Key points are not available for this paper at this time.
A retrospective review was conducted of the 2003 disciplinary records of close-custody inmates ( N = 24,514) in the Florida Department of Corrections. The frequency of various forms of violent misconduct was inversely related to the severity of this prison violence. A logistic regression analysis controlling for other factors demonstrated that younger age, shorter sentence, prison gang affiliation, prior prison violence, and prior prison term were predictive of violent institutional misconduct. Conversely, older age, longer sentence, and violent offense of conviction were associated with lower rates of violent prison infractions. Actuarial models constructed from the logistic regression analysis were modestly successful in predicting institutional misconduct (area under the curve = .717 to .738, p < .001), whether predicting all infractions or violent misconduct of varying severity.
Cunningham et al. (Fri,) studied this question.