Statistical Difficulties in Determining the Role of Race in Capital Cases: A Re-analysis of Data from the State of Maryland* |
| |
Authors: | Richard Berk Azusa Li Laura J Hickman |
| |
Institution: | (1) Department of Statistics, UCLA8125 Mathematical Science Building, 405 Hilgard Avenue, Los Angeles, CA 90095-1554, USA;(2) RAND Corporation, 1776 Main Street, Santa Monica, CA 90401-2138, USA |
| |
Abstract: | In this paper, we re-analyze data used to study the role of race in capital cases in the state of Maryland. We show that when
alternative, and arguably more appropriate, statistical procedures are applied, the racial effects reported in early work
turn out to be very fragile. The methodological point is more general: conventional causal modeling with observational data
is not likely to produce robust results for a variety of criminal justice applications.
Raymond Paternoster provided us with the data on which this paper is based and made a number of very helpful suggestions on
a earlier draft of this paper. We very much appreciate both. The research was supported by a grant from the National Institute
of Justice. |
| |
Keywords: | race death penalty logistic regression propensity scores classification and regression trees random forests |
本文献已被 SpringerLink 等数据库收录! |
|