Prediction is one research method of testing the validity of psychoanalytic hypotheses. It is a promising approach provided one does not attempt statistical analyses that are too sophisticated. It might be desirable to separate the roles of analyst, predictor, and judge of outcome to minimize the many existing opportunities for bias and distortion. Prediction should also rest on confirmatory evidence rather than on intuition. A chief aim is to assess the balance of defense and impulse and the influence of the environment on each. The most difficult aspect of prediction lies in the quantitative assessment of the balance of forces involved.

The present report concerns a four-year study based on some six hundred predictions of mood and behavior, including psychosomatic symptoms in the next psychoanalytic session. Four analytic patients, two with character disorders and