U.S. flag

An official website of the United States government, Department of Justice.

NCJRS Virtual Library

The Virtual Library houses over 235,000 criminal justice resources, including all known OJP works.
Click here to search the NCJRS Virtual Library

Why We Have Not Predicted Prison Admissions Well and Some Suggestions on How to Do Better

NCJ Number
140515
Author(s)
F Cheesman
Date Published
1992
Length
59 pages
Annotation
This study identifies the shortcomings of previous efforts to produce long-term forecasts of correctional institutional populations and recommends causal models as the most effective approach.
Abstract
The determinants of institutional population are the number of inmates in the population at the time of the forecast, the rate of intake of new inmates, inmates' length of stay, and the rate of parolees returning to institutions. Of the aforementioned factors, the rate of intake of new inmates is the most difficult to predict. The primary problem is how to improve the accuracy of long-term (over 2 years) forecasts of juvenile institutional intake. This is a significant problem because it bears upon planning for programming, bed capacity, and the safety of staff and inmates. Previous efforts to produce long-term forecasts of institutional populations have involved demographic disaggregation, linear regression, other time series techniques (exponential smoothing and ARIMA models), and causal models. Causal models offer the greatest promise for improving long-term forecasts of juvenile correctional populations. The promising causal models are multiple regression, constrained regression, queueing networks, and simulation models (discrete event and continuous). Simulation models have the greatest flexibility in modeling policy and incorporating and using widely diverse data. 40 references