92 Mark Chaves and Laura Stephens
pattern: Voting, attending a political meeting, attending any public meeting, serving
as an officer or committee member in any local club or organization, participating in a
local meeting of any national organization, attending a club meeting, joining a union,
participating in a picnic, playing sports, working on a community project.
The details vary for specific items, but the consistency – across many different indi-
cators drawn from many different sources – is impressive. For item after item, trend line
after trend line, decline starts sometime in the last third of the twentieth century and
continues into the present. This casts new light on the religious participation trend.
Religious participation, it seems, is a special case of something much more general:
Civic engagement. The newly reported findings of decline in virtually all sorts of civic
engagement since the 1960s, together with the direct evidence for decline in some of
the best data on religious participation itself, add weight to the notion that religious
participation in the United States has indeed declined in the last third of the twentieth
century. Seen in this context, it would be a great surprise indeed to learn that religious
participation, alone among all sorts of civic engagement, has failed to decline. Those
still wishing to maintain that religious participation has been stable over the last three
or four decades now must face the additional burden of explaining how it could be
that religious trends are so different from trends affecting virtually every other type of
voluntary association.
There is another important detail on which recent evidence is converging. Presser
and Stinson, among others, found that more recent generations attend religious services
at lower rates than did previous generations when they were the same age. Chaves
(1989, 1991) found this same pattern, and Putnam finds it as well across a strikingly wide
range of activities, including church attendance. Declining participation in all sorts
of voluntary associations, including religious ones, is not occurring so much because
individual people have become less involved over the last three or four decades. Rather,
more recently born cohorts of individuals do less of this activity than older cohorts,
and those born earlier are inexorably leaving the scene, being replaced by less civicly
engaged recent generations. Even if not a single individual changes his or her behavior
over time, it still is possible for widespread social change to occur via generational
turnover, and this seems to be largely what is happening with civic engagement in
general, and with religious participation in particular.
So, have U.S. church attendance rates been stable over recent decades, or have
they declined? The evidence is conflicting. Those wanting to argue in favor of stability
can point to traditional surveys, but they then need to explain why surveys using an
indirect approach, such as time-use studies, find decline. They also need to explain why
church attendance trends are different than trends in most every other type of civic
engagement. Those wanting to argue in favor of decline, by contrast, need to explain
why that decline is not evident in traditional surveys.
We can offer a plausible account for why traditional surveys might show stabil-
ity over time even if weekly attendance truly has declined Recall our suggestion that
survey respondents may perceive a direct question that is literally about religious ser-
vice attendance to be, instead, an inquiry about that person’s identity as a religious
or nonreligious person. It seems plausible to suggest further that the proportion of
Americans who truly attend religious services weekly might have declined at the same
time that the proportion who think of themselves as “church people” – and who may