Fifteen schizophrenics and 15 age-matched controls were compared on 3 auditory event-related potential (ERP) paradigms that elicited a variety of components. In one paradigm, tones were given at 0.75, 2.25 and 6.75 sec interstimulus intervals; in another, infrequently occurring targets in a reaction-time task were interspersed with frequent background stimuli; and, in a third, noise bursts or tones were delivered in a random sequence at either 70 or 100 dB SPL. The sensitivity of some of the ERP components in distinguishing schizophrenics from controls depended on the conditions under which the component was elicited. N1 amplitude was smaller in the schizophrenics than in the controls after longer interstimulus intervals. P2 amplitude was smaller in the schizophrenics than in the controls after longer interstimulus intervals. P2 amplitude was smaller in the schizophrenics only at higher stimulus intensities. P2 latency was shorter in schizophrenics except in the paradigm that varied interstimulus intervals. P3 amplitude, however, was much smaller in schizophrenics than controls ragardless of whether P3 was elicited by targets in a task or was elicited by 100 dB SPL stimuli. The loud stimuli also elicited blink reflexes that coincided with N1, but these reflexes did not vary by clinical group. Neither the amplitude of the slow wave following targets nor the sustained potential that accompanies prolonged auditory stimuli differed between schizophrenics and controls.