90 Mark Chaves and Laura Stephens
The bottom line here is that recent research has overturned an earlier conventional
wisdom about the level of weekly religious participation in the United States. To the
best of our current knowledge, the weekly attendance rate in the United States is closer
to 20 percent than to 40 percent.
From a broader perspective, it should not be surprising that individuals overreport
their religious service attendance when they are directly asked. We know that other
sorts of socially desirable behaviors are overreported, and we know that socially un-
desirable behaviors are underreported. For example, more people claim to have voted
than actually did (Parry and Crossley 1950; Traugott and Katosh 1979; Silver, Anderson,
and Abramson 1986, Presser and Traugott 1992). Presser and Traugott (1992) report
that about 15 percent of voters report their voting activity inaccurately. Furthermore,
since almost all of this error comes from people who have not voted claiming that they
have, about 30 percent of nonvoters are misclassified as voters. Similarly, young people
tend to underreport undesirable behaviors such as drug use (Mensch and Kendel 1988).
In the light of this broader phenomenon, well-known in survey research, it would be
surprising if religious service attendance wasnotoverreported in conventional surveys.
Overreporting socially desirable activity probably is not the only mechanism lead-
ing people to exaggerate their religious service attendance. The fact that overreporting
is reduced when religious service attendance is asked about indirectly (as in the time-
use diaries) rather than directly suggests that something else might be going on. We
speculate that survey respondents may perceive a question that is literally about reli-
gious service attendance to be a request for information about the person’s identity as a
religious or nonreligious person. On this scenario, respondents who inaccurately report
their literal church attendance may be intending toaccuratelyreport their identities as
religious individuals who attend services more or less regularly, even if not weekly.
From this perspective, one plausible interpretation of the attendance rates generated
by conventional surveys is that they are picking up the percentage of Americans who
think of themselves as “church people,” even if they attend less than weekly.
Although weekly attendance at religious services now appears to be less frequent
than previously believed, it still is the case that Americans attend religious services at
higher rates than people in most of the industrialized West. A recent study of sixty-five
countries, for example, found that 55 percent of Americans said they attend religious
services at least once a month, compared with 40 percent in Canada, 38 percent in
Spain, 25 percent in Australia, Great Britain, and West Germany, and 17 percent in
France (Inglehart and Baker 2000). Additionally, among advanced industrial democ-
racies the United States still stands out for its relatively high level of religious belief.
Fifty percent of Americans said “10” when asked to rate the importance of God in their
lives on a scale of 1 to 10. That’s compared with 28 percent in Canada, 26 percent in
Spain, 21 percent in Australia, 16 percent in Great Britain and West Germany, and
10 percent in France. Among advanced industrial democracies only Ireland, at
40 percent, approaches the U.S. level of religious belief. As in other arenas, a kind
of American exceptionalism holds when it comes to religion.
WHAT IS THE TREND IN ATTENDANCE AT RELIGIOUS SERVICES?
Some researchers have argued that religious participation has increased over the long
haul of American history (Finke and Stark 1992). This claim is based on increasing rates