The Structural Contexts of Civic Engagement: Voluntary Association Membership in Comparative Perspective

by Evan Schofer, Marion Fourcade-Gourinchas
The Structural Contexts of Civic Engagement: Voluntary Association Membership in Comparative Perspective
Evan Schofer, Marion Fourcade-Gourinchas
American Sociological Review
Start Page: 
End Page: 
Select license: 
Select License


University of Minnesota Princeton University and New York University

Voluntary association membership varies dramatically among nations, by both the number and the type of associations that people join. Two distinctions account for much of this variation: (1) the distinction between statist versus nonstatist (some- times called "liberal") societies, and (2) the distinction between corporate versus noncorporate societies. These two dimensions summarize historically evolved differ- ences in state structure, political institutions, and culture of nations that channel, legitimate (or deligitimate), and encourage (or discourage) various types of associa- tional activity. Membership in associations in 32 countries is examined using data

from the 1991 World Values Survey; hierarchical models estimate the effects of indi-

vidual-level and country-level factors on individual association membership. Results

show that statism constrains individual associational activity of all types, particu-

larly in "new" social movement associations. Corporateness positively affects mem-

bership, particularly for "old" social movements. Finally, temporal trends indicate

some convergence toward Anglo-American patterns of association.

I N CONTEMPORARY nation-states, vol- untary associations are important bodies that mediate between the individual and the broader societal environment. Following de Tocqueville's ([I8621 1945) early statements on the different political organization of America and Europe, political scientists

Direct all correspondence to Marion Fourcade- Gourinchas, Department of Sociology, Princeton University, Princeton NJ, 08540 (fourcade@ or Evan Schofer (schofer@soc. The authors contributed equally to this research. A version of this paper was presented at the 1999 annual meeting of the American So- ciological Association. The authors thank Ron Jepperson for extremely valuable insights and advice, as well as Frank Dobbin, Marshall Ganz, Joe Galaskiewicz, Kieran Healy, Ann Hironaka, Michkle Lamont, John W. Meyer, Kimberly Mor- gan, Virag Molnar, Francisco 0. Ramirez, Abigail Saguy, Theda Skocpol, Dietlind Stolle, and the members of the Stanford Comparative Workshop. We also thank the ASR Editors and anonymous reviewers for their helpful comments and suggestions.

and sociologists have noted that people of different countries and regions vary in their involvement in associational activity (Al- mond and Verba 1963; Putnam 1993; Wuthnow 1991). The United States, for in- stance, is traditionally described as a "nation of joiners," while some European countries (e.g., France, Italy) and Japan seem to have a much less developed civic orientation. In sum, "country of residence" appears to be "an important predictor of voluntary asso- ciation joining" (Curtis, Grab, and Baer 1992: 150).

Many scholars attribute this variation in civic involvement to the different value sys- tems internalized by members of each soci- ety (Almond and Verba 1963; Inglehart 1997). Final explanations often recognize that these value systems may be rooted in larger institutional and ideological structures. These explanations usually emphasize how such structures are mediated at the in- dividual level to produce particular attitudes and behaviors (e.g. "post-materialist val- ues," "trust," and "social capital") that are themselves conducive to the formation of voluntary associations, and (ultimately) to the prosperity of democratic institutions (see Putnam 1993, 2000).

This "bottom-up" view of the relationship between civic life and political institutions has been criticized. First, "social capital" or "trust" are, at best, elusive concepts that are not easily connected to observable empirical realities (Paxton 1999; Portes and Landolt 1996; Tarrow 1996; Wuthnow 1999). Sec- ond, some authors have contested the "di- chotomous thinking that counterposes civil society to the state" (Cohen 1999:283). In- stead, they argue that political institutions play an essential role in shaping civic activ- ity-not only the other way around (Levy 1999; Skocpol 1996, 1997; Skocpol and Fiorina 1999; Skocpol, Ganz, and Munson 2000; Tarrow 1996).'

We draw on a conceptualization of politi- cal structure originating in the work of insti- tutionalist sociologists and political scien- tists (Meyer 1983; ~e~~erson

and Meyer 1991; also see Birnbaum 1988; Dyson 1980; Schmitter 1974). More specifically, we de- pend on a synthetic typology developed by Jepperson (1992, forthcoming) to argue that institutionalized patterns of political sover- eignty and organization-what Jepperson calls the degree of statism and the degree of corporateness-are associated with distinc- tive patterns of civic engagement.' Involve-

For similar arguments about nonprofit orga- nizations, see Anheier (l990), James (1989), and Salamon and Anheier (1994, 1997). Related analyses have also been developed in the social movements literature. See, for instance, Kitschelt (1985), Klandermans, Kriesi, and Tarrow (1988), Kriesi et al. (1995), McAdam, McCarthy, and Zald (1996).

Jepperson (1992) prefers the term corporateness to the more familiar one of corporatism. "Corporatism" in the twentieth cen- tury refers canonically to the Italian Fascist state's practice of managing society via mixed syndical organs; or, in its modern forms, to pat- terns of institutionalized "peak-bargaining" between economic groups, this time not necessarily subordinate to state oversight. (Scandinavian wage management is a good example.) "Corporateness," on the other hand, is more neu- tral and simply refers to the degree to which po- litical representation and incorporation is typi-

ment in volunteer activities does not simply spring from already constituted social groups or from aggregated individual char- acteristics. Rather, the cultural and organi- zational dimensions of political institutions are, to a large extent, constitutive of the groups themselves and of the civic activities their members engage in.

The World Values Surveys constitute a unique dataset for testing hypotheses about the structural basis of individual value ori- entation and behavior. The surveys cover a large sample of countries and a broad set of variables over multiple points in time, in- cluding variables relating to the participation of individuals in voluntary associations. These data show that people in different na- tions differ dramatically in their level of in- volvement in volunteer activities and that these differences are stable over time. The percentage of individuals claiming member- ship in at least one voluntary association ranges from about 70 percent in the United States and in most Scandinavian nations, to less than 30 percent in Japan and the south- ern European nations (see Table 1). Also, people in the United States and Scandinavia tend on average to join a greater number of associations.

Countries also vary in the types of associa- tional activity their citizens engage in (see Table 2, data described on pp. 8 15-16). Note that membership levels in the United States are particularly high for religious associa- tions and for categories we identify as "new" social movements (such as environmental or human and women's rights organization^).^ By contrast, associational activity in a coun- try like Germany is more centered on "old" social movement associations, such as unions and traditional political parties.

cally located and organized at the group level (as opposed to the individual level).

See, for instance, Klandermans and Tarrow (1988) and Melucci (1980). We use a categoriza- tion similar to Wessels (1997): "New" social movement associations include environmental, women's, peace, and development associations; "old" social movements include unions, political, and professional associations.

Table 1. Membership in Voluntary Associations for Selected Countries: Percentages and National Ranks from Various Data Sources

Average Individual

Percentage of Individuals Membership score

Reporting Membership in (Out of 10 Categories

Any Association, and Country Rank of Associations)

World Values World Values World Values Almond Survey, 1981 Survey, 1991 Surveys

and Verba Country (1963) Percentage Rank Percentage Rank 1981 1991

Iceland -82 1 86 1 1.64 2.00

United States 57 72 2 68 6 1.40 1.48

Sweden -67 3 77 2 1.10 1.47

Denmark -64 4 73 4 .96 1.19

Netherlands -6 1 5 75 3 1.20 1.80






West Germany





Spain -3 1 16 17 19 .45 .26




Average -49 -49 -.81 .93

Note: Countries are presented in order of rank for the 1981 World Values Survey.

Despite large differences among countries, Inglehart 1990:48-65, 1997:99). Another individual-level variables continue to pro- "value" frequently referred to is "trust." vide the main frame of reference for under- Drawing on attitudinal surveys in five na- standing patterns of civic participation, both tions, Almond and Verba (1963) showed that within and across nations. Previous research high levels of civic participation in the has established a strong correlation between United States and Britain (as opposed to volunteering and association membership on Italy, Mexico, and Germany) were associ- one hand, and church attendance, religious ated with high degrees of interpersonal trust. orientation, education, income level, gender, Following on this argument, Inglehart (1990, and marital status on the other hand (Curtis 1997) found that as societies industrialize, 197 1; Cutler 1976; Greeley 1997; Knoke individuals get more education and become 1986; Knoke and Thomson 1977; Scott 1957). wealthier, and therefore emphasize the

Civic participation is also frequently ex- "postmaterialist" values of well-being, toler- plained in terms of specific value orienta- ance, and trust-values which in turn sup- tions rooted in the larger social system. Re- port the development of associations (espe- ligion, particularly Protestantism, has been cially "new" social movement associations) found to play an important role in fostering and other democratic institutions (Inglehart civic orientation (Curtis et al. 1992:149; and Baker 2000).
Values Survey, 1991
All Types of Associations     Social     Social     Religious     Number of
Country     (of 10 Type~)~(of 16 TYP~S)~ Movements     Movements     Associations     IndividualsC
United States         
East Germany         
West Germany         

aNumbers represent the country average membership for the 10 types of associations (summed) that were included in both the 198 1 and 199 1 surveys: welfare, religion, education, union, political, community, Third world/development, environment, professional, youth.

Also includes the six additional types of associations included in the 1991 survey: sports, women, peace, animal rights, health, and "other."

Total number of individuals = 36,724.

These explanations are partly rooted in a revival of the Parsonian approach of the 1950s, in which political culture is equated with subjectively internalized values that are themselves analytically separate from politi- cal structures and institutions (Somers 1995). In this framework, value orientation determines behavior, and thus voluntary as- sociations develop as a consequence, at the aggregate level, of the actions of numerous individuals sharing similar attitudes or char- acteristics. Yet historical scholars have shown that such attitudes and practices do not exist, and cannot be thought of, indepen- dently from their "dialectical" and histori- cally grounded relationship with institutions (Sewell 1992; Steinmetz 1999:20). Knowl- edge is "internalized," and values are formed only if there are institutions that channel them in certain directions (Berger and Luck- man 1966). On one hand, political structures constrain the institutional means to pursue civic engagement and thereby shape the pos- sibilities for individual action (Clemens 1997; Skocpol 1985; Skocpol et al. 2000); on the other hand, they serve as social sites where perceptions and ideas about actorhood and sovereignty are played out, institution- alized, and constructed as "legitimate" (Meyer and Jepperson 2000; Steinmetz 1993). As an example, civic engagement may not be particularly prominent in Anglo- Saxon societies simply because people share liberal values. But the dialectical operation of enabling political structures, arm's-length state policies, and available cultural models that emphasize actorhood may produce a so- cial environment in which civic engagement simply "makes sense."

Following neoinstitutionalist theorists, we contend that individual behavior is chan- neled or "scripted" by institutionalized cul- tural frames (Berger and Luckman 1966; Friedland and Alford 1991; Meyer, Boli et al. 1997; Thomas et al. 1987). These frames, which have their roots in the political, reli- gious, and economic histories of nations, op- erate at both the organizational and cogni- tive levels. First, they shape the develop- ment of national systems of rules and insti- tutions (e.g., administrative practices, legal and rights systems, and so on). Comparative research supports the idea that broad institu- tional factors are responsible for observed cross-national differences in various do- mains of social activity. For instance, labor market regulation (Western 1997), the pro- vision of social goods (Amenta, Bonastia, and Caren 2001; Esping-Andersen 1999), or the organization of blood-giving (Healy 2000) are patterned systematically around large ensembles of nations sharing similar cultural and institutional characteristics.

Cultural frames also operate at a disaggre- gated, cognitive level (Dobbin 1994; Friedland and Alford 1991; Meyer, Boli et al. 1997; Thomas et al. 1987). They provide a lens through which individual actors ap- prehend the world and act within it, defining what Swidler (1986) calls "repertoires of ac- tion" (also see Lamont and ThCvenot 2000). As such, cultural frames should not be re- garded simply as "internalized" value sys- tems that rational individuals use to form their preferences. Rather, they are cognitive scripts, embedded in long institutional tradi- tions and organizational frameworks that shape the social behaviors and practices that are deemed legitimate, even "thinkable." In this way the cultural frames themselves be- come constitutive of those individuals and groups. Jepperson (1992), for instance, has argued that polity characteristics shape po- litical opinion and representations of the self in vastly different ways across countries. Jepperson and Meyer (1991) have shown that formal models of organizing are closely related to institutionalized models of the pol- ity. Boyle (2000) has suggested a similar in- terpretation for cross-national differences in legal activity.

In sum, both theory and empirical evi- dence indicate that political culture shapes individual action. Thus we expect institu- tionalized scripts about political behavior to affect the level and character of associational activity across nations.

Historical analyses of societal developments over long periods of time have identified two

fundamental dimensions of variation of po- litical structures and institutions: "statism" and "corporateness." Both concepts, and the two-by-two typology they give rise to are specifically elaborated in the work of Jepperson and Meyer (Jepperson 1992, forth- coming; Jepperson and Meyer 1991; Meyer

1983; also see Birnbaum 1988; Birnbaum and Badie 1983; Dyson 1980; Lipset 1985; Nett1 1968; Schmitter 1974). Briefly, in this view societies differ from one another pri- marily in the location and organization of po- litical sovereignty. Jepperson (1992, forth- coming) refers to the first dimension as the degree of "statism," and to the other dimen- sion as the degree of "corporateness" of the political structure.

Although these dimensions have much in common with other well-known comparative typologies, they are unique in two important ways: First, they provide a general conceptual tool for understanding cross-national polity variation, rather than a description of a specific set of differences (e.g., Esping- Andersen's [I9901 "three worlds" of welfare capitalism, which are based on cross-na- tional differences in political resource^).^ Second, they are grounded theoretically in the historical analysis of the long-term macro-evolution of societies, rather than be- ing derived from clusters of observed varia- tion (e.g., Inglehart's culture group^).^ These differences avert the possibility of tautology or reverse causality between our analytical framework and the outcome we seek to explain.

We now describe statism and corporate- ness in detail, summarizing the typifications advanced by Jepperson (forthcoming) and providing empirical details about the vary- ing shape of civic engagement across polity forms. We then formulate testable hypoth- eses about the plausible impact of statism

Esping-Andersen's (1990) three-regimes ty- pology (1) does not explain the historical sources of each regime (liberal, social-democratic, and "conservative"), and (2) does not properly account for southern European cases, including France (Morgan 2000; see Esping-Andersen 1999 for an attempt to deal with these issues).

'Inglehart's (1997) "cultural ensembles" are based mainly on national differences in dominant religious orientations (also see Inglehart and Baker 2000).

and corporateness on associational activity in different nations.

STATISM. In modern societies, political sovereignty historically has been derived from two major institutions: the state and civil society. From an analytical point of view, these institutions define a continuum between two ideal types, with a centralized and totally autonomous state apparatus at one end and a form of political power to- tally decentralized within an active and or- ganized society at the other (Jepperson forthcoming; Jepperson and Meyer 1991). Existing polities fall somewhere within this contin~um.~

France and Germany exemplify high stat- ism, although most continental European countries, particularly those with an absolut- ist legacy, are also examples. In such coun- tries, the state constitutes a separate and su- perior order of political governance that de- rives much of its legitimacy from a well-de- veloped bureaucratic elite, as well as from a long history of authoritarian political rule (esp. Germany, Austria, Russia, and Japan). Civil society, on the other hand, is regarded as a source of chaos and anomie (Jepperson and Meyer 199 1 :216) and is therefore often subject to some form of central state con- trol-from outright oppression in the earlier periods to administrative supervision and guidance in more recent times.

Anglo-Saxon countries, by contrast, are situated toward the low end of the "statism" scale. Bureaucratic development emerged relatively late, and political culture remains firmly centered on the idea of a self-govern- ing society, largely autonomous from the state (Birnbaum and Badie 1983). The state derives its legitimacy from its function as the representation of civil society, which is considered to be the principal locus of pub- lic life. The public bureaucracy is much less

Scholars often have noted that religious doc- trines and trajectories partly account for the de- gree of "statism" of political structures (Meyer and Jepperson 2000). For instance, Durkheim ([I8931 1984) and Weber ([I9221 1978) both saw the Protestant sects as the early carriers of liberal individualism; Zaret (1989) interprets English democratic discourse as a consequence of reli- gious sectarianism; Gorski (1993) emphasizes the role of disciplinary revolutions in explaining state strength.

elaborate and rationalized than it is in statist systems.7

Intuitively, the statist political form should discourage voluntary activism and the nonstatist form should facilitate it. In statist systems, the persistence of a centralized ide- ology of decision-making has traditionally kept associations at bay from the true centers of power (Veugelers and Lamont 1991). France, for instance, never completely parted from its long tradition of civil society sur- veillance and the centralization of associa- tional activity under the tutelage of the state.8 Since the time of absolutism, all po- litical regimes have regarded collective or- ganization with suspicion, and treated local and intermediary institutions as potentially dangerous. Freedom of association in France remained subject to restrictions until 1901, far later than did other elements of demo- cratic rule (e.g., universal suffrage for men), and even after that date associational activi- ties continued to be hampered by complex administrative procedures. Still today, "mi- nority" identities (e.g., regional and ethnic identities) are associated with factionalism, which conflicts with the universalistic frame- work of incorporation promoted by the state.

Because it is less culturally legitimate in statist countries, civil society receives little institutional encouragement. In Italy, for in- stance, the legal environment within which associations operate lacks coherence, as does the pattern of financial support (Perlmutter 1991 :178). Consequently, volun- tary organizations are poorly equipped to be effective actors in the public sphere. In spite of the global renewal of civil society ideolo- gies, governments in statist polities have found it difficult to "empower" associations to take over responsibilities traditionally shouldered by the state. The failed strategy of "associational liberalism" promoted by French governments in the 1980s in their at- tempt to liberalize the economy serves as an

'This point is exemplified by the tradition of amateurism in the British civil service (Heclo and Wildavsky 1974) and the weak boundaries be- tween the inside and the outside of government in the United States (Heclo 1988).

Rosanvallon (1990), for instance, character- izes the French state as the "tutor of society" (lit- erally, "1'Ctat instituteur du social").

example of the extent to which institutional legacies continue to shape state-society re- lations (Levy 1999).

In nonstatist societies, by contrast, both prevailing cultural frames and institutions historically have actively promoted civic en- gagement. Nineteenth-century Britain bur- geoned with a mosaic of local civic institu- tions-voluntary associations, communities, clubs, and trade unions. The state, by con- trast, "existed mainly to serve the conve- nience and protect the rights of individuals in private life" (Harris 1990:67)-especially property rights and the individual pursuit of commercial activities (Mann 1986: 107). As- sociations often worked in symbiosis with administrative institutions, rather than against them (Morris 1990:440). Similarly, political culture in the United States was forged through the experience of community self-government and has remained fiercely defensive of local autonomy and initiative since then (Bellah et al. 1985; Dobbin 1994). Even as federal economic and social respon- sibilities grew during the twentieth century, the state came to rely on civil society's ac- tivism and encouraged its expansion, for ex- ample, by involving voluntary groups in the implementation of welfare policies (Skocpol 1992, 1996, 1997; Skocpol and Fiorina 1999).

The Scandinavian states also exhibit a culturally supportive and benevolent atti- tude toward associations, albeit in a differ- ent manner from Anglo-Saxon countries. Voluntary action in Anglo-Saxon countries is still cast in a powerful liberal ideology that continues to celebrate voluntarism as autonomous and jealously defends its arm's-length relationship from government (Wuthnow 1991:300-301).9 In Scandinavia, on the other hand, the boundaries between the state and civil society are more blurred. Boli (1991:101) points out that in Swedish, "both terms ['state' and 'civil society'] are often used synonymously." In spite of a

This relationship is illustrated by the fact that the voluntary sector in both Britain and the United States still relies mainly on voluntary in- come (Beckford 1991:43). This situation differs markedly from arrangements in more corporatist countries, where "voluntary" associations derive most of their finances from state grants.

considerable role in society and the economy, Scandinavian states typically do not appropriate political sovereignty. Rather, political sovereignty is vested in so- ciety as a collection of organized, legiti- mate interests that are then orchestrated or "mediated" by the public authority (a struc- tural feature discussed below as "corporate- ness"). Indeed, because of the relatively egalitarian context in which social privi- leges were abolished early,1° a "strong" civil society was allowed to develop (espe- cially around the free church and labor movements). By the nineteenth century in Sweden, the practice of "consultation" between central administrative institutions and civil society groups, which later evolved into a full-fledged system of pa- tronage of the voluntary sector through fi- nancial backing and integrated participation in public decision-making, was already well institutionalized (Heclo and Madsen 1987; Micheletti 1995). With the development of the social-democratic welfare state, which motivated constant interactions, the embed- dedness of the state in society became rou- tinized as a legitimate mode of economic and social governance in which "consensus- oriented" policies were centrally negotiated among all interested parties (Schmitter

1974; Schmitter and Lehmbruch 1979).

The degree of statism, in sum, should have profound effects on patterns of civic activ- ity. Some polities produce cultural models and institutions that de-emphasize involve- ment in voluntary associations as a legiti- mate (and effective) mode of political action and concentrate decision-making authority in the hands of the state. Others have a more fluid demarcation between the public and the private spheres that allows "private" actions to be more legitimate in the "public" context (Jepperson 1992: 165-66). Thus we formu- late our first hypothesis:

Hypothesis 1: Polities with a low degree of statism will exhibit higher overall lev-

lo At the end of the nineteenth century, Swe- den, like most other small European states, did not have a powerful landed class. Rather, the ar- istocracy had long been co-opted by the state and was primarily concentrated in the bureaucracy (Anderson 1974: 171-9 1 ; Heclo and Madsen 1987; Stephens 1995:171-72).

els of associational membership than will statist polities.

While statist polities should generally dis- courage engagement in all forms of associa- tional activity, the persistence of traditional cleavages in civil society, as well as the lack of institutional legitimacy and support, should make it even harder, comparatively, for "new" associational forms to emerge and gain momentum-especially those associ- ated with specializing claims and identities. Previous empirical work likewise suggests that new social movement organizations are much weaker in France and in the southern European countries than they are elsewhere in Europe (Duyvendak 1995). This leads us to Hypothesis 2:

Hypothesis 2: Statism will have a strong negative effect on membership in "new" social movement associations.

Along a second di- mension, polities vary in the way in which social actors are incorporated-what Jepperson (1992, forthcoming) calls the de- gree of "corporateness." Some social sys- tems assign sovereign "actorhood" to private persons and typically locate sovereignty for interest representation in individuals-with group action being legitimate only as the embodiment of individual wishes. Other sys- tems assign a higher moral purpose to orga- nized groups, empowering individuals chiefly as members of broader collectives that have specific "rights and functions" (Jepperson and Meyer 1991 :214-17). In such countries, society is organized along "corporate" lines, that is, around collectives united (most of the time) by a particular eco- nomic project. Historically derived from a feudal and patrimonial past in which society was organized by estates, corporate institu- tions (from the old guild forms in the former

I eastern and central European empires to the modern "peak associations" of Scandinavia) still represent the main channels of public activity.

Our suspicion is that stronger corporate organization, because of its "intermediary" status, should encourage civic activities and thus lead to a more developed voluntary sec- tor. Historically, the state in corporate poli- ties has played a supportive role toward col- lective institutional arrangements as a way of promoting economic and political "order" (e.g., in Germany) (Anheier 1991:68) or so- cial "consensus" (e.g., in Scandinavia). The military tradition in the German, Austro- Hungarian, and Russian empires gave rise to an ordered conception of society as a collec- tion of separate groups with distinct at- tributes and functions. In the absence of po- litical rights, the old German and Austrian status groups (Stande or estates) gave birth to the modern corporate associations (Verbande), which then gathered into pow- erful federations. These large, centralized as- sociations became part of a mode of gover- nance largely orchestrated from above, yet incorporation of such groups into the public sphere is neither taken-for-granted nor auto- matic. Rather, the state and public adminis- tration usually retain discretionary authority to decide which associational groups may have access."

In a country like Sweden, by contrast, corporateness is more "functional" (Jepper- son 1992:121). Its origins are to be found in historical patterns of state-society alliance and in the unified, broad-based class move- ments that developed during the twentieth century. As a result, its representative di- mension is better established than it is in Germany or Austria, where corporateness emerged from guilds and other hierarchical orders and was more "status-oriented."

In corporate countries, the state encour- ages all forms of collective organization as the main channel for political incorporation and usually provides generous support- provided associations are large, nationwide, democratically run, and structured in a cen- tralized way that authorizes negotiation and bargaining with administrative institutions.12 The carriers of corporate interests play substantive social roles, often being closely integrated with policymaking insti- tutions and assuming broad administrative

l1 A similar case can be made for the ex-so- cialist nations in which "the auxiliary institutions of the party state (trade unions, youth organiza- tions, professional associations)" served to orga- nize the incorporation of civil society during the one-half to three-quarters of a century of Com- munist rule (Ekiert 1991 :286).

l2 On migrant associations in Sweden, see Soysal (1994:91).

responsibilities, for example in the manage- ment of welfare provisions. Social regula- tion is thus mainly ensured by cooperation, both between groups and the state, and among corporate groups themselves.

This mode of governance differs from countries with individualist political cul- tures. In Anglo-Saxon nationi, individuals, rather than groups, are supposed to be the best judges of their own interests and are consequently empowered as their own legiti- mate representatives (Jepperson and Meyer 1991). Although important variations remain as to "which" individuals constitute the ulti- mate source of political legitimacy-Ameri- can universalism and veneration of the self- made man, for instance, contrasts sharply with the British traditional deference to "gentlemenu-these nations share a focus on decentralized decision-making authority and representation (Lipset 1963).

France, and to a certain extent the southern European and Latin American nations gener- ally,13 falls in the same category, albeit for different reasons. In France, the emuower- ment of the individual as opposedAto the group is the result of a revolutionary past di- rected at the abolition of feudalism and privi- leges, which fostered a strong cultural aver- sion to any form of "corporatism." The word even has a pejorative connotation in French, "denoting parasitic, protected groups that re- ceive undeserved advantages at the expense of the common group" (Levy 1999: 10).Consequently, relations among social units, and between social units and the state, tend to be more conflictual, as evidenced by the high levels of institutionalized class struggle, the

l3 There is a legitimate concern that the Catho- lic nations in our sample (Italy, Spain, Portugal, and most Latin American countries) could be in- cluded in the "corporate" category. They all have a powerful estate tradition (the "latifundia"). In addition, many of these countries have undergone prolonged periods of authoritarian rule during the twentieth century that attempted, in various ways, to organize society along corporatist lines through "state-licensed intermediaries," often re- lying on traditional authoritative orders like the church and the army (Williamson 1985). Yet by and large these efforts by dictatorships were only partially successful, and few of these "corporate" elements remain embedded in the social structure (for a discussion of fascist Italy, see Dyson 1980:59).

radicalism of social movement organizations (most prominently unions), and center-pe- riphery cleavages in France and Italy.

The degree of "corporateness" thus repre- sents a second important structural dimen- sion for understanding variation in associa- tional activity. Because they legitimate cen- tralized incorporation, universalism, and col- lective organization, corporate social institu- tions should increase the level of associa- tional activity. Western (1997), for instance, has shown that institutions such as the "Ghent system," in which unions are respon- sible for the provision and administration of unemployment benefits, generally boost membership levels and encourage the devel- opment of collective (class) consciousness. Our third hypothesis generalizes this argu- ment:

Hypothesis 3: Because they promote collec- tive, inclusive forms of political incor- poration, corporate polities will foster higher levels of associational member- ship than will noncorporate polities.

We also can formulate hypotheses about the type of civic engagement that corporateness encourages. Historically institutionalized patterns of societal organization in corporate polities should lead us to expect a greater emphasis on activities linked to economic sectors, such as unions and professional as- sociations, but also to political parties, which are partly aligned with economic in- terests. (These have been labeled "old7' so- cial movement associations because they typically developed during the early part of the twentieth century.) Furthermore, the con- siderable legitimacy of these institutions as the primary mode of political incorporation, and the highly centralized nature of gover- nance, should allow them to become more encompassing and expand into new areas of social action, thereby preventing the erosion or archaism common in other countries. In- terestingly, Scandinavian unions have been relatively immune to the sharp downward trend in membership most of their western European counterparts have experienced over the last three decades (Western 1997). Hence, our fourth hypothesis:

Hypothesis 4: Associational activity in highly corporate nations should dispro- portionately emphasize "old" social movement associations.

Finally, we consider global trends in asso- ciational activity over time. Some theorists have argued that social life is now increas- ingly organized at the international level (Boli and Thomas 1999; Meyer, Boli, et al. 1997). The organizations that comprise this "world society" are largely modeled after liberal polities, especially the United States, and therefore actively promote the two main institutions on which such social systems rest (i.e., strong markets and civil societies) (Meyer, Frank, et al. 1997; Somers 2001). We thus expect that the expansion of the world society in the recent period has in- duced a global shift toward liberal models of political organization, typified by high lev- els of association and the growth of "new" social movements.

Our primary aim is to discern the effects of country-level polity characteristics (statism and corporateness) on individual associa- tional membership within a given country, net of other relevant individual-level and country-level factors. Data on associational activity are derived from the 198 1 and 199 1 World Values Surveys (WVS), which in- clude nearly 90,000 respondents from 43 countries (World Values Study Group 1994). Because certain questionnaire items are not available for every country, our main analy- ses of 1991 data contain information on roughly 37,000 individuals in 32 countries. Respondents were asked whether they "be- longed to7' (i.e., held membership in) volun- tary associations of different types (e.g., re- ligious organizations, sports groups, envi- ronmental associations, and so on). Ten cat- egories of association were measured in the 1981 survey, and six were added in the 1991 survey.14 We combined these variables to

l4 The categories included in the WVS are: (1) social welfare services for the elderly, handi- capped, or deprived people; (2) religious or church organizations; (3) education, arts, music, or cultural activities; (4) trade unions; (5)political parties or groups; (6) local community action on issues like poverty, employment, housing, ra- cial equality; (7) Third World development or

construct scores reflecting overall member- ship, as well as membership in associations of particular broad types.15 The dependent variables are measured as follows:

OVERALL ASSOCIATIONAL MEMBERSHIP. Measured by summing for each indi- vidual all 10 associational categories avail- able in both the 198 1 and 1991 surveys. For example, an individual who is a member of at least one environmental association and at least one religious association would receive a score of 2 on this measure. Individuals with no memberships in any category would score

0. The maximum possible score is 10.16


Old social movements are measured by the sum of memberships in categories for trade unions, political parties, and professional as- sociations.

New social movements are measured by the sum of memberships in categories for envi- ronmental associations, Third World devel- opment associations, women's organiza- tions, and peace organizations."

human rights; (8) conservation, the environment, ecology; (9) professional associations; (10) youth work (e.g., scouts, guides, youth clubs, etc.); (1 1) sports or recreation; (12) women's groups; (13) peace movement; (14) animal rights; (15) volun- tary organizations concerned with health; (16) other groups.

l5 Note that these scores are not the actual number of memberships an individual holds. The survey data indicate membership in one or more associations of each category. The membership score variable, which is a sum of categories, un- derestimates the number of memberships in cases where an individual is a member of more than one association in a given category. Still, it is the best measure available for a large sample of in- dividuals and countries, and it is likely to be highly correlated with actual membership data.

l6 We observed nearly identical results when we used the full array of categories available in the 1991 survey. We use this smaller set of cat- egories to allow the comparison between the 1981 and 1991 time periods.

l7 There is some debate as to which of the available categories reflect new social movement associations. To ensure that our results were not an artifact of the particular categories chosen, we reconstructed our indicator using various combi- nations of associational types that are considered "new" social movement associations (e.g., using only environmental associations and Third World

We consider a series of individual-level and country-level factors that may affect these measures of associational activity. Variable descriptions and descriptive statis- tics for all variables are listed in Table 4. Variables warranting additional explanation are discussed here. Country-level indepen- dent variables are measured as follows:

STATISM.This concept is measured di- chotomously based on Jepperson and Meyer's (Jepperson 1992, forthcoming; Jepperson and Meyer 1991) description of state structure and polity characteristics. For countries not discussed in those sources, we created codes by applying their definition (statist polities = 1).

This measure is di- chotomous, again based on research by Jepperson and Meyer (corporate polities = 1). Table 3 summarizes the classification on the dimensions of statism and corporateness for selected countries.


Measured by real gross domestic product (GDP) per capita, logged (Summers and Heston 1991). The idea that national devel- opment strengthens democratic institutions and behaviors goes back to modernization theory and is a core tenet of political sociol- ogy (Lipset 1960). On one hand, societal wealth is associated with the collective re- sources required to support associations. On the other hand, societal wealth is associated with enhanced education, more leisure time, and other individual-level characteristics that may increase association membership. We leave it as an empirical question whether there are direct effects of societal-level de- velopment, controlling for mediating indi- vidual-level factors (i.e., individual socio- economic indicators).

Democracy is measured as a 10-point scale reflecting the institutional- ization of political democracy (Jaggers and Gurr 1995).18 In addition to providing the

development associations). Results were consis- tent.

Is Wessels (1997) and others measure democ- racy as the number of years of continuous de- mocracy, arguing that long periods of uninter- rupted democracy are required for the formation of societal associations. To address this question, we tried using two other measures of democracy: a dummy variable indicating nations that were
Table 3. Variation in National Polity Structure: Statism versus Corporateness

Degree of Statism

Degree of

Corporateness Low High

Low United States, France, Italy, Spain,
Britain, Canada Portugal, Latin America

High Scandinavian countries Wilhelmine Germany, postwar Germany, Austria, Central and Eastern Europe, Japan

Source: Adapted from Jepperson (1992, chap. 3).

Note: Countries shown in bold type are closest to the ideal type.

political freedoms to engage in civic activi- outcome as a function of individual-level and ties, democracy is thought to foster a partici- country-level variables using OLS regression patory political culture that leads to the for- would overlook characteristics of the error mation of voluntary associations. It remains structure resulting from the commonalities of to be seen whether these arguments, rooted individuals within countries, which violate primarily in studies of American democracy, the assumptions of the OLS regression can be generalized to democracies elsewhere model. Hierarchical models, on the other in the world.19 hand, explicitly incorporate both individual-

Individual-level characteristics may also level and group-level error. affect membership in associations. We em- A multi-level model consists of an indi- ploy a standard array of such indicators, at- vidual-level equation and one or more tempting to maintain consistency with other group-level equations. We specified an indi- empirical work on this topic (e.g., Curtis et vidual-level (level-1) equation with its own al. 1992; Moyser and Parry 1997) (see error term, much like an ordinary regression. Table 4). In addition, the constant of the individual-

We use hierarchical models to analyze as- level equation is modeled as a function of sociation membership (Bryk and Rauden- country-level properties (level-2) and a sec- bush 1992). Multi-level models are appropri- ond "group-level" error term. The equations ate in this case because we are interested in estimated for our base model are: an individual-level outcome that is affected by both individual-level and country-level Membership = Po+ P1(Age) + /&(Male)variables. To simply aggregate individual-

+ /&(Education)

level membership scores into a country-level dependent variable would overlook the indi- vidual-level processes that affect association membership.20 To model an individual-level

continuously democratic since 1900, and a con-
tinuous measure reflecting the number of years PO= + yOl(Corporateness)

of continuous democracy. Results (not presented here) were similar to those from models using the 10-point democracy scale.

l9 Unfortunately, only a few nondemocratic countries are included in the World Values Sur-

We employ a nonlinear Poisson model be-

vey. A larger sample of nations would be needed

cause our dependent variable, the number of

to draw confident conclusions about the effects

memberships held by an individual, is a

of democracy on associational activity.

"count" (i.e., a nonnegative integer). A lin-

20 Such aggregate models produce results gen-

ear relationship is not appropriate given the

erally consistent with the hierarchical models shown below. Such a result makes sense in this highly skewed nature of the data and the fact case, where the country-level predictors are ex- that a linear model might nonsensically pre- tremely strong. dict "negative" memberships for some cases.

Table 4. Definitions and Descriptive Statistics for Variables Used in the Analyses: 32 Countries from the World Values Survey, 1991

Standard Variable Definition Mean Deviation

Country-Level Variables

Corporateness Measured dichotomously (1 = nation is high
on corporateness dimension) (Jepperson 1992).

Statism Measured dichotomously (1 = nation is high
on statism dimension) (Jepperson 1992).

Democracy Measured by 10-point democracy index (Jaggers
and Gurr 1995).

National economic Measured by Real GDP per capita (log)
development (Summers and Heston 1991).

Individual-Level Variables

Overall association Measured by the number of categories of
membership association (of 10) in which respondent is
a member.

"New" social movements Measured by memberships in categories
membership reflecting old social movements associations.

"Old" social movements Measured by memberships in categories
membership reflecting new social movements associations.

Age Age measured in years
Gender Measured dichotomously (male = 1).

Education Measured using a 4-point index (higher = more

Marital status Measured dichotomously (married = 1)

Religious belief Measured on a 4-point scale (4 = religion is
"very important", 1 = "not at all important").

Employment status Measured dichotomously (part-time or full-time
employee = 1).

Trust Measured dichotomously (1 = respondent agrees
that "most people can be trusted").

Post-material values Measured on a 3-point scale (see Inglehart 1997).

Note: Number of individuals = 36,724.

Individual-level coefficients are constrained of associations (Tables showing random ef-

to be constant across groups and are as- fects for models are available from the au- signed an error variance. We do not specify thors on request). cross-level interactions among variables. Model 1 is our base model, which leaves Multi-level Poisson models were estimated out individual-level attitude variables-trust

using restricted penalized quasi-likelihood and post-materialism. We present models (PQL) estimation with the program HLM 5 without these variables to avoid possible (Raudenbush et al. 2000; also see Guo and concerns about the direction of causality. Zhao 2000). (Because of their theoretical interest, how-

ever, we include these variables in Model 2 despite these caveats.)

Individual-level variables have effects that are consistent with the literature, whether the

prior studies were single-country analyses Table 5 presents the results of hierarchical (e.g., Moyser and parry 1997) or cross-na- Poisson models predicting the level of indi- tional studies (e.g., Curtis et al. 1992). Age, vidual association membership in all types education, and religiosity have positive and

Table 5. Hierarchical Poisson Regression Coefficients Showing the Effects of Selected Independent Variables on Individual's Overall Level of Association Membership: 32 Countries from the World Values Survey, 1991

Model 1 Model 2 Independent Variable (Base Model) (Full Model)

Country-Level Variables

Statism Corporateness Democracy GDP per capita (log)

Individual-Level Variables


(male = 1)

Marital status
(married = 1)
Religious belief

Employment status
(employed = 1)
Trust in others



Log-likelihood (x -.643+++ -.580+++ (.136) (.118)

.354++ .178+ (.110) (.098)

-.013 -.0008 (.024) (.022)

.16 1 .054 (.146) (.137)

.392+++ .387+++ (.039) (.043)

- .225+++ (.025)

-.178+++ (.020)

-3.14* -2.57*

(1.32) (1.25)

-5.85 -5.73

Note: Numbers in parentheses are standard errors. Number of individuals = 36,724.

+p< .05 ++p < .01 +++p < .001 (one-tailed tests)

*p< .05 **p< .O1 ***p< .001 (two-tailed tests)

significant effects on association member- ship, as do the dichotomous variables indi- cating married, male, and employed indi- viduals. The coefficients for employment and education are particularly large. Hierar- chical Poisson coefficients can be inter- preted by exponentiation, which yields a multiplier on the rate of membership (Raudenbush et al. 2000). The coefficient of



.392 for employment corresponds to a mul- tiplier of l .48 (exp[.392] = l .48), represent- ing a 48-percent increase in association membership for employed individuals. A single point increase on the education scale reflects a 34-percent increase in member- ship. Moreover, college-educated individu- als (scoring 4 on the index) have more than twice the level of association membership compared with those having only a few years of education.

Country-level variables also have large ef- fects on association membership. As hypoth- esized, statism has a negative and significant effect on association membership: The stat- ism coefficient of -.643 corresponds to a 47- percent decrease in association membership (exp[-.643] = .526). In other words, mem- bership levels in statist countries are roughly half those in nonstatist countries. Hypothesis 1 is thus supported. In addition, as predicted, corporateness has a positive and significant effect on membership, supporting Hypoth- esis 3. The coefficient of .354 reflects a 42 percent higher rate of association member- ship in corporatist countries (exp[.354] = 1.42). When combined, the effects of statism and corporateness produce dramatic differ- ences: Association membership in corporat- ist, nonstatist countries like Sweden is 2.70 times that of noncorporatist, statist nations like France (inverting the .526 multiplier for statist, corresponds to a 1.90 multiplier for nonstatist; [1.90][1.42] = 2.70).

In Model 1, neither GDP per capita (logged) nor democracy has significant ef- fects on association membership. GDP per capita tends to be positive across various model specifications, but generally does not come close to statistical significance. The coefficient for democracy is near zero. These findings may be a result of the relatively small range of countries, and thus lack of variation in these measures. Effects of GDP and democracy might be more pronounced in a larger sample of countries that includes many nations other than the industrialized Western democracies.

Model 2 includes two individual attitude variables thought to affect political partici- pation: "trust in others" and post-materialist values. Both have positive, significant ef- fects on association membership. Associa- tion membership, however, is sometimes

Table 6. Hierarchical Poisson Regression Coefficients Showing the Effects of Selected Independent Variables on Individual's Level of Membership in "Old" and "New" Social Movement Associations: 32 Countries from the World Values Survey, 1991

"Old" Social Movements "New" Social Movements

Model 3 Model 4 Model 5 Model 6 Independent Variable (Base Model) (Full Model) (Base Model) (Full Model)

Country-Level Variables

Statism -.414+++ -.356+++ -.982+++ -.922+++ (.095) (.092) (.238) (.184)

Corporateness .529+++ .249++ ,015 .080 (.093) (.079) (.169) (.142)

Democracy ,0003 ,010 .019 .002 (.020) (.021) (.024) (.022)

GDP per capita, log ,019 -.013 ,276 ,215 (.116) (.116) (.235) (.202)

Individual-Level Variables

Age .009+++ .009+++ .007++ .009+++ (.001) (.001) (.002) (.002)

Gender (male = 1) .227+++ .194+++ -.432+++ -.448+++ (.041) (.040) (.075) (.076)

Education .274+++ .226+++ .357+++ .267+++ (.027) (.023) (.039) (.039)

Marital status (married = 1) .179+++ .174+++ ,008 -.0001 (.019) (.019) (.041) (.042)

Religious belief -.046++ -.041f+ .105+++ .I 1 l+++ (.014) (.013) (.027) (.027)

Employment status .819+++ .798+++ .244+++ .242+++ (employed = 1) (.057) (.057) (.047) (.05 1)

Trust in others -.144+++ -.333+++ (.022) (.043)

Post-materialist values -.130f++ -.378+++ (.022) (.039)

Constant -3.08** -2.85** -5.92* -6.07**

(1.01) (1.02) (2.27) (1.91)

Log-likelihood (x -4.78 -4.65 -5.70 -5.74

Note: Numbers in parentheses are standard errors. Number of individuals = 36,724.
+p < .05 ++p< .O1 +++p< .001 (one-tailed tests)
*p< .05 **p< .01 ***p < ,001 (two-tailed tests)

thought to affect social capital and trust tions: political parties, trade unions, and pro- (Orum 1989; Paxton 2001), and thus reverse fessional associations. Individual-level vari- causality may be clouding the results. They ables have effects consistent with models of should be interpreted with caution. In any overall association membership, with the ex- case, our main findings are unchanged. ception of religious belief, which has a sig-

nificant negative effect. Statism has a negative and significant ef-


fect on old social movement membership, Table 6 presents the results of models pre- although the effect is smaller than in models dicting membership in specific types of as- of overall level of membership. Corporate- sociations. Models 3 and 4 predict member- ness has a positive, significant effect on ship in "old" social movement organiza- membership in old social movements that is larger than the coefficient in the analyses of overall membership (Model 1). In Model 3, the coefficient of corporateness is .529, in- dicating a 70-percent higher rate of member- ship compared to noncorporatist nations (exp[.529] = 1.70). Hypothesis 4 is sup- ported.

Models 5 and 6 contain results for mem- bership in new social movement associa- tions. Individual-level coefficients again de- part little from those for models of overall membership. Marital status ceases to have an effect, and the coefficient for gender be- comes negative and significant, indicating higher levels of membership among women than men. This result is partly due to the fact that women's organizations are included in the "new" social movements ~ategory.~' "Trust" and "post-materialist values," added in Models 4 and 6, have positive and statisti- cally significant effects on membership for both old and new social movements.

Statism has a strong negative effect on new social movement membership, consistent with Hypothesis 2. Indeed, statism has a larger negative effect on new social move- ments than on any other organization type, corresponding to a 63-percent lower level of membership (exp[-.982] = .374). Corporate- ness, on the other hand, has a slight positive but nonsignificant effect on new social movement membership, in contrast to its strong positive significant effect on old so- cial movement and overall membership. The general positive effect of corporateness on association membership does not extend to new social movements-perhaps because old social movements are so well institutional- ized that they preempt or crowd out new forms of associational activity.


We conducted several methodological checks on our results. First, we conducted analyses across a range of statistical models and methods of estimation to ensure that our results were not an artifact of our methods. Our main findings were consistent across

2' When "women's organizations" are excluded from the category, the coefficient for gen- der remains negative but is no longer statistically significant.

OLS regression models, hierarchical linear models, ordinary Poisson models, and nega- tive binomial models. Likewise, minimal differences were observed between robust and ordinary standard errors, or when using different methods of e~timation.~~

Second, issues of model specification and omitted variable bias are always a concern. To address this, we incorporated a variety of other country-level and individual-level vari- ables in our models, including: national-level educational expansion, political regime char- acteristics, individual socioeconomic status, political views, television-viewing habits,23 and many others. (Additional tables are available from the authors on request.) These variables were not included in our main analyses for the sake of parsimony and to avoid the loss of cases because of missing data. None of these variables affected the findings regarding statism and corporateness.

Third, Curtis, et al. (1992) suggest that re- ligious organizations are distinctive in terms of membership patterns compared with other types of associations. To ensure that religious organizations do not bias our results, we es- timated our models omitting religious orga- nizations in our overall association member- ship indicator. Findings were consistent.

Fourth, we also conducted similar analy- ses based on a sample including only industrialized democracies-eastern Europe, Latin America, and Asia were excluded. The main findings again were unchanged.


INTERACTIONS AND MEANING OF MEMBERSHIP. Our discussion of corporateness suggests corollary hypoth- eses regarding cross-level interaction effects that yield insights into the meaning of asso- ciational membership in different nations. Our earlier description of the various polity types suggests that the encompassing, inclu- sive nature of corporate institutions might

22 This addresses concerns voiced by Guo and Zhao (2000) regarding PQL estimation.

23 This variable is suggested by Putnam (2000). We find, however, a significant effect in the op- posite direction than Putnam would predict. Tele- vision-viewing has a positive, significant effect on association membership.

foster an "automatic," less voluntaristic form of civic engagement-reflecting an indivi- dual's location in society and the economy (e.g., a worker joining his or her industrial union) rather than the more proactive behav- ior of a participant in the public sphere (e.g., a mother mobilizing against drunk driving). In noncorporate nations, membership is less taken-for-granted and might thus depend more directly on individual attitudes and values. Consequently, one would predict that individual-level attitudes and capacities would have a smaller effect on association membership in corporate nations and a greater effect in noncorporate nations.

The exploratory analyses we conducted on this issue support these claims. First, we found that cross-level interactions between corporateness and individual-level variables such as education tend to be negative and significant. Education has a significantly smaller coefficient in corporate societies where membership tends to be taken for granted. Where association membership is optional (and thus dependent on individual initiative), education is a more important de- terminant of membership; where member- ship is "automatic," education matters less.

Second, preliminary examination of data on "active participation" in voluntary asso- ciations (as indicated in the World Values Surveys by individuals "doing unpaid work" for an association) also supported this point. While corporateness is highly predictive of association membership, we found that it is uncorrelated with active participation in an association. This issue is extremely impor- tant and deserves more systematic study. The fact that "membership" and "active partici- pation" in associations might diverge consti- tutes a powerful reminder of our main theo- retical argument-that the shape and struc- ture of "civil society" varies across nations and that the "meaning" of civic activities is highly differentiated depending on the soci- etal and cultural context. For instance, many authors have noted that unions in corporatist countries are much more "consensual" than they are in noncorporatist countries. Also, typical members in corporatist countries are more "passiveu-leaving negotiations to a professionalized class of experts. By con- trast, although they have much lower mem- bership levels overall, similar organizations in statist countries have different "reper- toires of collective action" and carry out more protest activities (Therborn 1995; Tilly 1986). A similar point can be made about the disconnection between church membership, which is pervasive in corporate societies like Sweden and Germany, and religious practice (e.g., praying and church attendance), which is low in those countries (see Gustaffson 1982).24 Polity characteristics such as stat- ism and corporateness may prove helpful in understanding such variation.


199 1. Given recent concerns about a secu- lar decline in voluntary association member- ship, we briefly explore national trends from 1981 to 1991.25 Again taking cues from neoinstitutional theorists, we argue that the postwar dominance of liberal ideologies in world society (i.e. ideas that emphasize the authority of individuals as legitimate social actors) (Frank and Meyer forthcoming; Frank, Meyer, and Miyahara 1995; Meyer, Boli, et al. 1997; Meyer, Frank et al. 1997; Meyer and Jepperson 2000; Ramirez, Soysal, and Shanahan 1998; Thomas et al. 1987) might generate convergence around "American-style" visions of associational life resulting in higher levels of association over time and more emphasis on "new" social movement membership. The 1980s wit- nessed considerable progress of the liberal order internationally, in domains as varied as economic organization, trade, education, and individual rights. Also, the collapse of the Eastern bloc and several Latin American dic- tatorships has generated tremendous interest in civil society and substantial efforts (on the part of international organizations, for in- stance) to engineer a revival of grassroots in- stitution~.~~

Therefore, we expect that, from

24 We thank John Meyer for drawing our at- tention to this point.

25 Unfortunately, we cannot extend analyses to the 1995 World Values Survey. Changes in the survey rendered the question noncomparable.

26 Both the World Bank and the United Na- tions, for instance, started substantial programs of collaboration and financial sponsorship of civil society organizations in the 1980s. These programs received further impetus in the 1990s and are thriving today. (For the World Bank, see http://wbln0018 . NGOsIhome; for the the United Nations, see

198 1 to 199 1, patterns of associational ac- tivity worldwide will have shifted toward higher levels of associational activity over- all, as well as a greater emphasis on the new social movements that are characteristic of liberal, noncorporatist societies.

Table 1 (see p. 808) shows trends from 1981 to 1991 in the average number of memberships for selected countries. It is clear that average membership increased for most countries. A few nations (e.g., Argen- tina and Japan) faced a decrease in associa- tional memberships. Most, however, experi- enced expansion-sometimes quite dra- matic expansion (e.g., Finland and Bel- gium). Among the 19 countries with mem- bership data for both 1981 and 1991, the national membership average increases from .81 to .93. A paired sample t-test con- firms that the overall increase from 1981 to 1991 is statistically significant (t-value = 1.87, 19 countries, a < .05, one-tailed test). The increase appears substantial, especially considering that it occurred over a brief 10- year period. A t-test also confirms a statisti- cally significant increase in new social movement activity throughout the world, from .054 to .092 (t-value = 2.68, 19 coun- tries, a < .05, one-tailed test).27 That is, 9.2 percent of people in the average country claim membership in a new social move- ment association in 1991. At the aggregate level, countries thus appear to be shifting toward the "liberal" model of associational activity, perhaps as a result of the global dominance of such models and institutions in world society. However, multivariate analyses are needed to draw firm conclu- sions about the causes of these trends.

Polity characteristics strongly influence how people associate in different nations. Statism has a deterrent effect on involvement in as- sociational activities that is especially strong for new social movement activities. Cor- home.htm. Accessed October 2001.)

27 TO allow a comparison between 1981 and 1991, new social movements are measured by environmental and Third World development asso- ciations.

porateness, on the other hand, encourages membership in associations, especially in unions and other old social movements. These polity effects operate strongly over and above individual-level variables such as individual education, employment and mari- tal status, and so on, as well as other coun- try-level variables such as economic devel- opment and democracy. Thus, polity charac- teristics shape not only the level of involve- ment in associational activities across coun- tries, but also its social modalities-the types of associations joined, and possibly other outcomes, such as whether participa- tion is active or passive.

Finally, our results contain some sugges- tive evidence about what the future might hold. On one hand, observed trends over time seem to lend some credence to the idea of a long-term convergence toward a "lib- eral" model of political incorporation. Cer- tainly the worldwide diffusion of a powerful liberal vision, in which "civil society" is re- garded as the most important agent of a suc- cessful democracy, is not irrelevant to this transformation. Indeed, direct interventions to promote civic engagement (especially in developing nations and former socialist countries) have become somewhat of a pana- cea for international actors and professional communities in search of new models after the failure of state-centered forms of soci- etal development (e.g., see Van Rooy 1998).

On the other hand, our general argument warrants a more cautious assessment. The fact that associational activity is firmly set in broad, historically evolved social struc- tures and cultural frames also suggests that such forms of involvement may not be so easily "engineered." To the extent that the nature and possibilities of civil society are shaped by these patterns of polity organiza- tion, then, actions that seek to enhance it may have only a limited effect.

Our research reflects a different approach to the comparative study of associations than has been commonly used. We have argued that associations ought to be studied using a structural framework that focuses on histori- cally evolved patterns of polity organization and that stresses the relevance of large- sample comparative analysis for testing in- stitutional and historical hypotheses. These patterns are well known to historical schol-

ars and social movement theorists but are of-

ten overlooked in large cross-country stud-

ies, which typically focus on individual-

level processes.

Our findings, moreover, suggest the need for new images of voluntary association. People do not "just join" voluntary associa- tions because they are wealthy, educated, or trusting, or have particular interests or social problems to address. The act of joining, and the particular types of organizations people join, are embedded in cultural and institu- tional arrangements defined at the level of the national polity. These national character- istics shape whether voluntary involvement is rational, legitimate, or simply possible, and the way in which it will occur. Thus, where political institutions and culture en- courage centralized and consensual volun- tary action, mobilization typically occurs within large, strong, established channels; where they discourage association, mobili- zation is likely to be fragmented and perhaps more antagonistic. Finally, where institu- tions encourage decentralized access to the political sphere, mobilization will be wide- spread and competitive. The individual "choice" of civic engagement makes sense only in relation to these highly structured contexts of action, which define both its lim- its and its strengths.

Evan Schofer is Assistant Professor of Sociology at the University of Minnesota, and is currently a National Academy of Education/Spencer Foun- dation postdoctoral fellow. His cross-national research examines education, civic engagement and social movements from a macro/institution- alist perspective. He also studies the expansion and global spread of education and science and their effects on society. His research shows how these institutions provide authoritative, rational- ized depictions that influence political actors, transform economic activity, and provide fram- ing for social movements.

Marion Fourcade-Gourinchas is a lecturer in sociology at Princeton University and a Post- doctoral Fellow at the Institute of French Stud- ies, New York University. Her main interests lie in the comparative-historical study of political and economic cultures. She studies the develop- ment of the economics profession in the United States and Europe. Currently, she is working on cross-national comparisons of cultural models for civic action, transitions to neoliberal policies

in developed and developing countries, and rep- resentations of the environment in the United States and France.


Almond, Gabriel and Sidney Verba. 1963. The Civic Culture. Princeton, NJ: Princeton Uni- versity Press.

Amenta, Edwin, Chris Bonastia, and Neal Caren. 2001. "U.S. Social Policy in Comparative and Historical Perspective: Concepts, Images, Ar- guments, and Research Strategies." Annual Review of Sociology 27:213-34.

Anderson, Perry. 1974. Lineages of the Absolut- ist State. London, England: New Left.

Anheier, Helmut K. 1990. The Third Sector: Comparative Studies of Non-Profit Organiza- tions. Berlin, Germany and New York: DeGryuter. . 1991. "West Germany: The Ambiguities of Peak Associations." Pp. 64-93 in Between States and Markets: The Voluntary Sector in Comparative Perspective, edited by R. Wuthnow. Princeton, NJ: Princeton University Press.

Beckford, James A. 199 1. "Great Britain: Volun- tarism ansd Sectional Interests." Pp. 30-63 in Between States and Markets: The Voluntary Sector in Comparative Perspective, edited by

R. Wuthnow. Princeton, NJ: Princeton Univer- sity Press.

Bellah, Robert N., William M. Sullivan, Ann Swidler, and Steven M. Tipton. 1985. Habits of the Heart. Individualism and Commitment in American Life. Berkeley, CA: University of California Press.

Berger, Peter L. and Thomas Luckmann. 1966. The Social Construction of Reality. A Treatise in the Sociology of Knowledge. New York: Anchor Books.

Birnbaum, Pierre. 1988. States and Collective Action: The European Experience. New York: Cambridge University Press.

Birnbaum, Pierre and Bertrand Badie. 1983. The Sociology of the State. Chicago, IL: Chicago University Press.

Boli, John. 1991. "Sweden: Is There a Viable Third Sector?." Pp. 94-124 in Between States and Markets: the Voluntary Sector in Com- parative Perspective, edited by R. Wuthnow. Princeton, NJ: Princeton University Press.

Boyle, Elizabeth H. 2000. "Is Law the Rule? Us- ing Political Frames to Explain Cross-National Variation in Legal Activity." Social Forces 78:1195-1226.

Bryk, Anthony S. and Stephen Raudenbush. 1992. Hierarchical Linear Models: Applica- tions and Data Analysis Methods. Newbury Park, CA: Sage.

Clemens, Elizabeth S. 1997. The People's Lobby: Organizational Innovation and the Rise of In- terest Group Politics in the United States, 1890-1925. Chicago, IL: Chicago University Press.

Cohen, Jean L. 1999. "Does Voluntary Associa- tion Make Democracy Work?" Pp. 263-91 in Diversity and Its Discontents: Cultural Con- flict and Common Ground in Contemporary American Society, edited by N. J. Smelser and

J. C. Alexander. Princeton, NJ: Princeton Uni- versity Press.

Curtis, James. 197 1. "Voluntary Association Joining: A Cross-National Comparative Note." American Sociological Review. 36:872-80.

Curtis, James, Edward Grab, and Douglas Baer. 1992. "Voluntary Association Membership in Fifteen Countries: A Comparative Analysis." American Sociological Review. 57:139-52.

Cutler, Stephen J. 1976. "Age Differences in Voluntary Association Membership." Social Forces. 55:43-58.

Dobbin, Frank. 1994. Forging Industrial Policy. Cambridge, England: Cambridge University Press.

Durkheim, Emile. [I8931 1984. The Division of Labor in Society. New York: Free Press.

Duyvendak, Jan W. 1995. The Power of Politics: New Social Movements in France. Boulder, CO: Westview.

Dyson, Kenneth. 1980. The State Tradition in Western Europe. Oxford, England: Martin Robertson.

Ekiert, Grzegorz. 199 1. "Democratization Pro- cesses in East Central Europe: A Theoretical Reconsideration." British Journal of Political Science 21:285-313.

Esping-Andersen, Gosta. 1990. The Three Worlds of Welfare Capitalism. Princeton, NJ: Princeton University Press. . 1999. The Social Foundations of Postindustrial Economies. Oxford, England: Oxford University Press.

Frank, David and John W. Meyer. Forthcoming. "The Profusion of Individual Roles and Identi- ties in the Post-War Period." Sociological Theory 19.

Frank, David J., John W. Meyer, and David Miyahara. 1995. "The Individualist Polity and the Prevalence of Professionalized Psychology: A Cross-National Study." American So- ciological Review 60:360-77.

Friedland, Roger and Robert R. Alford. 1991. "Bringing Society Back In: Symbols, Prac- tices, and Institutional Contradictions." Pp. 232-66 in The New Institutionalism in Orga- nizational Analysis, edited by W. W. Powell and P. J. DiMaggio. Chicago, IL: Chicago Uni- versity Press.

Gorski, Philip S. 1993. "The Protestant Ethic Re- visited: Disciplinary Revolution and State For- mation in Holland and Prussia." American Journal of Sociology 99:265-3 16.

Greeley, Andrew. 1997. "The Other Civic America: Religion and Social Capital." The American Prospect 32:68-73.

Gustaffson, Goran. 1982. "Popular Religion in Sweden." Social Compass 29: 103-12.

Guo, Guang and Hongxin Zhao. 2000. "Multi- level Modeling for Binary Data." Annual Re- view of Sociology 26:441-62.

Harris, JosC. 1990. "Society and the State in Twentieth Century Britain." Pp. 63-1 17 in The Cambridge Social History of Britain, 1750- 1950, vol. 3, Social Agencies and Institutions, edited by F. M. L. Thompson. Cambridge, En- gland and New York: Cambridge University Press.

Healy, Kieran. 2000. "Embedded Altruism: Blood Collection Regimes and the European Union's Donor Population." American Journal of Sociology 105: 1633-57.

Heclo, Hugh. 1988. "The In-and-Outer System: A Critical Assessment." Political Science Quarterly 103:37-56.

Heclo, Hugh and Henrik Madsen. 1987. Policy and Politics in Sweden: Principled Pragma- tism. Philadelphia, PA: Temple University Press.

Heclo, Hugh and Aaron Wildavsky. 1974. The Private Government of Public Money. Commu- nity and Policy Inside British Politics. Berkeley, CA: University of California Press.

Inglehart, Ronald. 1990. Culture Shift in Ad- vanced Industrial Society. Princeton, NJ: Princeton University Press. . 1997. Modernization and Post-Modern- ization: Cultural, Economic and Political Change in 43 Societies. Princeton, NJ: Princeton University Press.

Inglehart, Ronald and Wayne Baker. 2000. "Modernization, Cultural Change and the Per- sistence of Traditional Values." American So- ciological Review 65: 19-5 l.

Jaggers, Keith and Ted Robert Gurr. 1995. Polity 11I: Regime Change and Political Author- ity, 1800-1994. Ann Arbor, MI: Inter-univer- sity Consortium for Political and Social Re- search.

James, Estelle. 1989. The Nonprofit Sector in Comparative Perspective: Studies in Com- parative Culture and Policy. Oxford, England: Oxford University Press.

Jepperson, Ronald. 1992. National Scripts: The Varying Construction of Individualism and Opinion across the Modern Nation-States. Ph.D. dissertation, Department of Political Sci- ence, Yale University, New Haven, CT. . Forthcoming. "Institutional Logics: On the Constitutive Dimensions of Modern Na-

tion-State Polities." Sociological Theory 20(1).

Jepperson, Ronald and John W. Meyer. 1991. "The Public Order and the Construction of Formal Organizations." Pp. 204-31 in The New Institutionalism in Organizational Analy- sis, edited by P. J. DiMaggio and W. W. Powell. Chicago, IL: Chicago University Press.

Kitschelt, Herbert. 1985. "New Social Move- ments in West Germany and the United States." Political Power and Social Theory 5: 273-324.

Klandermans, Bert, Hans-Peter Kriesi, and Sidney Tarrow, eds. 1988. From Structure to Action: Comparing Movement Participation Across Cultures. Greenwich, CT: JAI.

Klandermans, Bert and Sidney Tarrow. 1988. "Mobilization into Social Movements: Synthe- sizing European and American Approaches." International Social Movement Research 1: 1-

38. Knoke, David. 1986. "Associations and Interest Groups." Annual Review of Sociology 12: 1-21.

Knoke, David and Randall Thompson. 1977. "Voluntary Association Membership Trends and the Family Life Cycle." Social Forces 56:48-65.

Kriesi, Hanspeter, Ruud Koopmans, Jan Willem Dyvendak, and Marco Giugni. 1995. New So- cial Movements in Europe: A Comparative Analysis. Minneapolis, MN: University of Minnesota Press.

Lamont, Michkle and Laurent ThCvenot, eds. 2000. Rethinking Comparative Cultural Soci- ology. Repertoires of Evaluation in France and the United States. New York: Cambridge University Press.

Levy, Jonah. 1999. Tocqueville's Revenge. State, Society and Economy in Contemporary France. Cambridge, MA: Harvard University Press.

Lipset, Seymour Martin. 1960. Political Man; the Social Bases of Politics. Garden City, NY: Doubleday. . 1963. "The Value Patterns of Democ- racy: A Case Study in Comparative Analysis." American Sociological Review 285 15-3 1. . 1985. "Canada and the United States: The Cultural Dimension." Pp. 109-60 in Canada and the United States, edited by C. F. Doran and J. Sigler. Englewood Cliffs, NJ: Prentice-Hall.

Mann, Michael. 1986. The Sources of Social Power. Vol. 2, The Rise of Classes and Nation- States, 1760-1914. New York: Cambridge University Press.

McAdam, Douglas, John D. McCarthy, and Mayer N. Zald, eds. 1996. Comparative Per- spectives on Social Movements: Political Op- portunities, Mobilizing Structures, and Cul-

tural Framing. Cambridge, England and New York, NY: Cambridge University Press.

Melucci, Alberto. 1980. "The New Social Move- ments: A Theoretical Approach." Social Sci- ence Information 19: 199-226.

Meyer, John W. 1983. "Institutionalization and the Rationality of Formal Organizational Structure." Pp. 261-82 in Organizational En- vironments: Ritual and Rationality, edited by

J. W. Meyer and W. R. Scott. Beverly Hills, CA: Sage.

Meyer, John and Ronald Jepperson. 2000. "The "Actors" of Modern Society: The Cultural Construction of Social Agency." Sociological Theory 18: 100-20.

Meyer, John W., John Boli, George Thomas, and Francisco 0. Ramirez. 1997. "World Society and the Nation-State." American Journal of Sociology 103: 144-8 1.

Meyer, John W., David Frank, Ann Hironaka, Evan Schofer, and Nancy B. Tuma. 1997. "The Rise of an Environmental Sector in World So- ciety." International Organization 5 1:623-5 l.

Micheletti, Michele. 1995. Civil Society and State Relations in Sweden. Aldershot, England: Avebury.

Morgan, Kimberly J. 2000. Whose Hand Rocks the Cradle? The Politics of Child Care Policy in Advanced Industrialized States. Ph.D. dis- sertation, Department of Sociology, Princeton University, Princeton, NJ.

Morris, R. J. 1990. "Clubs, Societies, Associa- tions." Pp. 395-444 in The Cambridge Social History of Britain, 1750-1950, vol. 3, Social Agencies and Institutions, edited by F. M. L. Thompson. Cambridge, England and New York: Cambridge University Press.

Moyser, George and Geraint Parry. 1997. "Vol- untary Associations and Democratic Participa- tion in Britain." In Private Groups and Public Life. Political Involvement in Representative Democracies. Edited by J. van Deth. New York: Routledge.

Nettl, J. P. 1968. "The State as a Conceptual Variable." World Politics. 20559-92.

Orum, Anthony M. 1989. "Citizen Participation in Politics." Pp. 244-63 in Introduction to Po- litical Sociology: The Social Anatomy of the Body Politic. Englewood Cliffs, NJ: Prentice Hall.

Paxton, Pamela. 1999. "Is Social Capital Declin- ing in the United States? A Multiple Indicator Assessment." American Journal of Sociology 105:88-127. . 2001. "Social Capital and Democracy: An Interdependent Relationship." Working pa- per, Department of Sociology, Ohio State Uni- versity, Columbus, OH.

Perlmutter, Ted. 1991. "Italy: Why No Voluntary Sector?" Pp. 157-88 in Between States and

Markets: The Voluntary Sector in Comparative Perspective, edited by R. Wuthnow. Princeton, NJ: Princeton University Press.

Portes, Alejandro and Patricia Landolt. 1996. "The Downside of Social Capital." American Prospect 26 (MayIJune): 18-21.

Putnam, Robert D. 2000. Bowling Alone: The Collapse and Revival of American Community. New York: Simon and Schuster.

Putnam, Robert D. 1993. Making Democracy Work: Civic Traditions in Modern Italy. Princeton, NJ: Princeton University Press.

Ramirez, Francisco O., Yasemin Soysal, and Suzanne Shanahan. 1998. "The Changing Logic of Political Citizenship: Cross-National Acquisition of Women's Suffrage." American Sociological Review 62:735-45.

Raudenbush, Stephen, Anthony Bryk, Yuk Fai Cheong, and Richard Congdon. 2000. HLM 5: Hierarchical Linear and Nonlinear Modeling. Lincolnwood, IL: Scientific Software Interna- tional, Inc.

Rosanvallon, Pierre. 1990. tat en France de 1789 b nos jours (The state in France from 1789 to the present). Paris, France: Seuil.

Salamon, Lester M. and Helmut K. Anheier. 1994. The Emerging Sector: The Nonprofit Sector in Comparative Perspective: An Over- view. Baltimore, MD: Johns Hopkins Univer- sity Institute for Policy Studies.

. 1997. "The Civil Society Sector." Society 34:60-65. Schmitter, Philippe C. 1974. "Still the Century of Corporatism?" Review of Politics 36:85


Schmitter, Philippe C. and Gerhard Lehmbruch, eds. 1979. Trends towards Corporatist Inter- mediation. London, England and Beverly Hills, CA: Sage.

Scott, John C. 1957. "Membership and Participa- tion in Voluntary Associations." American So- ciological Review 22:3 15-26.

Sewell, William H., Jr. 1992. "A Theory of Structure: Duality, Agency, and Transforma- tion." American Journal of Sociology 98: 1-29.

Skocpol, Theda. 1985 "Bringing the State Back In: Strategies of Analysis in Current Research." Pp. 3-43 in Bringing the State Back In, edited by P. Evans, D. Rueschemeyer, and T. Skocpol. New York: Cambridge University Press. . 1992. Protecting Soldiers and Mothers: The Political Origins of Social Policy in the United States. Cambridge, MA: Harvard Uni- versity Press.

. 1996. "Unraveling from Above." The American Prospect 25:20-25. . 1997. "The Tocqueville Problem." Social Science History 24:455-79. Skocpol, Theda and Morris Fiorina, eds. 1999.

Civic Engagement in American Democracy.

Washington, DC: Brookings Institute.

Skocpol, Theda, Marshal Ganz, and Ziah Munson. 2000. "A Nation of Organizers: The Institutional Origins of Civic Voluntarism in the United States." American Political Science Review 94527-46.

Somers, Margaret. 1995. "What Is Political or Cultural about Political Culture and the Public Sphere? Toward an Historical Sociology of Concept Formation." Sociological Theory 13: 1 13-44. . 2001. "Romancing the Market, Reviling the State: Historicizing Liberalism, Privati- zation, and the Competing Claims to Civil So- ciety." Pp. 23-48 in Citizenship, Markets, and the State, edited by C. Crouch, K. Eder, and

D. Tambini. Oxford, England: Oxford Univer- sity Press.

Soysal, Yasemin. 1994. Limits of Citizenship: Migrants and Postnational Membership in Eu- rope. Chicago, IL: Chicago University Press.

Steinmetz, George. 1993. Regulating the Social. The Welfare State and Local Politics in Impe- rial Germany. Princeton, NJ: Princeton Uni- versity Press. . 1999. "Introduction: Culture and the State." Pp. 1-49 in State/Culture: State For- mation After the Cultural Turn, edited by G. Steinmetz. Ithaca, NY: Cornell University Press.

Stephens, John D. 1995. "The German Path to Modern Authoritarianism: Germany, Britain and Sweden Compared." Pp. 161-82 in Politics, Society, and Democracy: Comparative Studies, edited by H. E. Chehabi and A. Stepan. Boulder, CO: Westview.

Summers, Robert and Alan Heston. 1991. "The Penn World Table (Mark 5): An Expanded Set of International Comparisons, 1950-1988." Quarterly Journal of Economics 106:327-68.

Swidler, Anne. 1986. "Culture in Action: Sym- bols and Strategies." American Sociological Review 5 1 :273-86.

Tarrow, Sidney. 1996. "Making Social Science Work Across Space and Time: A Critical Re- flection on Robert Putnam's Making Democ- racy Work." American Political Science Re- view 90:389-97.

Therborn, Goran. 1995. European Modernity and Beyond: the Trajectory of European Societies, 1945-2000. London, England and Thousand Oaks, CA: Sage.

Thomas, George M., John W. Meyer, Francisco

0. Ramirez, and John Boli. 1987. Institutional Structure: Constituting State, Society, and In- dividual. Newberry Park, CA: Sage.

Tilly, Charles. 1986. The Contentious French: Four Centuries of Popular Struggle. Cambridge, MA: Harvard University Press.

Tocqueville, Alexis de. [I8621 1945. Democracy

in America. New York: Vintage Books.

Van Rooy, Alison, ed. 1998. Civil Society and the Aid Industry. London, England: Earthscan Publications Ltd.

Veugelers, Jack and Michkle Lamont. 1991. "France: Alternative Locations for Public De- bate." Pp. 125-56 in Between States and Mar- kets: the Voluntary Sector in Comparative Per- spective, edited by R. Wuthnow. Princeton, NJ: Princeton University Press.

Weber, Max. [I9221 1978. Economy and Society. Berkeley, CA: University of California Press.

Wessels, Bernhard. 1997. "Organizing Capacities of Societies and Modernity." Pp. 198-219 in Private Groups and Public Life. Political 1n- volvement in Representative Democracies, edited by J. van Deth. New York: Routledge.

Western, Bruce. 1997. Between Class and Mar- ket: Postwar Unionization in the Capitalist Democracies. Princeton, NJ: Princeton Uni- versity Press.

Williamson, Peter. 1985. Varieties of Corpora- tism. A Conceptual Discussion. Cambridge, England and New York: Cambridge University Press.

World Values Study Group. 1994. World Values Survey, 1981-1984 and 1990-1993. Ann Ar- bor, MI: Inter-university Consortium for Po- litical and Social Research.

Wuthnow, Robert. 1991. "Tocqueville's Ques- tion Reconsidered: Voluntarism and Public Discourse in Advanced Industrial Societies." Pp. 288-308 in Between States and Markets: The Voluntary Sector in Comparative Perspec- tive, edited by R. Wuthnow. Princeton, NJ: Princeton University Press. . 1999. "The Changing Character of So- cial Capital in the United States." Working Pa- per, Department of Sociology, Princeton Uni- versity, Princeton, NJ.

Zaret, David. 1989. "Religion and the Rise of Liberal-Democratic Ideology in 17th-Century England." American Sociological Review 54:163-79.

  • Recommend Us