Self-serving biases shape the relationship between future thinking and remembering of elections.

While there is a strong relationship between remembering and future thinking, it has been unclear whether this persists when constraining participants to one specific significant public event. We employed a unique longitudinal approach to uncover how the differences and similarities between remembering and imagining are influenced by self-serving biases evoked by the event itself. Across three longitudinal questionnaire studies testing participants before and after 2024 elections in Germany (N = 136), the UK (N = 89), and the USA (N = 243), we found evidence for self-serving biases in the congruence between future thinking and remembering. Election winners robustly remembered the election as more important and more vivid than they had imagined it before. In the US study, the inconsistency in attitudes across time caused by this shift was resolved by also misremembering the prediction given before the election, with Harris voters thinking they had predicted a less fair, and Trump voters thinking they had predicted a fairer election than they actually had. Additionally, there was an overestimation of pre-election optimism among Harris voters, possibly to help explain current feelings about the outcome, and an underestimation of optimism for Trump voters, making the win more significant. The results reveal that phenomenological differences between remembering and future thinking are contingent on self-serving biases and indicate that participants misremember previous future thoughts in accordance with current needs and attitudes. These mechanisms can lead to entrenched polarization, as partisan beliefs are reinforced by biased future thinking and remembering.