More American voters than ever say they are not religious, making the religiously unaffiliated the nation’s biggest voting bloc by faith for the first time in a presidential election year.

This marks a dramatic shift from just eight years ago, when the non-religious were roundly outnumbered by Catholics, white mainline Protestants, and white evangelical Protestants.

These numbers come from a new Pew Research Center survey, which finds that “religious ‘nones,’ who have been growing rapidly as a share of the U.S. population, now constitute one-fifth of all registered voters and more than a quarter of Democratic and Democratic-leaning registered voters.” That represents a 50 percent increase in the proportion of non-religious voters compared to eight years ago, when they made up just 14 percent of the overall electorate.

“In 2008, religious ‘nones’ were outnumbered or at parity with white mainline Protestants and white Catholics,” the survey’s lead researcher, Greg Smith, said in an interview. “Today, ‘nones’ outnumber both of those groups.”

The growth of the non-religious – about 54 percent of whom are Democrats or lean Democratic, compared to 23 percent at least leaning Republican – could provide a political counterweight to white evangelical Protestants, a historically powerful voting bloc for Republicans. In 2016, 35 percent of Republican voters identify as white evangelicals, while 28 percent of Democratic voters say they have no religion at all.

But while the religiously unaffiliated are making up a larger share of American voters, that doesn’t necessarily mean that will translate into actual votes. Exit polls of people who actually cast votes – as opposed to pre-election polls of registered voters – have traditionally shown that the unaffiliated underperform at the ballot box relative to their raw numbers.

For instance, in the 2012 election, the unaffiliated made up 18 percent of registered voters in pre-election polls, but only 12 percent of the people who actually voted, as measured in post-election exit polls. Some of this difference may be because of the different ways the two polls ask religious affiliation questions, but Pew’s researchers say that the underperformance of the unreligious is a real phenomenon.

“While the group is growing rapidly in the general public, its growth has been much less dramatic in the electorate,” Pew’s Greg Smith said. “It could be the ‘nones’ are not connected, almost by definition, to religious institutions which can play an important role in spurring turnout and interest in politics.”

Smith also points out that the unaffiliated tend to be younger than the religious, and that young people tend to vote less than older people.

Still, the Pew study finds other evidence that religion may be a less potent force at the ballot box. In 2008, 72 percent of voters said it was important for a president to have strong religious beliefs. That number is down to 62 percent today.