since most contributing agencies had previously submitted data directly to the FBI. Monthly reports were received from 209 agencies throughout 1975 and 1976. From 1977 to 1990, the number of agencies submitting reports slowly grew, reaching a total of 225 in 1990. With very few exceptions, the reporting agencies submitted data for every month from 1977 to 1990.

This summary-based system was used in Iowa until implementation of the National Incident Based Reporting System (NIBRS). Conversion to incident-based reporting (IBR) in Iowa was completed January 1, 1991.⁷ Iowa was the fifth state to be accepted as a certified "reporting state" of incident-based crime data to the national system.

Planning for IBR in Iowa began in 1986, with the impetus coming from the federal government. State and local officials saw the utility of an incident-based system, and the Department of Justice was providing funds for states to develop incident-based systems compatible with NIBRS. A statewide steering committee was formed which included representatives from sheriffs’ offices and police departments varying in size and region of the state. Having had experience working in and with local law enforcement agencies, steering committee members and DPS were aware that some incentives had to be provided to encourage participation in a new system which would require more work at the local level. The incentives were provided in the form of new incident, arrest, and supplementary reporting forms which could be used by local agencies in their daily operations.

The draft designs of the forms were taken to five regional meetings. Changes were made to the forms based on the feedback received in the meetings. A check box format was adopted that included a fairly large number of elements in an effort to reduce the need for lengthy narrative.

Forms were initially provided to departments at no cost. Use of these forms permitted small departments in particular to report data to DPS and obtain summaries back for local consumption. Many smaller departments had no other vehicle for the development of local reports, and their participation enabled DPS to provide a service to the local agencies. Although there was never any requirement to use the forms, some departments began using the first draft forms as soon as their existence became known, and usage became much more widespread after the forms were finalized.

Local departments can report data to Iowa’s IBR system in one of three basic ways. Those that have in-house computer systems capable of reporting data do so electronically. Other departments either report online through the IOWA system or dial into the system via telephone. Finally, some smaller agencies continue to report manually on data submission forms.

Domestic Violence Data Reporting in Iowa

Under Chapter 236 of the Code of Iowa, the Department of Public Safety is also charged with the responsibility of collecting information on incidents of domestic abuse. DPS began collecting domestic abuse data on July 1, 1985, assuming a function previously performed by the Iowa Department of Human Services (DHS). The transfer to DPS was due to legislative action which stemmed from dissatisfaction with domestic violence reporting to DHS; local law enforcement agencies simply weren’t reporting incidents of domestic violence. After data collection responsibility was transferred to DPS (to whom the local agencies already reported crime data), the number of reported domestic violence incidents increased dramatically, from 3,501 incidents in 1986 to 6,199 in 1990. It is the opinion of observers in the Department of Public Safety that these increases were due both to better reporting and actual increases in domestic violence.

With the implementation of incident-based reporting in 1991, the Department incorporated incident-based domestic violence data and hate/bias crime data as part of the new system, housing all crime data in one computerized system. Including domestic violence data collection as part of the new IBR system was relatively straightforward, since the existing domestic violence data collection was already incident-based. None of the decisionmakers involved in the transition supported maintaining domestic violence data outside of the new incident-based system.

The data elements included in the new IBR system were compatible with those collected in the previous incident-based domestic violence system, resulting in comparable data. Both systems involved check-off boxes to simplify completion of forms. The IBR system does include demographic data pertaining to victims and offenders; this information was not part of the old system. The data elements collected on domestic violence cases under the IBR system include the following:

• reporting agency
• day, date, and time of occurrence
• name, sex, race, ethnicity, address, and age of victim
• name, sex, race, ethnicity, address, and age of offender
• relationship of victim and offender
• type of injury
• presence of children at the time of abuse
• identity of person reporting abuse
• weapons used
• referrals made
• alcohol/drug involvement
• arrests made
• offender present at scene upon police arrival

Sexual Assault Data Reporting in Iowa

Data on sexual assault in Iowa are collected in the same manner as domestic violence data as part of the incident-based system. Sexual assault data collection has been part of Iowa’s UCR system since it originated in 1975. Most of the data elements collected for domestic violence are also collected for sexual assault, with the exception of the following:

• referrals made
• presence of children at time of abuse
• identity of person reporting abuse
• name, address, and date of birth of offender and victim.

Issues Associated with Implementation of the IBR System

Loss of Reporting Agencies and Data

One decision made by the steering committee during the implementation of the IBR system was to avoid parallel data collection systems. Under the rationale that there is little incentive for agencies to use a new system if a comfortable old system exists, the committee opted to discontinue Iowa’s old summary-based UCR system on December 31, 1990, and begin the new system the following day. Representatives from the DPS report that there was never any thought of operating dual systems and that they do not regret having made the transition in this manner.

At the same time, DPS representatives also admit that moving to a system which required more work on the part of local law enforcement agencies led to a decrease in reporting. In the final year of the summary-based system, all 225 eligible agencies in the state reported crime figures directly to the Department. In 1991, the first full year of reporting under the new IBR system, only 61% of eligible agencies reported data. This proportion increased to 72% the next year and 78% in 1994. In 1996, at the end of its sixth year of operation, 185 departments, or about 80% of eligible agencies, were direct contributors to the state. Some of the current non-reporters are among the largest departments in Iowa: Cedar Rapids, the state’s second-largest city, and Council Bluffs, the sixth-largest, are among them.

The impact of this loss of reporting agencies can be seen in the change in the number of reported domestic violence incidents under the new IBR system. In 1990, under the old system and with all agencies reporting, there were 6,199 reported domestic violence incidents. In 1991, the first year of the new IBR system, there were only 2,986 incidents. While reported incidents under the new system increased steadily after 1991, it was not until 1995 that reports under the new system reached the level they were during the last year under the old system. Reports of sexual assault also showed this same drop from 1990 to 1991.

Reasons for Non-Reporting

When questioned about the reasons that local agencies do not participate in Iowa’s incident-based reporting program, DPS personnel report that either the lack of compatible software or insufficient personnel typically are responsible for non-participation. These observations were supported by the findings of a survey of local law enforcement agencies. The 25 non-reporters who responded to the survey cited lack of compatible software and lack of data entry personnel as the two most common reasons for non-reporting. Other reasons given for not participating included lack of compatible hardware and having no computer system appropriate for UCR participation. Only two respondents, both sheriffs, mentioned any philosophical disagreement with the incident-based reporting system.

It is interesting to note that slightly over half the non-reporters (13 of 25 agencies) indicated that they did not collect any domestic violence data at all. The remaining agencies collected data through manual systems (6 agencies), automated systems (5 agencies), or a combination of the two (1 agency).

Variations in Reporting Domestic Violence Cases

One of the questions addressed by the survey of local law enforcement agencies is the degree to which reporting criteria for domestic violence cases vary among agencies. If policies pertaining to reporting domestic violence are not uniform from one agency to another, even agencies which strive to be complete in their reporting will report results different from other agencies with similar domestic violence problems but different policies. To address this issue, law enforcement agencies were asked the following question: "In your agency, what information is considered sufficient to make a determination of domestic violence (i.e., what minimum threshold must be met before a case can be defined as domestic violence)?" The responses of the 156 agencies which answered this question are shown in Table 8. As the table shows, there is substantial variation in the number and combinations of factors which departments consider sufficient for filing a domestic violence case. For example, responding to a domestic disturbance in which there is only the victim’s statement regarding what occurred would result in a domestic violence case being reporting in some agencies, but not others.

Another factor thought to influence domestic violence reporting is the individual actually making the final determination of domestic violence for purposes of reporting to the state system. Thus survey respondents were asked, "Who makes the final determination of domestic violence for purposes of UCR reporting?" Again, there was considerable variation among the 131 agencies who responded to this question (see Table 9). Although the investigating officer made this determination in most agencies, UCR clerks or officers also received frequent mention, along with a variety of other individuals. These differences in who makes the determination of whether an incident is classified as domestic violence are another potential source of variation in reporting cases to the statewide system.

Table 8. Factors Required for Determining a Domestic Violence Case

Table 9. Individuals Responsible for Identifying Domestic Violence Cases

Uses of Domestic Violence Data

One of the other questions the survey addressed was how incident-based data on domestic violence were used by law enforcement agencies. The responses to this question are shown in Table 10. The most common use of domestic violence data by police agencies was for reports and presentations: to boards, the community, coalitions, other criminal justice agencies, and the media. Fewer than one in four departments indicated that they used the data for planning or deploying personnel. Almost one in four departments reported that they did not use the data at all. In an additional question which asked why departments did not use the state-reported domestic violence data, most responded that there was no need to use the data, although many indicated that their departments collected their own data which were used for various purposes.⁸

Table 10. Law Enforcement Uses of Domestic Violence Data

Service providers were also asked about their use of domestic violence data. The first question asked providers to indicate the source of the domestic violence data they used. The responses of the 37 agencies who answered this question are shown in Table 11. Statewide data and local data from the state incident-based system were the most commonly used. Many programs reported using their own data as well. Only two programs reported that they did not use any domestic violence data.

Service providers were also asked how they used domestic violence data. The responses to this question are shown in Table 12. The most common responses were "presentations to the community" and "grant applications." About one in three programs reported using the data for planning purposes, and about 1 in 5 used the data for staffing. About 1 in 3 programs reported not using the IBR data at all. When questioned about their lack of use of the state IBR data, most programs reported that they generated and used their own domestic violence data. Four programs indicated that they did not use the data reported to the state because they had no confidence in the data.

Table 11. Domestic Violence Data Used by Service Providers

The other source of information about the uses of domestic violence and sexual assault data generated by Iowa’s IBR system comes from interviews with state and university analysts and legislative staff members. In general, the users were satisfied with the IBR data collected, and expressed particular satisfaction with the responsiveness of DPS to requests for information. The concerns raised by the interviewees could best be classified as issues endemic to any system which generates domestic violence and sexual assault data from more general reporting of law enforcement agencies regarding offenses and arrests. For example, interviewees expressed concern about the accuracy of the IBR data, since many domestic violence and sexual assault incidents go unreported to the police. Several interviewees also noted their desire for additional types of data related to domestic violence and sexual assault that are not collected in a law enforcement-based system, such as information about protection order violations and case dispositions.

Summary and Conclusions

This study of Iowa’s incident-based reporting system as a source of data on domestic violence and sexual assault incidents illustrates several issues typical of the implementation and use of such systems. First, not all local law enforcement agencies will be willing or able to convert from an aggregate reporting to a more complex incident-based system. In implementing the new IBR reporting systems, the IBR Steering Committee in Iowa wisely incorporated incentives for local agencies to participate by providing reporting forms which would be allow those agencies to collect information which would they would find useful. Despite this, however, about 1 in 5 eligible agencies still do not report data some seven years after the IBR system was implemented.

Table 12. Service Provider Uses of Domestic Violence Data

The problem of non-participation appears to be resource-related. Especially in the case of NIBRS-compatible systems, which have specific reporting requirements, some departments do not have the computer software or hardware, or the appropriate personnel, to report to the system.

Assuming that a state does not maintain dual systems, one aggregate and one incident-based, the result of this loss of reporting agencies will be a drop in the number of reported domestic violence and sexual assault incidents. Thus, the price paid by the conversion to incident-based reporting, and the subsequent loss of reporting agencies is the inability to track the number of domestic violence cases over an extended period of time. Iowa’s experience shows that even after a number of years, there would still be agencies which do not report data to the system. This means that even now it is impossible to obtain an accurate assessment of the incidence of domestic violence in Iowa.⁹

This issue, of course, is not specific to Iowa’s implementation of IBR, but rather is shared by those states attempting to implement NIBRS-compatible systems. These issues have been documented at the national level by the SEARCH Group in its report on NIBRS implementation for the Bureau of Justice Statistics.¹⁰

Another issue related to incident-based crime reporting systems as a source of domestic violence and sexual assault data is variability in reporting practices among reporting agencies. The survey of Iowa’s law enforcement agencies showed differences in the criteria used by agencies to report a domestic violence incident to the system, and variability in which individual was primarily responsible for making this determination. This inconsistency of reporting seems to be one of the factors that led some data users to report a lack of confidence in the accuracy and completeness of the domestic violence data in the system.

This variability in reporting can be addressed to some degree through law enforcement training which specifically addresses domestic violence and sexual assault reporting. Standardized definitions and criteria for identifying incidents as domestic violence, along with guidance regarding who at the local level should make this determination, would result in more uniform reporting across the state.

Despite these issues, IBR data appear to be a useful source for information on statewide domestic violence and sexual assault. Results of interviews and surveys of data users and data and service providers indicate that the IBR data are being widely used for a variety of purposes. State analysts, local law enforcement agencies, and service providers all report using the data generated by the IBR system.

_____________________

7 As is the case with states that have converted to NIBRS, incident-based reporting now forms the basis of Iowa's Uniform Crime Reporting (UCR) system.  However, since the term "Uniform Crime Reporting system" is used in this report to refer to the ummary-based data system used throughout the country, the term "incident-based reporting system" will be used to refer to Iowa's systems.

8 Departments were asked the same questions regarding sexual assault data; the responses were similar to those reported for domestic violence data.

9 It is possible to mitigate these effects to some extent through various analytical strategies, the most straightforward of which is to present the incident data in the form of rates, using the populations of the jurisdictions covered by only those agencies reporting to the data system.

10 Bureau of Justice Statistics, Implementing the National Incident-Based Reporting System: A Project Status Report, July 1997.

CONNECTICUT'S SPECIALIZED DATA COLLECTION SYSTEM

History of the Family Violence Reporting Program

In response to a number of federal and state reports and initiatives, as well as the urging of victim advocates, the Connecticut Office of Policy and Management contracted with St. Joseph's College in 1984 to conduct a study of family violence. Family Violence in Connecticut: A Preliminary Study of Official Reporting, Under Reporting and Incidence Rates, 1979-1984 was published in February 1985. The study included spouse and partner abuse, child abuse, and elder abuse in its definition of family violence (Connecticut continues to include this broad definition of family violence in its statutes, rather than the more specific spousal or partner violence). The most significant finding of the study was extreme underreporting, particularly for spousal or partner abuse. The primary recommendation was the appointment of a task force to review data and develop more effective programs, legislation, and policies.

Governor William A. O'Neill appointed the Governor's Task Force on Family Violence in September 1985. Members represented private industry, the medical and mental health communities, the academic community, law enforcement, the courts, the Connecticut State Labor Council, AFL-CIO, victim advocates, and the state agencies dealing with domestic violence, child abuse, and elder abuse. The task force's Final Report and Recommendations was published in January 1986, and included recommendations for landmark legislation which would change the way law enforcement and the judiciary handle family violence cases, increase services for family violence victims, and increase public awareness of family violence as a crime. With regard to reporting issues, the task force found that underreporting of family violence incidents was a serious obstacle in dealing with abuse in Connecticut.

Public Act Number 86-337, An Act Concerning Family Violence Prevention and Response, was passed during the 1986 legislative session and became law effective October 1, 1986. The legislation was comprehensive. Besides creating the reporting program, it required mandatory, uniform arrest policies, next day court arraignment, a family violence intervention unit within the courts, an education program for offenders, a criminal protective order and training for judges, prosecutors, and police. The reporting requirements of the legislation included the following components:

• Completion of a family violence offense report was required for each family violence incident regardless of whether or not an arrest occurred;

• All family violence incidents which resulted in an arrest were to be reported to the Department of Public Safety (DPS), which would compile and report statistics regarding family violence crimes;

• DPS was responsible for developing a reporting form for family violence offenses to include name, relationship, gender, age, time and date of incident, whether children were involved or whether the alleged act of family violence was committed in the presence of children, type and extent of the alleged abuse, existence of substance abuse, number and types of weapons involved, existence of any prior court orders, and any other data that they decided would be necessary for a complete analysis of all circumstances leading to the arrest.

The Family Violence Offense Report

The Connecticut Department of Public Safety, Division of State Police, Crimes Analysis Unit created the Family Violence Offense Report, SPC-230-C in July 1986 to capture the legislatively mandated law enforcement data. The forms (in triplicate) are printed and distributed to all law enforcement agencies in the state. Each law enforcement officer in the state must fill out a reporting form (SP-230-C) when making a family violence arrest. The arrest must meet the definition of family member and be for an offense which involves violence or the immediate threat of violence. One copy of the reporting form is sent to the Crimes Analysis Unit (the same unit which handles the Uniform Crime Reporting Program), and one copy is sent to the appropriate prosecutor.

The form collects both case identification and incident information. Case identification information includes:

• name of the local police department;
• department case number for the arrest incident;
• arresting officer's name, rank and badge number;
• date of the report.

Data elements which identify or describe the incident or event include:

• town in which the offense took place;
• date and time of the offense;
• offense code for the most serious offense committed;
• number and type of weapons involved;
• seriousness of injury;
• whether or not alcohol or drugs were involved;
• whether or not there was a prior court order;
• whether children were present or involved.

For each participant (victim, offender, or both), the name, sex, date of birth and the relationship to the victim is recorded. For the offense code, the officer must choose among homicide, assault, kidnapping, sexual assault, criminal mischief, risk of injury to a minor, breach of peace, disorderly conduct, or other. The four choices for weapon codes are 1) gun, 2) knife, 3) other dangerous, and 4) hands, fists, etc. For injury codes, the choices are serious, minor, or non-physical. The five relationship codes are: 1) spouse, 2) former spouse, 3) other family member (relative residing in home), 4) other relative (not residing in home) and 5) live-in or companion (living together, having lived together, never lived together but had a child in common). Involvement of alcohol or drugs and existence of a prior court order are yes/no choices, based upon the officer's knowledge.

Connecticut's law enforcement agencies are currently in the process of converting from summary-based crime reporting to incident-based reporting (NIBRS). Approximately 30 of Connecticut's 99 law enforcement agencies are currently collecting NIBRS data. The data components of the Family Violence Reporting Program are being incorporated into the NIBRS reporting program. Connecticut will continue to collect data using the current reporting program until NIBRS becomes operational statewide.

Processing of Data

The Family Violence Offense Reports arrive at the Crimes Analysis Unit by mail. They go through a series of reviews and checks to insure the accuracy of the data before they are input into the database. The first review involves checking to see if the arrest box is checked "yes" or "no." If no arrest was made, or the arrest is pending, the original report is sent back to the records department with a note asking to resubmit the same form after the arrest has been made. If an error or omission is found on the form, a copy is returned to the department for correction. A face sheet with a checklist of the most common errors has been developed for simplicity in requesting corrected data. Some of the more common errors are omission of birth date or relationship code, or recording risk of injury to a minor as an offense code, but not having a minor listed as a victim. Originals of incomplete or incorrect forms are held until the corrected copy is resubmitted by the appropriate department. If no correction has been submitted by the end of the year, a second request for the information is sent out.

A number of edits are done at the end of the year, particularly with homicides. Staff compare homicides listed in the Family Violence database with Uniform Crime Reports. Generally, there are more homicides reported under the UCR program because the Family Violence Reporting Program only produces an incident report when an arrest is made. Other edits performed include checking for accuracy when kidnapping or sexual assaults result in dual arrest, and the inclusion of children as victims if risk of injury to a minor is the offense.

Data from the Family Violence Reporting Program are published in special section of the annual UCR Report, Crime in Connecticut. The Family Violence Data is also distributed separately as an excerpt of the annual UCR Report. The excerpt is distributed in-house to state police commanding officers, the Commissioner, and all heads of units. It goes to all police chiefs, and to others upon request. Many of the recipients are employed by state agencies such as the Department of Mental Health and Addiction Services, the Department of Children and Families, and the Permanent Commission on the Status of Women.

The Family Violence Reporting Unit receives requests for data primarily from individuals seeking grant funding for various local programs, from the Judicial Branch's Office of Victims Services, from victim advocates, from family violence shelter staff and other service providers, from state police troops and other police departments, and from reporters, students and the general public.

Survey of Law Enforcement Reporting Procedures

After consulting with the Family Violence Reporting Program Staff, a list of survey questions was developed which would help to clarify the reporting process that law enforcement agencies use when submitting forms to the Department of Public Safety, Crimes Analysis Unit. A total of 14 police departments were surveyed, including the seven largest cities in the state, as well as seven smaller departments distributed geographically around the state. The localities represented 29% of the total state population, but accounted for 49% of family violence arrest incidents.

The responsibility for forwarding the SP230-C forms to the Department of Public Safety falls primarily upon the records clerks or records department staff. The two largest departments however, used crime analysis staff or had each individual officer submit the forms. In two of the departments, the submission responsibility fell upon specialty divisions such as the Domestic Violence Officer in the Youth Division or Victims Services Office.

One of the concerns of the crimes analysis staff was that not all forms would be submitted if a significant period of time elapsed between receipt of an arrest warrant and the actual arrest of the offender. Departments were questioned regarding their procedures in such instances. Most of the departments had a tracking method in place for arrests made by warrant. The methods varied considerably, the most common being that a form is filled out and held with the police report until the warrant is served. One of the largest departments had no formal method for tracking whether or not the forms were being filled out after arrest by warrant. No department reported a time limit for submitting older reports if an outstanding warrant was executed, as long as the warrant was still active. However, one large department indicated that the warrants were not identified as family violence cases and could not guarantee that a family violence reporting form would be submitted upon arrest of the offender.

Departments were asked to describe their internal review process for the reporting forms. All departments indicated that forms are routinely checked by the shift supervisor or shift commanders before being reviewed by the Records Supervisor prior to submission to the Department of Public Safety. They felt that their review processes were adequate to determine the validity of a form resulting in a "no arrest" response. The officers who complete the forms are generally responsible for making their own corrections. In some departments, however, the records staff made minor corrections.

Departments were asked about the length of time they kept reports for which no arrest was made. Responses to this question varied, with departments keeping reports which do not result in an arrest for 5 years, 7 to 10 years, 10 years, or indefinitely. One department microfilmed all reports older than 2 years, but kept the microfilmed reports indefinitely.

Training of Police Officers with Regard to Reporting Issues

The Police Officer Standards and Training Council (POST) is responsible for all aspects of the training of police officers in Connecticut, including basic and in-service training, certification and recertification. POST sets entry-level educational requirements, develops training curricula, and accredits training programs run by the larger police departments. By statute, police officers at all levels from patrol through chief are mandated to receive two hours of training on family violence every three years as part of their recertification process.

The family violence curriculum, Police Response to Crimes of Violence-A Training Manual for Connecticut Law Enforcement, was updated in 1997 by POST and the Connecticut Coalition Against Domestic Violence (CCADV). CCADV employed outside contractors to develop the curriculum manual, a training video to accompany it, and a set of updated model policies for police response to domestic violence. There have been no changes to the police reporting requirements since the first curriculum was developed in 1991. Therefore, there is no specific section devoted to reporting to the Family Violence Reporting Program. Instead, the actual reporting guidelines, along with a copy of a completed sample form, are included in the Appendix to the curriculum guide. The guidelines include definitions of family violence, and instructions for each block on the reporting form.

According to a POST staff member, accurate reporting depends upon each individual officer making a correct determination of who is subject to the law, what constitutes a family violence crime and probable cause for an arrest, and how to handle dual complaints. Therefore, the training focuses on issues such as these, rather than the technicalities of filling out the forms.

Uses of the Data

The information obtained through the Family Violence Reporting Program has become a valuable asset to all three branches of government in assessing and analyzing the problem of family violence in Connecticut. The data have been used for preparing training materials for training prosecutors, judicial personnel, probation officers, police officers, victim advocates, and other service providers. Data from the program have also been used to support statutory and regulatory changes and to support budget options by various criminal justice agencies. In addition, the data have been used to support research, evaluation, and policy decisions. Some of the specific data uses are discussed below.

Research and Evaluation

The Family Violence Reporting Program has supplied the necessary data for various research and evaluation projects:

Family Violence Cases in Connecticut - The Decision to Nol-Pros

This was a research project undertaken by the Research and Evaluation Subcommittee of the Inter-agency Family Violence Response Coordinating Committee, established to oversee the implementation of the new family violence legislation. A sample of 2,000 persons arrested from August through October 1987 was selected from the family violence arrest database. Some of the recommendations made by the study which were adopted include developing a wider array of alternative sanctions for family violence offenders, and developing a separate computer code for family violence cases in judicial disposition records to make data available for administrative and research purposes.

An Evaluation of Connecticut's Family Violence Education Program

Arrest data from the Family Violence Reporting Program were one of four sources of data used in the evaluation of Connecticut's Family Violence Education conducted by the University of Massachusetts in 1990. The study found that the rearrest rate for those who successfully completed the program was significantly lower than the rate for those who did not complete the program. The recommendations made by the study to develop gender-specific, bilingual, bicultural classes in locations where the need exits and to develop uniform program goals and completion criteria were adopted by the Judicial Branch and funded by the Connecticut General Assembly.

Study of Family Violence Incidents Which Result in the Arrest of Both Parties

This research was done by the Connecticut Department of Public Safety, Research and Planning Section, in response to unacceptable dual arrest rates of 18-20 percent recorded for 1987-1989. The study provides general information on dual arrests in family violence incidents in Connecticut from 1987-1989. Data were used from the Family Violence Reporting Program, police incident reports, and a survey of police attitudes toward dual arrest. As a result of this study, the dual arrest problem in Connecticut was handled through additional police training rather than through statutory changes attempting to establish primary or secondary aggressors.

Evaluation of the Court's Screening Tool for Family Violence Cases, the K-SID

The Kingston Screening Instrument for Domestic Violence Offenders (K-SID) is a screening instrument for probation officers, judges, family relations counselors, family violence victim advocates, and other trained court personnel to use with alleged or convicted domestic violence offenders for planning services and case disposition. It is designed to help the court determine the dangerousness level and risk of recidivism of the offender. The data from the Family Violence Reporting Program has been used to evaluate the effectiveness of the K-SID instrument for predicting future family violence offenses.

Judicial Branch, Family Division

Family Division Managers have relied upon the data provided by the Family Violence Reporting Program to assist them in designing and modifying programs which address victim or offender needs as they are processed through the court. The Family Division has used the data to identify special needs populations for the Family Violence Education Program, a diversionary program for domestic violence offenders. Spanish-speaking programs, evening programs, and programs for female offenders have been added in locations where they were needed. Information on dual arrests has helped the Family Division to develop education programs for female offenders. Finally, the Division used the Family Violence Reporting Program data in designing and locating its Children Impacted by Family Violence Programs throughout the state.

SAC Use of the Data

The Statistical Analysis Center (SAC) has been an important user of the Family Violence Reporting Program's data. While the Department of Public Safety publishes and distributes the numbers and basic analysis for each year, the SAC takes the analysis further and looks at trends over time in as much detail as the published data will permit. The SAC has prepared and published Connecticut's Family Violence Reporting Program - Summary of Incidents Involving Arrest 1987-1994, 1987-1995, 1987-1996, 1987-1997. This report is distributed to those known within the state to be involved with family violence issues and upon request.

Based within the state planning and budget agency, the SAC is constantly called upon to provide data to support the Governor's budget process and to supply any data necessary for supporting and managing all of Connecticut's federal and state criminal and juvenile justice grant programs. The SAC has also responded to requests for family violence data, including the Family Violence Reporting Program data, from a broad group of other state agencies, victim advocacy groups, the bar association, the media, the Governor's office and legislature, and local officials.

The Violence Against Women Grant Activities

During the first year of funding for the Violence Against Women Grant Program (1995), the SAC Director was the grant coordinator, coordinating the planning committee, preparing the state plan and managing the grants for that year. The data was used extensively to prepare the plan and make funding decisions by the Violence Against Women Act (VAWA) Committee.

Discussion and Conclusions

One of the main advantages of Connecticut’s Family Reporting Program is its ability to provide consistent data on family violence over a long period of time. Since the program began 12 years ago, and since the reporting procedures have not substantially changed during that period, researchers and policymakers have access to information on long-term trends in domestic violence in their state.

One of the advantages of specialized data collection systems is their ability to collect more detailed information on domestic violence than can be collected under more general crime reporting systems. Connecticut’s reporting system might be considered a "first generation" system, given when it was initiated. As a result, the reporting form used provides less detailed information than some that have developed in other states more recently. Potentially useful information, such as more detailed relationship and offense codes, are not available with this system.

One of the drawbacks of the Connecticut program is the unavailability at the state level of information on family violence cases which did not result in an arrest. Since the decision was made for police departments to retain the data forms for incidents in which no arrest was made, there is no central data collection point for this information. Thus, incidence data are not available for reported cases of domestic violence, but only for family violence arrests.

The many examples of research and analysis projects which have utilized the Connecticut Family Reporting Program data attest to the value of the program and the usefulness of the data for state policymaking. The fact that the state is moving to incorporate its family violence data into the NIBRS system being developed illustrates an important caveat regarding the future of specialized systems for collecting domestic violence and sexual assault data. As more and more states adopt NIBRS for general crime reporting, they may find that they have little reason to maintain a separate incident-based reporting program just for domestic violence or sexual assault offenses. More likely, states will follow the example of Connecticut (and Iowa, when they implemented NIBRS) and incorporate their domestic violence data collection into their NIBRS systems.

ILLINOIS' SERVICE PROVIDER DATA COLLECTION SYSTEM

History of Domestic Violence and Sexual Assault Data Collection in Illinois

There are three primary sources for domestic violence and sexual assault information in Illinois: the Illinois State Police, the Illinois Coalition Against Domestic Violence (ICADV) and the Illinois Coalition Against Sexual Assault (ICASA). The Illinois State Police maintains four datasets that archive information on the offenders, victims and incidents of domestic violence and sexual assault: the Orders of Protection data and the Criminal History Records contain offender data, and the Uniform Crime Reporting (UCR) and the Domestic Violence datasets keep information primarily on incidents of domestic violence and sexual assault.

In addition to these sources, several agencies that provide funding for services to victims also collect summary information on the services to victims of domestic violence and sexual assault. The largest of these funding agencies are the Illinois Department of Human Services, the Illinois Attorney General’s Office, and the Illinois Criminal Justice Information Authority.

The Illinois Coalition Against Sexual Assault and the Illinois Coalition Against Domestic Violence have developed their data collection efforts under similar circumstances. Both coalitions have established funding relationships with the state to support their services to victims in Illinois. Each coalition has recognized from inception the importance of data collection in relation to both funding and legislation.

Sources of Data

Illinois Coalition Against Sexual Assault

In 1982, ICASA began reporting summary information to its funding agency, the Illinois Department of Public Health (IDPH). In order to submit these reports, ICASA had to collect information from each of its member service programs and calculate the total number of clients served as well as the hours and type of services provided. In 1987, the Victims of Crime Act (VOCA) provided ICASA with additional funding to collect information on the arrests and prosecution of sexual assault offenders. The collection of this data began in 1988. With additional legislation and funding available for victim services, ICASA began to receive increased funding from a variety of sources. Each funding agency requested a different combination of information to describe both the program services and the victims served. As a result, ICASA’s data collection efforts rapidly expanded to gather the assorted pieces of information and monitor the increased spending of each service program. The Coalition’s programs were now submitting several forms to the ICASA office quarterly and annually. ICASA, in turn, struggled to respond to the changing and time-consuming demands for data in order to guarantee funding from a growing number of sources.

Illinois Coalition Against Domestic Violence

In 1978, ICADV was formed with the understanding that both funding and legislation for domestic violence cases depended heavily upon the evidence provided from data. The Coalition began its data collection efforts by designing a six-page intake form to gather information on the abuse, the abuser, the history of abuse, and the history of the abuser. For several years ICADV published an annual brochure with the analysis of this information. In addition to the adult intake form, data were also collected on an intake form for the victim’s children, an evaluation of services form, and a departure form. All of these paper forms were submitted monthly to the Coalition’s office and entered into a database.

In 1989, the Coalition used funds from the marriage license and divorce fees to build a computer network and a uniform data collection system. Each ICADV program received a computer and began entering data at the program site. During this development process, ICADV dropped much of the information collected on the six-page intake form, leaving only two pages of descriptive information to be collected on each client. Through the new system, ICADV downloaded all the data once a month through a network. A few years after this system was in place, the state reported that financial support was no longer available for the network. Fortunately, ICADV was able to keep most of the computers and the uniquely developed program and not return to paper reporting.

Illinois Criminal Justice Information Authority

In 1985, the Victims of Crime Act (VOCA) made federal funds available for victim service programs. From that time to the present, the Illinois Criminal Justice Information Authority (ICJIA) has been monitoring the distribution and spending of these funds. As a funder of victim services, ICJIA accepted the responsibility of both collecting and using victim data to improve the services to victims. To do so effectively, ICJIA needed reliable sources for information on the victims of crime in Illinois. ICJIA was able to access the UCR, the National Crime Victimization Survey, and the Department of Children and Family Services as sources of victim information, but each of these sources suffered from various limitations.

ICJIA funds several service providers who have recently grown into agencies with a variety of funding sources. As a result, the service providers must be accountable for increasing budgets which provide services to a growing number of clients. In an effort to support and effectively coordinate data collection by domestic violence and sexual assault service provider agencies, ICJIA has worked closely with them to develop one automated information collection tool - the InfoNet - that can meet all of the agencies’ needs for accurate and timely data.

The InfoNet System

Overview of the System

In 1996, the Illinois Criminal Justice Information Authority wrote a proposal for the InfoNet, a new system to collect victim data statewide. The InfoNet is a tool to automate the required victim and service information that is reported by ICJIA-funded service agencies. This tool also allows each service provider to easily collect a variety of case level information, including the victim’s circumstances, the court proceedings, and the services provided to the victim, and to create reports for other funding agencies. All of the information recorded by the service providers is kept confidential using a unique identifier for each client.

The InfoNet was designed to work as a network system with all of the information linked and stored in a central location. In early 1997, ICJIA staff met with ICADV and ICASA to present a prototype of the new system. After this initial meeting, ICJIA recognized and responded to the different needs of the two coalitions. This meant working individually with each coalition to create data entry screens which were customized to the differences in victims and services. Pilot testing continued throughout 1997. Data entry using the InfoNet software began at ICADV’s local service agencies in October 1997 and ICASA’s agencies in July 1998. Currently the local service providers submit all of their funding reports via e-mail to coalition offices. In the year 2000, the InfoNet should be running as a network system. Once linked by an intranet system, both coalitions’ total service calculations can be managed entirely within the InfoNet.

Both ICADV and ICASA are interested and excited about the capabilities of InfoNet data. The InfoNet will allow the coalitions to answer questions about the amount and nature of victim services provided by their member agencies. Additionally, the coalitions will be capable of investigating the effects that legislation has on the services desired by victims and how programs can most efficiently recognize and respond to these needs.

The victim data collected throughout the state will eventually be stored on the InfoNet network system and only used with the expressed consent of the reporting agencies. The development of the InfoNet has become a significant investment for ICJIA, ICADV, ICASA, the Illinois Department of Human Services, the Illinois Attorney General’s Office, and the individual service providers throughout Illinois. These principal players are combining their efforts and philosophies to provide improved services to victims of crime by increasing the quality of available data. As the funding dollars continue to grow, the funders and legislators have asked difficult questions regarding the impact of money spent on victim services. The future collection of domestic violence and sexual assault data in Illinois will utilize tools such as the InfoNet to answer these questions while providing improved service to victims.

Description of the System

The InfoNet database was designed to link a program’s entire structure in order to both record and calculate a variety of administrative and service information. The database is made up of relational tables that are linked by primary keys that run throughout the database. The tables include information on victims who are currently receiving services and those who have been previously receiving services, staff, volunteers, and financial information. The primary keys that link the information contained in the tables consist of separate identification numbers for the program, the staff member or volunteer, and the client.

Throughout the InfoNet software, one record is defined as all of the information relating to one client. The descriptive information for each client is contained in fields or columns of the table and varies with each record. In the client-level table, fields contain that client’s: status (new or returning), demographics, significant others, alleged offender (relationship to the victim and offender demographics), medical advocacy (treatment, serious injury, hours of advocacy), police involvement (advocacy hours, arrests, charges, order of protection), prosecution involvement (advocacy hours, charges, trial, verdict), counseling by reporting agency (in person, telephone, group, family) and other advocacy or support services provided. Client data can be stored in an active or archived table depending on the client’s status.

The tables with program-level data include staff, financial and service information. The program service data includes: institutional advocacy (contacts and number of hours), professional training (number of participants, hours of preparation, and hours of training), public education (number of participants and hours of preparation, training, and travel), information and referral (number of contacts and hours), media contacts, and a variety of administrative information, including lists of staff, volunteers, and board members.

The financial table includes data on each program activity’s and staff member’s source of funding. For example, one employee may be funded 60 percent by VOCA funds and 40 percent by private donations. The agency activities are recorded in the same way; for example, a staff training may be documented and paid for using VAWA funds and IDHS funds. Again, this table contains key identifiers that link the services with an employee and a funding source.

While both ICADV and ICASA have all of the above-listed InfoNet components, the actual data screens and fields of information vary to most appropriately fit the needs of these separate service agencies.

Data Entry

Data entry for the InfoNet system is completed at the reporting agency’s site. Staff from ICJIA created the manual to guide the agencies as they set up the software and security systems of the InfoNet. ICJIA staff also held user group meetings to train and pilot the InfoNet system. Several training sessions were held about one month before ICJIA’s release of the InfoNet software to ICADV and ICASA. This allowed time for the administrative arm of each agency to set up the staff and financial information as well as install the password-protected security system before beginning to enter client data. ICJIA was available for technical support by telephone and, in some cases, in person. ICJIA supported the setup and will continue to support the utilization of the InfoNet software.

Throughout the pilot and training process ICJIA compiled the opinions and reactions of users. The results from individual agencies have been overwhelmingly positive despite the difficulties of learning this new and complex automated system of data collection. Both ICADV and ICASA have reported hearing of the local agencies’ frustrations as users struggle with the new system and their reporting needs, but these difficulties were expected and seem reasonable in light of the complexity of the system.

Data Submission and Analysis

The data collected using the InfoNet system is currently being submitted to both ICADV and ICASA from their member agencies. The InfoNet is not yet on a network or intranet system and, as a result, routine reports must be sent from local agencies via e-mail to the coalitions. The coalitions are continuing to perform calculations for various quarterly and annual reports. This data has continued to be submitted by the coalitions to their respective funding agencies with relatively few delays.

To date, none of the data collected using the InfoNet software have been released to ICJIA for analysis. The primary reason for this is that both the users and ICJIA are working without a network system and continuing to adapt the InfoNet software to the needs of the coalitions. In the meantime, the process of data collection and report calculations is still relatively time-consuming. ICJIA has been working primarily to relieve the coalitions of these complications.

For all involved, the focus has been on the need for InfoNet users to be able to provide information for funding sources.

Advantages of the InfoNet System

Several complications of current domestic and sexual violence data collection will be resolved through the InfoNet software. Past data collection efforts of victim information were based on the requirements of the funders. As a result, data entry was often redundant, included complex calculations, and caused confusion over definitions of terms. A new client, for example, could be a client with no previous services or a returning client who appears for the first time in a particular reporting period. These problems have led to poor data quality and ultimately resulted in very limited information describing victims of domestic violence and sexual assault.

The InfoNet system eliminates redundant data entry in a number of ways. The system compiles all of the information on one client under a unique identifier, continuing to add to that information over time and essentially creating a history of service. In the past, the service agencies reentered intake information on clients who reappeared in each new reporting period. In addition, the InfoNet automates calculations required for reports. Once users enter the dates for a reporting period, the system will calculate the specified information on the types of services provided, the number of clients served (distinguishing new from ongoing), the employees who provided the service and the funding source for the service.

An additional benefit of the InfoNet reporting system is the ability the agencies have to archive data beyond the short reporting periods defined by funders. These client histories will be a tool for service providers to investigate how additional funding has allowed them to service victims most effectively.

One of the key elements of the InfoNet software is its emphasis on ease of data entry. Pop-up and drop-down lists were created for any questions with a specified list of responses. When feasible, these lists were linked to specific administrative information for the agency. For example, when responding to the question, "Which staff member provided this service?" the data entry person sees a list of all the staff members for that particular agency. This list is automatically created by linking to the administrative information that has been previously entered into another section of the database. These database links make recording client or service information easy and fast. The InfoNet also automatically assigns date variables wherever possible, thereby helping to reduce errors in the calculations of client services over particular reporting periods. Automatic calculations were included whenever possible to assist local agencies in tracking their active clients and remaining budgets. The InfoNet software also distinguishes between fields that are stable, such as intake information, and fields that will be updated, such as service and court proceedings information. Screens for these latter fields are enabled immediately when client identifiers are entered, to remind the data entry person to update the appropriate information.

All of the required reporting by local agencies has been organized into InfoNet report files that will automatically calculate the specified information from a local database. The InfoNet report files were created with input from both the funders and the service providers. Thus the InfoNet system is a practical tool for service providers, helping them to collect and report necessary information quickly and more accurately. In addition, the agency will also have the ability to query their data using variables that they specify; in this way they can use all of the information collected for internal purposes.

As the InfoNet develops the ICJIA hopes to work closely with the individual agencies in order to create simple and useful reports that are unique to each service program’s goals. The programs can then use this information to support or restructure particular projects, which may be entirely separate from the statewide data initiatives or funding requirements.

The InfoNet continues to be developed under the guidance of an advisory committee. The committee members include representatives of both the service providers and the funding agencies. Members of the advisory board, in particular representatives from the Department on Human Services and the Illinois Attorney General’s office, have played a valuable role recently by providing extensions for quarterly data reports as the local users adjust to the latest revisions of the InfoNet system.

Development Issues

A number of issues were raised in the development of the InfoNet that needed to be addressed in order for it to be successfully implemented. Several of these issues are discussed below.

Confidentiality

Confidentiality was a serious concern of both the participating coalitions and ICJIA when considering a system to collect and store information that describes victims. A victim’s safety can depend upon the confidentiality of support services. The InfoNet data collection system has been designed with four distinct layers of security. First, the intranet design links only specified users to the data system using technology similar to that of the internet. The second level of security uses encryption software and a password to secure each user’s link to the intranet. Each agency will also have internet access to obtain or publicize general information regarding services to victims. The third and fourth layers of security are at the local agency level. Each local agency will use encryption software and a password to protect the agency’s local database. Passwords will be used to separate data entry and administrative information for each agency. The result is that only designated local staff can access and change data entry and /or administrative information. The agency will also use unique identifiers in place of the names of clients whose information is entered into the system. Within the local agency these final two layers of security will protect both the administrative information and victim data contained in the database.

Identifying Individual Clients

Service provider agencies require a way to track services to clients, and to distinguish between services provided to the same client over time, as opposed to those provided to new clients. A unique case identifier code serves to protect the victim while allowing individual case information to be accurately documented for agency reporting. In addition, with the increase in funding sources, it is now important to know how much staff time was spent providing various services. An individualized client identifier limits the complications of matching services provided to a particular client by a particular staff member.

ICJIA has offered recommendations on creating and storing unique identifiers but ultimately each reporting agency is responsible for this information. Unique identifiers will not be designed by or known to staff at ICJIA; instead, each local agency will develop a method for creating a unique identifier for every victim served. The records containing a name and matching identifier will only be stored at local agencies and will be kept separate from the InfoNet system.

Hardware and Software

ICJIA’s original proposal for the InfoNet included the financing and distribution of both the equipment and software to users. ICJIA staff contributed technical expertise in selecting the appropriate equipment, developing the system, and eventually training the new users on the InfoNet. The hardware, which has been provided to each reporting agency, includes capabilities for word processing, database applications, automated presentations, and electronic mail.

The InfoNet was designed and equipment was purchased using the most recent technological information available. Unfortunately, there have been unscheduled delays in the acquisition of equipment necessary for the network system. ICJIA has provided the InfoNet software to both the domestic violence and sexual assault coalitions, so that they could begin using the data entry system. Until the network becomes operational, the data reports are being run by the reporting agencies and submitted via e-mail to coalition offices.

Ongoing Issues

Several issues continue to be the subject of discussions among those involved in the development and use of the InfoNet. One such issue is how to define new, returning, ongoing and archived clients. This will continue to be an issue for training as local agencies translate those definitions into their data entry routines. A related topic for discussion is the question of how long to keep a client in the database and ultimately the network system. In the past, the coalitions used fiscal years to measure client histories. The InfoNet offers the potential to expand outcome measures over multiple years. This issue will become more relevant once the InfoNet is a network system, housed and supported by ICJIA.

The timing of data entry and reporting will be topics for discussion as users become more comfortable with the InfoNet system and more interested in data quality issues. The data entry by local agencies should be either continuous or at regular intervals, such as weekly. Similarly, reports should be run routinely to familiarize both the reporting agencies and the coalitions with trends in the available information.

Future of the InfoNet

The increase in VOCA and VAWA funding over the last few years has made the InfoNet project possible through the Illinois Criminal Justice Information Authority. ICJIA plans to continue providing the technical support and collaborative efforts that are necessary to sustain this statewide collection of victim information. ICJIA recognizes that the InfoNet program will need constant updates to assure that it is useful to the local service providers. As funding sources and requirements change, the local agencies will rely on the InfoNet to meet their needs for information. The technical equipment for both the network and the users will also need continual support and regular updates. The advisory board and ICJIA are planning to expand this network system to include information collected by the victim-witness programs as well as providers of services to child victims. The funding for this work will be continually reevaluated so that ICJIA can continue to provide the necessary support to this resource for information on victims.

RECOMMENDATIONS

Based on the assessment of current state efforts and the case studies of three different state systems, the following are recommendations for the states with regard to domestic violence and sexual assault data collections systems:

1. States should implement incident-based reporting systems which use offense and relationship codes that are compatible with the National Incident-Based Crime Reporting System (NIBRS).

Most states are in the process of planning or implementing NIBRS or NIBRS-compatible incident-based crime reporting systems. NIBRS provides enough information about offenses, relationships, victims, and offenders to allow states to conduct detailed analyses of domestic violence and sexual assault issues. NIBRS also represents the best opportunity for the development of national estimates of the incidence of domestic violence and sexual assault, to the extent that these offenses are reported to the police.

Connecticut’s plans to develop a NIBRS-compatible system and then eliminate its long-standing specialized domestic violence reporting system may be indicative of future developments in the states. Given the movement toward NIBRS and its ability to provide domestic violence and sexual assault information, it seems unwise for states to expend the resources to implement or maintain specialized domestic or sexual violence incident-based systems.

NIBRS implementation in the states has been problematic for a number of reasons, some of which were documented in Iowa’s study of its NIBRS implementation. Local law enforcement agencies, especially large ones, have maintained that they do not have the personnel and other resources to implement the system. As a result, few states have comprehensive reporting, and currently only one large police department is represented in the national data. States need to take heed of Iowa’s experience in converting to a statewide incident-based system when all local departments are not ready to report. This results in the state losing valuable information about the number of domestic violence and sexual assault incidents. States which already collect domestic violence data through another method should consider continuing to collect this data from agencies not yet ready to report incident-based data, so they can continue to accurately track the total number of incidents.

States should also consider adding fields to their incident-based reporting forms which would allow them to identify domestic violence cases as defined by their state statutes. This is important for several reasons. First, NIBRS relationship codes do not include all of the possible relationships relevant to domestic violence situations, so that if cases are identified based on relationship codes, some will be missed. In addition, requiring each case to be identified as domestic violence by the police officer who responded to the call may result in more accurate classification of cases.

2. States should move toward implementing incident-based service provider domestic violence and sexual assault data collection systems.

There are several good reasons for states to implement service provider reporting systems. Such systems will provide useful information for the programs, advocacy groups, state decisionmakers, and funding agencies regarding the types of services being provided. In terms of incidence data, service provider systems can complement the data collected by crime reporting systems. While the latter provide information on incidents reported to the police, the former capture information on incidents for which victims seek assistance. States which have both systems in place should be able to develop more accurate assessments of the incidence and nature of domestic violence and sexual assault.

Incident-based reporting is recommended here because it produces much richer data for analytical purposes. Summary-based systems are of little analytical use, and are even of limited use in providing incidence data. Service provider data collection systems must have methods built in for distinguishing between new and returning clients, primary and secondary victims, and the nature of the services received. Not only is this information essential for program planning, it is also necessary to maintain accurate data on the extent and nature of the victimization.

Illinois’ InfoNet is an excellent example of the type of incident-based service provider system being recommended. By working closely with the programs and advocacy agencies, the Illinois Criminal Justice Information Authority has developed an automated system that will meet the needs of the programs, the funding agencies, and researchers and analysts. As shown in Table 1, several other states have developed, or are in the process of developing, similar systems. The federal government is playing an important role in the development of these systems by providing funding and technical assistance through programs such as the Violence Against Women Act, the STOP grant program, and the Grants to Encourage Arrest Policies initiative.

3. States should develop guidance and implement training on how to identify and report cases of domestic violence and sexual assault.

One of the issues which became apparent in the Iowa study of its NIBRS system is the degree to which local law enforcement agencies vary in the criteria they use to report domestic violence cases to the database. States should develop and disseminate clear policies regarding how incidents are to be classified, and should regularly provide training to law enforcement officers on how to identify and report such cases to the data system. Legislation which clarifies the definition of domestic violence may also be helpful in some cases.

4. States, with assistance from the federal government, should develop initiatives to analyze and validate domestic violence and sexual assault data being collected by statewide incident-based systems.

Many states are now in a position to begin to analyze and validate the domestic violence and sexual assault data being collected through incident-based reporting systems. These data collection systems could provide the basis for some interesting analyses. For example, several states which report domestic violence and sexual assault data using NIBRS offense codes also use a flag to identify domestic violence cases. It would be interesting to describe the characteristics of those cases which are flagged as domestic violence in terms of the actual offense and relationship codes involved. Similarly, several states have domestic violence and sexual assault data from multiple reporting systems available for analysis. It would be interesting to compare the incidence of these offenses as reported by the various data systems. The results of these analyses would not only be of importance to the individual states themselves, but would allow other states to gauge the validity and usefulness of these data collection systems.

5. States, with assistance from the federal government, should begin developing linkages among the various state data systems that collect information relevant to domestic violence and sexual assault incidents.

While this report has focused on law enforcement and service provider systems, there is information regarding domestic violence and sexual assault incidents available from other sources in the states. Once domestic violence and sexual assault law enforcement databases are firmly established, states should consider integrating law enforcement and service provider databases with other data collection systems, including courts, corrections, health, mental health and social services systems. Information on case dispositions and services provided in non-criminal justice settings would considerably enhance states’ ability to conduct meaningful analyses of domestic violence and sexual assault incidents, and would improve the coordination of service delivery to victims of domestic and sexual violence.

APPENDIX A. CRITERIA FOR DOMESTIC VIOLENCE FLAGS AND CASE SELECTION