Abstract
In U.S. presidential elections, voter turnout dropped about 14% between 1960 and 1988. This drop in turnout has inspired many reform proposals to increase turnout. One such proposal allows voters to register on election day. Prior research has estimated the impact of day of registration voting using multivariate models with the individual as a unit of analysis; however, day of registration voting is no longer a hypothetical reform since it has been implemented in three states. A pre-post quasi-experimental design revealed that states changing their state laws to allow day of registration voting significantly increased their turnout compared to a control group of states that did not allow day of registration voting. The author estimated that a nationwide law allowing day of registration voting would increase turnout by 5%. The increased electoral participation found for day of registraton voting held for both presidential elections and midterm elections. However, the turnout increase for presidential elections was more stable across different elections than the turnout increase for midterm elections.

This publication has 10 references indexed in Scilit: