There's a lot of complexity in that question. There's the distinction between born-again Christians and evangelical Christians and other varieties of conservative Christianity. There is the difficulty in evaluating the population size of that group, much less the number who do or don't vote. And, then, there's comparative turnout -- how that group votes relative to others.
After the 2000 election, a survey from the University of Akron put evangelical Protestant turnout at 50 percent, as Cruz stated. But by 2004, evangelical turnout was at 63 percent, according to the next iteration of that same survey. What's more, that turnout isn't abysmally low relative to other groups. In the 2000 survey, Catholics also turned out at 50 percent. In 2004, mainline Protestants turned out at 69 percent -- the same as the more conservative subset of evangelical Protestants in the survey.
In the wake of the 2012 Mitt Romney loss, there were various analyses of Christian turnout that pointed to weak conservative Christian turnout in some states as being a cause of the loss. Our crack pollster team looked at data from the 2012 American National Election Studies to estimate evangelical turnout. More people in the study report voting than actually do. Once corrected for that over-reporting, we estimate that 62 percent of evangelicals went to the polls in 2012. It's down from the 2004 estimate cited above, but it's also using different data. And that 62 percent is about four points higher than the overall population.
The idea of low turnout in the group isn't unique to Cruz. Last year, former Arkansas governor Mike Huckabee -- who would very much like to absorb those same religious votes in next year's primaries -- articulated a more extreme version of Cruz's math. At the Value Voters Summit, Huckabee said that there are about 80 million evangelicals, half of whom are registered to vote -- and a quarter of whom actually vote in presidential elections. And in midterms? Only half of those who vote in presidential races!
That's pretty clearly not true. A survey from Pew Research puts the percentage of the population of U.S. adults that is evangelical at 26.3 percent. (This may be high, as is argued here.) The Census Bureau tracks registration and voting over time, allowing us to see what the numbers would look like if Huckabee were correct.
Turnout according to Mike Huckabee
||Adult U.S. population
||Evangelical adults (per Pew)
||Percent of all voters
Before we get any further, notice the overall turnout rates in 2008 through 2012, according to the Census Bureau: 58.2 percent, 41.8 percent, 56.5 percent. Turnout of 50 percent, then, isn't anything particularly unusual, depending on the election.
But we can evaluate Huckabee's claim against exit polling. Since 2004, 25 percent of the electorate has identified itself as evangelical in exit polling, voting heavily for Republicans. In 2008, the figure was 26 percent; in 2010, 25; in 2012, 26 again. That's much higher than Huckabee's estimates, and it's pretty consistent.
You might notice that it's also about in line with the Pew findings that 26 percent of the country is evangelical. That's a different thing than saying that a certain percentage of the evangelical population came out to vote, but it does mean that turnout is about where you'd expect.
Which wasn't Cruz's point, really. In the same way that the 2008 and 2012 campaigns of Barack Obama put an emphasis on boosting African-American turnout -- successfully -- Cruz's goal isn't necessarily to say that evangelicals are under-performing; rather, that he needs them to over-perform, and preferably on his behalf. Today, roughly half of everybody isn't voting. But Cruz is not terribly interested in seeing more turnout from moderate Republicans or Democrats.