FiveThirtyEight Dives into the Messy World of Risk Assessment

They created a parole simulator to show the tightrope parole boards try to balance on.

The simulator can be found here, in the gray box with the subtitle “Who Should Get Parole?”

Part of a larger story on prison sentencing that is also worth a read, the parole simulator stuck out as an interesting, interactive feature that lets readers visualize the odds of a parole board’s risk assessment process leading to the correct parole decision.

As with most FiveThirtyEight content, the statistics are beyond most of us, but looking at the simulator for a few minutes begins to give you an idea about how prisoners might be denied parole when they would not repeat again, or being given parole when they would repeat again.

The simulator breaks prisoners up into three categories, low, medium, and high risk. The issue becomes apparent as you watch the percentages fluctuate, the highest disparity we noticed while we had it running on default settings was 37% of parolees reoffending while 7% of those denied parole would not have reoffended.

Tweaking the sliders to change the thresholds of each category to strike a balance, you might end up with a complete flip-flop on the numbers and all of a sudden you have a big disparity of rehabilitated folks being denied parole.

This is where FiveThirtyEight points out that there is no right answer, first because everyone has differing opinions on acceptable recidivism and incarceration rates, and second because this is tremendously hard to predict as there is a mountain of variables to consider in these decision making processes.

See for yourself and as they ask on the simulator, are you ok with the results?

For further analysis on how the simulator works, see here.