While doing research for a book, I met a woman in Mississippi who had been beaten by her boyfriend. She was in the hospital for days. But her case was never prosecuted. After some digging, I discovered that the local prosecutor hadn’t prosecuted a single domestic violence case in 21 years.
“Has it been that long?” the prosecutor asked after I told him what I had found. Yes. That long. Resources were scarce, he explained, so he focused on prosecuting cases that he thought he could win. But in the process, he had let an entire category of cases fall by the wayside.
That prosecutor is far from alone. I spent eight years sitting in courts throughout the United States, and I saw how easy it is for well-meaning people to turn a blind eye to systemic problems. The US criminal justice system includes 3,143 counties, each with at least one state criminal trial court. Today, there is no accurate way to compare county-level performance to determine which courts need attention, or to evaluate which court practices work and which don’t.
In April 2011, I founded Measures for Justice (MFJ), an organization that aims to create a Justice Index that will measure how criminal justice systems across the country are delivering basic services. For example, we created a measure to show prosecution rates of domestic violence arrests—in Mississippi and elsewhere. Court insiders, criminal justice activists, journalists, and community members will be able to use data from the Justice Index to shed light on a given county’s performance.
Nearly every public institution or resource in the United States—our schools, our hospitals, our water supply—is accountable to a watchdog system that provides comparative performance data. In education, measures such as student-teacher ratios and college admission rates have become standard tools for assessing the levels of opportunity that are available to students in different communities. In medicine, the Dartmouth Atlas of Health Care brings to light variations in resource allocation and in professional decision making, and it has become a common source of data for policy makers.
Criminal justice lags behind these realms, largely because there are no common measures that would allow one jurisdiction to compare its performance with that of other jurisdictions at each stage of the criminal justice process—from arrest to post-conviction. Federal, state, and local governments actually collect plenty of data. But there is no uniformity in the kinds of data that they collect or the way that they report that information. A county, for example, will typically count the number of people who enter its system. But its officials generally have no idea where the county ranks in terms of practices that prevent innocent people from being convicted, or practices that help to decrease recidivism. Nor do officials know how their county compares with others in money spent on holding people in jail. Without comparative data, we lack a crucial lever that would help us improve our system of due process.
County Versus County
Our goal at MFJ is to create a wide array of measures that gauge performance in the areas that victims, defendants, legal professionals, and community members care most about. We are focusing on three areas in particular: public safety, equality and fairness, and efficient use of resources.
To develop indicators for each area, we worked with a multidisciplinary group of researchers and practitioners to draft a set of 50 measures. We then began to experiment with those measures in a series of pilots. In our first pilot, the Illustrative Measures Study, we selected 12 measures that we could populate with publicly available data. (We drew the data from the US Bureau of Justice Statistics and the FBI’s Uniform Crime Reporting program, among other sources.) For many measures, we were able to find data for about half of the 100 largest US counties.
We found enormous differences between counties. Consider the example of average bail amounts. In 2004, Los Angeles County had an average bail of $169,047, whereas Broward County, Fla. (home to Ft. Lauderdale), had an average bail of $4,749. The purpose of bail is to ensure that a defendant will reappear in court. Even taking into account possible differences in patterns of criminal behavior, an average bail that’s 35 times as high as that of another major metropolitan area counts as a big red flag.
Eventually, the Justice Index will become a system that's able to gather data from most, if not all, counties. It will create a picture that no single data point can provide.
Advocacy groups in California have long identified use of bail as a pressing issue. In the Illustrative Measures Study, five California counties were among those with the highest proportion of felony defendants who could not make bail or who were not assigned bail. Many counties in the state still rely on “bail schedules” to help judges process defendants quickly. San Bernardino County (whose average bail is $80,958) has a schedule that calls for assigning a bail of $25,000 to a defendant charged with delivering a check with insufficient funds. So a person who writes a bad check and who can’t afford to pay that high bail amount might sit in jail for days or weeks or months. For years, legal experts have said that one-size-fits-all bail schedules are ill advised: They undermine judicial discretion and promote unnecessary detention. What’s more, keeping people in overcrowded jails is costly—both for defendants, who lose valuable work time (and sometimes even lose their jobs), and for taxpayers.
Wouldn’t it be beneficial to people in California to learn how other large US counties have tackled the challenge of setting bail? It’s too soon to say exactly which counties have better practices than others. But of the counties in our study, Fairfax County, Va., stood out by having the lowest proportion of felony defendants in custody who were unable to make bail or who were not assigned bail. Here’s one possible explanation for that result: Virginia has poured resources into a risk assessment tool that determines pre-trial release on the basis of a background investigation that screens for risk of flight and the risk of rearrest.
We shared the results of our Illustrative Measures Study with officials from counties that ranked either high or low on certain measures. Many officials, we discovered, are hungry for this information. Court administrators in St. Louis County, Mo., for example, asked to share our preliminary and incomplete data with their county’s governing board. To them, these data were part of a success story: Among counties included in the study, St. Louis had the lowest percentage of defendants (4 percent in 2006) who were rearrested while awaiting trial—a sign that it had identified people with a high risk of reoffending and had taken necessary steps to reduce that risk.
King County, Wash., fell on the opposite end of the spectrum. It had the second-highest percentage of felony defendants (34 percent) who were rearrested while awaiting trial. But Dan Satterberg, the King County prosecutor, asked us for permission to share data from our study as well. “This could shame us into action,” he said. Satterberg also highlighted the need for data that would allow counties to learn from one another about the design of effective re-entry programs, among other issues.
Better Data, Better Results
With the completion of the Illustrative Measures Study, our organization has moved beyond the proof-of-concept phase. We are now testing the measures to ensure that they are meaningful, accurate, and replicable. With support from the US Bureau of Justice Assistance, we have begun our first local pilot in Milwaukee County, Wis., where we will study the data that state and local officials collect and work to refine the measures. Then we will test those measures in other counties nationwide.
Eventually, we hope, the Justice Index will become a universal system that’s able to gather data from most, if not all, counties. It will offer a bundle of measures that, taken together, will create a picture of the US justice system that no single data point can provide. We will use these data to identify best practices and to enable county officials to learn from each other. In short, better data will create better systems, which in turn will create better results.
We can’t predict exactly how this initiative will play out. Take the example of the gross domestic product (GDP). An economist first unveiled the GDP model in 1937, and only after more than two decades of controversy did it become an established official metric. No one foresaw that it would become an established measure of national wealth. Our index will take time to develop, too. But its impact is potentially enormous. An estimated 65.7 million Americans—20 percent of the population—have a criminal record. And the number of Americans who are victims of crime is at least as high.
Bill Gates recently called for the adoption of a new, outcomebased model of philanthropy—one that is based on “setting clear goals, choosing an approach, measuring results, and then using those measurements to continually refine our approach.” That model, he argued, will “[help] us to deliver tools and services to everybody who will benefit, be they students in the US or mothers in Africa.”
Why not apply that model to the US criminal justice system? At MFJ, we are developing the tools to do so.