how statistical literacy can turn conventional wisdom on its head

A thorough study of prison records spanning 15 years shows that recidivism is actually a lot lower than the numbers commonly cited in the media.

jail cell

Pretty much every documentary about jails and prisons proclaims that three out of four inmates you’ll see released on your screen will be incarcerated again within five years, often citing many possible reasons as to why. Criminal records keep them from finding legitimate work, lack of an effective rehabilitation program didn’t help their drug habits, their mental illnesses weren’t really addressed and are dooming them to vagrancy and crime, overzealous policing and sentencing that takes a disproportional toll on the poor and minorities, and so on. All of these may be good reasons why so many people studied by the Bureau of Justice Statistics end up in jail and each should be talked about and addressed to drive down recidivism. But the public at large and the politicians they elect have taken this information to mean that our justice system is not only very badly broken, but has become little more than a rotating door for 75% of American criminals, so all sorts of reforms must be promptly instituted to end the constant cycle of incarceration.

Right now, as Congress is trying to figure out how to fix today’s overburdened justice system to make sentencing guidelines saner, and keep people who don’t need to be imprisoned for minor crimes out of jail and into rehabilitation and parole programs, knowing who is in jail, why, and if they’re likely to find themselves back there, is crucial for effective, smart changes. And against this backdrop comes a study showing that the public and the government misunderstood what the data collected by the BJS actually measures and what it means for us. The BJS just wanted to know what happens to prisoners year over year. As they do that, the statistical picture skews very heavily to frequent offenders because they’re the most likely people to be caught in those samples. The analogy the study’s lead author, public policy analyst William Rhodes, likened the difference in his methodology and that of the BJS to trying to build the demographics of who is visiting shopping malls over a week rather than over a year. The longer the period studied, the more accurate gauge you have over long term trends, and the more likely you are to catalog all those people who analysis over much shorter windows of time will more than likely miss.

Looking at records spanning 15 years across the country, indexed by prisoners’ ID numbers, it became very clear that 68% of people who serve a prison sentence do not re-offend, and of all those who do, only one third will re-offend more than once. So the chronically incarcerated who become the focus of BJS reports because they keep on getting caught in the sample again and again, actually represent a little more than a tenth of all prisoners. This means that far from the revolving doors for cons which do nothing to help curb crime, American prisons are actually an effective deterrent from further offenses. That said, the fact that 32% of people who did serve a sentence are ending up back behind bars is a problem which points to the need to address the aforementioned challenges of leaving jail with few resources to get one’s life back on track. For Rhodes, this is a clear indicator that there needs to be an investment in rehabilitating those with the highest risk for return, and considering that instead of having to focus on every prisoner we actually need to work with as little as a quarter of them, that investment is manageable.

This is why statistical literacy is so important. If we didn’t assess whether the methods we have been using accurately reflect long term trends, we’d end up having to take a wrecking ball to all the jails and prisons to truly fix things. Considering how much was spent on building them, it’s a politically and economically painful proposal no one would want to back. No one willing to fix the entirety of a broken system over decades means that much needed reforms would’ve died very slow and painful deaths while more and more prisons were being built on the incorrect notion of having to keep people sentenced for longer and longer stretches of time so they don’t re-offend as the BJS predicts most of them will. Now, we can take this policy advice to focus on reforming the chronic, most dangerous offenders to keep as many of them out of jail for good as possible and rehabilitating minor ones who are only being sent to jail now because the judges and most of the public are convinced that they’ll keep going right back to crime unless they’re jailed.

# politics // crime / prison / scientific method / statistics

  Show Comments