Data on variables often used in recidivism studies were gathered from state hospital records of 129 patients released within a two-year period to a three-county area in West Virginia. The variables were categorized as prehospital, inhospital, or posthospital; the relationship between each variable set and recidivism was determined. The three criteria of recidivism were readmission within one year following discharge, number of days in the community, and number of days to first readmission. Inhospital variables were found to be the best predictors for all three outcome measures. When the number of variables in the set was controlled for, the posthospital set and the inhospital set had the same predictive power for readmission within one year of discharge and number of days to readmission. Posthospital variables were poorest at predicitng, among recidivism criteria, the number of days in the community.