In today’s elections, abundantly available polls inform voters what parties lead and what parties trail. This allows voters to accurately predict the likely outcomes of elections before the final results are in. Voters may react to these ex-ante election outcomes by shifting their votes either toward leading parties, often termed the "bandwagon effect" or toward trailing parties, often termed the "underdog effect". The published literature presents different perspectives on the strength of both effects and the underlying psychological processes. Three preregistered studies (total N = 1,424) test the psychological causes of both effects. Exploratory Study 1 relates differences in interpersonal, moral, strategic, and epistemic psychological factors to shifts in voting intentions before the 2019 Polish parliamentary elections. Results suggest that the bandwagon effect reflects a lack of political expertise, whereas the underdog effect reflects fairness concerns. To provide experimental evidence, Studies 2a and 2b manipulate these two factors in a simulated election design. The results confirm that low expertise increases the bandwagon effect and that fairness concerns increase the underdog effect.