Folk wisdom counsels voters to stick with leaders in time of a war, though recent experience suggests otherwise. Taking advantage of largely unexplored polls from the 1940s (Gallup and National Opinion Research Center), this research probes vote choices in the World War II elections. Of particular interest is a counterfactual question that asked respondents how they would have voted if there was no war. In addition, the aggregate vote in presidential contests over time is used to estimate how the White House party fares in wartime elections. President Roosevelt received a special premium in electoral support from the wartime condition in both 1940 and 1944. This premium was earned through popular support for war-related issues. Hence, in wartime elections where such support is lacking voters may be inclined to change horses.