The difficulties of estimating the outcome of any slow process of maturation apply equally to education and to all psychiatric therapies. Among the reasons for this is the fact that the number of variables to consider are so many as ultimately to require computer handling. Furthermore, any process of change which may be initiated during a period of active therapy or education may not become active until the period of direct intervention has terminated. Yet the problems of when to make an evaluation of the results of psychiatric therapy (whether repeatedly, during the course of, or immediately afterwards, or at variable and increasing intervals after it is over) have been almost wholly neglected. This communication illustrates this problem and its importance, but offers no solution.