Schizophrenic (n=21) and manic (n=19) patients were followed up an average of 8 months after an index assessment during an acute admission. These patients were tested at both assessments with laboratory tasks measuring distractibility and reality monitoring and were examined with clinical ratings of positive and negative thought disorder. For manic patients, none of the measures predicted the patients' clinical state of followup, while negative thought disorder, although rare, was temporally stable. For the schizophrenic patients, both negative thought disorder and distractibility were temporally stable, and more severe negative thought disorder was found at index assessment in patients who were psychotic at followup. The differential utility of laboratory and clinical indices for the prediction of overall clinical state is related to these data.