Length of Incarceration and Recidivism is the seventh publication in the Commission’s recent series on recidivism. This study examines the relationship between length of incarceration and recidivism, specifically exploring three potential relationships that may exist: incarceration as having a deterrent effect, a criminogenic effect, or no effect on recidivism. (Published April 29, 2020)
- Full Report
- Key Findings
Other Recidivism Reports:
- Recidivism Among Federal Firearms Offenders (June 2019)
- Recidivism Among Federal Violent Offenders (January 2019)
- The Effects of Aging on Recidivism (December 2017)
- The Past Predicts the Future: Criminal History and Recidivism of Federal Offenders (March 2017)
- Recidivism Among Federal Drug Trafficking Offenders (February 2017)
- Recidivism Overview (March 2016)
The Commission consistently found that incarceration lengths of more than 120 months had a deterrent effect.
- Each of the research designs estimated that offenders incarcerated for more than 120 months were less likely to recidivate eight years after release. In the two models with the larger sample sizes, offenders incarcerated for more than 120 months were approximately 30 percent less likely to recidivate relative to a comparison group receiving less incarceration. In the third model, offenders incarcerated for more than 120 months were approximately 45 percent less likely to recidivate relative to a comparison group receiving less incarceration.
In two models, the deterrent effect extended to incarceration lengths of more than 60 months.
- Specifically, offenders incarcerated for more than 60 months up to 120 months were approximately 17 percent less likely to recidivate relative to a comparison group sentenced to a shorter period of incarceration.
For incarceration lengths of 60 months or less, the Commission did not find any statistically significant criminogenic or deterrent effect.
- When focusing on the shortest period of incarceration studied (12 to 24 months), the research designs yielded varying results, neither of which were statistically significant nor sufficiently reliable to make evidence-based conclusions.