A Contribution to the Study of Actuarial and Individual Methods of Prediction

Are predictions of conduct more accurate when made by case-study methods than by actuarial methods? This study is an experimental rather than a polemical attempt to determine the relative accuary of the two modes of prediction. A restricted form of behavior-academic success-was predicted from clinical (case-study) material and from a previously derived regression equation. The correlation coefficients demonstrate that the case-study method which presumably accounts for an innumerable asortment of variables is no more accurate than a simple statistical method which accounts for only two variables. Analysis of the predictions suggests that the case-study method takes behavior segments with known predictive weights and applies other weights which are less efficient. Case-study predictions-at least on the grounds of efficiency-should not be substituted for actuarial predictions. As a complement to the actuarial predictions, the clinical predictions add nothing. The posibility of generelizing to the prediction of other kinds of critera is discussed.