Electoral College, 1804 – 2008: Winner's Propensity and Electoral Entity Vote Size

Abstract

The Electoral College two electoral vote add-on and winner-take-all methodologies are generally thought to give less populous and more populous states an advantage in U.S. presidential elections. This study examined whether electoral entities with the least and most electoral votes were more likely to have cast their votes for the winning candi- date; and if so, whether this winner's propensity has determined the outcome. In the 52 U.S. presidential elections from 1804 though 2008, electoral entities with the least (Quartile 1) and most (Quartile 4) electoral votes were significantly (p < 0.001) more likely to have cast their votes for the winning candidate than were electoral entities in the two intermediate electoral vote quartiles (Quartiles 2 and 3). However, in the 17 close U.S. presidential elections (defined as when the wining candidate received less than 60 percent of the total potential electoral votes), electoral entities in Quartiles 1 and 4 were not significantly more likely (p = 0.7339) to have cast their votes for the winner than were electoral entities in Quar- tiles 2 and 3. Although electoral entities with the least and most electoral votes were significantly more likely to have voted for the winner in all U.S. presidential elections, this propensity was present only in landslide elections and was not present in close Electoral College elections. These findings suggest that the net effect of the Electoral College has not been to give the least and most populous states an advantage in determining the winner in U.S. presidential elections since any such winner's propensity was only observed in landslide victories.