Studies like this are frustrating. The conclusion confirms my own biases, so I want to respect the study - but there are, invariably, issues that make it diffic

Studies like this are frustrating. The conclusion confirms my own biases, so I want to respect the study - but there are, invariably, issues that make it difficult to take their conclusions at face value. Put another way, it's entirely possible somebody looking at the data a different way could present an entirely incompatible conclusion. For instance an issue this study faced was that it's difficult to control for the fact that extenuating circumstances and other factors help determine the sentencing for individuals. A judge, presumably, will make some effort to attempt to sentence individuals who are less likely to reoffend to probation rather than prison, and vice versa. The study argued that the random assignment of judges to cases, each of whom has their own sentencing tendencies, works as a control here but I think that is less than compelling. A judge may be more or less lenient on average, but that does not change the fact that all judges are going to be aiming to achieve the same thing. You're not comparing apples to apples. It seems the only solution would be genuine random sentencing in a controlled experiment, but there are obviously certain ethical issues there. It's unfortunate since our system is completely broken. You create environments where individuals are treated like dehumanized animals and pushed towards greater degrees of sociopathy with no meaningful effort made towards rehabilitation. And suddenly after being in this situation for years, they're supposed to come out better than ever ready to integrate and become a productive member of society? Obviously the idea is that fear of being put in that circumstance is supposed to be a deterrent. But given an overall recidivism rate now of 76.6% [1] over 5 years, it should be self evident that this idea is quite obviously not working. [1] - https://www.bjs.gov/index.cfm?ty=pbdetail&iid=4986 https://www.bjs.gov/index.cfm?ty=pbdetail&iid=4986