Three experiments investigated the effects of participants' mood during exposure to target information on delayed judgments of the target. Participants were exposed to a mood induction immediately before they acquired information about a political candidate and then reported their evaluation of the candidate at a later time. Effects of mood on judgment were moderated by 2 individual-differences measures that can be interpreted in terms of processing efficiency. These were political expertise and total recall for the candidate information, with higher scores on these indices interpreted as reflecting more efficient processing. Among low-expertise (or low-recall) perceivers, mood produced an assimilation effect on evaluative judgments. Among high-expertise (or high-recall) perceivers, mood produced a contrast effect on judgments. When pooling across these individual differences, mood exerted no influence on judgments. These findings are consistent with an on-line model of mood misattribution and overcorrection.