CONTINUING CRIMINAL HISTORY RECORDS IMPROVEMENT EVALUATION: FINAL 1994-1998 Report ------------------------------------------ This file contains the Highlights of this report. To see the full report go to: http/www.ojp.usdoj.gov/bjs/ascii/cchrie98.txt ------------------------------------------ April 1999 NCJ 179768 U.S. Department of Justice Bureau of Justice Statistics Jan M. Chaiken, Ph.D. Director Acknowledgments. This report was prepared by Queues Enforth Development, (Q.E.D.) Inc. The project director of the report was James M. Tien, Ph.D, Vice President; other authors include Michael F. Cahn, Vice President, and Robin C. Neray, Senior Analyst. The Federal project monitor was Carol G. Kaplan, Chief, Criminal History Improvement Programs, Bureau of Justice Statistics. Report of work performed under BJS Cooperative Agreements #95- RU-RX-K002 and #98-RU-VX-K003 awarded to Q.E.D., 432 Columbia Street, Cambridge, Massachusetts 02141-1048. Contents of this document do not necessarily reflect the views or policies of the Bureau of Justice Statistics or the U.S. Department of Justice. Copyright/1999 by Queues Enforth Development, Inc. The U.S. Department of Justice authorizes any person to reproduce, publish, translate or otherwise use all or any part of the copyrighted material in this publication with the exception of those items indicating they are copyrighted or printed by any source other than Q.E.D. Foreword Since 1995, with support from the Bureau of Justice Statistics (BJS), Queues Enforth Development, Inc. (Q.E.D.) has been evaluating the impact of three federal programs on state criminal history records; the BJS-funded Criminal History Records Improvement (CHRI) Program, the BJS-funded National Criminal History Improvement Program (NCHIP), and the five- percent set-aside of the Edward Byrne Memorial State and Local Law Enforcement Assistance Program, funded by the Bureau of Justice Assistance. This report brings Q.E.D.’s evaluation findings current through 1998 and addresses the extent to which the efforts funded under the records improvement programs have brought the states closer to meeting the mandates of the federal statutes concerning the quality and uses of criminal history records. In the course of their evaluation, Q.E.D. has analyzed over 1,500 federally funded criminal history records improvement activities undertaken by the states. State responses to questions on data quality and improvement strategies, and user perceptions of record accessibility, timeliness and utility are reported herein. The evaluation findings identify areas of progress, including promising approaches for continued records improvement. BJS hopes that the information contained in this report will inform state officials about records improvement initiatives across the nation and assist them in their future planning efforts. Jan M. Chaiken, Ph.D. Director Highlights 1. Making the Case The importance of accurate and complete criminal history records to a smoothly functioning and secure society cannot be overstated. These records are critical to decision making at virtually every juncture in the criminal justice system, and beyond. Police officers, prosecutors, defenders, judges, and other court officials, corrections officers, probation officers, and parole officers depend on timely, complete, consistent, and accurate criminal history information. This information provides the glue for holding together a coordinated and effective criminal justice system. Moreover, criminal justice records are being accessed increasingly for official purposes outside the criminal justice system, including establishing qualifications for employment, volunteer programs, and professional licensing. Each state maintains criminal history records in a central repository, coordinating and providing them in response to requests from legitimate users. Repositories process hundreds of thousands of fingerprint and arrest records from local arresting agencies, identify offenders, process disposition reports, and attempt to match disposition reports to arrests in their databases. Federal, state, and local criminal justice officials have long recognized problems in the quality of criminal history records. This issue was first widely discussed in 1967, with publication of the Report of the President’s Commission on Law Enforcement and the Administration of Justice, which noted that these records were frequently inaccurate, incomplete, and inaccessible. Over the past three decades, numerous workshops have been held and reports published on the quality of such records, and strategies have been devised for improving them. Further, federal and state statutes have increased the importance of criminal history records in such areas as eligibility to buy firearms, felony convictions of illegal aliens, sentencing guidelines, employment, and licensing. Federal agencies–in particular, the former Law Enforcement Assistance Administration, the Bureau of Justice Assistance (BJA), and the Bureau of Justice Statistics (BJS)–have funded state programs designed to enhance data quality. Key efforts include the BJS-funded Criminal History Records Improvement (CHRI) program, the BJA-funded Byrne 5% set-aside program, and the National Criminal History Improvement Program (NCHIP). In 1995, BJS, in conjunction with BJA, authorized Queues Enforth Development (Q.E.D.) Inc., to continue the BJA-funded Criminal History Records Improvement Evaluation (CHRIE) effort. Q.E.D’s current project, entitled “Continuing Criminal History Records Improvement Evaluation” (C-CHRIE), assesses the CHRI program, the Byrne 5% set-aside program, and the NCHIP through 1998. Findings in this report point to areas where progress in records improvement has been substantial, as well as those requiring greater effort, and identify promising approaches for improving data quality. Justice Department officials should find the report useful in assessing how federal funds are being spent; state officials can use it to find out what’s going on in other states. Federally Funded Programs The CHRI, Byrne 5% set-aside, and NCHIP programs seek to improve the quality of criminal history records. In 1989, the US Attorney General recommended using $9 million of Anti-Drug Abuse Act Discretionary Funds in each of fiscal years 1990, 1991, and 1992 to fund the CHRI program. The three overall objectives of this program are to: * enhance state criminal history records to accurately identify convicted felons; * meet the new FBI/BJS voluntary reporting standards for identifying such individuals; and * improve quality and timeliness of criminal history records information. An amendment to the Crime Control Act of 1990 required that states spend at least five percent of their annual Edward Byrne Memorial State and Local Law Enforcement Assistance formula grant funds– ordinarily intended for initiatives to control violent and drug-related crime–on improving quality of criminal history records. This amounts to a total of approximately $156 million from fiscal years 1992-98. The objectives of the Byrne 5% program are similar to those of the CHRI program–specifically, to: * enhance completeness of criminal history records, especially including final disposition of all felony arrest offenses; * fully automate all criminal justice histories and fingerprint records; * improve frequency and quality of criminal history reports to the FBI; * improve state record systems and sharing with the Attorney General of all records described above, in order to implement the Brady Act; and * improve state record systems and sharing with the Attorney General of all records described above, in order to implement the National Child Protection Act. Three key federal statutes were also enacted–the Brady Handgun Violence Prevention Act of 1993, the National Child Protection Act of 1993, and the Violent Crime Control Act of 1994; these three have led to certain pertinent actions. * First, the Brady Act–in an effort to identify ineligible prospective firearm purchasers–requires establishing a National Instant Criminal Background Check System (NICS), to be contacted by dealers before they sell any firearm. States are to make criminal history records available to NICS through the Interstate Identification Index (III), a decentralized index-pointer system maintained by the FBI, and containing personal identifiers of offenders and “pointers” to states that maintain criminal history records on these offenders. (NICS checks can also access records maintained by the FBI.) NICS, which became operational on November 30, 1998, also includes limited data on persons other than felons who are ineligible to purchase firearms. (Of the $100 million appropriated for Brady in Fiscal Year 1995, BJS transferred $6 million to the FBI for NICS development.) * Second, the National Child Protection Act (NCPA) of 1993 requires that records of child abuse be transmitted to the FBI’s national records system and encourages states to adopt legislation requiring background checks on individuals before they assume responsibility for the care of children, the elderly, or the disabled. In the context of NCPA, background checks are restricted to prospectively disqualified care providers, but state legislation varies and may have a broader scope, including the performance of routine background checks of many categories of potential employees, volunteers, and licensees. * Third, the Violent Crime Control Act of 1994 and the Lautenberg Amendment of 1996 added the eighth and ninth firearm ineligibility categories–namely, persons who are “subject to a civil restraining order arising out of domestic or child abuse” and those convicted of “domestic violence misdemeanors.” The seven other categories of persons ineligible to purchase firearms under the Gun Control Act are: people under indictment for or convicted of a felony, fugitives from justice, unlawful drug users or addicts, mental defectives, illegal aliens, those dishonorably discharged from the military, and those who have renounced US citizenship. To implement these statutes, BJS established the National Criminal History Improvement Program (NCHIP); from its inception through fiscal year 1998, the program has awarded $206 million to fund state activities in records improvement. NCHIP implements grant provisions of these statutes and thereby improves the nation’s public safety by: * facilitating accurate and timely identification of people ineligible to purchase a firearm; * ensuring that people responsible for the care of children, the elderly, or the disabled do not have disqualifying criminal records; * improving access to protection orders and records of people wanted for stalking and domestic violence; and * enhancing the quality, completeness, and accessibility of the nation’s criminal history records systems and the extent to which such records can be used for criminal justice-related purposes. More specifically, NCHIP helps states: * expand and enhance participation in the FBI’s Interstate Identification Index (III) and the National Instant Criminal Background Check System (NICS); * meet timetables for achieving criminal history records completeness and participating in III, as established for each state by the Attorney General; * improve level of criminal history records automation, accuracy, completeness, and flagging; * develop and implement procedures for accessing records of people other than felons who are ineligible to buy firearms; * identify–through interface with the National Incident-Based Reporting System (NIBRS), as necessary–records of crimes involving use of a handgun and/or abuse of children, the elderly, or the disabled; * identify, classify, collect, and maintain–through interface with the National Crime Information Center (NCIC) and III, as necessary– protection orders, warrants, arrests, and convictions of individuals violating protection orders (to protect stalking and domestic violence victims), and support development of state sex offender registries and an interface with a national sex offender registry); and * ensure that states develop the capability to monitor and assess state progress in meeting legislative and program goals. Common Goals Many states view the various federal grant programs for improving the quality of criminal history records as one large pool of funds. This makes it difficult, if not impossible, to separate “CHRI data quality impact” from “Byrne 5% data quality impact” or from “NCHIP data quality impact.” Still, each program plays a synergistic and complementary role in improving criminal history records, and as such must be part of the overall evaluation. Fortunately, the substantial overlap among goals makes it possible to formulate one set of common improvement goals for federally supported criminal history records. These six goals make it easy to discern how well federally funded activities undertaken by the state align with program goals to: * provide resources to establish the infrastructure for improving criminal history records and related systems; * improve criminal history records quality (completeness, accuracy, consistency, timeliness, and accessibility) * improve interstate, intrastate, and federal criminal history records-related reporting; * automate systems for creating, storing, and sharing criminal history records; * identify ineligible firearms purchasers; and * identify individuals disqualified from caring for children, the elderly, or the disabled. Study Approach This report is the third major deliverable in a multi-year Q.E.D. effort to evaluate the impact of federally funded criminal history records improvement programs. The first and second deliverables, Preliminary Assessment and 1994-96 Report, built upon a 1994 Q.E.D. study evaluating the Bureau of Justice Statistics-funded Criminal History Records Improvement (CHRI) Program. To accomplish our study objectives, Q.E.D used the two-pronged evaluation approach, conducting both an overall impact evaluation of all states and a more focused evaluation of a handful of states. The overall evaluation should benefit Justice Department officials and members of Congress (who need to know how well program funds have been spent) and individual states committed to improving their current criminal history records. The focused evaluation enables a deeper analysis of selected issues. After reviewing 56 states’ and territories’ NCHIP plans and their Criminal Justice Record Improvement (CJRI) plan updates, we developed a scaleable classification scheme to categorize the 1,552 identified record improvement activities. This scheme categorizes activities that mirror the flow of data as they are captured, updated, and used throughout the criminal justice system, and identifies funding sources and expected and actual schedules, as available. The scheme is designed to accommodate diverse activities and help understand relationships among activities, funding sources, and timeframes. Based on a model we developed and then fed sample state data, we examined issues of linking arrests and dispositions, including appropriateness of national linkage goals. We concluded with the design of a set of measures which can be used to objectively assess overall data quality over time. The scheme is three-tiered: categories 1-19 constitute what we refer to as “Level 1” and are subdivided into 50 more specific Level 2 sub- categories. Level 3 is a further sub-division of Level 2, and offers the greatest specificity; it contains 171 subcategories which ultimately “house” the specific improvement activities. For example, Level 1. System Improvements consists of 1.1 Conduct study/develop plan, 1.2 Conduct audit, and 1.3 Establish infrastructure. Continuing the example, 1.2 Conduct audit, in turn, consists of 1.2.1 Audit criminal history data quality, 1.2.2 Conduct legislative audit, 1.2.3 Audit superior court. In this way, the classification scheme permits a consistent comparison of activities across states. The table below shows the number of activities in each of the 19 Level 1 categories. More than half the activities fall into the System Improvements and Criminal History Records categories; this is understandable, since they reflect the initial two stages of developing an effective criminal history records system. Interestingly, these types of activities are as prevalent under NCHIP as they were under CHRI; however, fingerprinting-related activities such as AFIS or livescan implementation (not funded by CHRI) also prevail. The number of activities undertaken by a state ranges from two to 63, with an average of 28.2 per state. The variability in number of activities indicates that some states engage in a small number of costly improvements, while others undertake less-expensive activities. Activity Categories by Prevalence While both viable and robust enough to permit an expanding C-CHRIE effort, the classification scheme is limited in several respects, as with all such schemes or taxonomies. Two limitations merit discussion. First, the scheme categorizes improvements by choosing the one category–from a hierarchical list of categories–that best represents that activity. This approach is somewhat analogous to the Uniform Crime Reporting (UCR) system, which captures only the most serious charge for each arrest. Classifying information in this way biases results towards those categories at the top of the hierarchy. Second, activities are not comparable in either cost or benefit and should not be weighted as such. While we count each activity as if all activities were equivalent, they are not; thus, an audit activity, while critical, is less costly than the purchase of an AFIS system. However, notwithstanding these typical limitations, the classification scheme and the resultant findings form a sound basis for understanding the status of criminal history records and for funding their improvements. 2. Findings To evaluate the impact of the three federally funded programs on criminal history records improvement, we considered the extent to which state efforts have helped accomplish the six common goals. While it would be ideal to assert that each goal has or has not been met, this is not yet possible. Improving criminal history records is a lengthy process, best assessed with the aid of national aggregate measures, which can quantify the state of data quality over time. Until these measures are established (see Remaining Issues, below), an evaluation must be based on activities being undertaken by the states. Goal 1: Provide Required Resources Provide resources to establish the necessary infrastructure for improving criminal history records and related systems. By providing ongoing funding since the beginning of the CHRI program, the Department of Justice has demonstrated a commitment to improving criminal history records. Between FY 90 and FY 98, the federal government awarded a total of $389M–$27M through the CHRI program, $156M through the Byrne 5% program, and $206M through the NCHIP program. This represents an annual average of $0.77M in federal funds awarded to each state, over the past nine years. The chart below depicts the level of awards over time. Total Federal Program Awards by Year Specific financial assistance has also been targeted to states at both ends of the criminal history records automation spectrum. “Priority” states (Maine, Mississippi, New Mexico, Vermont, and West Virginia) each received a supplementary grant of up to $l million in NCHIP funds to spend on basic activities to enhance automation of criminal history records. Similarly, the 18 NCHIP “advanced” states, a subset of III states, were eligible, under the Advanced State Award Program (ASAP), to collectively spend an additional $5 million on extended core activities that would enhance the interface of their computerized criminal history systems with databases of persons other than felons who are ineligible to purchase a firearm. Finding 1.1: The establishment of federal programs has helped states place a high priority on criminal history records improvement. Byrne 5% and NCHIP program requirements have heightened awareness of the importance of improving criminal history records. As part of the Byrne 5% requirement, states must: (1) develop a Criminal Justice Records Improvement (CJRI) Plan and update it annually in order to expend their 5% funds, (2) convene a multi-agency criminal justice records improvement task force and, (3) as part of NCHIP, coordinate Byrne 5% and NCHIP funds. In addition, states have target dates for meeting the Attorney General’s timetable for current and sharable records as well as dates for III participation, where applicable. Further, federal funds have helped leverage state and local funds, targeted at improving the quality of criminal history records. One-third of the states expended more Byrne funds for criminal justice records improvement than the federally mandated 5% set-aside– evidence that states recognize the need for improving criminal history records. Moreover, states indicate that flexibility in the administration and use of Byrne 5% funds is helpful: it does not require that all projects be equally subsidized and allows the states to put funds to best use. Finding 1.2: The amount of available federal funds is not excessive. The question of whether some states may be unable to handle additional workloads associated with a large infusion of funding–precipitating a so-called “saturation phenomenon”–has been raised; for example, between FY 95 and FY 98 states drew down only 36% of their NCHIP awards, on average. There are, however, other possible explanations. First, the typical NCHIP-and Byrne- funded activities (e.g., an AFIS effort) take considerable time to complete; this is to be encouraged, since states may otherwise be unsuccessful in undertaking such major efforts and explains why funds are not being spent. Second, some states strategically accumulate their Byrne funds over several years to purchase “big ticket” items. Third, no state has requested to waive compliance with the requirement to allocate at least 5% of its Byrne funds for improving criminal history records. Finally, new programs, such as the State Identification Systems, come into existence, necessitating additional funding. Finding 1.3: While there is synergy among the CHRI, Byrne, and NCHIP programs, an attempt should be made to improve coordination with the newer DOJ initiatives and with other federal and state programs that have implications for criminal history. Byrne 5% and NCHIP funds are coordinated, in the sense that they complement each other in related efforts, rather than supplement one another in the same efforts. A state may fund improvement activities in the judicial branch with one of these two sources, while activities in the executive branch could be underwritten by the other source. Although logically, the Byrne and NCHIP funds could be commingled to implement an interface between a courts information system (judicial) and a computerized criminal history records system (executive), this does not occur because Byrne, unlike NCHIP, requires a match and local pass through. Commingling the two sources would introduce complexities in administrative and funds tracking. CHRI and NCHIP also complement each other in related efforts. While any leveraging of NCHIP and CHRI funds to support the same activity is negligible, the two funding sources overlap in the kinds of activities they support, namely, those falling into the System Improvements and Criminal History Records categories. Interestingly, these types of activities are as prevalent under NCHIP as they were under CHRI, implying a continuing need for funding these initiatives. The difference in allocation of NCHIP and CHRI funds is also understandable. Because the average NCHIP award is much greater than the average CHRI award, only 16% of NCHIP-funded activities leverage state and/or local funds, compared to over 41% of CHRI- funded activities. By the same token, 41% of all activities are partially funded by NCHIP, whereas the analogous percentage for CHRI is only 17%; this can be attributed to the narrower CHRI focus. Some state officials feel the greatest barrier to effective coordination of the increasing number of records-related programs is institutional. At the federal level, programs are administered by multiple organizational units within BJA and BJS; this occurs more disparately at the state level, where the respective administrators may be not only in separate agencies but even in different branches of government (i.e., executive vs. judicial). As new programs emerge (e.g., State Identification Systems, which supports AFIS development) and integration initiatives proliferate across agency lines (e.g., Health and Human Services programs requiring selective access to criminal history information), it will become more crucial than ever to coordinate the various federal and state criminal justice programs with federal and state non-criminal justice programs. Organizational changes are being considered at the state level to address this need. Finding 1.4: The majority of records improvement activities are initiated and completed on schedule. An overwhelming 75% of activities start on time, and some 70% of activities are completed on time, based on an analysis of activities that included planned and actual start and completion dates. This is commendable, given myriad possible delays–attributed to contractor problems, personnel changes, and political difficulties–not within the control of the department implementing the initiatives. Ongoing activities–including training and auditing–comprise 7% of the total. Only 19 activities experienced starting lags exceeding two years, while only 14 activities experienced completion lags of two years or more. The average criminal history records improvement activity takes 2.7 years to complete. These statistics should help guide states through future planning efforts. Goal 2: Improve Records Quality Improve the quality (i.e., completeness, accuracy, consistency, timeliness, and accessibility) of criminal history records. To gain insight into the states’ perspective, we administered a questionnaire to state officials, requesting their views on the relative importance of data quality issues and data quality improvement activities. We also conducted telephone interviews with 50 users of criminal history information in both the criminal justice and non- criminal justice communities and asked for their views on changes in the quality of records between 1992 and 1997. While our sample is limited and somewhat biased–38% of criminal justice users were from local law enforcement–we find that these users are generally content with records quality; although, to the degree they could recollect, they believe that improvements since 1992 have been modest. Our findings indicate that while federal funds have been instrumental in progress towards improving the quality of criminal history records, more work needs to be done. Finding 2.1: The automation of criminal history records systems– especially their interfaces–has made records available on a more timely basis. Eighty-eight percent of users interviewed see access to criminal history records as either being timely or very timely; 30% perceive that access was either more timely or much more timely in 1997, compared to 1992. Attribution for the improvement was evenly split between improvements in automated systems and in data entry protocol. Reduction in disposition submission times is one factor contributing to the greater timeliness of record accessibility. Responses to our questionnaire indicate that disposition submission times–deemed problematic by the states in 1994–are no longer a concern. Thanks to the CHRI emphasis on increased automation of disposition reporting, submission times have been successfully reduced. In cases where there is no difficulty linking a disposition to its arrest, the improved disposition submission times lead to the timely availability of a complete record. However, the troubling fact that arrest-to-disposition linking problems remain suggests that automation alone is insufficient to alleviate poor linkage, which is usually a symptom of a more structural problem (e.g., pertinent tracking or control numbers not entered on the arrest/disposition record). Finding 2.2: More federal funds are needed to substantially improve the quality, and particularly the completeness, of criminal history records. While availability of federal funds has enhanced quality of criminal history records, there is still substantial room for improvement. Completeness–the extent to which the criminal history record contains available disposition information–remains an acute problem. The degree to which arrests in the criminal history database have a final disposition was cited by states as being the most critical and most problematic issue they face, in both 1994 and 1997. The past decade has witnessed a major increase in automated disposition reporting, but states still find it challenging to link dispositions to associated arrests and charges. While automated disposition reporting has accelerated the rate at which dispositions are received at the repository, this does not necessarily guarantee the linking of a disposition to its corresponding arrest. The linking task can be especially difficult in states where dispositions are matched to corresponding charges, since charges can be often dropped or modified anytime following an arrest. One manifestation of this linking problem is the increase in suspense files–that is, repository files containing dispositions that cannot be linked to arrests. A procedural change, such as implementing unique identifiers, or Offender Based Transaction Statistics (OBTS) numbers, should be encouraged, since it has been shown to help states alleviate the problem. States should also continue to locate and process disposition reports not submitted to the repository–an activity which many states have cited as improving the quality of records, and which should be implemented on a wider scale. States assert that upgrading the AFIS and CCH systems and implementing livescan will yield the greatest improvement in data quality, and as such, are among the most frequently undertaken activities. Federal funds have played a key role in subsidizing these costly initiatives (see Finding 4.2). The importance of these efforts is understandable, since the AFIS and CCH are necessarily the two critical components of an efficient repository. Further, legacy AFIS and CCH systems installed in the 1980s need to be replaced with state-of-the-art hardware and software. Livescan, on the other hand, is a newer technology that should be fostered, since it improves arrest reporting and helps build towards a paperless system. The timely focus on livescan and automated arrest reporting is likely related to the fact that automated disposition reporting has made major strides since CHRI, allowing more emphasis on the front end of the records process. The fact that the average time to complete an improvement activity exceeds two-and-one-half years explains why the need for supplemental funding can also be expected. Finding 2.3: Records are more accessible and more useful as a result of improvements to criminal history records. Eighty-five percent of users interviewed feel that records were either accessible or very accessible in 1997; 34% feel that they were either more or much more accessible in 1997, compared to 1992. This latter low percentage may be due to the fact that local law enforcement–a third of the users we interviewed–traditionally has had greatest access to the records, and hence no substantial difference is apparent to them. The majority attributed the improvement to changes in their automated systems, which, as in Goal 5, has been a focal point of federal funds. Seventy-nine percent find records information useful or very useful, and 34% feel it was either more useful or much more useful in 1997, compared to 1992. The predominant reason for increased usefulness was seen to be the greater completeness of the information. Goal 3: Improve Reporting Improve interstate, intrastate, and federal criminal history records- related reporting. Finding 3.1: Linking dispositions to their associated arrests poses a number of lingering problems. Four issues inherent in linking arrests and their dispositions are: * The delay in rendering a disposition pursuant to a felony arrest could be due to prosecutor or defense postponements, and/or to court backlogs. * The delay in entering a rendered final court disposition could be due to a communication delay between the court and the central repository and/or processing backlogs at the central repository. * The long-term difficulty in obtaining dispositions for 100 percent of felony arrests could be due to problems in tracking arrest cases through the criminal justice system as charges are modified and plea bargaining occurs. Prosecutorial dispositions may also not be readily available to the repository. * The long-term difficulty in entering all rendered final dispositions could be due to problems in linking dispositions to appropriate arrests. National goals of making arrest-to-disposition linkage raise concern about state-to-state comparability and data availability. A preliminary list of questions that should be addressed: * Is a disposition required for every charge, or is one per arrest enough? States which post dispositions for every charge–compared to those that post one disposition for each arrest–are at a disadvantage in attempting to dispose of an arrest. * How does a state determine whether a disposition is linked to an arrest (or charge)? Is there a field indicating that the disposition has been received and entered, or is a proxy used, such as the date of entry of the disposition? If neither of these data elements exists, how does the state know this information? * Does the criminal history records database identify disposed arrest/charges? In some states, prosecutorially disposed arrests are not consistently reported, if at all, to the repository. * Does the state expunge old, undisposed arrest records? There may be points in time after which “old” arrests whose dispositions have not yet been received by the repository are no longer counted in the arrest base against which the degree of linkage is measured. States that engage in this practice would obviously have better arrest/disposition linking track records than states that do not. Finding 3.2: Setting realistic standards for linking arrest and disposition records remains a challenge. On average, states continue to view the linking of a disposition to an arrest as problematic. Not only is this troubling for the states, which require complete and accurate records to make informed decisions on bail setting and sentencing, for example, but also because standards helpful in measuring record completeness are difficult to establish. For example, the National Child Protection Act and the Brady Act’s Attorney General’s timetable each refers to objectives in linking dispositions to their corresponding arrests, but a statistical model we formulated showed these to be unrealistic. Specifically, our model examined the relationship between the average percent linkage required and the average elapsed time (in weeks) between arrest and disposition linking. An assumption of even modest variability in the elapsed time between arrest and linkage to a disposition suggests that a typical objective of having 80% of criminal history records be “current and shareable” is in practice unattainable. Moreover, our model showed that for that goal to be achievable under even a modest variability assumption would require the average elapsed time between arrest and disposition linking to be less than 10 weeks. Finding 3.3: The infusion of federal program funds has increased the ranks of III membership, albeit slowly. In contrast to CHRI and Byrne 5% efforts, a key goal of NCHIP in support of NICS is participation in the FBI’s Interstate Identification Index (III). As such, since the start of the program ten states have become III members under NCHIP–Alabama, Arizona, Arkansas, Indiana, Iowa, Maryland, Mississippi, Nebraska, New Mexico and West Virginia–bringing the total to 39. While states report that they do not believe that a major effort in III participation will improve data quality, they continue to use federal funds to accomplish this goal, suggesting the importance of federal funds as an incentive for III participation. From a records quality perspective, joining III should be encouraged, since state-supported records are more complete than FBI-supported records. Goal 4: Automate Systems Automate systems for creating, storing, and sharing criminal history records. Finding 4.1: Federal funds are responsible for major automation improvements in criminal history records throughout the states. The importance of automation in improving data quality cannot be overemphasized; the states obviously concur. The three highest ranked federally funded improvement activities are upgrading CCH software, installing livescan, and electronically transmitting dispositions to the repository. Each of these activities falls into the category of automation; collectively, they account for over 11% of all activities. In particular, livescan implementation and electronic disposition reporting are critical in helping states in their efforts to achieve “data entry at the source”–and ultimately a paperless record system. In addition, the number of NCHIP-funded flagging activities is up over 50%, as compared to those funded by CHRI. This is clearly beneficial, and not only for identifying felons. Eighteen percent of activities focus on flagging disqualifying crimes, such as child abuse, which may include misdemeanors. Finding 4.2: Without federal funding, the states would not have achieved their current levels of AFIS and livescan implementation. In 1994, states asserted that livescan implementation was the activity with the greatest potential for improving criminal records. Since then, federal funds have played a major role in the increased levels of livescan implementation. In addition to improving quality of fingerprints, livescan also improves arrest reporting. Implementation of livescan, especially at high-volume arresting agencies and central booking sites, should be fostered. AFIS-related activities undertaken by 50 states account for over 8% of all activities; NCHIP funds half of these. This level of interest is evidence of the rapidly burgeoning pace of AFIS technology. The large number of AFIS-related activities also reflect the greater-than-ever need in states to store civilian prints in their AFIS, in response to the proliferation of fingerprint-based background checks. In some states, the volume of civilian fingerprint checks surpasses criminal checks. As noted earlier, future planning of these initiatives should leverage other DOJ funding sources, such as SIS. Finding 4.3: Integration of automated justice systems is becoming increasingly important in improving data quality. While integration poses formidable challenges, it is critical as we move toward a paperless system, in which data is entered only once at the source (thus reducing the possibility of human error and inconsistent data). Because integration efforts cross agency, and often jurisdictional, lines, their success depends on a top-down commitment from heads of participating agencies. Consensus building is also needed to overcome “turf” issues and to coordinate resource utilization. The most prominent shift towards integration shows up in the increase in new prosecutor information systems, coupled with an increase in prosecution/repository interface activities. Traditionally, court dispositions have been the funding focus for disposition reporting, and rightfully so. Moreover, in some states prosecutors are elected and may not be eager to report cases that are not being prosecuted because their constituents would be displeased. Tracking prosecutorial declinations, which will improve completeness of criminal history records, should be fostered. Goal 5: Identify Ineligible Firearm Purchasers Identify persons ineligible, for criminal and non-criminal reasons, to purchase firearms. Finding 5.1: More firearm sales to ineligible purchasers may occur under NICS than during the interim provisions of Brady. During the interim provisions of Brady from 1994-1998, all states checked their own records when performing firearm eligibility checks. Under NICS, however, which began in November 1998, state-level checks are performed only by states serving as so-called Points of Contact (POCs)–in which case, a federal firearms licensee (FFL) contacts the state prior to the sale of a firearm. Unfortunately, the majority of states are not POCs–in which case the FFL contacts the FBI, whose criminal records are not as complete as state records. This is particularly an issue in non-POC and non-III states. Further, NICS may not be able to verify certain non-felon information: some state repositories may be permitted access to mental health information for the purpose of conducting a firearm eligibility check, but that same information would be prohibited from being passed on to populate the NICS index. Another artifact of NICS is the absence of a “cooling-off” period prior to the purchase of a firearm. The interim Brady five-day “waiting period” was effectively a “cooling off” period for an individual who wished to buy a gun with the intent to harm. For the state, it was a “maximum response” period, since a firearm purchaser did not have to wait five days before buying a handgun, but had to allow up to five days for the CLEO to check his/her records to determine purchase eligibility. Interestingly, even with NICS, there is a feeling in the current federal administration that the five-day waiting period should be reinstated to allow law enforcement officials more time to check noncomputerized records and to help prevent rash acts of violence. Finding 5.2: The identification of non-felons ineligible to purchase firearms is expected to remain problematic. As noted earlier, the seven categories of individuals prohibited from purchasing a firearm listed in the Gun Control Act, the Anti-Drug Abuse Act, and the Brady Act are: (i) persons under indictment for or convicted of a felony; (ii) fugitives from justice; (iii) unlawful drug users or addicts; (iv) mental defectives; (v) illegal aliens; (vi) dishonorably discharged; and (vii) citizenship renunciates. The eighth and ninth firearm ineligibility categories–namely, persons who are “subject to a civil restraining order arising out of domestic or child abuse” and those convicted of “domestic violence misdemeanors”–were added as part of the Violent Crime Control Act of 1994 and the Lautenberg Amendment of 1996, respectively. Identifying non-felons ineligible to purchase firearms is challenging since non-felon information is not readily available to state criminal history record repositories. Also, the dissemination of mental health and drug abuse information raises legal and ethical questions about the rights to privacy and presents new security challenges. It is understandable, therefore, that two of the three dominant NCHIP- funded Advanced State Award Program (ASAP) activities aimed at identifying non-felons are establishing access to mental health records and establishing access to drug abuse records, undertaken by nine and seven states, respectively. (The third most popular ASAP activity, undertaken by 12 states, is incorporating civil protection orders in the repository database, as discussed below.) The challenges include determining whether databases maintaining this type of non-criminal information exist and, if so, the feasibility and legality of accessing them, especially if they belong to private institutions. New enabling statutes may be required to overcome these obstacles. Two other ineligibility categories present unique implementation challenges: subjects of restraining orders and domestic violence misdemeanants. States cannot reliably identify individuals for whom Gun Control Act-compliant restraining orders–among the plethora of restraining order categories–have been issued. For this reason, some states deny firearms to subjects of all restraining orders. The challenge with domestic violence misdemeanor convictions is that the law is retroactive; but domestic violence incidents have historically been categorized as assaults, making it difficult to segregate them from other criminal history records. Goal 6: Identify Disqualified Care Providers Identify individuals disqualified from caring for children, the elderly, and the disabled. Finding 6.1: The passage of federal and state legislation has precipitated growth in the volume of requests for background checks of employees, volunteers and licensees–the challenge is how to meet the subsequent demand placed on the resources of state repositories. Although practices (e.g., statutory mandates and regulations concerning inquiries) vary from state to state regarding background checks, careful planning and explicit procedures are needed to support the high volume of such inquiries, which in some cases surpasses that of criminal checks. The volume of civilian fingerprints is now overwhelming AFIS storage capacities. Moreover, the volume of inquiries can be expected to increase as states continue to pass laws that increase the scope of background checks. In addition, the Volunteers for Children Act, passed as part of the Crime Identification Technology Act of 1998, amends NCPA to authorize qualified volunteer organizations to contact authorized state agencies (e.g., the repository) to request national criminal fingerprint background checks, in the absence of state procedural requirements. Not surprisingly, the increased volume of fingerprint-based applicant background checks has resulted in longer response times in a number of states. Based on our interviews, we noted a heightened frustration on the part of agencies waiting for responses. Obviously, the demand placed on state repositories for background checks must be appropriately met. Finding 6.2: There are problems associated with acquiring and interpreting information needed to disqualify prospective care providers. Incomplete records are especially a problem in states that release conviction-only data to authorized agencies requesting background checks. For example, if the subject of a background check has been arrested and convicted of a disqualifying offense, but the disposition has not yet been received at the repository or has not been linked to its arrest, the conviction will not appear on the record. The repository will not release any information, and the agency will not know that there has been a conviction. The agency will not even know that there has been an arrest, which could otherwise be followed up with the court of jurisdiction. In addition, agencies requesting background checks do not always know if a particular conviction is disqualifying for employment. Agencies are not necessarily qualified to understand the plethora of violation and conviction codes contained in the reports they receive. For example, sometimes they cannot distinguish whether a felony violation involved a child, and hence whether it is disqualifying. 3. Remaining Issues Future evaluation efforts should build on findings in this report, seeking closure on outstanding issues and assessing more recent BJS and BJA initiatives to further improve criminal history records. More specifically, they should: 1. Continue to assess the impact of federally funded activities. This report’s timeframe precedes the FY 98 NCHIP and Byrne 5% awards and many of the CHRI-, Byrne 5%-, and NCHIP-funded activities are still in progress. Moreover, an evaluation of the State Identification Systems (SIS) and National Sex Offender Registry (NSOR-AP) programs should be initiated. SIS and NSOR-AP are new programs which have yet to be assessed–SIS enhances states’ ability to identify offenders by upgrading their information systems and DNA analysis capability, and NSOR-AP promotes establishment of a national sex offender registry. Thus, formal monitoring of all federally funded activities should be ongoing. 2. Continue to develop a measures framework. Measures must continue to be identified, building on the C-CHRIE study, in which we develop a framework that incorporates a core set of input, process, and outcome measures with which to assess records quality, over time. A related issue is the identification of a set of desirable attributes for pertinent records quality measures which, in the aggregate, can be used to assess the state of records quality over time. As part of the C-CHRIE study, we have identified such attributes as understandability, measurability, availability, consistency, validity, reliability, stability, accuracy, independence, robustness, and completeness. 3. Create a computer-based simulation model of the criminal history records process from arrest-to-disposition linkage. Building on the measures framework, a simulation model of the arrest- to-disposition linkage process should be developed, using actual system data from a set of focus states. Results would shed light on the interaction and relevance of measures, as well as their impact on national goal setting. 4. Define a set of pertinent measures to assess the aggregate improvement of records quality, over time. In partnership with BJS, develop a set of pertinent measures to determine the nation’s progress in improving criminal history records. These measures should reflect common goals of federally funded criminal history records improvement programs, capture progress over time, and to the extent possible, have the above-mentioned attributes. Once developed and tested–perhaps using the simulation model described above–measures should be updated regularly to determine the extent to which federal goals are being met, to indicate where deficiencies lie, and to point to activities which could mitigate such deficiencies. 5. Expand the assessment of user perceptions about the value of criminal history records. We have learned a great deal about the ultimate usefulness of criminal history records by speaking with a small group of records users. They were anxious to share current perceptions of the quality of the records, as well as expectations and concerns for the future. Data quality improvement will benefit from interviews with a larger, more diverse set of users from both the criminal justice and non-criminal justice communities. Ultimately, user perceptions are key to understanding the true value of criminal history records and thus the ultimate success of federally funded improvement programs.