NCJ Number
223599
Journal
Journal of Experimental Criminology Volume: 4 Issue: 2 Dated: June 2008 Pages: 83-115
Date Published
2008
Length
33 pages
Annotation
This paper explores problems in criminology field experiments and possible reasons for the questionable validity of results.
Abstract
The paper provides helpful suggestions to focus attention on experimental design issues with the purpose of promoting development of corrective strategies. These include: reporting and transparency; prior explanation or anticipation of attrition; and multimethod approaches and replication. As was noted, with the use of field experiments in criminology increasing, previous discussions have focused on methods for achieving randomization, correcting for flawed randomization, increasing power and addressing treatment implementation problems in the interventions under study. The paper also explores what may be a fairly common and basic challenge for field experimentation: the problem of attrition before random assignment in the targeted population--when an experiment may be well conducted, except that it ‘misses the target’ for a variety of possible reasons and produces experimental results that may be of questionable validity. To illustrate this basic, but seldom discussed phenomenon, the work drew on the examples of 13 field experiments by the author and colleagues as a body of case studies to assess the nature and magnitude of preassignment attrition and its possible implications for experimental findings. Table, figure, references