A probabilistic association task that manipulated the necessity to temporarily store information was combined with the recording of event-related potentials. In Experiment 1, scores obtained from a positive schizotypy scale were used to categorize participants as either low or high schizotypal individuals. Low, but not high, schizotypal individuals displayed evidence for selective associative learning in the working memory dependent (“trace”) version of the task. The amplitudes of the occipito-temporal N150 were attenuated in high schizotypal individuals. In Experiment 2, the intravenous administration of a single dose of haloperidol (0.04 mg/kg), but not of a placebo, strengthened the selectivity of associative learning in the trace version of the task. The amplitudes of the occipito-temporal N150 were augmented by haloperidol. Psychometrically identified schizotypal individuals and normal individuals under mild stress demonstrated defective prioritization of information in working memory and deficient visual encoding. These neurocognitive effects of schizotypy and stress seem to be mediated by the D2 family of dopamine receptors.