Objective: Abnormalities of the auditory P300 are a robust finding in patients with psychosis. The purposes of this study were to determine whether patients with a psychotic disorder and their unaffected siblings show abnormalities in P300 and N100 and to establish test-retest reliabilities for these ERP components. Methods: Using an auditory oddball paradigm, P300 and N100 latency and amplitude were acquired from 19 patients with a psychotic disorder, 28 unaffected siblings, and 37 healthy controls, on two separate occasions. ERP components were compared between groups, using multilevel random regression analyses. Intraclass correlations were used to determine consistency of ERP components between the sessions. Results: A delayed target N100 latency was found in unaffected siblings. Patients showed significantly delayed P300 latency and diminished P300 amplitude compared to controls. Most ERP parameters showed good test-retest reliability. However, patients did not show sufficient reliability for N100 latency for standard stimuli. Conclusions: The present study failed to find significant P300 abnormalities in unaffected siblings. However, N100 latency is delayed in siblings and can be reliably measured in all groups for target stimuli, suggesting that this component, rather than P300, may serve as liability marker. Significance: N100 latency is a promising biomarker for psychosis liability.