82 Michael Hout
deal with age: The mortality rate is much lower for people in their twenties than for
people in their sixties. Then a change in the population that increases the number
of twenty-somethings while the number of sixty-somethings stays the same or goes
down will decrease the overall mortality rate. The number of deaths in Florida rose
dramatically in the 1950s and 1960s despite improved overall longevity in the United
States because so many people retired to Florida during those years, not because the
environment in Florida suddenly became hazardous.
The religious connection here is in the relationship between age and religiosity.
The currently aging American population will probably increase the church attendance
rate because church attendance is also lower for twenty-somethings than it is for sixty-
somethings. We may never see this change, however, because rising immigration and
falling marriage and fertility counteract it. The analysis of the heterogeneity in all these
rates is grist for the demographer interested in religious behavior.
DEMOGRAPHY AND RELIGIOUS RESEARCH
Religion has long been recognized by demographers as an important factor in fertility
and migration. More recently, demographers have become aware of important religious
differences in mortality. Hummer et al. (1999) published life tables for the religiously
active and inactive that show the advantage that the religious enjoy. McCullough et al.
(2000) compiled forty-two independent studies of religious involvement and mortality.
Not only did researchers consistently find that involvement in religion prolongs life,
but they also found that religion adds to the effects of things – like stable marriage –
that often go with religious involvement.
An earlier line of research documented large differences between the fertility of
Catholics and Protestants during the baby boom (e.g., Westoff and Jones 1979; Mosher
and Bachrach 1996). At the point of peak difference (in the late 1950s), Catholic women
were averaging one more birth than Protestant women were having. By 1970 – a span
of just fifteen years – the difference was gone. Although most researchers gave scant
attention to differences among Protestant women of different faiths, recent work shows
that they were just as large as the Protestant-Catholic gap (Hout et al. 2001). Women
from evangelical and fundamentalist denominations were averaging one birth more
than women from mainline denominations were having. This gap, too, was gone by
the early 1970s. Another way to summarize this pattern is to note that women from
mainline Protestant denominations contributed what amounted to a baby blip; the
baby boom was concentrated among Catholic, evangelical, and fundamentalist women.
These studies view religion as the cause of important demographic differences. The
persistence of religion from one generation to another means that demographic dif-
ferences based in religion in one generation show up as religious differences based
in demography a generation later. I have already referred to the recent work my col-
leagues and I have done on the role of fertility differences in the decline of mainline
Protestant denominations. In a companion paper we ask why Catholics’ demographic
advantages – higher fertility from 1920 to 1975 and greater immigration in both the
first twenty and last twenty years of the last century – did not raise the Catholic share of
the U.S. population above 25 percent. Without a demographer’s sensibility, of course,
the nearly constant share of the population that is Catholic is not problematic in the
least. Who worries about nontrends? But this is an interesting puzzle. The Catholic