It seems nearly impossible to do this study ethically and also remove all confounding variables. The best way I could think of would be to compare the results o

It seems nearly impossible to do this study ethically and also remove all confounding variables. The best way I could think of would be to compare the results of sentences given by different judges or jurisdictions who have different tendencies to imprison for the same exact offense. The problem is that there’s not really any such thing as “the same offense.” If you talk directly to people who work in law enforcement they indicate sentencing is much less about the precise crime and much more about whether all involved perceive the offender as a “bad guy.” If the people involved (da, police, judge, jury) are in any way accurate in that assessment, it will stand to reason a person incarcerated will be more likely to reoffend. That’s what everyone was perceiving which is why they incarcerated the person in the first place.