130 Patricia M. Y. Chang
symbolic processes relative to organizations. Since the centrality of culture and symbol
are precisely what makes religious organizations different from secular organizations,
this approach has naturally elicited the attention of religious scholars but it has also
frequently misled them.
Religious organizations are distinct in that they are usefully conceived as having an
internalculture that intentionally sets them apart from other communities of religious
believers. Indeed, their very identity rests on this distinctive culture. This culture guides
the blueprint of their formal structure, flavors the meaning of their behaviors, and forms
a reservoir of experience that they draw on when making difficult decisions. This culture
manifests itself most directly in boundary setting behaviors that distinguish religious
insiders from religious outsiders.
Neoinstitutionalists, by contrast, focus entirely on howexternalcultural processes
affect organizational behavior. Neoinstitutionalists explicitly ignore the internal cul-
ture of organizations that religion scholars focus on as an important determinant of
behaviors. Neoinstitutionalists treat organizational leaders as automatons who reflex-
ively respond to environmental cues. They see cultural processes in the environment
as exerting a homogenizing influence while cultural differences among organizations
are virtually ignored (DiMaggio 1988).
This lack of fit between neoinstitutional and religious approaches is not merely one
of a difference in the locus of analysis. Real empirical differences exist between the
sectors neoinstitutionalists have tended to study, and the religious sector. Empirical
investigations that have supported neoinstitutional theory have all been conducted
in social sectors that are highly “institutionalized,” that is, where social networks are
already dense through the effects of federal regulation, technological standardization,
or financial centralization. In comparison, the religious sector is very weakly institution-
alized, showing little evidence of centralization, standardization, or regulation (Scott
and Meyer 1991). In fact, no study using neoinstitutional theory has been able to show
the effects of institutional isomorphism in the religious sector to the extent found in
other organizational populations. The most rigorous empirical attempt to apply neoin-
stitutional theory to an organizational population in the religious sector found, in fact,
that neoinstitutional hypotheses predicting the rapid and universal diffusion of orga-
nizational practices related to the ordination of women were not supported (Chaves
1997).
Rather than being an uncomfortable anomaly, however, the weakness of institu-
tionalizing processes is a revealing insight that allows one to usefully compare the dif-
ferences in the institutional patterns of strong and weak institutional sectors. Strongly
institutionalized sectors tend to be highly integrated by institutional practices and
norms brought about by technological standardization, centralization, or government
regulation found in the health, education, technology, and the arts sectors (DiMaggio
1991; Meyer and Scott 1992; Scott 1995). Each of these sectors is distinguished by strong
organizational rules that permit the easy identification of a population of organizational
actors, the clear definition of many normative practices, and the easy measurement of
organizational outcomes.
By contrast, the religious sector is highly decentralized, organizational practices
vary broadly, and a number of differing organizational forms can be identified. The
field is not regulated by federal or industry rules or standards and there is no centralized