NIJ researchers examined the impact of the rate of crime prior to prison and how prison affected crime post release. The method was applied to the same datasets used by the Bureau of Justice Statistics for its special report, Recidivism of Prisoners Released in 1994. NIJ's researchers found that—
- Criminal history prior to incarceration reliably predicted whether or not incarceration would deter reoffending within three years after release.
- For 56 percent of the offender sample, incarceration had the predicted deterrent effect (that is, they did not recidivate within the three-year period).
- Forty percent of the sample reoffended as predicted from their criminal history before incarceration.
- For a small percentage of offenders (4 percent), incarceration had a criminogenic effect, increasing the rate of crime after release from prison.
- Supervision after release did not seem to lower likelihood of re-arrest.
These findings suggest that an analysis of criminal history prior to incarceration may help corrections practitioners identify who is and is not likely to be deterred from post-release reoffending. 
Effects of security level assignment in prison. Researchers have established a relationship between the security level inmates are assigned during incarceration and the recidivism rate after they are released from prison.
The theory is that differential placement may affect post-release crime rates (recidivism) but not necessarily as intended. Higher security prisons are more punitive and, therefore, should decrease recidivism among inmates who have equivalent propensities to commit crime. Research shows, however, that being exposed to inmates who have higher propensities to crime may increase criminal behavior or reinforce antisocial attitudes. 
[note 1] Studying the Effects of Incarceration on Offending Trajectories: An Information-Theoretic Approach, by A.S. Bhati, July 2006, NCJ 216639.
[note 2] For example, see "Do Harsher Prison Conditions Reduce Recidivism? A Discontinuity-Based Approach," by M.K. Chen and J.M. Shapiro, American Law and Economics Review 9 (1) (2007): 1-29, and The Relationship Between Security Level Placement, Prison Conduct and Post-Release Recidivism, by G.G. Gaes and S.D. Camp.