E pluribus unum: data and operations integration in the California criminal justice system.

• Author: Ball, W. David

INTRODUCTION

There is no single criminal justice "system" in California--just hundreds of separate agencies, each with its own culture, organization, and data. Offenders who are brought into "the system" are actually brought into several separate systems as they are arrested by law enforcement, jailed by sheriffs, tried by courts, sentenced to prison, and released into community supervision under probation or parole. Given that each offender will travel through a number of different agencies, integrated criminal justice seeks, at a minimum, to ensure that information about each offender will travel along with him. For example, once one agency learns that an offender needs psychiatric medications or that he has a particular gang affiliation, other agencies who subsequently gain custody of him should not spend time and resources rediscovering the same information. Integration should ideally involve coordination of programming and other interventions as well, so that an offender undergoing, say, drug and alcohol treatment in one part of the system can simply resume his treatment when he is transferred to a different part of the system.

It is difficult for the casual observer to understand just how hard it can be for criminal justice agencies--even ones in regular contact--to currently get information from one another. Part of the problem of information management is due to its scale: criminal justice is a core function of state and local governments employing tens of thousands of people. Part of it is due to inertia: different agencies have already developed their own recordkeeping programs and their own idiosyncratic ways of coding standard types of information, and those agencies that are connected typically do so via single-use databases, which do not allow easy searches involving multiple parameters. (1) In a Google era, where information on almost everything is easy to find, it can be shocking to consider that answering basic questions about criminal populations (such as how many people are serving what sentence for what crime) is not as simple as logging into a computer, entering a password, and running a search. Given how often basic questions need to be answered--and how important it is to know the answers--the scope for improving the efficiency of criminal justice administration is vast. Better information can make all interventions more efficient and more targeted. Information can also help us, crucially, to analyze which programs are working in the first place, and how, if at all, they can be improved. (2)

This Article reflects some of the insights from the Stanford Criminal Justice Center's (SCJC's) year-long project on data and operations integration in California's criminal justice system. The project employed the Executive Sessions model first popularized by the Kennedy School of Government at Harvard, where leaders from all area of criminal justice get together "to take joint responsibility for rethinking and improving society's responses to an issue." (3) Every few months, judges, district attorneys, sheriffs, police chiefs, public defenders, probation officers, court administrators, formerly incarcerated persons, and nonprofits from across the state came to Stanford to participate in a series of loosely moderated one-day sessions. Each session focused on a particular subset of integration issues. Studying the problem from a combined academic and practical perspective pointed out gaps between the theory and practice. Our prescriptions--which were made in concert with practitioners-reflected (or at least attempted to reflect) practical changes that were feasible within the context of how criminal justice agencies actually conduct business.

Five sessions focused on integrated criminal justice (4): a session introducing the topic, (5) a session on the importance of the pre-sentence report, (6) a session on regional information sharing, (7) a session on the ways in which integrated criminal justice might improve outcomes for an inmate's first seventy-two hours of release after prison, (8) and a final session on the use of Risk-Needs Assessments (RNAs) in managing local custodial populations. (9) Each session involved a multitude of participants speaking on complex topics; the conversations did not limit themselves to a single valence or a single conclusion. Because the reports from each session were posted to the SCJC website, (10) I will not recount all of the specific contours and conclusions of each session.

This Article will, however, develop the fundamental takeaway from the Executive Sessions: that better integration is not, primarily, a technology problem. Although there are infrastructural and technological barriers to integration, the main obstacles are organizational and conceptual. We cannot build an integrated criminal justice system without first answering some thorny questions. How do we define the boundaries of the network: by geography, population, or patterns of criminal activity? What is the role of state and local cooperation, and how can we overcome the mistrust between California's state government and the governments of its fifty-eight counties? What metrics are we using to evaluate programmatic success, and does everyone agree on the policy implications of a particular metric? More broadly, how will the collection and use of contemporaneous information shift criminal justice policy and practice from a static to a dynamic model? All of these issues relate to more than just criminal justice: what we learn from criminal justice integration can guide us in other kinds of multiagency, multijurisdictional problems. (11)

This Article will proceed in three parts. Part I lays out some of the benefits of an integrated system as a means of illustrating why law enforcement agencies across the state are actively pursuing data integration. Part II discusses three organizational and political obstacles to creating an integrated system: defining what we mean by the criminal justice "system," drawing boundaries of relevant networks, and resolving tensions among state and local agencies with concurrent jurisdiction. Part III then discusses three ways in which integration might have far-reaching implications on criminal justice policies, processes, and principles: by defining metrics of success, by improving organizational learning, and by isolating the goals of risk reduction and punishment.

1. THE BENEFITS OF AN INTEGRATED CRIMINAL JUSTICE SYSTEM

   Because the need for and scope of an integrated criminal justice system is not readily apparent, this Part will outline the potential ways in which integration might improve policies and procedures.

   A typical offender passes through many different agencies after breaking the law: law enforcement arrests him, a jail holds him pending bail or trial, the courts accept his plea or try him, and he is sentenced to probation, jail, or prison. Each agency typically has exhaustive policies and procedures governing the physical transfer of the offender, but less thought is given to transferring the information accompanying each person. Offenders are more than just bodies, of course: agencies need to know who an offender is, what he has done, and what particular risks and needs he presents. Integration could ensure that each agency has common knowledge about an offender's risks and needs and a common approach for addressing them that avoids needless duplication of effort--or, even worse, counterproductive changes in approach.

   In this Part, I will sketch out the ways in which integration could improve the efficiency and accuracy of several core public safety functions: investigation, sentencing, incarceration, and release. I conclude with a discussion of the ways in which integration could lead to better system-wide analyses and policies.

   A. Investigation

   Criminal activity is not confined to a single jurisdiction--an offender doesn't have a map she uses to confine herself to a particular city or county when committing crimes. The Bay Area, for example, has a number of what might be called "BART crimes," where offenders take the BART subway system to another jurisdiction, commit crimes, and return home. With better information sharing among the counties BART connects, law enforcement could, for example, more quickly and easily link a pawned piece of jewelry in one county with a police report describing that jewelry in another county. Currently, information sharing requires the establishment of joint task forces, or, in individual cases, shot-in-the-dark phone calls to other agencies. Systematizing data sharing would make local law enforcement more effective in investigating and preventing criminal activity.

   The state does collect and disseminate information, but the architecture is based on a fundamental design that is more than thirty years old. The California Law Enforcement Telecommunications System (CLETS) began operation in 1970 and traffic has grown enormously since then. (12) The system was designed in an era before the Internet and, to some extent, suffers from the legacy of its architecture, which requires special equipment to access databases that are hosted separately from one another. A more modern approach would use regular, multiuse computers over a secure connection, with databases that can be more easily integrated with each other. Coplink (13) is a private, peer-to-peer type system that allows users to search the contents of record management systems (RMSs) of neighboring agencies. (14) While state databases allow users to search a wider geographic area, the information in each of them is more narrow, related to discrete offenses (e.g., domestic violence) or types of offenders (e.g., gang members). (15) These databases do not allow users to search easily across databases with multiple variables.

   More detailed information can be invaluable, however, even when it does not lend itself readily to a particular category. For example, an...