[77] ISSUES OF DESIGN AND ANALYSIS IN EVALUATION RESEARCH

This paper addresses the question of internal validity in quasi-experimental and nonexperimental social policy research. We focus primary attention on problems of selection in research designs where assignment to experimental conditions occurs on a nonrandom basis. Five different solutions to the problems of selection are discussed-randomization, covariance adjustment, gain scores, matching, and explicit selection-and we examine the conditions under which these solutions are useful. The problem of random measurement error is discussed within the context of our elaboration of these techniques. n recent years considerable attention has focused on the analysis of data in social policy evaluation where true experimental designs for research are either impossible or impractical. Analysts of the effectiveness of social amelioration programs frequently engage in debates over the issues involved in making policy inferences from data gathered using nonexperimental and quasi-experimental research designs (Bowle