What’s the appropriate role of our prison system? Depending on who you talk to, it’s supposed to function as punishment for criminal activity, a deterrent to future crimes, and an opportunity for rehabilitation. It’s often possible to find people arguing that an existing prison system is already playing more than one of these roles, which raises questions about how well we understand a system that US society has committed to in a big way.
Fortunately, some researchers decided to view this question as an opportunity and put some hard numbers to what, exactly, our prison system is doing. Using a data set covering more than 100,000 convicted criminals, the researchers compared the outcomes of people sentenced to prison and a similar population that was given probation instead. The results suggest that prison does limit future violent crime by keeping criminals out of the general population, but the experience of prison provides little deterrence for future crime.
Violence in Michigan
A team of social scientists had access to data on everyone who committed a felony in Michigan between 2003 and 2006. This included follow-up data running through 2015, allowing the scientists to track whether any of this population committed additional crimes.
Michigan provides an interesting opportunity because its judges aren’t limited by the state’s sentencing guidelines—the guidelines are meant to be advisory, not mandatory. This means that people convicted of similar crimes under similar circumstances could potentially receive very different sentences depending on a judge’s inclinations. And, in fact, the team’s statistical analysis showed that many judges consistently deviated from their peers in handing down harsher or more lenient sentences. This allowed the scientists to use a random aspect of the process—the assignment of a judge—as a tool to look at the impact of sentencing on the convicted population.
This doesn’t entirely get rid of the issue that the details of these sentences depended in part on the circumstances of the crime itself. But it does allow the researchers to mathematically control for that and focus on the influence of the judge’s sentencing, specifically the parole versus prison decision. This let them get at their real area of interest: the impact of imprisonment on future violence.
As they note, there are three ways that imprisonment can have an effect on convicts. One is deterrence, as time in prison may cause people to avoid actions that could put them back into jail. A second is what they call “incapacitation,” meaning that people in prison are less able to engage in violent behavior toward the rest of the public. And finally, there’s the possibility that rehabilitation in prison will allow those released to better integrate into society afterward. Of course, the researchers acknowledge, time in prison could also enhance future violent behavior, from exacerbating mental health problems to disconnecting people from their communities to immersing them in a violent environment.
So what does the data tell us about what happens in prison?
Isolation but not deterrence
The authors tracked arrests and convictions at one, three, and five years after both sentencing and after those who were sentenced to prison were released.
There was a pretty substantial difference between the two. If you simply include the post-sentencing periods, being sentenced to prison lowered the risk of participating in violent crime. And the impact was noticeable: “For example, at five years since sentencing, a prison sentence reduced the probability of arrest for a violent crime by 8.4 percentage points.” That says that prison is serving at least one of its proposed functions: by isolating those convicted from the general population, it’s keeping the general population a bit safer.
If you shifted to examining post-release time among those sentenced to prison, however, then things were a bit more complicated. In raw terms, people sentenced to prison were more likely to be arrested for and convicted of additional crimes. But once additional factors were controlled for—like eliminating parole violations that people who weren’t in prison couldn’t commit—this difference vanished. There was no statistically significant effect of imprisonment, suggesting that it did little to deter people from further criminal activity.
Separating out whether the original sentence had been for violent or non-violent crimes made no difference. But, because of things like parole violations, being sentenced to prison made it 20 percent more likely that someone would end up back there during the study period.
The study has some limits, which the researchers were more than happy to detail in their paper. Michigan, as they note, may not accurately reflect conditions elsewhere, as the state’s Rust Belt economy has struggled to provide opportunities for its citizens. The researchers also suggest that prison release dates may be influenced by in-prison behavior, i.e, convicts with more violent tendencies tended to be kept in prison and not influence the five-year statistics.
It also, quite reasonably, does not provide answers on all the questions people have about the function of prisons in US society. But by putting some hard numbers on what prisons are actually doing currently, it has the potential to allow the debates over that function to be somewhat more informed.