A study conducted about the State of Kentucky’s use of the risk assessment tool concludes the tool has not lived up to its promises
The Texas Alliance for Safe Communities is today continuing their series of releases outlining knee-jerk, California/New Jersey-style bail reform efforts that have failed, been rolled back or repealed, or undermined the criminal justice system in a way that adversely affects community safety and the rights of crime victims.
Over the past several years, a movement to overhaul state bail systems has spread across the United States, leading cities, counties and states to pass legislation that eliminates commercial bail and implements risk assessment tools in its place. This practice has led to a massive increase in defendants being released on their own recognizance, with many later failing to appear for their scheduled court hearing. The pitfalls of such reforms are plenty, and research shows that many states are beginning to witness the devastating consequences of bail reform on public safety and reversing course.
So far, the Texas Alliance for Safe Communities has outlined the pitfalls of certain criminal justice reforms in the states of Alaska, California, New Mexico and New Hampshire.
Today, TASC shares the story about the risk assessment tool in Kentucky:
In June of 2018, Megan Stevenson, a researcher at George Mason University’s Antonin Scalia Law School, conducted a study for the Minnesota Law Review analyzing the results of the risk assessment tool in Kentucky, one of the first adopters of the tool and a state that is often considered to be the gold standard for its use.
According to the study, the use of the risk assessment tool and the releases it led to dramatically increased failure-to-appear rates and pretrial crime.
- “Using rich data on more than one million criminal cases, this Article shows that a 2011 law making risk assessment a mandatory part of the bail decision led to a significant change in bail setting practice, but only a small increase in pretrial release. These changes eroded over time as judges returned to their previous habits. Furthermore, the increase in releases was not cost-free: failures-to-appear and pretrial crime increased as well. Risk assessment had no effect on racial disparities in pretrial detention once differing regional trends were accounted for.” (Assessing Risk Assessment in Action, Minnesota Law Review, June 2018)
The study goes on to outline that there is virtually no empirical evidence proving the effectiveness of the risk assessment tool.
- “The empirical research evaluating whether outcomes are improved by incorporating algorithmic risk assessment into the decision-making framework is beyond thin; it is close to non-existent…Somehow, criminal justice risk assessment has gained the near-universal reputation of being an evidence-based practice despite the fact that there is virtually no research showing that it has been effective.” (Assessing Risk Assessment in Action, Minnesota Law Review, June 2018)
As time wore on, the effectiveness of the risk assessment tool wore thin.
- “Thus, while there was a change in the type of defendants released, as well as the conditions of release, the net effects on the overall release rate were small. Furthermore, they were not permanent: the sharp change in practices and outcomes that occurred right after the law was implemented eroded over time as judges returned to their previous bail-setting practices.25 Within a couple of years, the pretrial release rate was lower than it was before the bill, and lower than the national average.” (Assessing Risk Assessment in Action, Minnesota Law Review, June 2018)
The study goes on to outline that risk assessment tools were found to be no more effective in predicting recidivism than a human being with little to no prior criminal justice experience.
- “Using an experimental method that was explicitly set up as a horse race between survey respondents and algorithmic risk assessment models, they found no evidence that algorithms were more accurate in predicting recidivism than human beings…The authors found no evidence that any of the algorithms could outperform the predictions of a random group of online respondents.”(Assessing Risk Assessment in Action, Minnesota Law Review, June 2018)
Similar to this piece by ProPublica, the study mentions that the risk assessment tool is often biased against minorities.
- “There are a number of reasons why risk assessment tools could be biased against blacks. The most common argument is that inputs to risk assessment – prior convictions, prior incarceration sentences, education, employment, etc. – are themselves the result of racially disparate practices…Similarly, a risk algorithm that is trained to predict an outcome that is the result of racially disparate law enforcement or prosecution practices also incorporates bias into the algorithm.” (Assessing Risk Assessment in Action, Minnesota Law Review, June 2018)
Given what we know about the risk assessment tool, why would Texas and its counties even consider its use in the criminal justice system? When it comes to trying out unproven bail reforms, other states have done the dirty work. The State of Texas should heed the warnings of other states before moving to completely overhaul our justice system, leading to potentially dangerous side effects for true victims of crime and for the safety of our communities.