Alabama Sentencing Commission: Data Analysis and Simulation Enhancement

• NCJ Number: 225390
• Author(s): John Speir; Lynda Flynt; Bennet Wright
• Date Published: 2008
• Length: 66 pages
• Grant Number(s): 2004-DD-BX-1014
• Agencies: NIJ-Sponsored
• Publication Type: Program/Project Description

Annotation

This report reviews statistical simulation models used by the Alabama Sentencing Commission to forecast correctional populations and assesses the advantages of a microsimulation model and how the Commission applied this forecasting tool.

Abstract

Microsimulation is designed to mimic the flow of offender populations over the course of a specified timeframe. This is achieved by culling historical data and reviewing trends in the criminal justice system, while adjusting the underlying assumptions of the model. Microsimulation enables users to test "what if" scenarios by altering actual or proposed policy and practice changes that influence the path of individuals through the criminal justice system. The Alabama Sentencing Commission's decision to use a microsimulation model to project correctional populations was based on the model's flexibility in incorporating anticipated changes; the Commission's access to accurate, detailed individual offender records; and the ability to incorporate core assumptions. The development of the simulation model was undertaken in a three-stage process. The first stage involved the development of a baseline projection of current practices for later comparison with projections made following implementation of the sentencing standards. The second stage incorporated the initial sentencing standards into the simulation model; and the third stage integrated disparate modules together into a user friendly model interface. The development of the simulation model along with the hiring of an in-house statistician has been used by the Commission to develop initial sentencing standards, as well as to produce impact projections to the Governor and the legislature. In the course of this process, data-collection and data-entry problems were resolved. Future projects include developing and implementing "truth in sentencing" standards based on time-served data; incorporating the new classification system into the simulation model; and expanding the simulation model to include alternative sentencing programs. 6 figures, 13 references, and appended supplementary information