Convictions - Mapping Pretrial Risk

The majority of RATs in our research include prior convictions in their calculations to predict future risk.

Using prior conviction in RATs to assess risk may seem more fair or accurate than arrests. However, over 90% of convictions come not from a jury at trial but from a guilty plea.¹ Including convictions in RATs is not an objective or fair measure to predict failure to appear and re-arrest.

Many plead guilty for a host of reasons besides actual guilt.

Because of mass incarceration and huge numbers of arrests, plea bargains have become the norm in the criminal legal sphere.² Pleas can lead innocent people to accept a conviction to get rid of onerous stacked charges against them, especially in an overcrowded criminal legal system that barely has time for each case.

Legal scholars have argued that prosecutors deliberately overcharge accused people, bringing more severe charges against them in order to push them to accept guilty pleas for lesser charges.³

Many feel pressured to accept pleas because the consequences of trial could be much worse; indeed “plea bargaining has become so coercive that many innocent people feel they have no option but to plead guilty.”⁴

The plea bargaining process is rife with racial bias as well.

One study found that there were serious racial disparities in plea bargaining for lower-level cases where the accused had no prior convictions. White individuals were over 74% more likely than Black individuals to have misdemeanor charges with the possibility of imprisonment dropped, dismissed, or reduced. White individuals were also 14.5% more likely to have felony charges reduced to misdemeanors.⁵

The study found that, because the racial differences in plea bargaining were so stark for those who had no prior convictions and where prosecutors had less information about the accused, “race perhaps is being used as a proxy for a defendant’s latent criminality and likelihood to recidivate.”⁶

Studies have also shown that when someone is detained pretrial, which often happens if they cannot afford bail, they are much more likely to plead guilty so they can get out of jail and go home before their trial.⁷

Scholars argue that those who are detained pretrial even “accept pleas they may not otherwise accept if they were not in custody,”⁸ meaning convictions may simply be the direct result of lack of means to pay a bond.

Even when a case does go to trial, the accused have to contend with the racial bias of both judge and jury⁹ that can impact their likelihood of conviction and the severity of their sentences.

One study found that Black individuals are more likely to be wrongfully convicted than white individuals and represent almost half of innocent individuals wrongfully convicted and later exonerated.¹⁰

It is clear that racial bias in plea bargaining and at trial can lead to conviction, whether or not a person has actually committed any crime — and that conviction will show up in a risk assessment.

Further, most tools do not consider how long ago a conviction was when calculating a predicted risk score. Many tools use “prior convictions” as a factor, or even count up the number of previous felony and misdemeanor convictions to give a point for each one.

However, older and newer convictions do not have the same impact on likelihood of rearrest. One study found that the differences between more and less recent convictions were such that for those whose most recent conviction was six or seven years ago, they were about as likely to have a new offense as someone with no criminal record.¹¹

RATs might view all of these convictions, even if they are from a plea or in some cases, from 10 years ago, the same way.