Two independent groups of high-risk children for schizophrenia and their matched control children were submitted to the following experiments: an auditory oddball paradigm registrating late event-related potentials (ERPs) and a psychometric test battery including the assessment of Wechsler Intelligence Scales, reaction times (after regular and irregular preparatory intervals), and the d2-attention test. The study was intended to clarify whether long-latency ERPs and the selected psychometric tests would contribute to reliably differentiating between these groups. The results showed significantly prolonged latencies of the P3 component of the ERPs to rare, task-relevant target stimuli in both high-risk groups compared with the controls. Similarly, the N2 latencies were delayed in both groups. By contrast, ERP patterns to frequent, nontask-relevant stimuli were very similar, with no significant differences between high-risks and normals; nor did any ERP amplitudes show significant differences. The data are interpreted as a reflection of a subtle deficit in maintaining attention and a subsequent impairment of stimulus discrimination in high-risk children. This is consistent with the psychometric findings of higher error scores in target counts and d2-test, and significantly prolonged reaction times after regular preparatory intervals (Pls) in the high-risks. The findings may hint at a vulnerability for schizophrenia in high-risk children. Given the high prevalence of the attentional dysfunctions in both high-risk groups, however, it is hypothesized that their presence does not necessarily imply an unequivocal manifestation of schizophrenia.