NCJ Number
137284
Journal
Journal of Quantitative Criminology Volume: 8 Issue: 2 Dated: (June 1992) Pages: 189-215
Date Published
1992
Length
27 pages
Annotation
This study developed a prison-population-projection model that extends previous work by considering the impact of limited prison capacity on time served, releases, and future admissions.
Abstract
The authors continued the demographically disaggregated projection approach of Blumstein et al. (1980) but adapted their model to consider explicitly the effect of recidivism and capacity constraints on prison population forecasts. The model was demonstrated for the State of North Carolina. Forecasts were made for males in 14 age/race classes. Commitments were assumed to be generated by first-time commitments, produced from estimates of future values in the general population, and recidivist commitments, generated by predicting failures from the population of prison releases. A split-lognormal failure model, which allows estimation of the timing of recidivism, was used to predict failure. The capacity constraint was incorporated into the model by assuming that mean time served was a function of available capacity. The model was more successful at predicting commitments than prison population. This result is not unexpected, since the prediction of population is sensitive to assumptions about the sentencing and time-served distributions. What is unique to this model is its ability to generate additional commitments as a result of early release policies engendered by crowding. The model reveals the tradeoff that is inherent among prison capacity, degree of punitiveness as reflected in time served, and future criminality. 4 tables, 6 figures, and 33 references