Emotional memories are vivid and lasting but not necessarily accurate. Under some conditions, emotion even increases people's susceptibility to false memories. This review addresses when and why emotion leaves people vulnerable to misremembering events. Recent research suggests that pregoal emotions - those experienced before goal attainment or failure (e.g., hope, fear) - narrow the scope of people's attention to information that is central to their goals. This narrow focus can impair memory for peripheral details, leaving people vulnerable to misinformation concerning those details. In contrast, postgoal emotions - those experienced after goal attainment or failure (e.g., happiness, sadness) - broaden the scope of attention leaving people more resistant to misinformation. Implications for legal contexts, such as emotion-related errors in eyewitness testimony, are discussed.

Memory for feelings is subject to fading and bias over time. In 2 studies, the authors examined whether the magnitude and direction of bias depend on the type of feeling being recalled: emotion or mood. A few days after the U.S. Presidential elections in 2008 and 2012, participants reported how they felt about the election outcome (emotion) and how they felt in general (mood). A month after the elections, participants recalled their feelings. The intensity of past emotion was recalled more accurately than the intensity of past mood. Participants underestimated the intensity of emotion but overestimated the intensity of mood. Participants' appraisals of the importance of the election, which diminished over time, contributed to underestimating the intensity of emotion. In contrast, participants' strong emotional response to the election contributed to overestimating the intensity of mood. These opposing biases have important implications for decision making and clinical assessment.

Elaborating on misleading information concerning emotional events can lead people to form false memories. The present experiment compared participants' susceptibility to false memories when they elaborated on information associated with positive versus negative emotion and pregoal versus postgoal emotion. Pregoal emotion reflects appraisals that goal attainment or failure is anticipated but has not yet occurred (e.g., hope and fear). Postgoal emotion reflects appraisals that goal attainment or failure has already occurred (e.g., happiness and devastation). Participants watched a slideshow depicting an interaction between a couple and were asked to empathise with the protagonist's feelings of hope (positive pregoal), happiness (positive postgoal), fear (negative pregoal), or devastation (negative postgoal); in control conditions, no emotion was mentioned. Participants were then asked to reflect on details of the interaction that had occurred (true) or had not occurred (false), and that were relevant or irrelevant to the protagonist's goal. Irrespective of emotional valence, participants in the pregoal conditions were more susceptible to false memories concerning goal-irrelevant details than were participants in the other conditions. These findings support the view that pregoal emotions narrow attention to information relevant to goal pursuit, increasing susceptibility to false memories for irrelevant information.

In their comment on our article on affective forecasting (Levine, Lench, Kaplan, & Safer, 2012), Wilson and Gilbert (2013) criticized the meta-analysis, proposed alternative explanations for the empirical studies, and concluded that the impact bias is alive and well. Our reply demonstrates that, irrespective of the exclusion of effects and selective recoding of effects recommended for the meta-analysis, the pattern of results remains the same: Study participants' forecasts are more accurate when they report their feelings about a focal event, or immediately after a focal event, than when they report their feelings in general after a delay. New analyses rule out individual differences and focalism as alternative explanations for the results of our empirical studies. These studies show that people can accurately predict the intensity of their feelings about events. People overestimate in predicting the impact of events on their emotional state in general, but clarifying the meaning of the forecasting question reduces the magnitude of this bias. We conclude that the impact bias, which encompasses overestimating the intensity of feelings about events and overestimating the intensity of feelings in general, is both dead and alive. The importance of predicting feelings about events for decision making and the reasons people predict some features of emotion more accurately than others are discussed.