Beschreibung:
<jats:p> Internal validity refers to the confidence with which changes in a dependent variable can be isolated from the effects of extraneous factors and causally attributed to the effects of an independent variable. While a randomized design is commonly recommended as the best way to preserve internal validity, random assignment of subjects to groups is not always feasible in evaluations of correctional programs. This paper argues that the biased assignment design, whereby subjects are assigned to groups according to controlled criteria, is a viable alternative means of preserving internal validity which should be employed more frequently by corrections researchers. The steps for using the biased assignment design are presented, as is a strategy for proper data analysis. </jats:p>