The voter turnout in 2020 was a stunning 67%, according to one source.
Another had it at 94%.
A third fixed 2020 voter turnout at 63%.
All three are correct — because they do the math differently. They’re comparing actual voters with the number of eligible voters, registered voters and Americans of voting age, respectively.
The turnout for 2022 is still coming into focus as the last ballots continue to be tallied, but it looks to be unusually high for a midterm election. Those turnout figures are typically discussed as evidence of voter interest (or apathy) in an election. But the way that voter turnout is calculated reveals who counts, and who is left out, in American elections.
Generally, turnout considers “the number of registered voters who actually get to the polls or send in their mail-in ballots,” said Khalif Ali of Common Cause Pennsylvania.
The tricky part is determining what figure to divide that number of votes by. Election officials and journalists often use the number of registered voters. (That’s not static: 22 states allow same-day registration. And since the 2020 election, Republican-controlled legislatures in some states have made it more difficult to register to vote.)
Academics who study elections favor other measures. One, used by the U.S. Census, is the “citizen voting-age population,” which it determines after each national election. That figure grew from 145 million in 1980 to more than 230 million in 2020.
But not all U.S. citizens of voting age are eligible to vote. So the University of Florida’s Michael McDonald calculates the “voting-eligible population” for each state, subtracting people barred from voting because they are in prison, have a felony on their record, or are on probation or parole.
That type of disenfranchisement varies wildly across the country due to state laws. In Georgia, it blocked 4% of the voting-eligible population from casting a ballot in 2020, while in Massachusetts the figure was 0.1%, according to McDonald’s estimates. Calculating turnout using the voting-eligible population is a more precise accounting of the participation of Americans, since it considers those who could register and vote, but haven’t.
By that measure, turnout was 67% two years ago.
But even that doesn’t tell the whole story.
“Looking at aggregate or overall turnout is certainly important, but considering who turns out to vote is just as important,” said Zoltan Hajnal, a political science professor at the University California San Diego.
In 2020, the electorate was older, whiter, more female and more educated than the country as a whole. That’s a consistent pattern in American elections, even as the nation has become more diverse.
A slew of new laws that made voting more difficult in states that include Georgia, Iowa and Florida add to the challenge of addressing those disparities. The bills, passed in the last two years, have targeted methods disproportionately used by people of color and younger, more Democratic-leaning voters. The measures attacked mail voting, non-governmental funding for elections and ballot drop boxes.
For all the hand-wringing over national turnout, the figures are dramatically worse at the local level, where mayoral races see less than 25% of potential voters actually cast a vote. But there are ways to boost that. Hajnal found that in California, timing local elections to coincide with national ones led to a larger and more representative electorate, boosting turnout among Latino voters, Asian voters, and voters with less family wealth.
“By simply shifting the date of the election, we can basically double or triple participation in local democracy,” he said.
One other factor looms large over discussions about turnout: It’s much more difficult to register to vote in the U.S. than in other countries. That’s no coincidence. The Center for Public Integrity found that many Republican-controlled states have made it harder to register and vote since 2020.
“That additional hurdle lowers turnout in the United States relative to other countries by a significant margin,” Hajnal said.
Nations like Taiwan, Brazil, Sweden and Mexico all see higher shares of their voting-age population turning out in elections than the U.S., according to Pew Research Center.
Turnout takes time
The most intense interest in voter turnout comes on election night and the days after. But that’s a poor time to gauge turnout in states with widespread mail voting and postmark deadlines.
In those places, counting can take weeks.
Two days after the 2022 election, officials in the nation's largest local election jurisdiction, Los Angeles County in California, announced they had over 1 million ballots left to process. Those were a mix of vote-by-mail ballots, ballots cast by voters who registered "conditionally" and ballots that were damaged or required closer inspection from officials.
And some ballots were still on the way at that point, since California allows those postmarked by Election Day to arrive as late as seven days after the election.
The number of yet-to-be-processed ballots means that turnout can look lousy for several days after the election — and then steadily climb in the days that follow. L.A. County’s turnout climbed more than 20 percentage points two years ago, as more and more ballots were counted.
Turnout that initially appears weak but later grows is common in states with substantial amounts of mail voting. Local pundits bemoan the initial figures, then fail to point out the much higher final turnout.
The amount of time it takes all ballots to be counted does not affect only voter turnout and local media coverage — it has transformed into an explosive political issue. In 2020, it took four days for news organizations to call Joe Biden’s victory over Donald Trump in Pennsylvania. In the intervening time, as Trump’s lead shrank, the former president and other right-wing figures pushed false information about the election.
The delays happen, in part, because Pennsylvania is one of nine states that forbids election workers from processing ballots before Election Day. The state’s Democratic governor and Republican-controlled legislature haven’t been able to agree on a bill to change that. Most other U.S. states, including Florida, New Jersey and Wyoming, do allow ballots to be processed pre-election.
“Since we don't have that, then what is required of us is patience,” said Common Cause’s Ali. “I'm more concerned about an accurate count than I am a quick count.”