Religion Gap Swings New Ways
John C. Green
The news from the religion-gap front in the 2004
election was that George W. Bush was able to reach beyond his base of
frequent worshipers to pick up a majority of voters who said they attend
religious services “a few times a month.” This group, which he lost to Al
Gore in 2000, emerged as the religion gap’s new swing vote.
Bush also made small
gains among those who said they attend services a few times a year or not at
all. Since those claiming to attend once a week or more voted for Bush at
about the same rate as in 2000, the upshot was a slight decrease in the
overall religion gap, as measured by the difference between the percentages
of his vote coming from frequent as opposed to less frequent attenders.
Because these results
derive from the now notorious 2004 election-day exit polls, they must be
viewed with some caution. But given how well they track the results from
earlier, more accurate exit polls, there is reason to believe they provide a
broadly dependable portrait of the impact of religion on the election.
The first part of the accompanying table presents the two-party vote
in terms of frequent (weekly or better) and less than weekly worship
attenders. These figures show a now common pattern: Bush won the former and
Kerry the latter. Specifically, Bush’s margin in 2004 among the frequent
attenders increased by so small an amount that it’s hidden in rounding the
figures, while among less frequent attenders Kerry fell one percent behind
Gore’s performance in 2000.
The second part of the
table unpacks this attendance dichotomy into each of its component parts. It
shows that the “few times a month” attenders shifted their support from six
percentage points in favor of Gore over Bush to two percentage points in
favor of Bush over Kerry.
Representing 15 percent
of 2004 voters, the monthlies would have given Kerry more than two and a
half million extra votes nationwide had they voted for him at the same rate
they voted for Gore in 2000—enough to have brought him within one million
votes of the president. To have won, Kerry would have had to have found some
additional support elsewhere.
What about the dramatic
expansion of turnout in 2004? That increase occurred quite evenly up and
down the levels of worship attendance, with just a few minor fluctuations.
So the big turnout in 2004 did not result from a change in the relative
distribution of worship attenders. Because the electorate was so much bigger
in 2004, the actual number of ballots at each level of attendance expanded
As in 2000, the 2004
religion gap extended across the largest religious traditions, with Bush
doing best among regularly attending white Christians. But in most cases,
Bush gained more among the less frequent attenders. In the case of white
Catholics, for example, the Bush vote increased from 64 percent to 66
percent among those who went to church more than once a week, but from 51
percent to 59 percent among those who turned up a few times a month. The
Catholic swing vote swung in Bush’s direction.
A change in the exit
poll question wording between 2000 and 2004 renders comparisons among white
Protestants imprecise, but a similar pattern appears to have held there as
well.1 Bush gained slightly among the most active evangelicals
(80 percent to 81 percent), but improved more among the monthly attenders
(64 percent to 79 percent). He also experienced a small improvement among
Latino Catholics and African-American Protestants.
There were a few offsets
to the Bush gains. Not surprisingly, the president appears to have lost a
little ground among voters who said they were unaffiliated or secular.2
And although Bush once
again won a majority of the mainline Protestant vote, it looks as though he
made only a tiny gain among the less frequent attenders (from 52 percent to
53 percent) and, most significantly, lost substantial ground among their
frequent attending counterparts (from 62 percent to 54 percent). Mainline
Protestants may be the faith group to watch in the future. Because the exit
poll data has not yet been released in their entirety, it is hard to be any
more specific about the extent to which the 2004 religion gap was affected
by factors like gender, region, and turnout among particular religious
groups. But the published reports offer some suggestions.
Most importantly, the
Bush campaign’s extensive mobilization of religious communities maintained
the support of regular attenders and expanded their turnout in line with the
rest of the country. The greater prominence of social issues, especially
same-sex marriage, surely facilitated these efforts.
It is possible that this
mobilization of voters on religious lines—through an appeal to their “moral
values”—should be credited with enabling the president to gather in more
few-times-a-month attenders and minorities. It is, of course, also possible
that those groups voted for Bush in higher numbers for other reasons such as
the war on terror.
At least for this
election, the monthly attenders voted more like the weekly attenders, and
their shift narrowed the religion gap so as to give Bush a decisive
advantage in a very tight election. Closely divided between Republican and
Democrat, these pretty regular churchgoers may be the swing vote that
determines the outcome of future national elections. Under the
circumstances, the Democrats would be well advised to listen to those urging
them to do a better job connecting with people in the pews.
2000, it asked “are you a member of the religious right?” In 2004, it asked
“are you an evangelical or born again Christian?” To compare the two it was
necessary to estimate what the “born again” measure would have been in 2000
based on past surveys that asked both questions.
“unaffiliateds” and “seculars” need to be differentiated from those who say
they never attend religious services. The latter, who trended slightly
towards Bush in 2004, include those who, despite their lack of attendance,
nonetheless profess a religious identity.