The 2018 Midterm Elections are just two weeks away, but in some states early voting already started. If my social media feeds are any indication, nonreligious people are very motivated to vote this year. However, considering that people in my feeds are very engaged politically to begin with, they may not be the best indicator of political participation among the nones.

Common wisdom suggests that the nones are a population that punches well below their weight when it comes to voting. An oft-quoted figure comes from a 2016 PRRI report comparing the religious makeup of the electorate in Exit Polls and the religious makeup of the population.

In 2016, the nones accounted for 15 percent of voters in the National Exit Polls, while they represented one-quarter of the adult population. If we were to translate that into actual numbers (with the help of the United States Elections Project (USEP)), turnout was a dismal 34 percent. According to the USEP, there were about 250 adults in the United States in 2016. Twebty-five percent represent 62.5 million nones in the adult population. The USEP says that about 139 million people cast a vote in 2016. If 15 percent of those voters are nones that means some 21 million. Compared to the 54.7 percent overall turnout estiamted by the USEP, that’s awfully low.

That’s bad, though we shouldn’t use Exit Polls to estimate turnout. As the folks as Latino Decisions have argued for year, the Exit Polls are not a very trustworthy way of analyzing the electorate, especially when it comes to Latinxs. The Poll does give some useful information….but it has many issues. In our case is that (a) we don’t know how representative is the overall sample of none voters in the Exit Polls, and (b) the Exit Polls may be undercounting nones for a couple of reasons. (for more on Exit Poll methodology, read this entry from the Pew Research Center).

First, it only asks people if they have no religion: “none” is a response option (I don’t know if they have changed the way the ask the question recently, but working with some older one, that’s the way it was asked…if someone can find a link to the actual questionnaire and send it my way, I’d love that since I can’t find it online). Asking just for “nones” may leave some people who consider themselves atheist or agnostic. Second, people may be less willing to admit in a face-to-face survey that they have no religion.

Fortunately, the Cooperative Congressional Election Study (CCES) exists. The survey included a sample of about 65,000 adults and about 50,000 answered a post-election follow up. In the charts below I compare the nones (a combination of atheist, agnostic, and nothing in particular) with other religious groups. Born-again Protestants are those who identify as Protestant and evangelical, while Mainline Protestants are those who don’t identify as Born-again. Other Christians include Mormons and Orthdox, Non-Christian are those who are religious but are not Christian or nones.

The 2016 CCES validates voters with the Catalyst database and gives some idea of how groups voted.  In this definition of turnout I count those whose votes are validated as “voters” and those who are not validated as “nonvoters.”

Using this calculation, the nones punch below their weight, but not as bad as the Exit Poll/PRRI poll comparison suggests. They represent about 31 percent of the adult sample and 28 percent of voters. This suggests, as the second tab below shows, that about half of nones voted, compared to 54 percent of the public. They are still far below the turnout of most Christian groups, but not dramatically off from the general population numbers.

Nones probably have a turnout problem. But it may not be as bad as the Exit Poll analysis suggests. We should also take into account that nones are younger than any other religious group, so that’s probably something to be aware of. Young people (under 30)  have the lowest turnout of any age cohort. Hopefully campaigns like #Atheistvoter and #SecularVoter can help boost some of these numbers.