People learn more from new information when it leads to favorable future outlooks and thus can maintain optimism despite conflicting evidence. In two studies (N = 20 and 26), we investigated whether this optimism bias in belief updating is self-specific by modifying a recently introduced learning paradigm. In each trial, participants had to estimate the probability of experiencing a negative future event, were then presented with the population base rate of that event, and were subsequently asked for a second, updated estimation. In half of the 88 trials with varying events, estimations were made for oneself, in the other half for a similar other. We tested whether the updates (differences between the first and second estimates) following undesirable base rate were lower than those following desirable base rates, and whether this difference was greater for self relative to other. In both studies, the overall results support the presence of a self-specific optimism bias in belief updating. However, taking into account trait optimism (TO) as a moderator variable revealed that this was the case only in participants with high TO, whereas those with low TO showed optimistic belief updating for both self and other. In Study 2, we additionally controlled for possible misclassifications of base rate desirability. Taken together, the optimism bias in belief updating was demonstrated by a selective neglect of unfavorable information. A self-specific influence of this bias in individuals with high TO may ultimately cause the impression of a more positive future outlook relative to others. Copyright © 2014 John Wiley & Sons, Ltd.