In experimental design, determining statistical power and qualifying sample results require a formal definition and statistical measures of the magnitude of experimental effects. Criticisms of existing statistical inference techniques are noted, but it is contended that such criticisms are due primarily to the face that statistical power and effect size receive little attention in research techniques evaluation. Although no separate studies dedicated to statistical power exist in criminal justice research, there is no reason to assume that research in this field is more sensitive to possible treatment effects. For example, reduced recidivism rates constitute a frequently used measure of success but one which has not proved to be very sensitive. Problems related to the practical application of statistical power and effect size analysis are discussed, along with alternative statistical procedures. An empirical example taken from an evaluation study of social therapy confinement is cited to show the advantages of statistical power and effect size analysis. 121 references, 1 table, 1 figure.