Snap elections, often triggered by government dissolutions or leadership resignations, increase political uncertainty and disrupt voters’ decision-making processes. In this paper we investigate how such heightened political uncertainty necessarily leads to an increase in uncertainty for voting estimates through a rise in the share of “don’t know” answers and item nonresponse in vote intention questions. We also examine the durability of such phenomenon.

We analyze cases from Germany, Spain, and the UK – representing diverse political systems and different timings of snap election calls. Using individual-level panel data from the German Longitudinal Election Study, the Spanish Political Attitudes Panel and the British Election Study we causally identify a significant increase in “don’t know” responses and item nonresponse in vote intention questions following snap election announcements. To test the robustness of these findings and also to shed light on the durability of the observed effects, we employ an unexpected events during survey design (UESD) approach based on country-specific cross-sectional data fielded around each snap election call. The rapid convergence of uncertainty levels of voting estimates to pre-event baselines underscores the transient nature of this effect. The consistency of findings across countries may well highlight the systematic nature of the observed effects.

These results have methodological implications, highlighting the need to account for temporary estimation uncertainty following disruptive political events. Politically, our results emphasize that while snap election calls do introduce higher uncertainty to decision-makingvote-choice formation processes, voters’ adaptability is high enough for this effect to diminish by election day.