The belief that large prisons breed more violence and recidivism than smaller prisons may have been caused by the failure to separate the effects of size from the effects of other factors. For example, one study reported that the larger U.S. prisons tended to be maximum security institutions and that the security level of a prison was more closely related to the homicide rate than was its size. Furthermore, the larger prisons tended to be older, to have higher inmate to staff ratios, to be more overcrowded, and to be in urban areas. Unlike the analysis of prison offenses, an analysis of recidivism rates included a control for the kinds of inmates incarcerated in each institution. The results of the study of 2,000 men who were followed for 2 years after the release showed that prison size was not significantly correlated with correctional effectiveness. However, there was a high negative relationship between prison overcrowding and the recidivism rate. It was concluded that an important priority for governmental agencies should make the reduction of overcrowding in prisons an important priority. The use of a multivariate or partial correlation analysis to establish factors influencing behavior inside and after leaving prison, independently of the kinds of inmates assigned to it, was also recommended. Statistical and tabular data and references are included.