114 Philip S. Gorski
findings, the leading proponents of the REM claim to have disproven the secularization
thesis and argue that the term “secularization” should be “dropped from all theoretical
discourse” (Stark and Iannaccone 1994: 231). Is their claim justified?
The work of the religious economies school has been challenged on a number of
different fronts. Some scholars accepted the empirical findings, but questioned their
theoretical significance (e.g., Lechner 1991; Yamane 1997; Gorski 2000). They pointed
out that secularization theory is a theory, not only of individual behavior, but also, and
indeedprimarily, of social-structural change. In their view, secularization refers first
and foremost to an increasing differentiation between the religious and nonreligious
spheres of life, and only secondarily to its effects on individual behavior. Since the
REM focuses exclusively on individual behavior, they argue, it does not really address
the core claim of secularization theory and speaks only to those versions of the theory
that postulate a direct connection between increasing (social-level) differentiation and
decreasing (individual-level) religiosity.
Other scholars have challenged the reliability and validity of the findings them-
selves (e.g., Blau et al. 1992; Breault 1989a, 1989b; Olson 1998, 1999). Using new
datasets of their own, or reanalyzing REM data, these scholars often obtained null
or negative correlations between pluralism and vitality. Defenders of the REM then
challenged these results on methodological grounds (Finke and Stark 1988; Finke et al.
1996). The ensuing debate was long and complex, but the key issue was Catholics. Many
of the analyses that had yielded a positive correlation between religious pluralism and
religious vitality also included a statistical control for “percent Catholic.” Advocates of
the REM defended this procedure on the grounds that the Catholic Church displayed a
high degree of “internal pluralism,” and that treating it as a single denomination would
therefore distort the findings. Critics pointed out that the positive relationship between
pluralism and vitality usually disappeared or became negative when the control was
removed (see especially Olson 1999). More important, they showed that removingany
group with the characteristics of the (American) Catholic population – large in overall
size but varied in local presence – wouldautomaticallyresult in a positive finding, and
for purely arithmetic reasons!
On the whole, then, the REM’s claims to have disproven the secularization thesis
and laid the foundations for a “new paradigm” in the sociology of religion are somewhat
overblown. As we have seen, the central findings of the REM do not really address the
core concerns of secularization theory, and are themselves open to dispute. Indeed, in
a recent survey of the literature on “religious pluralism” and “religious participation,”
Chaves and Gorski (2001) found that the balance of evidence actually tipsagainstthe
REM, once we exclude analyses that employ inappropriate measures of competition or
statistical controls for percent Catholic.
Of course, it is possible that further research could tip the balance back the other
way. But this seems unlikely to me. For even a cursory review of the comparative and his-
torical evidence reveals two large and potentially troubling anomalies for the REM. The
first regards Catholic-Protestant differences. As we saw earlier, overall levels of Christian
practice in various Western countries are closely related to the proportion of the
members than majority religions, a finding that also has been replicated by scholars working
outside the religious economies perspective (e.g., Zaleski and Zech 1995; Johnson 1995; Perl
and Olson 2000; but see also Phillips 1998).