ABSTRACT
The process and consequences of randomization are considered only with relation to the simplest possible situation, that of a continuous dependent variable, a dichotomous independent variable, and one dichotomous confounding factor. The critical problem for sociologists, then, in Professor Hanan C. Selvin’s view, rests in the fact that the experimental procedure of randomization is impossible in most of their research. It is clear that the act of randomization plays a crucial part in Selvin’s argument. Thus, to understand his position it is necessary to have some knowledge of the effects of randomization and of the possible consequences of failure to randomize. The physical process of randomization consists of assigning subjects to classes of the independent variable by some random process. In general, the claim that all statistical tests of all hypotheses require the experimental procedure of randomization is unwarranted. If Selvin had said no more than that randomization is desirable in sociological research, social scientists should have concurred enthusiastically.
