August 10, 2004 An Election Year Puzzle
by Jeffrey M. Jones and Joseph Carroll
Recent Gallup analyses of 2004 voting preference data have revealed an interesting puzzle. Regular churchgoers are much more likely to support George W. Bush, while non-churchgoers overwhelmingly support John Kerry. Men -- particularly white men -- are much less likely to attend church than are women of any race or ethnicity (see "U.S. Churches Looking for a Few White Men" in Related Items). Despite this, the traditional gender gap still persists, with men overall more likely to support Bush and women more likely to support Kerry. So if men are less likely to attend church, and churchgoers overwhelmingly support Bush, how can men as a group still support Bush over Kerry? A closer look at the data reveal how.
The Basic Data
Much attention this year is being paid to the wide gulf in presidential preference between churchgoers and non-churchgoers. The latest CNN/USA Today/Gallup poll* shows Bush leading Kerry among registered voters who are weekly churchgoers by 57% to 39%, while Kerry has a nearly equally big lead among registered voters who seldom or never attend church, 57% to 41%. Registered voters who attend church on a semi-regular basis (nearly weekly or monthly) are divided in their preferences, with 49% supporting Bush and 47% Kerry.
Last week, The Gallup Poll Tuesday Briefing presented an analysis showing that men are much less likely to attend church than are women. Given these findings, it should be the case that women -- the more churchgoing of the two genders -- should show higher support for Bush than men do. But that is not the case.
According to Gallup's July 30-Aug. 1 poll, 52% of male registered voters say they will vote for Bush, and 45% for Kerry. Among female registered voters, 51% support Kerry, and 45% Bush. So Bush has a net advantage of 7 points among men, and Kerry an advantage of 6 points among women, for a gap of 13 points.
The answer to the riddle lies in three findings. First, both religiosity and gender have independent effects on the vote. Second, at the highest and lowest level of religious participation, men are more likely than women to support Bush than Kerry. Third, male regular churchgoers, though fewer in number than female churchgoers, support Bush by an overwhelming margin, and by the largest margin of any gender by religiosity subgroup.
Weekly Churchgoers
Among registered voters who attend church weekly, a majority of both men and women support Bush over Kerry. However, male regular churchgoers support Bush by a nearly 2-to-1 margin over Kerry, 64% to 34%. The gap among female regular churchgoers is much smaller, with 52% supporting Bush and 42% Kerry.
Among white men who attend church on a regular basis, the gap is even larger -- 70% of registered voters in this group support Bush, 27% Kerry.
Those Attending Church on a Monthly Basis
Clearly, those who attend church on a semi-regular basis -- at least monthly but not every week -- are most divided in their vote preferences. In fact, both men and women at this level of religious participation are equally divided in their vote preferences, with 49% of men supporting Bush and 47% Kerry, and 48% of women supporting both candidates.
The data suggest, however, that whites at this level of religious commitment show a noticeable preference for Bush over Kerry. Given the division in the overall numbers, this would indicate that nonwhites in this subgroup show a stronger preference for Kerry.
Those Attending Church Rarely or Never
Both men and women who are registered to vote and who seldom or never attend church are more likely to say they will vote for Kerry than Bush. Women in this religiosity subgroup, however, show much stronger support for Kerry than men do. A slight majority of men in this sample, 53%, say they would vote for Kerry, while 45% would vote for Bush. Roughly 6 in 10 women (61%) support Kerry, while 36% support Bush.
Bottom Line
Churchgoing men -- especially white churchgoing men -- may very well wind up being one of the strongest groups supporting Bush this year, and may show the strongest support for a particular candidate of any subgroup this year. Perhaps only blacks, of whom about 80% to 90% will probably support Kerry, will show a stronger level of support for a candidate this year.
*Results are based on a CNN/USA Today/Gallup poll of 1,366 randomly selected registered voters, conducted July 30-Aug. 1, 2004. For results based on this sample, the maximum margin of sampling error is ±3 percentage points.
For results based on the sample of 670 male registered voters, the maximum margin of sampling error is ±4 percentage points.
For results based on the sample of 696 female registered voters, the maximum margin of sampling error is plus ±4 percentage points.
For results based on the sample of 463 registered voters who say they attend church on a weekly basis, the maximum margin of sampling error is ±5 percentage points.
For results based on the sample of 351 registered voters who say they attend church on a nearly weekly or monthly basis, the maximum margin of sampling error is ±6 percentage points.
For results based on the sample of 532 registered voters who say they seldom or never attend church, the maximum margin of sampling error is ±5 percentage points.