Beware: Evangelical Polling Ahead

by Nathan L. Gonzales August 23, 2008 · 12:05 AM EDT

Evangelicals are in the spotlight for the thousandth time this election cycle, after both Sens. Barack Obama (D-Ill.) and John McCain (R-Ariz.) appeared at Rick Warren’s church on Saturday night.

But there is a significant problem with analyzing the evangelical vote. Pollsters can’t seem to agree on what constitutes an evangelical or how many evangelicals there are in the country.

The discrepancy in definition makes thoughtful analysis almost impossible because polling numbers can’t be used interchangeably. Unfortunately, the difference is lost on most reporters.

According to 2004 exit polling during the presidential race, 23 percent of Americans described themselves as “white evangelical/born again,” and gave President Bush 78 percent of their vote. Four years earlier, the exit poll showed that 14 percent of presidential voters were part of the “White Religious Right” and voted for Bush with 83 percent. The question was rightly modified between elections, because “Religious Right” is a narrower group, but the modification makes comparison difficult.

ABC News, in its polling, evaluates self-identified “white evangelical Protestants,” who constitute about 20 percent of American adults. The definition isn’t necessarily wrong, even though it excludes a small number of Catholics who might consider themselves evangelicals, but it’s another definition. Obama is winning about 25 percent of the white evangelical Protestant vote, according to ABC polling.

NBC News and the Wall Street Journal, in their polling, first ask respondents if they are Catholic or Protestant. Then they ask all Protestants and people who said they had no religion, “Would you describe yourself as a fundamentalist or an evangelical Christian, or would you not describe yourself that way?”

The result is yet another definition. This time it’s a group of 18 percent of Americans who describe themselves as “Protestant fundamentalist/evangelical,” who are voting for McCain 64 percent to 24 percent. First of all, it’s unclear how many people said evangelical versus fundamentalist, but just using the latter term probably turns off some respondents. Secondly, the NBC sample likely includes African-Americans, which makes it impossible to compare to the other samples.

Finally, the Barna Group, a religious research firm, recently released a poll that confuses the matter even further.

According to the Barna survey, self-identified evangelicals make up 42 percent of the population and are going for McCain by a narrow 39 percent to 37 percent. That is a much larger self-identified evangelical sample compared with other pollsters, and much more Democratic.

It is not entirely clear why the self-identifying Barna sample is almost double what is in the other polls since Barna didn’t release demographic profiles of the samples and sub-samples. It is likely that the sample includes a large number of African-Americans, who may have similar theological beliefs to white evangelicals but vote overwhelmingly Democratic.

But Barna doesn’t rely on self-identifying evangelicals. Instead, it asks a series of nine theological questions based on the National Association of Evangelicals’ Statement of Faith to find out who is really an evangelical.

After the test, Barna concludes that 8 percent of the population is actually evangelical. The definition may be the most pure, but it makes cross-analysis with other polls useless. Barna evangelicals were voting for McCain 61 percent to 17 percent in an August survey, compared with 78 percent for McCain in June.

An e-mail from the Obama campaign touted the survey as “great news” because the race was tightening. But 17 percent is statistically identical to the 15 percent Sen. John Kerry (D-Mass.) received from the group in the 2004 White House election, and it is irresponsible to make thoughtful conclusions from such a small sub-sample with a high margin of error.

Barna also separates out “non-evangelical born again” Christians with a less stringent theological test. Those voters, which would likely include African-Americans, went for Obama 43 percent to 31 percent. “All born again” Christians, presumably including evangelicals, African-Americans and some Catholics, went for Obama by a similar 42 percent to 32 percent, according to Barna.

The political behavior of evangelical voters is going to be a favorite media story for months and years to come. But as evangelicals come under the microscope more often, the more theological nuances and differences are laid bare. Until there is some sort of standard definition set in the polling world, analysis of the evangelical vote should be met with extreme caution.