Rorschach records of 109 unselected veterans who had received therapy were used to predict the results of the treatment. Psychologists experienced in the method did no better than chance in predicting outcomes. Combinations of various variables also failed to discriminate between the criterion groups; 99 Rorschach scores also failed to discriminate. Extensor M's, when appearing alone in a record, were associated with improvement in therapy. Rules for prediction obtained empirically and cross validated showed promise. In general, the authors conclude that "while the Rorschach may be used for prognostic purposes successfully by some psychologists, there is emphatic need for further empirical check." (PsycINFO Database Record (c) 2006 APA, all rights reserved)