On the Use of the Total Incarceration Variable in Sentencing Research

• NCJ Number: 205893
• Journal: Criminology Volume: 42 Issue: 1 Dated: February 2004 Pages: 211-240
• Author(s): David Holleran; Cassia Spohn
• Date Published: February 2004
• Length: 30 pages
• Publication Format: Article
• Publication Type: Report (Study/Research)
• Language: English
• Country: United States of America

Annotation

This study demonstrates the limitation and implications of measuring sentencing severity by means of a "total incarceration" or "in/out" response variable.

Abstract

The paper first develops the argument that the populations of prisons and jails in the United States are distinctive; i.e., offenders sentenced to prison have qualitatively different characteristics than those sentenced to jail. Also, the collateral consequences of a prison sentence are more severe than the consequences of a jail sentence. The study itself focused on a single Pennsylvania county with sizeable Hispanic populations, so as to have sufficient data on Hispanic as well as White and Black offenders. The unit of analysis was the individual offender. In all of the multivariate models, the study included controls for legally relevant case characteristics and for offender characteristics. The question the study sought to answer was as follows: "In analyzing the decision to incarcerate or not, is it appropriate to combine prison sentences and jail sentences into a single response category?" The decision to incarcerate was modeled by using the total incarceration response variable (prison/jail versus probation). The results of this analysis were then compared to the results of a series of regressions that used a trichotomous disposition response variable, i.e., prison, jail, and probation. This was followed by a test to determine whether prison and jail could be combined by using a simple test provided in most conventional statistical software. Study findings suggest that jail and prison sentences, which this study contends are qualitatively different types of punishment, should not be combined into a single response category, since their correlates are different. Moreover, combining jail and prison sentences masks important differences in the types of sentences imposed on Black, White, and Hispanic offenders, on male and female offenders, and on offenders convicted of different types of crimes. Separating jail sentences from prison sentences thus enhances understanding of the sentencing process and the factors that affect the sentences imposed. 4 tables, 1 figure, and 29 references