Emotional memories are useful for future decision-making. Despite their utility, memories for details about past emotional events, and memories for the emotions experienced during the event, are susceptible to bias over time. Whether emotion enhances or hinders memory depends on a number of factors, including the valence of the emotions experienced. The present dissertation implemented a novel emotion induction technique designed to test memory accuracy in a controlled yet ecologically valid situation: the positive Trier Social Stressor Task (TSST). Instead of giving a speech about their qualifications for a job in front of stern evaluators, participants in the positive condition discussed a highly enjoyable hobby for warm and supportive observers. A neutral TSST condition was also developed. The new TSST conditions, which were compared to the original negative TSST, were validated in Study 1 and replicated in Study 2. Participants in the positive condition increased in positive emotion from baseline to post-speech, participants in the neutral condition remained in a stable affective state, and participants in the negative condition decreased in positive and increased in negative emotion. The intensity of emotions was similar in the emotional conditions. Study 2 tested memory accuracy for the details of, and emotions experienced during, the speech task by giving participants a surprise memory test one week after their initial visit. The influence of memory bias on intentions for future behavior was also measured. Memory accuracy was similar across groups over a number of indices except one in which the negative group performed worse than the positive and negative conditions. The negative group overestimated negative and underestimated positive emotions experienced during the TSST when recalling them one week later. Regardless of condition, the more positive and the less negative participants recalled feeling during the TSST, the more willing they were to participate in a similar future study even after statistically adjusting for emotions experienced during the TSST. This dissertation contributes a new emotion induction technique that can be used in a variety of studies. The lack of group differences in memory for emotional detail and the importance of emotional memory for decision making is discussed.

Memories for emotional experiences are influenced by a person's mood at the time of recall. This makes evaluating memory accuracy particularly challenging when asking about the daily experiences of people with a history of depression. People who report depressive symptoms typically experience negative moods and report more negative life experiences. As a result, it is unclear the extent to which their memories reflect a negativity bias based on their negative mood at the time of recall, or instead accurately represent their prior experiences. Using data from a daily diary study (N = 1,657), the association between meeting criteria for depression in the past year and bias in memory for positive and negative emotional experiences that occurred during the prior week was examined. It was further tested whether trait negative affect (TNA) and trait positive affect (TPA) explained these associations. Overall, people recalled more negative emotional experiences from the past week than they had indicated in daily reports. Recalled negative emotions, but not positive emotions, differed more from daily reports for those with a recent history of depression. TNA fully explained this association, and both TNA and TPA influenced how emotions were recalled. Specifically, higher TNA was related to greater overestimation, and higher TPA was related to less overestimation, in recalling negative emotions.