| Abstract: | Supermax prisons have been advanced as means of controlling the “worst of the worst” and making prisons safer places to live and work. This research examined the effect of supermaxes on aggregate levels of violence in three prison systems using a multiple interrupted time series design. No support was found for the hypothesis that supermaxes reduce levels of inmate‐on‐inmate violence. Mixed support was found for the hypothesis that supermax increases staff safety: the implementation of a supermax had no effect on levels of inmate‐on‐staff assaults in Minnesota, temporarily increased staff injuries in Arizona, and reduced assaults against staff in Illinois. |
