Political identification is the basis of enduring conflict, suggesting thatpolitical attitudes are difficult to change. Here we show that in the 2016 U.S.Presidential Election, political identities underwent modification in responseto salient political events. We investigate these dynamics in detail bycollecting data at periodic intervals from mid-June 2016 through the generalelection (N = 3,958). We operationalize identification using prosocial giving inDictator Games played between supporters of competing primary candidatesrecruited from Amazon Mechanical Turk. The observed dynamics differed acrosspolitical parties. In-group bias among Democrats remained high until theDemocratic National Convention, disappeared shortly thereafter, and thenreturned during the final stage of the election. Bias among Republicans wasgenerally high until the final days of the election. The late resurgence of biasamong Democrats was not reflected in voting intentions, but may have presagedthe Democratic election loss.

Research suggests people try to protect their social self-esteem from threats from the ingroup or the outgroup. However, how members react to a threat to social self-esteem from a third party remains unclear. Three studies were conducted to examine the influence of a threat to social self-esteem from an authoritative third party on ingroup favoritism. We explored the effect of negative (versus positive) evaluation from the testing system on explicit and implicit ingroup favoritism in Study 1 and Study 2 respectively. We compared the effect of negative evaluation posed by the testing system or the competitive outgroup on ingroup favoritism in Study 3. Results suggested that individuals experiencing a threat to social self-esteem from an authoritative third party manifested less ingroup favoritism than those experiencing no threat or outgroup threat. The theoretical implications of this research on social identity theory and the practical implications of reducing intergroup bias are discussed.