In a column published last Friday on the Rothenberg Political Report, political editor Nathan Gonzales took aim at a poll we commissioned in two Super Tuesday states to demonstrate the need for exit poll surveys to ask all voters, not just Republicans, if they are evangelical. Gonzales claims that the poll failed to demonstrate a shift among evangelicals away from the Republican Party. This misses the point of our poll. We, along with the Center for American Progress Action Fund, commissioned the poll to demonstrate the political diversity of evangelical Christians, which the exit polls have chosen to ignore. Our poll accomplished four important things:
• It showed that evangelicals are voting in significant numbers in both parties’ primaries.
• It showed that evangelicals are broadening their issue priorities beyond the narrow culture war agenda.
• It showed the need for more thorough polling of evangelicals.
• It provided a meaningful baseline for future comparisons.
This is not the first time the Rothenberg Report has attempted to write off evangelicals as a locked down ideologically monolithic voting bloc. In a March 2007 column, Stuart Rothenberg advised Democrats not to "waste a lot of time trying to attract evangelical voters to their party," pointing out that Democratic evangelical gains were small in the 2006 elections despite evangelical outreach efforts. Unfortunately, Rothenberg ignored what John Green at the Pew Forum on Religion & Public Life took the time to notice: "Although Democrats as a whole made only moderate gains with most faith communities during the '06 election, that picture changes when you focus in on specific races where Democrats made a concerted effort to reach out to the faith community. In these races, Democrats...made impressive gains among evangelicals."
Last week, the Rothenberg Report took liberties in their analysis of the poll we commissioned in two Super Tuesday states to demonstrate the political diversity of evangelicals and the need for the presidential primary exit polls to ask all voters, not just Republicans, if they are evangelical.
The most blatant example of the Rothenberg Report’s unwillingness to fairly analyze the data was Nathan Gonzales’ dismissal of one of our central findings about the broadening evangelical agenda. We found that in both MO and TN, a majority of evangelicals (62 percent to 33 percent in MO; 56 percent to 35 percent in TN) from both parties want a broader evangelical agenda that goes beyond abortion and same-sex marriage. Gonzales dismissed these findings and others by accusing us of leading question-wording and by impugning the small sample size of the survey.
Gonzales charged that our "question's wording virtually guaranteed the desired result since a small number of people are against 'ending poverty.'" This misleading characterization selectively quotes our poll question (which is publicly available in the interest of transparency) and ignores the current debate within evangelicalism over the evangelical issue agenda.
Our poll did not just ask about poverty, it asked: "Recently, some evangelicals have embraced a broader issue agenda that goes beyond abortion and gay marriage to include ending poverty, protecting the environment, and tackling HIV/AIDS. But other evangelicals have argued for sticking to the more limited agenda of opposing abortion and same-sex marriage. Which do you agree with most..."
Gonzales is certainly right that had we asked only about poverty, the question would have been little more than a push-poll question. But Gonzales has done enough analysis of evangelicals to know that environmental protection and HIV/AIDS, which he inexplicably omits from his quote, have been divisive issues in evangelical circles. (For example, Religious Right leaders have harshly criticized Rick Warren for embracing the issue of HIV/AIDS and attempted to oust Richard Cizik from his position at the National Association of Evangelicals for his work on climate change.)
And Gonzales ignored that the double-digit margins (29 points in MO and 21 points in TN) by which evangelicals support this broader agenda far exceed the margins of error, even with a modest sample size. Our results on this question were based on 293 (78 Democrats, 215 Republicans) evangelical respondents in MO and 399 (117 Democrats, 282 Republicans) evangelical respondents in TN. The margins of error for these subsets were 6 percent in MO and 5 percent in TN.
Another of Gonzales's gripes was that we should not have made any comparison of 2004 general election data and our 2008 primary poll. We acknowledged in on our press teleconference that this comparison was not apples-to-apples and because it compares general election to primary data. Given that primary exit polls have not asked the questions necessary to construct a true benchmark, suggestive data is necessary to provide context, even if it is not ideal. Moving forward, our Tennessee and Missouri polls can provide a future basis for comparison. You might call it the first apple.
Gonzales also criticizes the Washington Post for publishing our finding that Sen. Hillary Clinton defeated Sen. Barack Obama among white evangelical voters, and makes a blanket declaration that "the sub-samples [of white evangelical Democrats] were so small (n=76 in Missouri and n=116 in Tennessee) that any conclusions are not statistically reliable." Although the samples of white evangelicals were small (actually 78 in MO and 117 in TN), the poll did produce results that exceeded the necessarily wider margins of error. (The white evangelical Democratic sub-sample in MO had an 11 percent margin of error and this TN sub-sample had a 9 percent margin of error.) While the tight Clinton-Obama race in MO (54 percent to 37 percent among white evangelicals) was too close to draw statistically significant conclusions with this sample size, our finding of Clinton’s decisive win in TN among white evangelicals (78 percent to 12 percent) far exceeded the 9 percent margin of error. The Washington Post correctly reported these findings.
As an organization that believes no party can own any faith and that no group of people should be dismissed or ignored on the basis of religion, we are pleased to have sought out this instructive data that the exit polls have thus far refused to provide, and we stand by our work. The full wording of questions and topline results are publicly available from Faith in Public Life.