We studied the interrater variability among three neurologists of the diagnosis of a seizure in 100 patients evaluated for a possible “first seizure.” We found that use of simple descriptive diagnostic criteria and discussion among the neurologists themselves improved the diagnostic agreement.