Natural experiment of 111,110 individuals suggests imprisonment is ineffective at deterring future violence

New research published in Nature Human Behaviour provides evidence negligible public safety gains are made from imprisoning individuals for crimes like assault or robbery.

“Our research team had been working for some time on a project looking at various effects of imprisonment, but we realized that a lot of the public debate about criminal justice reform and the use of prison as a form of punishment was focused around the threat of violence and what to do with individuals previously convicted of a violent crime,” said study author David J. Harding, a professor of sociology and faculty director of the Social Science D-Lab at UC Berkeley.

“Although the effects of imprisonment on recidivism had been the focus of much prior research (e.g. see our PNAS paper), no one had looked carefully at violent crime recidivism specifically before. This is particularly important because almost half of those currently in prison have been convicted of a violent crime, and such individuals have usually been excluded from criminal justice reforms due to public safety fears,” added Harding, who is also the author of “On the Outside: Prisoner Reentry and Reintegration“.

The researchers examined the records of 111,110 individuals sentenced for felonies in the state of Michigan between 2003 and 2006. The convicted individuals’ arrests and convictions for violent crimes were tracked from the time they were sentenced through June 2015.

The researchers’ analysis focused on cases in which judges had the discretion to either sentence defendants to prison or to probation.

Imprisonment did appear to reduce violent crime relative to probation — but the effect was marginal. After their release back into the community, those sentenced to prison were no more likely to be arrested or convicted for violent crimes compared to those sentenced to probation.

“Sentencing someone who has committed a common violent felony like assault or robbery to prison rather than probation does not prevent as much violent crime as the public policy debate seems to assume,” Harding told PsyPost.

“Preventing one person who was previously convicted of a violent crime from committing a new violent crime within five years of their sentence requires imprisoning 16 such individuals. And even that effect is a short-term one, as it is entirely due to removing them from the community during the time they are in prison. Once they are released, they are no more likely than someone sentenced to probation to commit a violent crime.”

The study — like all research — includes some limitations.

“The data for this study comes from only one state (Michigan). Also, the study specifically focuses on people who were convicted of felony crimes that are eligible for probation (who could have been sentenced to either prison or probation). That is why I specifically mention assault and robbery above. The most serious violent crimes like murder or rape are not covered here,” Harding said.

“An important strength of the study is that it is based on a ‘natural experiment,’ so we can be much more confident that the effects of imprisonment (or lack thereof) are causal rather than correlational. The study used the random assignment of criminal defendants to judges to mimic a randomized experiment in a context in which a true randomized experiment would be unethical and impractical,” Harding added.

“Different judges are more harsh or lenient in their sentencing, even within the same county courthouse. The study compared defendants randomly assigned to harsher judges to those who were randomly assigned to more lenient judges.”

The study, “A natural experiment study of the effects of imprisonment on violence in the community“, was authored by David J. Harding, Jeffrey D. Morenoff, Anh P. Nguyen, Shawn D. Bushway, and Ingrid A. Binswanger.