> deterrent effects after release These studies are always done on people who were undeterred by the possibility of incarceration in the first place. It’s har

> deterrent effects after release These studies are always done on people who were undeterred by the possibility of incarceration in the first place. It’s hard to imagine a bigger selection bias. It’d be like testing the efficacy of a drug, but only on patients for whom it didn’t work the first time.