Voter Turnout in Presidential Elections

This table last edited on 01/22/2021.  Corrected registration figures.

“Turnout” refers to efforts to measure the extent of popular participation in elections.  Turnout is usually discussed as a ratio although always based on a count of votes cast.  The numerator is the number of votes cast.  Various measures may be used as the denominator:  (1) The Voting Age Population—broadly speaking it is the population above the legal voting age;  (2) Voting Eligible Population—all citizens who are not excluded from voting because of some legal impediment;  (3) Registered voters.  Reported measures of each of these has varied somewhat over time as estimates have been revised and refined.

Three propositions underlie most research on turnout. 

• First: turnout may be a way of assessing the health of a popular democracy.  Well-functioning democracies are more inclusive and will have higher turnout. 
• Second:  Ease of registration should affect turnout.  In comparing two jurisdictions over time with comparable demographic caracteristics (education, age, income, etc.), turnout should be higher in the one with less restrictive registration requirements.
• Third:  Electoral competition should drive up turnout.  Other things equal, when the stakes in the election seem greater, turnout should increase.

Voting Age Population is typically calculated based on census data (“resident population [21 or 18] years and older”).  But before 1920 the numbers used are always adjusted for the shifting definition of citizens with voting rights.  So women are excluded prior to 1920.  The classic attempt to define the voting age population for the 19^th century is by Walter Dean Burnham, “The Turnout Problem” in Elections American Style ed., Reichley (Brookings:  Washington D.C., 1987)  Burnham published only the turnout ratio, not his actual estimate of the voting age population!

Voting Eligible Population is an attempt to make an even more precise definition of the population of people who have a legal right to vote—potential voters.  Making the estimates of noncitizens and disfranchized felons has been carried out mostly by Professor Michael McDonald and data are published in the U.S. Elections Project website. 

Registered voters counts the total number of eligible people who have taken the additional step of actually registering to vote.  We report here the estimates produced by the Census Bureau’s Current Population Survey. This revised table no longer reports registration totals before 1968, consistent with recent CPS publications.  Registration was not a universal requirement until “well into the twentieth century” (Ansolabhere and Konisky) and some states did not impose uniform registration requirements until the 1970s.  As of 2016, North Dakota had no registration requirement. So it is possible for the number of votes to exceed the number of registered voters. Scholars point out that census data may not be entirely accurate (see Bennett 1990).  A data source that may be of interest to many is the U.S. Election Assistance Commission which surveys county-level officials about voting and elections.  Among their data is a series reporting the total number of persons "registered and eligible to vote."  For 2016, the sum of the individual county numbers, for counties with data in the US is 185,714,229--a number 15% greater than the CPS estimate for the same year and election. 
We thank users who have taken the time to suggest specific modifications to our data on turnout:  Thomas Meagher and Phil Kiesling.

Number of votes cast in presidential elections is published by the U.S. House of Representatives, Office of the Clerk, Statistics of the Presidential and Congressional Election for various years starting with 1920.