Problems in Making Policy Inferences

This paper is a methodological critique of the Coleman Report (James S. Coleman ~ M. Equality 9!. Educational Opportunity, U.S. Office of Education, 1966). The Cain-Watts criticism is directed towards the statistical methods used, ~ to th~ report's substantive findings. The principal theme of Cain and Watts is that the analytical part of the Coleman Report bas such serious methodological shortcomings that it offers little guidance for policy decisions. They show, first, that the specification of the theoretical model is inadequate--without a theore.· tical framework'to.provide order.and a rationale for the large number of variables, there is no way to interpret the statistical results. They show, second, that in those instances where the Coleman Report makes clear the justification for the use of a variable in the regression model, the criterion used to assess or evaluate the statistical performance of the variable (namely, its effect on the coefficient of determination, or R2, of the regression) is inappropriate. Cain and Watts then go on to suggest a more meaningful approach to the problem of measuring determinants of educational achievement for policy purposes. They show (1) how the role of a variable in affecting objectives~ be interpretable in the context of a carefully specjjied, theoretically justified model; and (2) that when such a model is in the form of a regression equation, an appropriately scaled regression coefficient is the most useful statistic to measure the importance of the variable for the purposes of policy action. ._----.._.~-_._~~ ---~-----Problems in Making Policy Inferences from the Coleman Report. Glen Cain and Harold W. Watts