The purpose of this project was to determine if it is possible to predict from the information contained in the write-ups of the preliminary evaluations of patients accepted for Institute supervised analyses those patients who would successfully complete their analyses and those who would not. Three raters independently studied the write-ups of 183 cases treated at the Boston Institute during the period 1959-1966. Each rater filled out a questionnaire of 105 items which were thought to be of possible predictive value. Outcomes were sought from the analysts who treated the patients. 130 outcomes were obtained and revealed that the patients fell into 4 distinctly different outcome groups: (Group 1) those who completed analysis by mutual agreement between the analyst and patient; (Group 2) those who prematurely terminated their analyses against the advice of their analysts; (Group 3) those whose analyses were prematurely interrupted by their analysts; (Group 4) those whose analyses became interminable. Statistical analyses were done to see which of the 105 predictor items distinguished these four distinct outcome groups. Many of the predictor items were not useful because they either showed no variation among patients, or were too often left blank, or were rated with very low interrater agreement. Of the remaining items, we found ten items which did show a highly significant difference between Group 1 patients and those patients in at least one of the other three outcome groups. We found five other predictor items which showed large, although not quite statistically significant, differences between Group 1 and at least one of the other three groups. Of these 15 items, 7 dealt with family history. The others concerned the patient's past history (2), object relationships (2), patient's sex (1), symptomatic state (1), field of endeavour (1), and history of previous psychotherapy (1). We found it made no difference in terms of these 4 outcome groups what the patient's diagnosis was or whether he was a 1st, 2nd, 3rd or 4th supervised case. Among the other negative findings were patient's age and the ability to experience and tolerate felt anxiety. These did not distinguish any of the outcome groups. These findings show that additional predictively useful information is present in the preliminary evaluations of patients already screened and accepted for Institute analysis by trained evaluators. These results suggest that one particularly important area on which to focus future attention is family history.