NCJ Number
106525
Journal
Police Studies Volume: 10 Issue: 2 Dated: (Summer 1987) Pages: 67-71
Date Published
1987
Length
5 pages
Annotation
Quasi-experimental analysis is a useful methdology in research situations in which units larger than individuals are being studied and where it is, therefore, impossible to achieve the benefits of randomization.
Abstract
Two subjects that are appropriate for this kind of research are the effects of new policing strategies and the effects of new laws. Penal policies could also be studied using quasi-experimental methods. Randomized experiments are inappropriate for these topics because too few geographic areas are involved to permit randomization and because the researcher does not have control over the independent variable. In contrast, the techniques of quasi-experimental analysis require researchers to draw tentative conclusions about the effects of a variable and then systematically to investigate alternative possible explanations of the effect that was observed. Among possible explanations for a study of the effects of policing strategies are that the results occurred by chance, that the areas selected had preexisting differences that affected the results, that some other change produced the effects, that the effects were the result of a preexisting trend, or that measurement problems produced the results. The generalizability of the results must also be considered. The Kansas City police patrol study and the study of the United Kingdom's breathalyzer law of 1967 illustrate the uses and problems of quasi-experimental analysis. 14 references.