Is the Prosecutorial Black Box Finally Opening?

It’s not every day the headlines announce government officials releasing mounds of data about their own operations. Actually, the frequency verges on the unprecedented. Cook County State Attorney Kim Foxx’s decision to release information on years’ worth of her office’s decision-making is, in that light, astounding.

Legal scholars have lamented the existence of so-called “black boxes” within the criminal justice system. They arise whenever various dimensions of legal decision-making are unobservable to outsiders, whenever the reason for a decision is left unexplained. Whether it’s the jury or the prosecutor, the public rarely (if ever) has access to systematic information on their motivations and priorities. We can observe the outputs—who is charged, who receives a plea deal. But we only observe those outputs when the court official proceeds with the case. Without access to internal data, we cannot parse the determinants of, for example, different declination rates by race and crime type. For elected prosecutors, this information gap is thrown into sharp relief. The voting public is left to rely on their gut instincts and high-profile convictions when assessing an office’s track record in promoting public safety.

Data transparency is also relevant to the Access to Justice Lab’s work. Our expanding studies of pretrial release decision-making, although primarily focused on judicial discretion, can benefit from prosecutors’ data. Here’s how. The PSA’s developers define new criminal activity—a key outcome variable in our analysis—as any arrest for new charges. We can use the court’s database to locate all arrests that lead to filed charges. But what of the arrests that do not? We could merge the court data with jail booking information, but that comes at the cost of having to un-duplicate the dataset when a jail booking also leads to formal charges. The superior source would be the prosecutor’s own database; it contains information on all of the in-custody arrests and non-custodial citations regardless of eventual charge activity. It’s effectively a one-stop shop for identifying any new arrest following pretrial release.

Time will tell if Ms. Foxx’s bold action motivates other jurisdictions to follow her lead and facilitate evaluation of this relatively opaque player in the criminal justice system.

4 thoughts on “Is the Prosecutorial Black Box Finally Opening?

  • Thanks for flagging this release of data by a District Attorney’s office. Hopefully a trend that wil become an expectation for elected DAs.
    It would be good to define PSA and provide a link to the framework you comment on. Is this one? http://www.arnoldfoundation.org/wp-content/uploads/PSA-Risk-Factors-and-Formula.pdf

    Rate of declination–is that a new metric? or is it rate of decline? If new metric, can you define, or explain how it is constructed? Or provide a reference?

    • The Public Safety Assessment (PSA) is a pretrial risk assessment instrument that only relies on static criminal history (and age) to predict failures to appear and new criminal activity during the pendency of a case. You found the right link to the risk factors that are used to generate risk scores. More information about the PSA can be found at http://www.arnoldfoundation.org/initiative/criminal-justice/crime-prevention/public-safety-assessment/.

      Declination here refers to decisions not to file charges based on law enforcement’s recommendation or to dismiss charges as the case unfolds. One article discussing this form of prosecutorial discretion is Michael Edmund O’Neill, When Prosecutors Don’t: Trends in Federal Prosecutorial Declinations, 79 Notre Dame L. Rev. 221 (2003).

  • We’re rolling out a PSA score bond system in our County. It’s been a complete disaster so far – mostly because the data collection specialists are missing everything.

    For example, we had a guy come through with three bail jumping convictions and the PSA recommended OR’ing him because he had no failures-to-appear on his record.

    The PSA data people have only 8/ish hours to gather information when people are arrested before they appear before the judge setting bond. That’s not enough time. Also, they don’t check all the County databases – just the County of arrest. So some people have honest-to-god arson/murder felonies pending in other Counties that the PSA doesn’t find.

    Also, the judges aren’t obligated by statute to consider it.

    So far, the only effect its had is to give people who would have been OR’d anyways (on misdemeanors or minor felonies like DUI 4 etc.) extra reporting requirements (pretrial services etc.) that expose them to additional bond revocation risks.

    How much of this system is based on the assumption that judges are setting cash bonds on minor offenses? Is a PSA/pretrial services framework valuable in a County that already only hold serious felony offenders and serial absconders?

    • Cameron,

      I appreciate the concern over your jurisdiction’s experience with the PSA. As an evaluator, I can’t respond to the implementation issues you describe, with which every county/state contends. They are important but can’t ultimately speak to whether the underlying science behind the PSA is better than the unguided decision-making of judicial officials. On that point, it’s unlikely that most jurisdictions would have adopted the PSA if it led to statutorily binding release decisions. The preservation of discretion adds some complexity to any evaluation, even a randomized one; nevertheless, it’s important to know the extent to which judges concur with the recommendation.

      The PSA’s risk factors focus more on the fact that judges tend to set bail based on the severity of the charge rather than an objective demonstration that the defendant would pose a risk of failing to appear or committing new criminal activity if released. The PSA is, hypothetically, most valuable in jurisdictions where judges have a difficult time assessing those risk on their own and where judges rely on little more than a bail schedule to make release decisions.

Leave a Reply

Your email address will not be published. Required fields are marked *