DESIGN SENSITIVITY IN CRIMINAL JUSTICE EXPERIMENTS (FROM CRIME AND JUSTICE: A REVIEW OF RESEARCH, P 337-379, 1993, MICHAEL TONRY, ED. -- SEE NCJ-146350)

• NCJ Number: 146356
• Author(s): D Weisburd; A Petrosino; G Mason
• Date Published: 1993
• Length: 43 pages
• Grant Number(s): 88-IJ-CX-0007
• Sponsoring Agency: Rutgers University; National Institute of Justice (NIJ); US Dept of Justice NIJ Pub
• Sale Source: University of Chicago Press
• Publication Type: Survey
• Language: English
• Country: United States of America
• Note: Research and Justice Volume 17

Annotation

The statistical concept of power, that is, the likelihood that a test will lead to the rejection of a hypothesis of no effect, is used in a survey that examines the design sensitivity of experiments in sanctions.

Abstract

The results do not support the assumption that sample size is unrelated to effect size. The power advantages of large studies are often offset by the more serious problems that researchers encounter when trying to implement treatment among a large sample as compared to a smaller one. However, the results of this evaluation also demonstrate that the design advantages of small studies do not offset the power disadvantages inherent in smaller study samples. Future efforts should focus on developing strategies that allow researchers to expand sample size while maintaining the integrity of treatments and minimizing variability. Some of these strategies might involve the use of random observations, case flow studies, or randomization within blocks. 6 tables, 32 notes, 144 references, and 1 appendix