New Paths in Corrections Evaluations - Controlling Disruptive Variables in Time-Series Designs

• NCJ Number: 89974
• Author(s): W Blass-Wilhelms
• Date Published: 1981
• Length: 26 pages
• Publication Format: Document
• Publication Type: Program/Project Evaluation
• Language: German
• Country: West Germany (Former)
• Note: Edited, expanded version of a lecture given at the Colloquium 'Methods of Corrections Research' on August 11, 1981 at the Criminological Research Institute in Hannover, West Germany

Annotation

Noting that unmeasured factors are always present in social science experiments and that they constitute an initial difference between control and experimental groups which may ultimately distort the changes being measured over time, this article discusses methods for mitigating such distortive effects.

Abstract

It examines the extent to which controls against disruptive variables have been implemented in West German evaluation studies of correctional programs to ensure both the internal and external validity of the experimental results. Arguments are presented for specifying such controls into the evaluation methodology early in the planning stages, since interpretations of correctional evaluations frequently overlook this methodological requirement for political reasons. Models for the implementation of true experimental designs are then presented and consequences of their application to corrections evaluation are discussed. Over 50 references are given.