Epilepsy histories obtained by the computer interview were similar in substance and detail to histories recorded by house staff and attending physicians. In general, more information was obtained by the computer, and negative findings were consistently recorded by the computer. For potentially embarrassing questions, the computer may obtain more accurate information (Slack and Van Cura, 1968b). This might explain the contradictions concerning incontinence and medication. This preliminary study suggests that computer interviewing can provide a reliable method for obtaining epilepsy histories. The study also pointed out where changes in question wording and branching logic would improve the computer interview. These changes are being incorporated into the second version of the epilepsy interview. We hope computer-based interviewing will prove useful to patients and physicians in the diagnosis, treatment, and study of epilepsy.