PRINCETON, NJ -- A Gallup update based on more than 21,000 interviews conducted as part of Gallup Poll Daily tracking in October shows that registered voters' religious intensity continues to be a powerful predictor of their presidential vote choice. John McCain wins overwhelmingly among non-Hispanic whites who attend church weekly, while Barack Obama dominates among whites who seldom or never attend church.
The relationship between church attendance and voting has been well-established in previous presidential elections, and the analysis of Gallup data collected Oct. 1-26 suggests that this year is no different.
The swing in vote choice between groups of non-Hispanic white voters differentiated by their religious intensity (operationalized in this review as self-reported church attendance) is dramatic, ranging from a 37-point McCain advantage among whites who attend church weekly (about 32% of all non-Hispanic white registered voters) to a 19-point Obama advantage among those who seldom or never attend worship services (about 47% of white voters). The middle group of non-Hispanic whites who attend church nearly every week or monthly (20%) support McCain by a 12-point margin.
McCain's selection of Sarah Palin as his running mate just before the Republican National Convention might have been expected to bolster his standing among highly religious whites, given her credentials as a staunch conservative on moral or values issues.
However, a comparison of Gallup's October data to data collected in August -- for the most part before McCain's selection of Palin as his running mate -- shows that McCain's strength among highly religious white voters has not changed substantially over that time period. In August, McCain was beating Obama among weekly church-attending white voters by a 64% to 25% margin, or 39 points, compared to a similar 37 points today. Importantly, McCain was doing better in August among whites than he has been in October, winning in August by a 51% to 39% margin, compared to a narrower margin of 49% to 44% in the Oct. 1-26 data. The fact that support for McCain did not slip among highly religious whites, while slipping overall among whites, could be in part attributable to Palin's presence on the GOP ticket, although the precise impact of her selection is not clear. (Among those who seldom or never attend church, McCain was losing by 11 points in August, and is losing by 19 points today.)
Black voters overwhelmingly support Obama regardless of their relative level of religiosity. But highly religious blacks who attend church weekly (39% of non-Hispanic black voters) give Obama an 84-point margin, slightly lower than the 92- and 88-point margins for Obama among blacks who attend almost every week/monthly (33%) and those who seldom/never attend (26%), respectively.
Among Hispanics, the biggest difference in vote choice by religion is between those who attend monthly or more frequently, and those who seldom or never attend church. Hispanics who attend church weekly (36% of non-white Hispanic registered voters) favor McCain by a very narrow 3-point margin, and those who attend almost every week or monthly (23%) favor Obama by a similarly narrow 6-point margin. But among Hispanics who seldom or never attend church (40%), the margin for Obama swings to a highly significant 33 points.
A positive correlation between religious intensity and voting for the Republican candidate for president has been a part of the American political landscape for a number of years.
Early in this election cycle, there was discussion that McCain was perhaps not the favorite candidate of highly religious whites, and indeed former Arkansas Gov. Mike Huckabee (a former Southern Baptist seminarian and preacher) did well among religious whites in some GOP primaries.
But McCain appears to have had little problem in gaining the vote of religious whites. The data reviewed here show that McCain was doing very well among highly religious whites in August, and continues to do so in October, despite the fact that his overall standing has slipped some among whites.
Results are based on telephone interviews with 23,111 registered voters, aged 18 and older, conducted Oct. 1-26, 2008. For results based on the total sample of national adults, one can say with 95% confidence that the maximum margin of sampling error is ±2 percentage points.
Interviews are conducted with respondents on land-line telephones (for respondents with a land-line telephone) and cellular phones (for respondents who are cell-phone only).
In addition to sampling error, question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of public opinion polls.