A crossover pattern similar to that reported for inpatient schizophrenics (e.g., Bellissimo and Steffy 1972) has now been found for outpatients, and its frequency of occurrence in these outpatients is comparable to that reported for hospitalized patients by DeAmicas and Cromwell (1978). The process-reactive construct was unrelated to crossover in the current study, although crossover is found only in process patients in hospitalized samples. The difference in findings may be due to the insensitivity of the particular measure of this concept used here, or to the lack of validity for the construct in samples of chronic patients in the community and in treatment. The present results, together with DeAmicas and Cromwell's (1978) data suggestive of the presence of crossover in nondisturbed first degree relatives of schizophrenics, are consistent with an interpretation of this reaction time phenomenon as a behavioral marker of vulnerability. This hypothesis warrants study in a more extensive series of patients, and from a longitudinal perspective in which behavioral measures and clinical-social functioning variables are tracked across time.