Voter Turnout: Behind the Numbers

Posted by Haley Smith on December 23, 2015

Two friends sit in a New York coffee shop, casually discussing the upcoming presidential election. One friend says to the other, “I wonder if people will be more inclined to vote in this next election; last time only 53 percent of voters in New York turned out”. His friend replies, “I thought turnout was higher, I remember hearing it was something around 60 percent”.

In a large state like New York, the difference between 53 percent turnout and 60 percent turnout translates into roughly one million more (or less) people showing up to the polls. Needless to say, that’s quite a differential.

So which friend is right?

The answer is both. Like many social phenomena, voter turnout can be measured in different ways. In the above example, one friend quoted turnout as percentage of the number of eligible New York voters (“voting eligible population turnout”) and the other as the percentage of the number of registered voters in New York (“registered voter turnout”).

Without a label associated with turnout, it is quite easy to confuse the level of voter participation and make misleading inferences about our participatory health.




Registered Voter Turnout

For the most part, states report voter turnout as the percent of their registered voter list that voted on election day (“registered voter turnout”). Some states, such as Connecticut and Utah report their voter turnout as the percentage of “active” voter registration. Active voter registration refers to voters who have voted in at least two consecutive federal  election cycles prior to the current cycle.

In general, registered voter turnout produces inflated estimates of participation. Since voter registration in the great majority of the United States is opt-in (and not automatic, as in many other democracies), it requires citizens to take the initiative and put considerable effort into registering. Citizens who take the time to register are found to be those most committed to political participation and are more likely to vote (Highton (2004)). A measure of voter turnout as a percentage of registered voters shows higher turnout than other measures, since it is a measurement of citizens already inclined to participate. This inflation of voter turnout is even more true when we measure turnout as a percentage of active voters, as it is a percent of those who have already consistently voted.


Problems with Registered Voter Turnout

Reporting turnout in terms of registered voters is also confounded by state registration rules. For instance, states like Colorado, Connecticut, and Minnesota allow same-day registration. But most other states require voters to register weeks ahead of time in order to vote. Complicating registration data even more are the various voter identification requirements necessary to register.

In states like Florida, Indiana, and Kentucky, for instance, registration is restricted to those who are thinking about the election well in advance (and have internet access, means to get to the registration office, etc), as well as those who possess the necessary documents to satisfy voter ID laws (and the money to afford the necessary identification documents).

Several recent court cases (most prominently Crawford v Marion County Election Board 553 U.S. 181(2008)), and a slew of academic literature (Alth (2009), Cobb, Greiner, et al (2010), Barreto, Nuno, et al (2009) to name a few) have addressed the issue of voter ID requirements and its disproportionate impacts on poorer voters, particularly communities of color. Although the Supreme Court in Crawford upheld Indiana’s voter ID law as a legitimate means to prevent voter fraud, the Court acknowledged that strict voter ID laws disadvantage poor voters. Academic literature, largely stemming from the Court’s ruling in Crawford,further finds that voter ID requirements have disparate impacts on poor and racial minority voting sects. 

As Benjamin Highton (2004) finds, the most significant barrier to voting is registration. Where registration requirements are high (and a burden to many), registered voter turnout reflects the turnout of wealthier and typically higher educated Americans. This means registered voter turnout statistics reflect the turnout of a specific part of the population and greatly obscure the fact that there is a population of voters who are eligible to vote, but are unable to participate due to burdensome registration policies.   


Automatic Registration

Recently, momentum has built towards automatic voter registration. According to the Brennan Center, legislators in 18 states and D.C. have introduced some legislation pertaining to automatic voter registration. In 2015, Oregon and California passed automatic registration, where all eligible citizens with a driver’s license or state ID are registered to vote. Although automatic registration does not eliminate the burden of registration for those without a license or state ID, by increasing the pool of registered voters it does improve the accuracy or reflectiveness of voter turnout in terms of registered voters and the general population.

With automatic registration, however, turnout will most likely decrease; not because voters are less interested, but because the percent of the eligible population that is registered will become much greater. Certainly, there may be more voters casting ballots. Voters who could not vote in previous elections, because they did not register in time or they could not register, might increase the overall number of voters voting. However, the percentage of registered voters turning out could decrease since, under automatic registration, the denominator (the number of voters registered) is more expansive. 

The different policies, timelines, and requirements for registration means comparing registered voter turnout between states is not an apples to apples comparison. The move to automatic registration in some states further complicates comparisons of registered voter turnout. Ultimately, states with more restrictive requirements (such as early registration deadlines and stringent voter ID laws) will typically report misleadingly high turnout because those who clear the steeper barriers to voting are more likely to vote. States with more lenient registration requirements are likely to report misleadingly low turnout since their registered voter lists are more likely to incorporate a larger chunk of the population, since it is easier for people to end up on the voter register.  


infogram_0_registration_before_electionRegistration Before Election//

Voting Eligible Population Turnout

Measuring turnout through the voting eligible population (VEP) is one way to untangle the misleading statistics about participation that we end up with if we use registered voter turnout. Measuring turnout using VEP as the denominator allows turnout to be expressed as the percentage of the population that is eligible to vote. The voting eligible population is an estimate of the population that is over 18, a citizen, and has not been disenfranchised due to felony convictions.

Use of VEP turnout makes comparing participation between states more accurate. VEP turnout measures turnout as a percent of everyone in the population who could possibly (legally) vote, whether or not they have the means or inclination to do so. While different felon disenfranchisement laws muddy comparisons a bit, VEP more accurately reflects the level of democratic participation. In general, this makes VEP turnout a more meaningful measure of how many people show up to vote compared to registered voter turnout measures, and more indicative of our overall democratic health. The below table of Virginia turnout demonstrates how participation statistics change based on the measures used. As is commonly the case, the more inclusive the measure is of the overall population, the lower participation becomes.




International Context

Internationally, registered voter turnout is typically used to compare voter turnout in democracies across the globe. But, as IDEA International explains, there are complications with using registered voter turnout to compare voter participation across countries. These complications are similar to domestic American issues with registered voter turnout. For example, some countries like Australia and Italy have automated voter registration, while other countries have opt-in registration, meaning the denominator is taking in different samples of the overall population. This makes comparing voter turnout figures misleading, especially if comparing a country (like Australia) where voting is also compulsory for registered voters, against one where it is not.  

Adding to the complication, voting age population (VAP) is sometimes also used to compare turnout globally. Voting age population is an estimate of the total number of people who have reached the age of majority for voting (typically 18 years). It can lead to misleading comparisons  because it includes populations that are disenfranchised, including non-citizens. Some countries have a large non-citizen population (for example, 13 % of American residents are not citizens), while other countries have relatively few non-citizen residents. A measure of VAP turnout is likely to understate turnout in the former countries, while overstate it in the latter countries.  

VEP measures of turnout would fix inherent inaccuracies in international turnout comparisons. However, VEP is not available for quite a few countries.



Of the several different ways to measures participation, the VEP turnout measure is the best. It paints a more accurate picture of participation in the United States than using registered voter turnout (or VAP turnout, for that matter).

However, neither VEP or registered voter turnout is necessarily the wrong measure to use; both give important information about who is showing up to vote. Further, keeping track of both measures is important. Large mismatches between the two measures in a state can indicate that registration is proving too great a barrier to voting and  that laws need to be re-examined. Similarly, small differences in turnout might indicate that the barriers to registration in a state are low, and that state can perhaps serve as a model for other states.

Both measures can be considered useful, especially together. However, it is important to better qualify what voter turnout measure we are using to inform discussions of voter participation as accurately as possible.  


For a complete list of Registered Voter Turnout vs. Voting Eligible Turnout for the 2012 Presidential election, click here

Show Comments
comments powered by Disqus

Join Us Today to Help Create a More Perfect Union